Department of Mechanical and Manufacturing Engineering

Aalborg University, Denmark.

Special Report No. 71

Improving the Fatigue and Control

Performance of Loader Cranes

Ph.D. Thesis
by
Mikkel M. Pedersen
Department of Mechanical and Manufacturing Engineering, Aalborg University
Pontoppidanstræde 101, DK-9220 Aalborg East, Denmark
e-mail: mmp@m-tech.aau.dk

Copyright c 2011 Mikkel M. Pedersen

This report, or parts of it, may be reproduced without the permission of the author,
provided that due reference is given. Questions and comments are most welcome
and may be directed to the author, preferably by e-mail.

Typeset in LATEX and printed in Aalborg, 2011.

ISBN 87-91464-25-0
 iii

Preface

This dissertation has been submitted to the Faculty of Engineering, Science and
Medicine in partial fulfillment of the requirements for the degree of Doctor of
Philosophy in Mechanical Engineering.

The work has been carried out at the Department of Mechanical and Manufactur-
ing Engineering, Aalborg University and at Højbjerg Maskinfabrik A/S (HMF) as
an Industrial PhD project in the period from August 2007 to January 2011.

The work has been partially funded by the Danish Ministry of Science, Technology
and Innovation, under the Industrial PhD initiative as project no. 07-017594.

I would like to thank my supervisors Prof. Michael R. Hansen at University of

Agder and Assoc. Prof. Ole Ø. Mouritsen from Aalborg University. Also my
colleagues and supervisors from HMF are thanked; Morten Ballebye, Jes G.
Andersen, Jimmi Wenderby and Gudmund Braendgaard. Last but not least, I
thank my family and especially my girlfriend, Mie, for patiently listening to a lot
of talk about cranes.

Mikkel Melters Pedersen

Højbjerg, January 2011.
iv
 v

Abstract
This industrial Ph.D. project concerns methods for development of loader cranes
with higher performance than currently possible. It covers two topics which are
highly related when considering a fatigue prone welded structure subjected to
significant dynamic loading due to the control of the hydraulic actuation, i.e.:

• Fatigue of welded joints, and

• Control of mobile hydraulics

While these topics may seem quite distinct, it should be considered, that the first
governs the fatigue resistance and the second governs the fatigue loading. It is
therefore expedient to take a holistic approach and consider both topics within
the same project.

In regards to fatigue of welded joints, this project establishes methods for

improving the fatigue performance of welded joints by post weld treatment
and design optimization using the notch stress approach. Existing methods
are investigated and adapted to the specific conditions of loader cranes. These
include, very high strength steel, thin sections, complex geometry, high stress
ranges and low-to-medium cycle fatigue lives. Furthermore, a proposal is made
for extending the notch stress approach to consider post weld treatment in the
fatigue assessment. The fatigue resistance of the critical welded joints can thus be
optimized in the design stage.

In order to limit the dynamic peak loading during operation of the crane,
especially at increased speeds, it is expedient to use more advanced control
principles than the current direct control of the individual actuators. Therefore,
a tool point control scheme is investigated and further developed, considering
application related issues, such as joint limit avoidance, deflection compensation
and configuration control. A practical approach for control development is
established using interactive, real-time dynamic simulation in order to included
the effects of an active operator in the control loop. Additionally, in order to
verify correct implementation of such tool point control in the current controller,
a hardware-in-the-loop simulator is developed to satisfy future testing needs.

Based on the contributions of this work, it will be possible to develop high

performance, lightweight and durable loader cranes and the addition of tool point
control should pave the way for improved productivity.
vi
 vii

Resume
Dette ErhvervsPhD projekt omhandler metoder til udvikling af lastbilkraner med
højere ydeevne, end muligt for nuværende. Det omhandler to emner, som er
særdeles relaterede, når man betragter udmattelsesfølsomme svejste strukturer
udsat for dynamiske belastninger pga. styringen af den hydrauliske aktuering,
nemlig:

• Udmattelse af svejste samlinger, og

• Styring af mobilhydraulik

Umiddelbart kan disse emner virke fjernt beslægtede, men for lastbilkraner er
udmattelsesstyrken styret af de svejste samlinger og udmattelseslasten bestemmes
af styringen af den hydrauliske aktuering. Det er derfor hensigtsmæssigt, ud fra
en holistisk tilgangsvinkel, at bearbejde begge emner i det samme projekt.

Indenfor udmattelse af svejste samlinger etableres der i dette projekt metoder til at
forbedre udmattelsesstyrken ved efterbehandling og designoptimering vha. notch
metoden. Eksisterende metoder undersøges og tilpasses til de specifikke forhold der
gælder for lastbilkraner. Dette inkluderer højstyrkestål, tynde tværsnit, komplekse
geometrier, høje spændingsvidder og korte levetider. Desuden gives et forslag til
udvidelse af notch metoden til at inkludere efterbehandling i levetidsvurderingen.
Udmattelsesstyrken af de kritiske samlinger i en kran kan derved optimeres på
designstadiet.

For at begrænse de dynamiske spidsbelastninger kranen udsættes for, specielt

ved forhøjede hastigheder, er det hensigtsmæssigt at anvende mere avancerede
styreprincipper, end det nuværende med direkte styring af de individuelle aktua-
torer. Derfor undersøges og videreudvikles et tool point styreprincip, under hensyn-
tagen til applikationen, f.eks. udbøjningskompensation og arbejdsstillingskontrol.
Der anvendes en praktisk tilgangsvinkel til udvikling af det nye styreprincip, som
gør brug af interaktiv, dynamisk simulering i realtid for at inkludere effekten af en
aktiv bruger i kontrolløkken.

På baggrund af dette arbejde vil det være muligt at udvikle højtydende og holdbare
lette lastbilkraner og samtidig brug af tool point styring skulle muliggøre højere
produktivitet i kranarbejdet.
viii
 ix

Appended papers

This thesis is based on the following papers. All papers are reproduced as-
published, only adapted to fit this printing format.

Papers A and B have also been presented in the IIW as documents XIII-2272-09
and XIII-WG3-10-09, respectively. Paper D has also been presented at the 2010
NODE seminar, Grimstad, Norway.

Publications in refereed journals

• Paper A: MM Pedersen, OØ Mouritsen, MR Hansen, JG Andersen, J Wenderby
Comparison of Post Weld Treatment of High Strength Steel Welded Joints in
Medium Cycle Fatigue. Welding in the World, 54(7/8), pp. 208–217 (2010).
• Paper B : MM Pedersen, OØ Mouritsen, MR Hansen, JG Andersen, J Wenderby
Re-analysis of fatigue data for welded joints using the notch stress approach.
International Journal of Fatigue, 32, pp. 1620–1626 (2010).
• Paper D: MM Pedersen, MR Hansen, M Ballebye
Developing a Tool Point Control Scheme for a Hydraulic Crane Using Interactive
Real-time Dynamic Simulation. Journal of Modelling, Identification and Control,
31(4), pp. 133–143 (2010).

Publications in proceedings
• Paper C : MM Pedersen, OØ Mouritsen, MR Hansen, JG Andersen
Experience with the Notch Stress Approach for Fatigue Assessment of Welded
Joints. In: Proceedings of the Swedish Conference on Light Weight Optimized
Welded Structures, (ed. Z. Barsoum and J. Samuelsson), March 24–25, 2010,
Borlänge, Sweden, pp. 122–133.
• Paper E : MM Pedersen, MR Hansen, M Ballebye
A Cost-Effective Approach to Hardware-In-The-Loop Simulation. Submitted to
the 9th International Conference: Mechatronics 2011, September 21–24, Warsaw,
Poland, 2011.

Division of work between the authors

For all papers M.M. Pedersen did the main work, and the co-authors contributed
with valuable discussion and helpful comments.
x
 xi

Nomenclature
∆σ Stress range
σn Nominal stress
σhs Hotspot structural stress
σk Notch stress
σk,P S Notch stress, principal stresses
σc Characteristic fatigue strength at 2 · 106 cycles
σy Yield strength
Nf Fatigue life, cycles to failure
km Stress magnification factor considering misalignment
R Radius or Stress ratio
Rref Reference radius
PS Probability of survival
PF Probability of failure
Kt Geometric stress concentration factor
Kf Fatigue notch factor
Kw Notch factor (σk /σhs )
fR Factor considering stress ratio
jσ Factor considering probability of survival
m Inverse negative slope coefficient of SN curve
n Number of specimens
ṙT P Tool point velocity vector
θ Joint coordinates vector
J Jacobian matrix
J+ Moore-Penrose pseudo inverse of the Jacobian matrix
z Arbitrary vector in joint space
Fc Cylinder forces vector
u Control signals vector
xii

Abbreviations
DOF Degrees of freedom
HMF Højbjerg Maskinfabrik
HIL Hardware-in-the-loop
AW As-welded
BG Burr grinding
TIG Tungsten inert gas
UIT Ultrasonic impact treatment
UP Ultrasonic peening
HiFIT High frequency impact treatment
HFP High frequency peening, common for UIT, UP and HiFIT etc.
IO Input/output
OCV Over-center valve
CAN Controller area network
LAN Local area network
Contents

1 Introduction 1
1.1 Højbjerg Maskinfabrik . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Loader Cranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Background and Motivation . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 Research Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 Outline of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Fatigue of Welded Joints 7

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Post Weld Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.1 Post Weld Treatment Techniques . . . . . . . . . . . . . . . 9
2.2.2 Implementation of Post Weld Treatment . . . . . . . . . . . 10
2.3 Fatigue Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Notch Stress Approach . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 Fatigue Resistance in the Notch Stress System . . . . . . . 12
2.3.3 Mild Notch Joints . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.4 Extension to Post Weld Treatment . . . . . . . . . . . . . . 15
2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
xiv Contents

3 Control of Mobile Hydraulics 19

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Tool Point Control . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.1 Real-time Simulation . . . . . . . . . . . . . . . . . . . . . . 21
3.2.2 Control Scheme . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Hardware-in-the-loop Simulation . . . . . . . . . . . . . . . . . . . 23
3.3.1 Loader Crane HIL Simulator . . . . . . . . . . . . . . . . . 24
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4 Conclusion 27

Bibliography 28

Appended Papers

A Comparison of Post Weld Treatment of High Strength Steel

Welded Joints in Medium Cycle Fatigue 35

B Re-analysis of Fatigue Data for Welded Joints Using the Notch

Stress Approach 51

C Experience with the Notch Stress Approach for Fatigue Assess-

ment of Welded Joints 65

D Developing a Tool Point Control Scheme for a Hydraulic Crane

Using Interactive Real-time Dynamic Simulation 79

E A Cost-effective Approach to Hardware-in-the-loop Simulation 99

 1
Introduction

This thesis concerns improvement of the fatigue and control performance of truck
mounted loader cranes. The project is initiated by Højbjerg Maskinfabrik A/S and
therefore takes its basis in the loader cranes produced by them.

1.1 Højbjerg Maskinfabrik

Højbjerg Maskinfabrik A/S (HMF) is a Danish manufacturer of hydraulic loader
cranes for truck mounting. HMF is a middle sized company with production in
Denmark and sales in most of the world, however, main markets are Europe and the
USA. Over some 60 years, HMF has grown out of a bicycle repair shop to become
one of the largest manufacturers of loader cranes in the world, HMF (2010).

The loader crane market is dominated by few, relatively large companies. Palfinger
is the market leader, followed by Hiab, and then Fassi, Effer, HMF and others in
varying order depending of the parameter used for ranking.

HMF is currently implementing a new business strategy and is focusing activity

on development and production of high-end loader cranes. Especially considering
welded high strength steel components, which is one of the core competences of
the company. This project was therefore initiated to help secure the position of
HMF as a leading brand and increase the competitiveness of their products.

1.2 Loader Cranes

Truck mounted loader cranes, as shown in 1.1, are primarily used for loading and
unloading building material, machinery, sand, scrap, etc. to/from trucks. Today
these cranes are complex welded structures made from ultra high strength steel
(in the range S700-S1300) in order to limit their self weight and maximize the
lifting capacity. Mobile cranes in general represent the leading edge in what can
be accomplished within the field of lightweight welded structures. Newly developed
2 1.3. Background and Motivation

ultra high strength weldable steel grades are even sometimes referred to as crane
steels by steel manufacturers.

Figure 1.1: Two examples of HMF loader cranes.

Due to increased safety demands, e.g. EN12999 (2010), a typical loader crane now
includes an advanced controller and a large number of sensors. A loader crane
thus represents a so-called mechatronic system with highly interacting mechanic-,
hydraulic- and electronic systems. The control system is very important, since it
governs the end-user’s daily experience of working with the crane.

Whereas loader cranes have reached a dominant design mechanically speaking,

i.e. most manufacturers apply almost identical principles, the control systems are
still evolving quite diversely. There is a tendency - both regarding cranes, but also
components - to convert mechanical/hydraulic features into electronic features.
This facilitates late product configuration, larger volumes, less stock, etc.

1.3 Background and Motivation

The business segment of truck mounted loader cranes represents a highly
competitive field, where a continuous increase in product performance is required.
The two most significant measures of performance are the ratio of lifting capacity
to self-weight and the operational speed.

A high lifting capacity is desired for obvious reasons, but the low self-weight is
motivated by several factors. Firstly, at longer reaches, most of the lifting capacity
is consumed by the self-weight moment of the crane. Furthermore, the self-weight
of the crane limits the load-carrying capacity of the truck and increases its fuel
Chapter 1. Introduction 3

consumption. The challenge is therefore to achieve the highest lifting capacity

while at the same time minimizing the self-weight. This is achieved using very
high strength steel.

The attainable operational speed, on the other hand, is directly proportional

to the amount of work the end-user can perform and thus his/her achievable
profit. However, increased speed often leads to premature fatigue failures, since the
dynamic loads on the crane increases. Therefore, when adjusting the operational
speed, it’s essential to consider and maintain the fatigue performance.

Using high strength steel in welded structures poses several challenges. The reason
for using higher strength steel is obviously that it allows higher loads to be
transferred using less material, i.e. higher stresses in reduced plate thickness.
Unfortunately, the fatigue strength of welded joints do not follow the static
strength of the base material. The difference between the fatigue strength and
static strength of a welded structure in high strength steel thus increase with the
base material strength. It therefore becomes difficult to fully utilize the high static
strength of the new steels, since the fatigue strength remains relatively low for
welded joints irregardless of the base material strength. This problem is one of the
main barriers preventing further use of ultra high strength steels in development
of durable, high strength, lightweight welded structures.

On the other hand, if the fatigue strength of the welded joints can be increased,
further utilization of the high strength steel can be achieved. Fatigue strength
improvement of welded joints is thus one of the keys to achieve higher performance
of loader cranes.

A significant implication of high strength steel design with reduced plate thickness,
is increased structural flexibility. Loader cranes with an operating range of up
to 40m operate under considerable deflection (several meters) and reducing the
plate thickness will only increase this deflection. It is thus expedient to introduce
more sophisticated control principles such as tool point control, both in order to
ensure safe high speed maneuverability and limit dynamic peak loading, in order
to maintain a high fatigue performance.

Put shortly, the fatigue resistance of the crane is governed by the critical welded
joints and the fatigue loading is governed by the hydraulic control system.
Therefore, when considering a dynamically loaded, fatigue prone welded structure
as a loader crane, the fatigue- and control aspects are interconnected.

1.4 Research Objectives

Since this is an industrial Ph.D. project, the objectives are twofold; in addition to
generating scientific progress, it must also consider the applicability in industry.
The research presented here will therefore be very application-minded and will
4 1.5. Research Approach

address the main challenges in loader crane development, as discussed above. The
research objective is therefore defined as follows:

The project shall establish a set of methods enabling development of loader cranes
with higher fatigue- and control performance than currently possible, i.e.

1. Determine how the fatigue strength can be improved by post weld treatment,
and how this improvement can be included in the fatigue assessment of the
complex welded parts comprised by a loader crane?
2. Determine how more sophisticated control principles, i.e. tool point control,
can be implemented in loader cranes considering development, implementa-
tion and testing?

1.5 Research Approach

Different approaches will be taken within the two sub-topics as explained in the
following.

Fatigue of welded joints

This work will focus on post weld treatment and local fatigue assessment. Existing
methods will be investigated and adapted to the specific conditions of loader cranes
after careful analysis. The following aspects characterize the conditions that needs
to be taken into account when considering loader cranes:

• Ultra high strength steels, e.g. S700, S900 and S1100

• Complex geometries and relatively thin sections
• Medium cycle fatigue life (20.000 to 500.000 cycles)
• Very high stress ranges, ∆σn > 500M P a

Firstly, different post weld treatments will be subjected to experimental investigation in

order to compare them against each other under the above mentioned conditions. The
treatments will also be analyzed qualitatively in order to determine the most suitable
treatment, considering practical implementation in mass production of loader cranes.

The commonly used nominal stress approach is not well-suited for fatigue assessment
of the complex geometries that characterize loader cranes. The effective notch stress
approach appears more suitable, however, the amount of published experimental evidence
of the reliability of this approach is very limited. A systematic re-analysis of fatigue data
in the notch stress system will therefore be carried out.

Lastly, in order to account for post weld treatment during development and fatigue
assessment of loader cranes, a proposal is made for extension of the notch stress approach
to include effects of TIG dressing, burr grinding and high frequency peening in a practical
manner.
Chapter 1. Introduction 5

Control of mobile hydraulics

Current loader cranes are typically controlled by electro-hydraulic actuation via remote
control. One axis on the remote controls one hydraulic cylinder. In order to move the tool
point (hook) in a controlled way, it is normally necessary to use several DOF of the crane
simultaneously. This is often a highly nonlinear task, which only experienced operators
can do fast, safe and accurate. Hence, there is a potential for increased operational speed
and productivity by implementing semi-automatic, intuitive tool point control.

Practical implementation of tool point control in loader cranes will therefore be addressed,
based on existing methods from robotics. Control of mobile hydraulic machinery is very
different from that of robotics. Among others, the following conditions complicate the
control of mobile hydraulics:

• Reference input assigned continuously by operator

• Significant structural flexibility (up to several meters)
• Highly varying and unknown loading
• Complex interconnected dynamics of the mechanical and hydraulic systems
• Multiple saturation phenomena in the hydraulic actuation

The current simple control scheme where each axis of the crane is controlled directly
and individually from the remote control, is not suited for optimizing the operational
speed. It is assumed that using tool point control instead, it will be possible to increase
the operational speed and reduce the effort and skill needed to safely operate the crane,
while maintaining fatigue performance. Such tool point control scheme will therefore be
investigated and further developed.

Implementing a new control principle in the existing controller requires thorough testing
before releasing the software. Current testing facilities were developed for simple control
principles and are not adequate for the more complex controller software required for
tool point control. A hardware-in-the-loop simulator will therefore be developed to satisfy
future controller software testing needs.

1.6 Outline of Thesis

Following this Introduction, two technical chapters presents the state of the art and
summarizes the work carried out within the project.

Chapter 2 concerns fatigue of welded joints and focus on post weld treatment and local
fatigue assessment using the notch stress approach.

Chapter 3 is on control of mobile hydraulics, considering tool point control of loader

cranes and controller software testing using Hardware-in-the-loop simulation.

The conclusion in Chapter 4 reviews the main contributions from the present work and
gives suggestions for future work.
6 1.6. Outline of Thesis
 2
Fatigue of Welded Joints

2.1 Introduction
It is a well known fact that welded joints have a low fatigue strength compared to the
base material. This is mainly caused by local stress concentrations due to the presence
of notches and high tensile residual stresses due to contraction of the weld metal during
cooling and solidification. It is one of the dominating causes of failure in welded structures
and has thus received much scientific attention, Maddox (1991, 2003).

The fatigue strength of welded joints proves to be practically independent of the base
material strength, Fig. 2.1. Hence, when applying higher strength steels, the gap between
the static strength and fatigue strength increases. It is therefore expedient to improve
the fatigue strength of the welded joints, in order to fully utilize higher strength steel
when designing lightweight structures.

400 500
Stress range for life of

300
Stress range, MPa

10 cycles, MPa

200 400

100 300

50 200
6

100
10
5
10 106
10 7
10 8 400 600 800
cycles ultimate tensile strength , MPa

Figure 2.1: the low fatigue strength of welded joints is independent from the base
material strength, after Haagensen (1985).

Increasing the fatigue strength of welded joints can be achieved in several ways. This
work mainly concerns improvement by design through local fatigue assessment and
the application of post weld treatment. This chapter presents the state of the art
and summarizes the contributions of this work within these areas. For details, see the
appended Papers A, B and C.
8 2.2. Post Weld Treatment

2.2 Post Weld Treatment

Post weld treatment has typically been applied in repair or life extension of existing
welded structures, but it is getting more and more interesting for application to new high
strength steel structures.

Due to high tensile residual stresses and abrupt geometry changes in the weld toe, the
weld will experience an increased and elevated stress range in the as-welded condition,
as shown in Fig. 2.2(a). The stress ratio of the applied loading is to a large extend
insignificant to the local stress ratio, due to the high tensile residual stresses.

ı ı ıy ı ıy ı ıy

ı
a) As-welded b) Geometry c) Residual stress
condition improvement state improvement

Figure 2.2: Local effects of post weld treatment.

Improving the geometry of the weld toe leads to a reduction of the stress concentration
here. The weld will therefore experience a reduced stress range, as shown in Fig. 2.2(b).
Geometry improvement considers both reduction of the macroscopic stress concentration
due to the section change of the joint, but also reduction or removal of stress
concentrations on the microscopic level due to weld defects.

By improving the residual stress state, the local stress ratio is reduced, so that all or part
of the local stress range becomes compressive, Fig. 2.2(c). Local compressive loading
significantly reduces the crack growth rate.

Generally, geometry improvement can be considered more robust than residual stress
state improvement, in the low cycle/high stress area, since the latter improvement can
be destroyed by overloading. On the other hand, residual stress state improvement tends
to be more effective under the right conditions, i.e. high cycle/low stress range and low
stress ratio.

Improvement of the fatigue strength of welded joints by application of post weld treatment
has received much attention, especially in connection with high strength steel. Gurney
(1979) mentions a variety of post-weld improvement methods. Since then, several major
reviews of different post weld treatment methods have been published, e.g. Booth
(1983), Haagensen (1985), Kirkhope et al. (1999) and Haagensen (2007). The IIW
provides recommendations for post weld treatment considering grinding, TIG dressing
and hammer/needle peening, Haagensen and Maddox (2008).

Most recent research on post-weld treatment includes a relatively new method – high
frequency peening, e.g. ultrasonic impact treatment (UIT), because of very impressive
Chapter 2. Fatigue of Welded Joints 9

results. A number of Ph.D. theses have also been written on the subject, including Roy
(2006), Lihavainen (2006), Dürr (2007) and Weich (2008).

A large European research project, Lagerqvist et al. (2007) investigated improvement of

welded joints in S700, S960 and S1100, considering burr grinding and TIG dressing.
Samuelsson (2007) reports on a Nordic research project, which investigates fatigue
improvement of S355, S700, S900 and S1100, including UIT, LTT electrodes and grinding.
Results show a tendency, that the effectiveness of post weld treatment increases with
increased base material strength, however not fully consistent.

This work investigates three post weld treatments in the medium cycle area, under very
high stress ranges. The treatments will be subjected to both experimental investigation
and qualitative analysis, to determine the most effective and suitable treatment for mass
production of loader cranes.

2.2.1 Post Weld Treatment Techniques

A selection of the most widely applied post weld treatment techniques is illustrated in
Fig. 2.3, divided according to their working principle. A large number of techniques exist,
both generic, such as grinding, and proprietary, e.g. ultrasonic impact treatment (UIT).
Most peening techniques also provide some degree of geometry improvement, but their
primary improvement is related to residual stress state improvement.

RESIDUAL STRESS STATE IMPROVEMENT GEOMETRY IMPROVEMENT

• Shot peening/blasting • Hammer peening • Disc grinding

• LTT electrodes • Needle peening • Burr grinding
• Thermal stress relieving • High frequency peening • TIG dressing
(UIT, UP, HiFIT, etc.)

Figure 2.3: Selected improvement techniques.

In this project, three of the most common techniques are investigated; burr grinding,
TIG dressing and ultrasonic impact treatment. The main objective is to determine the
most suitable treatment for loader crane components, considering both its effectiveness
and implementation issues. Fig. 2.4 shows the resulting weld toe profiles due to the three
improvement techniques.

Figure 2.4: Micrographs of weld toes obtained by post-weld treatment, Paper A.

10 2.2. Post Weld Treatment

The level of improvement due to the investigated techniques will be discussed in the next
section using the notch stress approach. This allows for comparing more data, since the
effect of different joint geometries can be disregarded to some extend. Paper A gives the
details of the current experimental work and Paper C presents a re-analysis of published
fatigue data from investigations of post weld treatment.

Its is clear from the presented data, that the geometry improvement techniques, burr
grinding and TIG dressing supports parallel shifting of the SN curve upwards. How much
can always be debated, but here conservative suggestions are given, i.e. an approximate
30% increase in fatigue strength.

For high frequency peening treatment, the data supports rotating the SN curve, in
addition to parallel shifting it upwards. This technique thus appears to be much more
effective in the high cycle range, as opposed to the low/medium cycle area. Still Wiedner
et al. (2010), and the author (Paper A) found the UIT treatment to give significant
improvements in the low/medium cycle area, at least under bending loading.

2.2.2 Implementation of Post Weld Treatment

Post weld treatment is sometimes considered an expensive solution, but in many cases,
only a very small part of a component is stressed to such extent, that post weld treatment
is necessary. Thus, applying post weld treatment locally - where needed - can actually
be the cheaper solution, as opposed to increasing the plate thickness in order to reduce
stresses.

As an example, Fig. 2.5 shows a welded crane component, the main part of which consists
of a hexagonal beam-like profile made from high strength steel. During service, it is
subjected to very high cyclic loading, and in order to ensure the required fatigue life,
the designers have two options; 1) post weld treatment of the highly stressed area or 2)
increasing the plate thickness of the beam-like part.

Figure 2.5: Welded crane component locally subjected to TIG dressing.

The designated weld seam (red) is subjected to TIG dressing, in order to increase its
fatigue strength to a level closer to the remainder of the component. In this case, TIG
Chapter 2. Fatigue of Welded Joints 11

dressing of approximately 500mm of weld seam is clearly both the cheaper and better
solution, since it will allow a much lighter and thus more competitive product.

However, implementing e.g. TIG dressing in production takes time and effort. Instructions
were prepared and a short course was given for the operators and quality control
personnel. A new weld quality class was adopted for specification of post weld treatment
in technical drawings, similar to VB in the new Volvo CE weld quality system, described
by Jonsson and Samuelsson (2008). Fig. 2.6 (left) shows the initial TIG dressing
quality achieved, which featured both deep undercuts and slag inclusions. After further
instructions and more practice, however, a consistent high quality TIG dressing were
achieved, Fig. 2.6 (right).

Figure 2.6: Initial quality problems (left) and final quality TIG dressing (right),
Paper A.

2.3 Fatigue Assessment

Many approaches exist for fatigue assessment of welded joints. They are typically
categorized into nominal stress, structural stress, notch stress, notch strain and fracture
mechanical approaches, Radaj et al. (2006). The IIW continuously monitors the state of
the art and derives recommendations accordingly, Hobbacher (2008).

Traditionally, the nominal stress approach has been applied for fatigue assessment, but
more sophisticated approaches are now more widely used due to the increase in available
computational power. These so-called local approaches correlate local stresses in the
vicinity of the weld toe (or root), typically obtained using finite element analysis (FEA),
with the fatigue life, Fricke (2003), Nykänen et al. (2007) and Radaj et al. (2009).

The fatigue performance of a welded detail can then be improved during the design stage
by optimizing the joint geometry, including the weld shape, using local approaches.

In this work, the notch stress approach has been chosen for further investigation and
development, since it is found to be the most suitable approach for assessing the
welded components of loader cranes. This is in regard to the size and complexity of
the components, and it fits well into the current design approach using FEA.
12 2.3. Fatigue Assessment

2.3.1 Notch Stress Approach

The notch stress approach has become increasingly attractive due to the available
computational power and the need for assessing more complex geometries, see e.g.
Barsoum (2008), Fricke and Feltz (2009) and Fukuoka and Mochizuki (2010). In this
work, the notch stress approach is investigated regarding the reliability of the claimed
fatigue resistance, with special attention to mild notch joints. Furthermore, a proposal
for extending the notch stress approach to post weld treated details is presented.

The notch stress approach have evolved over some 40 years, as described by Radaj et al.
(2006) and different sub-approaches exist. In this work, we consider the notch stress
approach according to the IIW, Fricke (2008). The approach is based on the early work
by Radaj (1990), considering Neuber’s fictitious notch rounding concept, but modified
as described by Olivier et al. (1989, 1994). The reference radius of R1 is determined
as a mean value of measurements and the fatigue strength (FAT225) is derived from
experimental data.

The basic work flow of the notch stress approach is shown in Fig. 2.7. The weld is included
in a finite element model and rounded in the toe or root by a reference radius R1 (the
notch). Here, a very fine mesh is applied in order to ensure convergence of the local
stresses. Then, the 1st principal stress range (or other stress components) is determined
in the notch and assessed against a SN curve.

Reference Elements Principal ∆σ

radius R1 <0.25mm stresses

FAT225

Nf

Figure 2.7: Work flow of the notch stress approach, Paper C.

2.3.2 Fatigue Resistance in the Notch Stress System

The IIW recommends the use of FAT225 when considering the fatigue strength of welded
steel joints using the notch stress approach and the principal stress hypothesis. This value
was derived from the results of a large experimental investigation of different T- and Y-
joints Olivier et al. (1994) and further re-analysis of the data assembled by Olivier and
Ritter (1979), as shown in Fig. 2.8.

Kranz and Sonsino (2009) explain the assumptions that led to the derivation of the
FAT225 fatigue class in the IIW recommendations: The average mean local stress range
(R = 0, PS = 50%) in Fig. 2.8 (347MPa) is used as a starting point and recalculated to
fit the IIW system. This means dividing by fR = 1.10 to account for high tensile residual
stresses (R = 0.5) and dividing by jσ = 1.37 to account for a probability of survival
of PS = 97.7%, which leads to a value of FAT231. IIW uses a fixed spacing of 12.5%
between FAT classes, so the value is rounded down to FAT225 to fit the IIW numerical
system.
Chapter 2. Fatigue of Welded Joints 13

Welded joint type Fatigue notch Global fatigue Local fatigue strength
(structural steels ) factor K f strength (R=0) ∆σkE=2KfσnA [N/mm2]
σnA [N/mm2], PS=

360
400
440
480
240

280
320
Fracture initation 10% 50% 90%
∠30°: 1.89 294 log
∠45°: 2.27 61 78 99
Butt joint Weld toe 354

As-welded, N=2 x 106, R=0

2.45
52 69 91 338
Transverse stiffener Weld toe

2.50
54 67 83
Weld toe 336
Cruciform joint
3.20
47 55 65
344
Overlap joint Weld toe
4.03
32 43 57
Weld root 346
Cruciform joint

362
Results for crane As-welded
constructions Stress relieved
352
T + Y joints

Figure 2.8: Fatigue strength in the notch stress system, after Olivier et al. (1994).

The global fatigue strength in Fig. 2.8 is derived from a very extensive database
fatigue test results, containing several thousand individual results and must therefore
be considered very reliable.

The fatigue notch factors for each joint type in Fig. 2.8 is determined in general for
a thickness of 10mm and a flank angle of 45◦ , see Olivier et al. (1994). The fatigue
notch factor of the butt joints was initially determined for a flank angle of 30◦ , but then
increased by assuming 45◦ in order to account for generally larger flank angles. However,
the IIW recommendations suggest the use of a flank angle of 30◦ , Hobbacher (2008).

This approach of applying a single general fatigue notch factor for each joint type can be
problematic since the fatigue notch factor varies significantly with the dimensions of the
joint, especially the thickness, as shown in Fig. 2.10.

In Paper B, a systematic re-analysis of fatigue data is carried out with more accurate
determination of the fatigue notch factor for the individual specimens. Fig. 2.9 shows
some of the results, i.e. the same fatigue data for transverse stiffeners in the nominal and
notch stress systems, respectively.

When comparing the data in the two systems, it is seen, that a larger portion of the
data in the notch stress system lies below the recommended SN curve. It is thus found
14 2.3. Fatigue Assessment

that the notch stress approach can be slightly less conservative than the nominal stress
approach.

Nominal stress system Notch stress system

700 1500
600 R≥ 0, n=265 Lif1 R ≥0, n=265 Lif1
500 Fos1 Fos1
400 Fos2 1000 Fos2
Fos3 Fos3
300 Son1 700 Son1

[MPa]
Man1 Man1
∆σ [MPa]

200 Oht2 500

→ Oht2
→ Muc1 →
→ →→ 400 Muc1
→ → →

k PS
→→ → →
Hut2
n

Hut2
→ →→ → →
→

,
Kud1 300 → Kud1
→ → →

∆σ
100
80 225 →
→
→
→→ 200 →
→ →
50 →
→
→ →
30 4 5 6 7 8
100 4 5 6 7 8
10 10 10 10 10 10 10 10 10 10
Nf [cycles] Nf [cycles]

Figure 2.9: Fatigue data in the nominal and notch stress system, Paper B.

2.3.3 Mild Notch Joints

It is a known issue, that the notch stress approach can lead to non-conservative
assessments in connection with mild notch joints, and Fricke (2008) therefore proposes
a minimum notch factor of Kw = 1.6 as a preliminary approach. This solution is not
verified and hence there is a need for better understanding of the issue.

Paper B illustrates the problem of non-conservative assessment of butt joints - especially

thinner ones - when using the geometry idealization proposed by the IIW. Here, a remedial
action was proposed, i.e. requiring a minimum notch factor of Kw = 2.0 instead of the
current 1.6.

An alternative could be to model the weld with a flank angle of 45◦ instead of the 30◦
suggested in the IIW recommendations. Using 2D finite element analysis, the fatigue
notch factor of a butt joints can easily be determined for different thicknesses as shown
in Fig. 2.10. It’s clear that using a 45◦ flank angle would be a much more conservative
approach.

Fig. 2.10 furthermore shows a line indicating full correspondence between the nominal
and notch stress approaches, i.e. the notch factor, that will force the notch stress approach
to give the same result as the nominal stress approach. Since the nominal stress approach
does not include any thickness effects in this range, the line is horizontal. This notch factor
is determined from the characteristic fatigue strength at 2·106 cycles for the two systems,
including a stress magnification factor km to account for misalignment in the notch stress
system.

∆σc,k 225M P a
Kf = = = 2.27 (2.1)
∆σc,n · km 90M P a · 1.1
Chapter 2. Fatigue of Welded Joints 15

Kf Kf [-]
R1 30°/45°
F t F

factor,
2.5

e n otch fa ctor,
Correspondence between the
concentration nominal and notch
0.216stress approaches
K = 1.94⋅t -1.11
2 f,45
StressFatigu

30° FEM
1.5 30° Fit
K = 0.161⋅t0.566+1.07 45° FEM
f,30
45° Fit
1
5 10 15 20 25
Thickness, t [mm]

Figure 2.10: Notch factor of butt joints with different flank angles.

Using the notch stress approach, the estimated fatigue strength is proportional to the
fatigue notch factor and the approach will clearly lead to quite different results than the
nominal stress approach, depending on the thickness.

In Paper B, the thickness effect of butt joints is briefly investigated, and it is found that
experimental data suggests in favor of the nominal stress approach, i.e. not assuming any
thickness dependency in this range (5-25mm). Ohta et al. (1990) also found the thickness
effect of butt joints to be insignificant in this range. It thus appears, that the implicit
thickness effect of the notch stress approach is too sensitive to the thickness variation in
this range, regardless of the chosen flank angle.

In Paper C, a welded crane detail is investigated using the notch stress approach,
especially considering the flank angle. Using the measured flank angle of approximately
15◦ yields non-conservative results. If applying an idealized flank angle of 45◦ , instead
of the measured one, much safer results are obtained. The experimental investigation of
the crane detail was carried by Rasmussen (2008) in corporation with the author.

Sonsino et al. (2009) also report problems with fatigue assessment using the notch stress
approach considering thin/flexible welded joints. They observed shallower slopes of the
SN curves for these particular joints and therefore suggest the use of m = 5.0, while
maintaining the FAT225 value. This solution seems to solve the problem, but it appears
much too conservative in the low-to-medium cycle area, and the compatibility between
the nominal and notch stress approaches will be lost.

2.3.4 Extension to Post Weld Treatment

Applying the notch stress approach to post weld treated joints is a subject still under
debate. According to the IIW guideline, Fricke (2008), it is suggested to use the actual
weld toe radius, i.e. in the treated condition, increased by 1mm and evaluate the principal
stress range here against the FAT200 SN curve. However, it is also stated, that this
16 2.3. Fatigue Assessment

approach is only valid for relatively sharp notches (R1 − R3mm) and that it has not yet
been verified.

A problem in this approach is, that it is difficult to determine the actual weld toe radius
in a design situation. As shown in Fig. 2.11, the toe radius varies significantly from test
series to test series and is assumed to follow a normal distribution within each test series.
An unanswered question therefore remains; how to define the actual radius? Should it
be the mean or minimum or something else?

As welded Burr ground TIG dressed HFP treated

2
Probability density

[15] [13]
[14][20] [14]
[19] [13]
1 [14]
[18] [18] [17] [13]
[21][22] [17] [20] [21]
[18][19]
[16][17]
0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Toe radius, R [mm]

Figure 2.11: Measurements of weld toe radii, Paper C.

With the design situation in mind, a new proposal for fatigue assessment of post weld
treated details using the notch stress approach is presented in Paper C. Instead of altering
the idealized notch geometry (radius), it is proposed to use different SN curves for
different post weld treatments, and thus maintain the reference geometry (R1mm). Then,
a design engineer can use the same FE model to assess the fatigue performance - both in
the as-welded and post weld treated states.

In order to establish SN curves for post weld treated details in the notch stress system,
a large amount of fatigue data is collected and transformed into the notch stress system.
The transformation is carried out by scaling the nominal stress range with the fatigue
notch factor determined for each specimen using the standard idealized geometry, i.e.
identical to the as-welded case.

Fig. 2.12 shows the collection of fatigue data for TIG dressed welded details in the notch
stress system. Statistical analysis of these data collections is avoided, and a conservative
SN curve is suggested, e.g. FAT300 for TIG dressed joints. Paper C furthermore presents
suggested SN curves for burr ground and high frequency peened joints. For both burr
grinding and TIG dressing, FAT300 and m = 3.0 is suggested, whereas for high frequency
peening, FAT360 and m = 5.0 fits the data collection well.

This approach to some extend ignores the original intention of the notch stress approach,
that the stress range should be determined from the real (idealized) local geometry.
However, it is very practical and it allows the nominal, structural and notch stress
approaches to be compatible regarding post weld treated details, as is the case for as-
welded details in the IIW recommendations.
Chapter 2. Fatigue of Welded Joints 17

Notch stresses

1500 R ≥ 0, n=237 Haa1

Sta1
Ped1
Lif1
→
1000 → →
Fos1
Son1
→ Man1
m=3 →
→ →
700 → →→
→
Lif3

∆σk,PS [MPa] → →→ Lif4

→
→ →
→ →
Haa2
500 →→ Hut1
→
→ Wan1
400 →
→
→
→
Dah1
Lop1
300 300
→ →
→
225 →
200 TIG dres
sed

TIG dressed As-weld

ed
r ref=1mm
100 4 5 6 7 8
10 10 10 10 10
N f [cycles]

Figure 2.12: Fatigue data for TIG dressing in the notch stress system, Paper C.

2.4 Summary
This chapter have presented the work carried out in regards to fatigue of welded
joints with the aim to provide methods for the development of durable, high strength,
lightweight loader cranes. The following topics are covered:

• Evaluation of the three post weld treatments; burr grinding, TIG dressing and UIT
treatment by experimental investigation and qualitative analysis.
• Investigation of the fatigue resistance used in the notch stress approach based on
re-analysis of a large amount of published fatigue data.
• Investigation of the problems involved when applying the notch stress approach
for fatigue assessment of mild notch joints.
• Established a proposal for extension of the notch stress approach regarding fatigue
assessment of post weld treated details.
18 2.4. Summary
 3
Control of Mobile Hydraulics

3.1 Introduction
Earlier loader cranes had purely mechanical/hydraulic safety and control systems.
Typically, the operator controlled the flow to the cylinders directly by mechanical handles
linked to the hydraulic directional control valves and the lifting capacity was limited by
hydraulic devices.

Today however, advanced loader cranes are typically operated by a wireless remote control
which sends a flow reference to an onboard controller, which then evaluates the input
and, if acceptable, forwards the flow command to electro-hydraulic actuated directional
control valves.

From the control signal is generated in the operator interface, i.e. the remote control,
until it arrives in the valve, it passes several electronic modules, each with increasingly
advanced programmable micro-controllers. Additionally, much communication in vehicle
systems nowadays are transmitted by digital networks, mainly CANbus and the use of
analogue components are declining in favor of network attached components.

Using such complex systems allows for advanced control features, such as selective
stopping of movements and smart speed reduction depending on operating conditions.
And with the increased amount of sensors mounted and digitally controlled valves, it is
possible to implement very advanced control principles.

But when complexity increases, so does the risk of errors, and finding these errors becomes
more and more difficult, since the IO interface of the controller now includes network
busses, and not only A/D voltage signals. It is thus difficult to provide the stimuli to test
controller software without connecting it to the real machine.

In general, working with control of loader cranes using the real machine, it is neither safe
nor efficient. Model based development is therefore preferred and will be used extensively
in this work.

Within control of mobile hydraulics, this work concerns two topics; tool point control
and hardware-in-the-loop simulation, see the appended Papers D and E, respectively.
20 3.2. Tool Point Control

3.2 Tool Point Control

In this work, tool point control refers to a semi-automatic control scheme, assisting the
operator in smooth, efficient and intuitive operation of the crane. The operator controls
the speed and direction of the tool point (hook) of the crane directly, instead of the
individual actuators, as is currently the case, Fig. 3.1. The main advantages of tool point
control is the reduced skill level required and increased productivity.

Tool point control of mobile hydraulic manipulators has been subjected to research for
several years. Krus and Palmberg (1992) presented a simple vector control strategy for a
2 DOF hydraulic crane in the early nineties. Mattila and Virvalo (2000) describe a more
advanced control scheme for a similar crane; where an online model is used to reduce the
pressure levels and thus the energy consumption. Beiner (1997) solves the redundancy
in a 3 DOF hydraulic crane using the minimum norm of the actuator forces. Recently,
Yuan et al. (2009) presented a motion control for a 4 DOF aerial lift. An interactive
real time simulation was achieved by Esqué et al. (2003) for a simplified 2 DOF crane.
Tool point control of mobile hydraulic manipulators has been a key area of research at
Aalborg University for several years, see Pedersen and Nielsen (2002), Münzer (2003),
Hansen and Andersen (2005), Ebbesen (2007) and Kabus and Haastrup (2008).

The typical control scheme is based on open-loop velocity control of the crane, where
the operator ‘closes the loop’, i.e. he/she compensates for deviations in the tool point
trajectory using visual feedback. However, this concept needs to be verified, especially
considering the operator perception of e.g. saturation handling or joint limit avoidance.
In order to secure the usefulness and commercial value of a tool-point control scheme
it is important to have some kind of user-interaction at the initial stages. This work
therefore presents an approach to tool point control development for loader cranes, using
interactive real-time dynamic simulation.

x
z
3

2
Current control 1 2 3 4

Tool-point control x y z

Figure 3.1: Tool point control principle.

Chapter 3. Control of Mobile Hydraulics 21

3.2.1 Real-time Simulation

It is expedient to apply real-time simulation here, since the crane will be controlled online
by an operator, and his influence on the control scheme is difficult to include otherwise.

A simulation model is therefore developed in Simulink using the 3D animation toolbox.

It takes input from the operator continuously via joystick, calculates the resulting motion
and displays it in the form of a 3D animation. As shown in Fig. 3.2, the model consists
of a hydraulic model and a mechanical model, which together forms a virtual crane.

u u Virtual crane
θ Fc Fc
θ θ θ ∫ ∫ θ
Joystick Hydraulic θ
input model Mechanical 3D
model animation

Figure 3.2: Overview of the simulation, Paper D.

Achieving real-time simulation performance relies on the following two Simulink concepts;
rate-transition blocks and embedded Matlab blocks. Rate-transition blocks allows for part
of the simulation to run faster than the rest, e.g. the computationally heavy animation
update is only performed at 10Hz. Embedded Matlab blocks contain Matlab script code
that is automatically converted to C code and compiled before running the Simulink
model and thus executes very fast. Using the fixed step 4th order Runge-Kutta integrator
(ODE4) and a time step of 2.5ms yields real-time execution on a standard 2GHz laptop
computer for the simulation model presented here.

In the mechanical model, the equations of motion is established in the joint space, in
order to achieve the smallest possible system matrix for inversion. The mass and inertia
is furthermore lumped into few bodies and the telescopic extension system is modeled as
single body with variable inertia properties depending on the extension length. In total,
the model includes 6 DOF; 4 hydraulically actuated and 2 passive, which models the
in-plane and lateral flexibility.

The hydraulic system is modeled using the pressure build-up equations and thus takes
into account the flexibility of the oil in hoses etc. Loader cranes are equipped with over-
center valves (OCVs), which typically complicates simulation due to very fast dynamics.
Here, these are modeled according to a simple discharge characteristic which includes the
steady-state influence of the OCV, but neglects the dynamics of the OCV spool.

This model enabled investigation of the traditional tool point control approach for loader
cranes by interactively testing it on the virtual crane in real-time simulation. Several
issues where identified and the control scheme were further developed, as described in
the following section.
22 3.2. Tool Point Control

3.2.2 Control Scheme

The purpose of the tool point control scheme is to make it more intuitive for the operator
to control the crane and avoid potential excessive or abrupt motion, thus lowering the
dynamic loading on the crane even at increased velocities.

Since the crane is redundant, the Jacobian is not square and thus cannot be directly
inverted. A pseudo inverse approach is therefore used to find a set of joint velocities, that
will produce the desired tool point motion. The pseudo inverse returns the minimum
velocity norm in the joint space, as discussed by Beiner and Mattila (1999):

ṙT P = Jθ̇ (3.1)

The pseudo inverse is then

θ̇ = J+ ṙT P + (I − J+ J)z (3.2)

The first term is a set of joint velocities that will produce the desired tool point velocity.
The second term describes a set of joint velocities which have no influence on the tool
point, but can be used to control the so-called self-motion of the crane. Both of them are
controlled by the operator.

The pseudo inverse is furthermore weighted in order to implement joint limit avoidance
and configuration control. Joint limit avoidance is implemented according to Chan and
Dubey (1993), but the weighting matrix is extended in order to include automatic
optimization of the torque arm of the main lift cylinder. The configuration control thus
comprise both automatic optimization of the main joint lifting capacity, but also active
control by the operator allowing him/her to control the self-motion directly e.g. to avoid
obstacles.

The reference velocities determined above do not account for the significant flexibility of
loader cranes and a straight horizontal tool point trajectory will be difficult to achieve. A
deflection compensation scheme is therefore developed, inspired by the simple approach
presented by Yuan et al. (2009). Here a small upwards velocity is continuously added to
the tool point reference velocity, when the crane telescopic system is extending.

Flow sharing is necessary in order to avoid both the valve- and pump flow limits, otherwise
the reference joint velocities, and thus the desired tool point velocity, will not be realized.
A flow sharing scheme is therefore implemented, similar to that used by Pedersen and
Nielsen (2002), which scales down the reference velocity of all actuators if one reaches
saturation. Hence only the tool point speed is violated, but the direction is maintained.

During the development of this control scheme, the interactive real-time simulation model
proved to be indispensable. As an example, Fig. 3.3 shows the difference between a
predefined operator model without any feedback and an active operator using visual
feedback to adjust the input. It is clear that, when the operator capabilities are included
in real-time simulation, the control performance is excellent.
Chapter 3. Control of Mobile Hydraulics 23

Tool point trajectory

6
Reference
5.5 Predefined operator
5 Active operator

4.5

y [m]
3.5

2.5

1.5
4 5 6 7 8
x [m]

Figure 3.3: Influence of the active operator, Paper D.

3.3 Hardware-in-the-loop Simulation

Hardware-in-the-loop (HIL) simulation is a method for testing controller software,
without using the actual machine for which the controller is intended. Here, the machine
is simulated in real-time in a computer and the controller IO interface is connected to
the simulation.

Previously, testing were typically performed using the real machine, however, it is time
consuming and difficult to make accurate and repeatable tests this way. Recent advances
in computer hardware and software, have made it possible to implement very cost-effective
HIL simulation platforms for such machinery, and thus significantly improving testing
capabilities.

Early HIL simulation (HILS) platforms typically considered a single component or minor
isolated systems for testing controller software with limited visual output, e.g. Linjama
et al. (2000). However, advances in computational power have propelled a rapid evolution
of HILS platforms which now considers entire systems, such as heavy construction
machinery or offshore structures and offer rich 3D animation output.

A HILS system for a hydraulic excavator with both audio and visual simulation output
integrated in the operator cabin of a real excavator, was presented by Elton et al. (2009).
The system included a remote hydraulic load emulator for a variable displacement pump,
connected via Internet. The simulation are thus fed accurate (non-modeled) information
from the pump.

Ghabcheloo and Hyvönen (2009) discusses development of a motion control system

for a hydraulic wheel loader in a HILS platform called GIMsim. It consists of 5 PCs
communicating via LAN to accommodate animation, dynamics, tire contact, IO including
CAN and simulation control. The simulation is developed in Matlab/Simulink, compiled
and transfered to an xPCtarget for hard real-time operational speed.
24 3.3. Hardware-in-the-loop Simulation

Recently, Trydal (2010) presented a very advanced proprietary HILS platform from
National Oilwell Varco. It concerns an offshore drill-rig including the entire drilling
process equipment and displays a 3D animation in a dome in order to completely immerse
the operators. Instead of physical IOs, each controller are connected to a similar controller
which simulates the IO functionality.

The general trend is to develop the simulation in Matlab/Simulink, use several PCs
connected via LAN and use hardware IO connectivity from National Instruments, dSpace
or others. Simulation input are given from real operator interfaces and results are
presented using more and more realistic 3D animations. The animation software is often
developed specifically for the considered system e.g. in C++ using OpenGL.

In this work, a cost effective approach to HIL simulation is demonstrated using only
standard high-level engineering software. A truck and loader crane with a total of 21
DOFs is simulated with less than 10.000$ in total expenses for hardware and software.

3.3.1 Loader Crane HIL Simulator

The HIL simulator presented in this work is developed for the specific controller currently
used on HMF loader cranes, however, the same approach should be applicable for similar
machinery. The HIL simulator consists of a simulation PC running Matlab which is
connected to the CAN network including the controller through another PC running
LabView, Fig. 3.4. The two PCs communicate via LAN over the UDP/IP protocol, and
are denoted Hardware Gateway and Simulation PC, respectively.

LAN communication using the UDP protocol is an efficient approach to achieve co-
simulation with two different programs. Using several PCs gives a performance advantage
and enables additional monitoring possibilities and remote location of hardware. It is thus
possible to take advantage of the simulation features of Matlab and at the same time use
the excellent hardware connectivity features of LabView - in the same simulation.

Hardware Hardware Gateway PC Simulation PC

Simulated CAN hardware LabView Matlab
CAN loop Initialization

Simulation loop
Get inputs
Slew Sensor Inclinometer Valve actuation LAN loop (UDP) & Hydraulics model
Simulation control Kinematic model
Animation data
CANbus Time integration
Send outputs

Animation
Remote Fly-jib I/O Winch I/O Stab. I/O Controller
IO IO IO IO IO
A/D loop
A/D IO devices

Figure 3.4: Overview of the HIL simulation, Paper E.

Chapter 3. Control of Mobile Hydraulics 25

The Hardware Gateway runs the low-level communication with the controller hardware
and emulates a number of CAN devices, i.e. virtual sensors and virtual hydraulic valves.
As shown in Fig. 3.4, the Labview program is divided in three separate loops; a LAN
loop, CAN loop and A/D loop. Each loop covers communication between the gateway
and LAN network, CAN network or analog/digital hardware, respectively.

The simulation contains a full kinematic model of the truck and crane, including a number
of crane accessories, and a number of virtual sensors, which generates the necessary
input to the controller. The simulation is driven by the flow commands send from the
controller to the virtual hydraulic valves and a number of output signals. The results of
the kinematic analysis is presented as a 3D animation.

Using this HIL simulator, it has become possible to perform much more precise and
repeatable testing of the current controller software, and a platform is established for
development and testing of the next generation software.

3.4 Summary
This chapter presented the work carried out in this project regarding improvement of
control performance of loader cranes, including the following topics:

• An approach to interactive, real-time dynamic simulation of loader cranes is

established, with the aim to develop more advanced control principles.
• A tool point control scheme have been investigated and further developed in order
to adapt it to current HMF loader cranes, and verify its effectiveness for this
application.
• A demonstration of a cost effective approach to HIL simulation is presented with
the ability to simulate all necessary IO functionality in the current controller of
HMF cranes.
26 3.4. Summary
 4
Conclusion

This work spans two distinct topics in mechanical engineering; fatigue of welded joints
and control of mobile hydraulic systems. It was planned to devote more effort to
implementation of the tool point control strategy in a real crane using a specific newly
developed hydraulic servo valve from an undisclosed manufacturer. Unfortunately, the
final development of this valve was delayed and it was not ready in due time to be part
of this project.

The original plan of evaluating the fatigue performance of a crane with tool point control
against one with traditional control was therefore discarded. Other topics in the project
were then expanded and several contributions within each field are made.

The main contribution of this project is a set of tools enabling development of high
performance lightweight structures, such as loader cranes with increased fatigue- and
control performance, as compared to current standards. Due to the nature of this project,
the contributions presented are all directly applicable in industrial context, e.g. in the
sponsor company Højbjerg Maskinfabrik A/S.

Contributions
Regarding fatigue of welded joints, this project have contributed with:

• An evaluation of the post weld treatments; TIG dressing, burr grinding and high
frequency peening under the specific conditions of loader cranes, including the
aspects of practical implementation in mass production.
• A better understanding of the reliability involved in fatigue assessment of welded
joints using the notch stress approach, including the effect of mild-notch joints.
• A proposal for extension of the notch stress approach to include the post weld
treatments; TIG dressing, burr grinding and high frequency peening treatment in
an efficient and practical manner.

Within control of mobile hydraulics, this project have contributed with:

• An approach to real-time interactive simulation of loader cranes, including the

dynamics of the hydraulic and mechanical systems.
28

• A validation of a tool point control scheme with an operator in the loop, i.e. showing
the positive effect of operator visual feedback.
• An approach to cost-effective hardware-in-the-loop simulation for loader cranes
using only standard engineering software.

Recommendations for Further Work

A subject within post weld treatment of high strength steel, that has only been briefly
touched upon here, and in general, is the effect of high frequency peening treatment, e.g.
UIT, in the low to medium cycle fatigue area and the associated high stress ranges. There
is a need for thorough investigation of the stability of the treatment in this area, due to
the risk of relaxation of the compressive residual stresses introduced by the treatment.

In regards to fatigue assessment using the notch stress approach, there is a need for better
understanding of the effects of mild notch joints. Since the notch stress approach is an
ideal choice for optimizing the local geometry of a welded joint, including the weld shape,
it will often be applied in a context where the designer works towards mild notches. The
approach is thus intentionally used in the direction of the limit of applicability – which
is not well defined.

The implementation of tool point control in loader cranes in industrial context opens
up for a number of research topics. While much research – including this work – have
demonstrated the benefits of tool point control of mobile hydraulic manipulators, very
few actual implementations are found in a commercial context. Especially the task of
achieving a sufficiently accurate flow control in connection with the over-center valves
poses an interesting research topic.

While the developed HIL simulator possess a sufficient level of detail to test the current
controller used on HMF loader cranes, it would be interesting to include the dynamics
of both the truck and the crane in the simulation model. Computationally, this is a very
challenging task, but it should be possible by including one or more computers dedicated
to solving the dynamics and connecting these to the current setup via LAN.
Bibliography

Barsoum, Z. (2008), Residual stress analysis and fatigue assessment of welded steel
structures, Ph.d. thesis, KTH, Stockholm, Sweden.

Beiner, L. (1997), ‘Minimum force redundancy control of hydraulic cranes’, Mechatronics

7(6), 537–547.

Beiner, L. and J. Mattila (1999), ‘An improved pseudoinverse solution for redundant
hydraulic manipulators’, Robotica 17, 173–179.

Booth, G. (1983), Improving the Fatigue Strength of Welded Joints, TWI, Abington
Publishing, UK.

Chan, T. F. and R. V. Dubey (1993), ‘A weighted least-norm solution based scheme for
avoiding joint limits for redundant manipulators’, Robotics and Automation 3, 395–402.

Dürr, A. (2007), Zur Ermüdungsfestigkeit von Schweisskonstruktionen aus höherfesten

Baustählen bei Anwendung von UIT-Nachbehandlung [Fatigue Resistance of Welded
High Strength Steel Structures Using UIT Post Weld Treatment], Ph.d. thesis,
Universität Stuttgart, Germany.

Ebbesen, M. K. (2007), Optimal Design of Flexible Multibody Systems, Ph.d. thesis,

Institute of Mechanical Engineering, Aalborg University, Denmark.

Elton, M.D., A.R. Enes and W.J. Book (2009), A virtual reality operator interface
station with hydraulic hardware-in-the-loop simulation for prototyping excavator
control systems, in ‘IEEE/ASME International Conference on Advanced Intelligent
Mechatronics’, Singapore.

EN12999 (2010), ‘FprEN12999:2010, Cranes - Loader cranes’.

Esqué, S., A. Raneda and A. Ellman (2003), ‘Techniques for studying a mobile hydraulic
crane in virtual reality’, Int J Fluid Power 4(2), 25–34.

Fricke, W. (2003), ‘Fatigue analysis of welded joints: State of development’, Marine

Structures 16, 185–200.

Fricke, W. (2008), ‘Guideline for fatigue assessment by notch stress analysis for welded
structures’. IIW Doc. XIII-2240r1-08.
30 Bibliography

Fricke, W. and O. Feltz (2009), ‘Fatigue tests and numerical analyses of partial-load and
full-load carrying fillet welds at cover plates and lap joints’. IIW Doc. XIII-2278-09.
Fukuoka, T. and K. Mochizuki (2010), ‘Effect of plate thickness on fatigue strength of
typical welded joints for a ship structure’. IIW Doc. XIII-2333-10.
Ghabcheloo, R. and M. Hyvönen (2009), Modeling and motion control of an articulated-
frame-steering hydraulic mobile machine, in ‘17th Mediterranean Conference on
Control and Automation’, Thessaloniki, Greece.
Gurney, T.R. (1979), Fatigue of Welded Structures, Cambridge University Press, Cam-
bridge, UK.
Haagensen, P.J. (1985), Improving the fatigue strength of welded joints, in A.Naess, ed.,
‘Fatigue Handbook - Offshore Steel Structures’, Tapir, Trondheim, Norway, pp. 259–
289.
Haagensen, P.J. (2007), ‘Fatigue improvement techniques - advantages and limitations’,
Welding in the World 47, 43–63.
Haagensen, P.J. and S.J. Maddox (2008), ‘IIW recommendations on post weld improve-
ment of steel and aluminium structures’. IIW Doc. XIII-2200r1-07.
Hansen, M. R. and T. O. Andersen (2005), A method for deriving the optimal operation
of mobile hydraulic manipulators, in ‘Proc 9th Scandinavian International Conference
on Fluid Power, SICFP’05’, Linköping, Sweden.
HMF (2010), ‘www.hmf.dk’. HMF history.
Hobbacher, A. (2008), IIW Recommendations for Fatigue Design of Welded Joints and
Components, WRC Bulletin 520, New York, USA.
Jonsson, B. and J. Samuelsson (2008), ‘A new weld quality system’. IIW Doc. XIII-2235-
08.
Kabus, S. and M. Haastrup (2008), Simulations of flexible loader crane with designed tool
point control, Master’s thesis, Institute of Mechanical Engineering, Aalborg University,
Denmark.
Kirkhope, K.J., R. Bell, L. Caron, R.I. Basu and K.T. Ma (1999), ‘Weld detail fatigue
life improvement techniques. part 1 and 2’, Marine Structures 12, 447–496.
Kranz, B. and C.M. Sonsino (2009), ‘Verification of the notch stress concept for the
reference radii of rref=1.00 and 0.05mm’. IIW Doc. XIII-2274-09/XV-1312-09.
Krus, G. and J. O. Palmberg (1992), Vector control of a hydraulic crane, in ‘Proc
International Off-Highway and Power plant Congress and Exposition’, Milwaukee,
USA.
Lagerqvist, O., M. Clarin, J. Gozzi, B. Völling, D. Pak, J. Stötzl, H.P. Lieurade, B.
Depale, I. Huther, S. Herion, J. Bergers, R.M. Martsch, M. Carlsson, A. Samuelsson
and C. Sonander (2007), LiftHigh - Efficient Lifting Equipment with Extra High-
Strength Steel, European Commission, Technical Steel Research, EUR22569EN, Brus-
sels, Belgium.
Bibliography 31

Lihavainen, V.M. (2006), A Novel Approach for Assessing the Fatigue Strength of
Ultrasonic Impact Treated Welded Structures, Ph.d. thesis, Lappeenranta University
of Technology, Lappeenranta, Findland.

Linjama, M., T. Virvalo, J. Gustafsson, J. Lintula, V. Aaltonen and M. Kivikoski (2000),

‘Hardware-in-the-loop environment for servo system controller design, tuning and
testing’, Microprocessors and Microsystems 24, 13–21.

Maddox, S.J. (1991), Fatigue Strength of Welded Structures, 2nd ed., Abington Publish-
ing, Cambridge, UK.

Maddox, S.J. (2003), ‘Key developments in the fatigue design of welded constructions’,
Welding in the World 47, 6–40.

Mattila, J. and T. Virvalo (2000), Energy efficient motion control of a hydraulic

manipulator, in ‘Int Proc Robotics and Automation’, San Francisco, USA.

Münzer, M E. (2003), Resolved Motion Control of Mobile Hydraulic Cranes, Ph.d. thesis,
Insitute of Energy Technology, Aalborg University, Denmark.

Nykänen, T., G. Marguis and T. Björk (2007), ‘Simplified assessment of weld quality for
fatigue loaded cruciform joints’. IIW Doc. XIII-2177-07.

Ohta, A., T. Mawari and N. Suzuki (1990), ‘Evaluation of effect of plate thickness on
fatigue strength of butt welded joints by a test maintaining maximum stress at yield
strength’, Engineering Fracture Mechanics 37(5), 987–993.

Olivier, R., V.B. Köttgen and T. Seeger (1989), Schweissverbindung I [Welded Joints I],
FKM Forschungshefte 143, Frankfurt/M, Germany.

Olivier, R., V.B. Köttgen and T. Seeger (1994), Schweissverbindung II [Welded Joints
II], FKM Forschungshefte 180, Frankfurt/M, Germany.

Olivier, R. and W. Ritter (1979), Wöhlerlinienkatalog für Schweissverbindungen aus

Baustählen [SN curve catalog for welded joints in structural steels], DVS GmbH,
Düsseldorf, Germany.

Pedersen, H. C. and B Nielsen (2002), Resolved motion control of flexible hydraulic

manipulators, Master’s thesis, Institute of Mechanical Engineering, Aalborg University,
Denmark.

Radaj, D. (1990), Design and Analysis of Fatigue Resistant Welded Structures, Woodhead
Publishing, Cambridge, UK.

Radaj, D., C.M. Sonsino and W. Fricke (2006), Fatigue Assessment of Welded Joints by
Local Approaches, 2nd ed., Woodhead Publishing, Cambridge, UK.

Radaj, D., C.M. Sonsino and W. Fricke (2009), ‘Recent developments in local concepts
of fatigue assessment of welded joints’, Int. J. Fatigue 31(1), 2–11.
32 Bibliography

Rasmussen, L.V. (2008), Levetidsbestemmelse og analyse af typisk svejst krandetalje i

højstyrkestål (danish). [Fatigue assessment and analysis of typical welded crane detail
in high strength steel], Master’s thesis, Institute of Mehanical Engineering, Aalborg
University, Denmark.

Roy, S. (2006), Experimental and Analytical Evaluation of Enhancement in Fatigue

Resistance of Welded Details Subjected to Post-Weld Ultrasonic Impact Treatment,
Ph.d. thesis, Dept. of Civil and Env. Eng., Lehigh University, USA.

Samuelsson, J. (2007), Integrated Design and Manufacturing of Welded Structures,

Norden - Nordic Innovation Centre, Oslo, Norway.

Sonsino, C.M., T. Bruder and J. Baumgartner (2009), ‘Sn-curves for welded thin joints
- suggested slopes and fat-values for applying the notch stress concept with various
reference radii’. IIW Doc. XIII-2280-09.

Trydal, S. (2010), ‘HIL simulator. Hardware In Loop Simulator med fokus på integrerte
prosesser. [HIL simulator. hardware-in the loop simulator with focus on integrated
processes]’. www.nodeproject.no/index.php?articleid=393.

Weich, I. (2008), Ermüdungsverhalten mechanisch nachbehandelter Schweissverbindun-

gen in Abhängigkeit des Randschichtzustands [Edge Layer Condition and Fatigue
Strength of Welds Improved by Mechanical Post Weld Treatment], Ph.d. thesis,
Technischen Universität Brauschweig, Germany.

Wiedner, P., I. Weich and T. Ummenhofer (2010), ‘High frequency hammer peening of
LCF-stressed ultra high strength steels’. IIW Doc. XIII-2341-10.

Yuan, Q., J. Lew and D. Piyabongkarn (2009), Motion control of an aerial work platform,
in ‘Proc. American Control Conference’, St. Louis, MO, USA.
Appended Papers
 A
Comparison of Post Weld
Treatment of High Strength Steel
Welded Joints in Medium Cycle
Fatigue

M.M. Pedersen, O.Ø. Mouritsen, M.R. Hansen, J.G. Andersen, J. Wenderby

Welding in the World, 54 (7/8), pp. 208-217, 2010.

Abstract. This paper presents a comparison of three post weld treatments for fatigue
life improvement of welded joints. The objective is to determine the most suitable post
weld treatment for implementation in mass production of certain crane components
manufactured from very high strength steel. The processes investigated are; burr grinding,
TIG dressing and ultrasonic impact treatment.

The focus of this investigation is on the so-called medium cycle area, i.e. 10,000-500,000
cycles and very high stress ranges. In this area of fatigue design, the use of very high
strength steel becomes necessary, since the stress range can exceed the yield strength of
ordinary structural steel, especially when considering positive stress ratios (R ≥ 0).

Fatigue experiments and qualitative evaluation of the different post weld treatments
leads to the selection of TIG dressing. The process of implementing TIG dressing in
mass production and some inherent initial problems are discussed. The treatment of
a few critical welds leads to a significant increase in fatigue performance of the entire
structure and the possibility for better utilization of very high strength steel.

Introduction
It is a well known fact that welded joints have a low fatigue strength compared to the
base material. This is mainly caused by local stress concentrations due to the presence of
notches and high tensile residual stresses. Notches occur both because of the geometry of
the joint, but also because of weld imperfections such as undercuts and slag inclusions.
Tensile residual stresses arise from the contraction of the weld metal during cooling and
solidification.
36

The fatigue strength of as-welded joints proves to be practically independent of the base
material strength. Thus, when applying higher strength steels, the gap between the static
and fatigue strength increases. However, in many applications, e.g. mobile machinery,
there is a tendency to pursue an increase in the performance to weight ratio by applying
higher strength steels. In such cases, the fatigue strength of the welded joints becomes
the dominating factor and limits the benefit of the application of the high strength steel.

Improvement of the fatigue strength of welded joints by application of different post-weld

treatments has therefore received much attention lately e.g. [1-5], especially treatments
based on high frequency peening, e.g. UIT. These relatively new treatments consistently
provide very significant improvements in fatigue strength, particularly in the high cycle
regime.

Unfortunately, only little effort has been put into investigation of the medium cycle area,
i.e. the upper area marked in Fig. A.1. In this area of fatigue design, there is an obvious
reason for using high strength steels, since the stress range exceeds the yield strength of
ordinary structural steel, assuming R > 0. Ordinarily this area would be classified as low
cycle fatigue, and plasticity should be taken into account. When applying high strength
steel, on the other hand, the elastic region extends upwards in the Wöhler diagram, and
plasticity is not an issue.

1500
1000 Assuming R=0
800
High strength
600 steel necessary
500
400
Stress range [MPa]

300
200

100

50
Steel strength
insignificant

10 3 4 5 6 7 8
10 10 10 10 10 10
Fatigue life [cycles]

Figure A.1: Net of S-N curves accord- Figure A.2: A truck mounted loader
ing to the IIW [6]. crane is primarily manufactured from
S700, S900 and S1100.

Even though the fatigue strength of as-welded joints is independent of the base material
strength at lower stress ranges, this is not the case for the high stress / medium cycle
area. There is a need for special fatigue design guidelines in this area considering the
use of high strength steel. An example is the IIW parent material curve FAT160, see
Hobbacher [6], which is too conservative for high strength steel. The new crane code
EN13001 [7] for instance suggests FAT315 for high strength steel parent material.

Mobile lifting equipment, e.g. loader cranes as shown in Fig. A.2, is a good example
of structures designed for high stresses and medium cycle fatigue lives, e.g. 20,000 -
250,000 full load lifting cycles. The manufacturers strive to increase the lifting capacity
and reduce the weight of these structures using higher and higher steel grades and thin,
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 37

locally reinforced plate structures. This tendency inevitably leads to fatigue problems,
which is why post weld treatment becomes more and more interesting for these new
structures, instead of upgrading/repair of existing ones.

This work investigates three different post weld treatments in the medium cycle area,
under very high stress ranges, i.e. up to the yield strength of the base material. The
objective is to determine the most suitable treatment in this area when also considering
the practical issues of implementation in mass production of crane components.

Experimental
The fatigue investigation was carried out on simple T-joint specimens in four point
bending. The high stress ranges and the required accuracy are more easily achieved in
bending tests as compared to tests in tension. Furthermore, by testing in bending loading,
misalignment-induced secondary bending stresses are avoided. This is essential, since the
different post weld treatments cause different levels of distortion in the specimens.

Test specimens
The test specimens are produced by robotic MAG welding of Domex 700 by the Danish
loader crane manufacturer HMF A/S, using a matching filler material. The plate was
initially clean blasted and pre-bent 3 degrees to counteract welding distortion. The final
specimen geometry is shown in Fig. A.3 and the welding parameters are listed in Tab.
A.1.

Current 270A
6 Voltage 26V
50 40 4
Travel speed 440mm/min
6
Shielding gas 90% Argon, 10% CO2
Heat input 0.8kJ/mm
250
T-joint, S700
t8/5 time 6sec

Figure A.3: Test specimen geometry. Table A.1: Welding parameters.

The specimen edges were ground longitudinally and the corners slightly rounded using a
burr grinder. The weld quality was considered normal. The cooling rate, expressed as the
time for cooling from 800◦ to 500◦ (t8/5 time), was within the margins specified by the
steel manufacturer, SSAB, namely between 5 and 25 seconds for the Domex 700 steel.

Fatigue testing
Fatigue testing was performed using a 100kN Schenk hydraulic fatigue testing machine
at the Department of Mechanical Engineering at Aalborg University. The tests were
performed in four point bending, at constant amplitude, using the equipment shown in
Fig. A.4 and Fig. A.5. The stress ratio was kept at R = 0.1, and the frequency was varied
from 7 − 28Hz. The tests were performed at room temperature and no heating of the
specimens occurred.
38

Figure A.4: Fatigue testing machine. Figure A.5: Fatigue test method.

Post Weld Treatment

Three post weld treatments were selected for the investigation; burr grinding (BG), TIG
dressing and ultrasonic impact treatment (UIT). The first two processes mainly aim at
improving the weld toe geometry in order to reduce the stress concentration here. UIT
on the other hand primarily aims at improving the residual stress state in the weld toe
area.

Since the fatigue testing was carried out under very high stresses, it was anticipated that
the residual stresses would not influence the fatigue life significantly due to relaxation.
The majority of the improvement is therefore expected to stem directly from improvement
in the weld toe geometry. Fig. A.6 shows the specimens and Fig. A.7 shows micrographs
of the weld toe geometry in as-welded and post weld treated conditions.

All post weld treatments have potential drawbacks under the conditions of this
investigation. Burr grinding is difficult due to the high hardness of the material. TIG
dressing may lead to excessive softening of the high strength steel. The compressive
residual stresses introduced by UIT may be relaxed due to the high stresses.

Evidently, the weld toe stress concentration factor will be reduced by all the post weld
treatments considered. However, TIG dressing seemed to leave the weld toe with the
smoothest transition between the base material and weld face. Nevertheless, the transition
geometry of the TIG dressed weld toe varied significantly along the treated weld seam
due to minor variations in the welding torch angle and treatment speed.
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 39

Figure A.6: As-welded and post weld treated test specimens.

Figure A.7: Micrographs illustrating the different toe geometries obtained by post-
weld treatment.

Burr grinding
The purpose of burr grinding is to remove minor weld defects in the weld toe by
machining, while at the same time reducing the stress concentration factor of the weld
toe. The use of a burr grinder instead of a disc grinder leaves the grinding marks parallel
to the direction of loading, which prevents them from acting as crack initiation sites.

Hansen et al. [8] report significant improvements in fatigue life from a two-stage burr
grinding procedure, using first a rough tungsten carbide burr followed by a second run
with a finer grinder, see Fig. A.8. This procedure has therefore been adopted in this
investigation, together with the general guidelines provided by the IIW, see Haagensen
and Maddox [9].

A relatively small diameter burr (∅6mm) was used in order to limit the reduction of
section of the base plate due to excessive grinding. The treatment left a shiny smooth
surface in the weld toe. When inspected under microscope, there were no sign left of the
original weld toe.
40

Figure A.8: 6mm diameter tungsten-carbide burr (top) and fine grinder (bottom).

Burr grinding in high strength steel is rather slow because of the high hardness of the
steel, which also causes significant wear in the burr. The fact, that some of the base plate
must be ground away in order to remove weld defects furthermore limits the use of this
treatment in the very thin plates commonly used when applying high strength steel, e.g.
4 − 5mm plate.

High Frequency Peening (HFP) treatment

The following different types of high frequency peening treatments are described in the
literature; ultrasonic impact treatment (UIT) [10], ultrasonic peening (UP) [11], high
frequency impact treatment (HiFIT) [12] and ultrasonic needle peening (UNP) [13], see
Fig. A.9. UIT is based on magnetostriction, UP uses a piezoelectric transducer, whereas
HiFIT is pneumatically actuated. Although they are different processes, their properties
and resulting improvement in fatigue strength appear to be similar.

Figure A.9: Examples of four different tools for high frequency peening treatment.
From left; UIT [3], UP [4], HiFIT [5] and UNP [6].

They all operate in a manner similar to ordinary hammer peening equipment, only at a
considerably higher frequency, which reduces vibration and noise and is claimed to give
a high improvement in fatigue strength. The tip of the tool oscillates at about 200Hz
with a displacement of only about 40µm. Usually, the tip of the tool is fitted with several
hardened pins, e.g. 3-4 on a line, but many different configurations exist. The tools are
usually handheld, but a robot-mounted version of the UIT device has also been developed.

According to Lihavainen [14] only a force of approximately 30N is necessary on the UIT
tool, whereas ordinary hammer peening tools require a force of more than 200N against
the material being treated. The application of a HFP treatment is therefore much more
comfortable for the operator, which may lead to a better result.
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 41

The HFP treatments improve the fatigue strength of the welded joint in several ways
by plastic deformation of the weld toe. Firstly, the tensile residual stress state present
in the weld seam is relieved and beneficial compressive residual stresses are introduced.
Secondly, the sharp notch in the weld toe is blunted and the treatment leaves behind
a smooth trace with a radius of 2 − 3mm, see Fig. A.7. Finally, the surface material is
mechanically hardened, which locally increases the fatigue strength of the material in the
notch.

In this investigation, UIT was applied. Fig. A.10 shows the tool used for the treatment
- a single pin shaped to accommodate treatment along the length of a weld seam. The
treatment produced an even, smooth trace along the weld toe, approximately 3mm wide
and increased the weld toe radius to approximately R2.

25
Ø6.35
R2

R25
R1.5

Figure A.10: Double radius pin used for ultrasonic impact treatment.

The treatment can be performed at a speed of about 700 − 900mm/min for high-quality
even weld seams. Irregularities, e.g. start/stops, significantly reduce the treatment speed
though and may require grinding. UIT is therefore by far the fastest treatment of the
three investigated in this project, but at the same time the one that requires the largest
initial investment.

TIG dressing
The objective of TIG dressing is to remove welding imperfections, such as undercuts,
cold laps and inclusions by remelting the weld toe, thereby leaving the weld practically
defect free. Moreover, the treatment significantly reduces the stress concentration factor
of the weld toe by introducing a smooth transition.

Current 175A
Voltage 18V
Travel speed 100mm/min
Shielding gas 90% Argon, 10% CO2
Heat input 1.3kJ/mm
t8/5 time 16sec

Table A.2: TIG dressing parameters

TIG dressing was carried out according to IIW recommendations [9] and following the
steel manufacturer cooling rates. The weld seam and adjacent plate were thoroughly
42

cleaned by sand blasting prior to TIG dressing. Tab. A.2 lists the TIG dressing
parameters. The heat input and t8/5 time were calculated assuming an efficiency of 0.7
for the TIG process. Again, the cooling rate is well within the specifications of the steel
manufacturer.

The TIG dressing treatment was carried out manually using standard TIG welding
equipment without adding filler material. The TIG welding torch was aimed 1mm from
the weld toe, directing most of the heat towards the base plate. As seen in Fig. A.7, there
is a slight tendency towards an undercut in the TIG dressed welds; however, the very
generous transition radius limits the negative effect of this undercut.

A major concern when performing TIG dressing in high strength steel is whether it
will cause excessive softening of the material. In order to investigate this, hardness
measurements through the TIG dressed area were made, see Fig. A.11. The investigation
showed a resulting hardness drop of approximately 15 − 20% in the TIG dressed area.

Figure A.11: Hardness profiles through the thickness of a TIG dressed weld.

In this investigation TIG dressing could be performed at approximately twice the speed
of burr grinding, i.e. approximately 100 − 200mm/min. Observations and experience
clearly indicate that the TIG dressing process is less demanding for the operator than
the burr grinding process. Additionally, TIG dressing is found to be the best suited post
weld treatment for handling irregularities in the weld seam, e.g. a start/stop.

Fatigue Test Results

All specimens failed from a weld toe. Fig. A.12 shows a typical example of the fracture
surface. It is seen that the fatigue crack initiates in the middle of the specimen.
Investigations using the notch stress approach also suggest the middle as the weakest
point of the specimens.

The fatigue test results are listed in Tab. A.3 and plotted in Fig. A.13. Evidently, all three
post weld treatments led to significant improvements in fatigue strength. The degree of
improvement is comparable for all treatments; however, TIG dressing is the most effective
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 43

AW BG TIG UIT
∆σ N [cycles] ∆σ N [cycles] ∆σ N [cycles] ∆σ N [cycles]
619 48.246 623 144.801 675 84.527 732 48.833
525 58.992 735 44.373 709 93.788 659 72.024
454 136.462 757 39.727 631 156.266 666 73.094
660 26.231 714 72.084 548 352.235 648 166.486
702 28.558 651 74.943 524 666.989 778 27.413
527 78.961 625 76.851 712 91.578 774 15.562
696 20.159 561 172.042 638 112.553 776 35.199
411 164.733 685 102.958 598 179.863 708 45.417
561 43.719 715 100.913 616 136.497 668 83.970
515 96.224 817 25.283 725 36.907 547 244.956
441 128.156 803 25.292 656 94.745
632 35.701 796 48.665
401 305.201 811 44.062
432 154.325
683 19.303
776 12.663
778 15.205

Table A.3: Fatigue test results.

treatment in the medium cycle regime. UIT on the other hand seems to be the most
effective in the high cycle regime.

Contrary to what might be expected, the UIT treatment provided significant improve-
ment, even under the highest applied stresses, i.e. up to the yield point of the base
material. The compressive residual stresses introduced by the UIT treatment were
expected to relax, thus not improving the fatigue strength of the specimens. However,
the UIT treatment also improves the local geometry of the weld toe, which might explain
the improvement in fatigue life seen in this investigation.

Figure A.12: Example of the fracture surface.

S-N curves were fitted to each set of results by free regression analysis. Curves representing
characteristic fatigue strengths, set two standard deviations of logN below the mean
curves, are included in Fig. A.13. The characteristic fatigue strengths at 2 · 106 cycles
(FAT values in the IIW scheme [6]) obtained from these curves are summarised in Tab.
44

1000
900 R=0.1
800
700
Nominal stress range, ∆σ [MPa]

600
500
n

400

300 AW
BG
TIG
200 UIT
AW FAT217, m=4.2
S700, t=6mm BG FAT323, m=5.1
TIG FAT368, m=5.9
UIT FAT387, m=7.1
100 4 5 6 7
10 10 10 10
Cycles to failure, N
f

Figure A.13: Fatigue test results with characteristic fatigue strength S-N curves.

AW BG UIT TIG
Free slope 4.2 5.1 7.1 5.9
2,000,000 cycles 217 323 (+49%) 387 (+78%) 368 (+70%)
100,000 cycles 443 581 (+31%) 590 (+33%) 611 (+38%)

Table A.4: Characteristic fatigue strengths [MPa].

A.4. In addition, the fatigue strengths at 100,000 cycles, which seem more appropriate
in this investigation are also given.

It is clear that the improvement in fatigue strength is greater at 2,000,000 cycles than
100,000 cycles, because the S-N curves for the improved details are shallower than that
of the as-welded specimens. However, due to the limited number of test results in the
high cycle regime, the FAT values at 2,000,000 are quite uncertain. At 100,000 cycles, on
the other hand, the results are less scattered, the level of improvement being essentially
the same for all post weld treatments.
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 45

Comparison with published results

In Fig. A.14, the results obtained in this investigation are compared with existing results
extracted from the literature [4,15,16,17,19]. All results are for T-joints tested in four
point bending at a stress ratio of R = 0.1, but with differences in steel grade, thickness,
welding process etc.

As-welded Burr grinding

800 800
700 R=0.1 700 R=0.1
600 600
500 500
400 1 400 1
300 m=3 300 m=3
∆σn [MPa]

∆σn [MPa]
200 200

This inv., S700, t=6 133

Ref [17], S420, t=20
100 Ref [16], S420, t=20 96 100 96
Ref [4], S700, t=5 This inv., S700, t=6
Ref [4], S355, t=6 Ref [15], S420, t=20
Ref [15], S420, t=20 As-welded
50 4 5 6 7
50 4 5 6 7
10 10 10 10 10 10 10 10
N [cycles] N [cycles]
f f

TIG dressing High frequency peening

800 800
700 R=0.1 700 R=0.1
600 600 1
500 500 m=5
400 1 400 1
300 m=3 300 m=3
∆σn [MPa]

∆σn [MPa]

259

200 200
This inv., S700, t=6
140
Ref [4], S355, t=6
Ref [4], S700, t=5
100 This inv., S700, t=6 96 100 96
Ref [17], S420, t=20
Ref [16], S420, t=20 Ref [16], S420, t=20
Ref [15], S420, t=20 Ref [19], comp.preload
As-welded As-welded
50 4 5 6 7
50 4 5 6 7
10 10 10 10 10 10 10 10
N [cycles] N [cycles]
f f

Figure A.14: Results from this investigation compared with similar results from
the literature.

The scatter bands in Fig. A.14 were calculated from all results pooled together. They are
not meant for design purposes, only for comparison. The curves for AW, BG and TIG
series were calculated using m = 3, whereas the curves for the HFP series were based on
m = 5.

The fatigue data from this investigation are seen to be consistent with the published data
and provide a useful extension of each database to higher applied stresses. It will also
be evident that the present results provide no support for the current recommendation
[9] to limit the design curves for improved welds to the IIW design curve for plain steel,
FAT160 with m = 5, as long as a high strength steel is used.
46

The HFP treated specimens of [19] was preloaded in compression 5 times at a magnitude
of 85% of the yield strength, in order to study the effect of improvement of HFP after
potential relaxation of the introduced compressive residual stresses. These results are
therefore not included in the scatter bands in Fig. A.14. It is clear, that the degree
of improvement obtained in [19] is significantly reduced because of the compressive
preloading.

TIG Dressing in Mass Production

The objective of this work was to determine a suitable post weld treatment procedure
for implementation in mass production of crane components. TIG dressing was chosen
due to the large improvement in fatigue strength observed during tests, but also because
of the availability of the equipment and the flexibility of the process.

Many crane components are made from very high strength steel and therefore have a
very high static strength. The weld seams are placed in compressive zones or zones with
low stresses as far as practically possible. Still, the fatigue strength of the component is
governed by one or few critical welds which are highly stressed and cannot be moved or
improved by means other than increasing the plate thickness or by post weld treatment.
Fig. A.15 shows an example of such a critical weld.

Figure A.15: Welded crane component with illustration of highly stressed weld
subjected to TIG dressing.

The designated weld seam is subjected to TIG dressing, in order to increase its fatigue
strength to the level of the remainder of the component. Instructions were prepared and a
short course was given for the operators and quality control personnel. A new weld quality
class was adopted for specification of post weld treatment in technical drawings, similar
to VB in the new Volvo CE weld quality system, described by Jonsson and Samuelsson
[18].

As expected, taking the process from the laboratory to mass production caused problems.
Initially, operators failed to follow instructions and a very low quality treatment was
achieved. Fig. A.16 shows examples of the resulting deep undercuts and presence of slag.
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 47

Figure A.16: Initial quality problems with TIG dressing.

The following problems were identified

• Incorrectly inclined work piece caused the weld pool to run away from the base
plate.
• Mill scale on the base plate obstructed joining of base material and re-melted weld
metal.
• Insufficient de-slagging and cleaning after the primary welding.
• Too high heat input and wrong welding torch angles.
• Insufficient operator experience with the TIG welding process.

Many of the initial problems could be remedied with increased focus on thorough cleaning
of the work pieces. The mill scale on the base plate was removed by sand blasting prior
to the primary welding. Subsequently, using a chisel and an electric rotating wire brush,
the weld seam was carefully cleaned before TIG dressing. The operator was also given
more time to practise and instructed to level the work piece in order to avoid undercuts,
as illustrated in Fig. A.17.

Figure A.17: Correct levelling of the work piece helps avoid undercuts.

A proper heat input was obtained by letting the operator select current and weld speed
freely for the TIG dressing operation at hand, but always maintaining a weld pool of
48

Figure A.18: Examples of final production quality TIG dressing, free from
undercutting and significant imperfections.

6 − 7mm in diameter. Finally, a satisfactory and repeatable quality level was obtained,
as seen in Fig. A.18.

In such cases, where only a few critically loaded welds govern the fatigue strength of the
entire structure, post weld treatment can be a very economic alternative to increasing
the plate thickness. For generally highly stressed components, on the other hand, post
weld treatment may be too expensive compared to increased plate thicknesses.

On a typical loader crane, only a few welds are loaded to such extend that post weld
treatment is necessary. In this example, only approximately 500mm weld seam is treated,
but this small effort significantly increases the fatigue performance of the crane and allows
for a very lightweight and durable design.

Conclusions
On the basis of this investigation of the benefit of selected weld toe improvement methods
on the fatigue performance of S700 grade high-strength steel T-joints, the following
conclusions were drawn:

• High strength steel can be used very advantageously in the medium cycle fatigue
regime when used in combination with local post weld treatment.
• At a fatigue life of 100,000 cycles, the fatigue strength improvement levels obtained
from the three methods investigated were almost identical, i.e. 31% for burr
grinding, 33% for UIT and 38% for TIG dressing.
• The UIT process showed a consistent high level of improvement, even under stress
ranges up to the yield strength of the base material.
• TIG dressing was found to be the best suited post weld treatment for implemen-
tation in mass production of crane components, because of the large improvement
observed in the experiments, the availability of equipment and the flexibility of the
process.
• However, practical experience highlighted the importance of correct work-piece
preparation prior to the application of an improvement technique and operator
training.
Paper A. Comparison of Post Weld Treatment of High Strength Steel Welded
Joints in Medium Cycle Fatigue 49

Acknowledgements
Mr. Thaddeus Stam and Applied Ultrasonics Europe are gratefully acknowledged for
performing the ultrasonic impact treatment. The research work is funded by an Industrial
PhD-grant from the Danish Agency for Science, Technology and Innovation under the
Danish Ministry for Science, Technology and Innovation.

References
1 Marquis G. and Björk T. Variable Amplitude Fatigue Strength of Improved HSS
Welds, IIW Doc. XIII-2224-08, 2008.
2 Lieurade H.P., Huther I., Lefebvre H. Effect of Weld Quality and Post Weld
Improvement Techniques on the Fatigue Resistance of Extra High Strength Steel,
IIW Doc. XIII-2184-07, 2007.
3 Lagerqvist O., Clarin M., Gozzi J., Völling B., Pak D., Stötzl J., Lieurade H.P.,
Depale B., Huther I., Herion S., Bergers J., Martsch R.M., Carlsson M., Samuelsson
A. and Sonander C. LiftHigh - Efficient Lifting Equipment with Extra High-
Strength Steel, European Commission, Technical Steel Research, EUR22569EN,
Brussels, Belgium, 2007.
4 Galtier A. and Statnikov E.S. The Influence of Ultrasonic Impact Treatment on
Fatigue Behaviour of Welded Joints in High-Strength Steel, Welding in the World,
48, no. 5/6, 2004.
5 Manteghi S. and Maddox S.J. Methods for Fatigue Life Improvement of Welded
Joints in Medium and High Strength Steels, IIW-Doc. XIII-2006-04, 2004.
6 Hobbacher A. (editor), Recommendations for Fatigue Design of Welded Joints and
Components, IIW Doc. IIW-1823-07, 2008.
7 DS/CEN/TS EN13001-3-1, Cranes - General Design - Part 3-1: Limit States and
Proof of Competence of Steel Structures, 1st edition, 2005.
8 Hansen A.V., Agerskov H. and Bjørnbak-Hansen J., Improvement of Fatigue Life
of Welded Structural Components by Grinding, IIW Doc. XIII-2051-05, 2005.
9 Haagensen P.J. and Maddox S.J., IIW Recommendations on Post Weld Improve-
ment of Steel and Aluminium Structures, IIW Doc. XIII-2200r1-07, 2008.
10 Dürr A., Zur Ermüdungsfestigkeit von Schweisskonstruktionen aus höherfesten
Baustählen bei Anwendung von UIT-Nachbehandlung [Fatigue Resistance of
Welded High Strength Steel Structures Using UIT Post Weld Treatment], PhD
thesis, Universität Stuttgart, 2007.
11 Kudryavtsev Y., Kleiman J., Lugovskoy A. and Prokopenko G., Fatigue Life
Improvement of Tubular Welded Joints by Ultrasonic Peening, IIW Doc. XIII-
2117-06, 2006.
12 REFRESH Project, http://outpost.stahl.bau.tu-bs.de/refresh/index.php?
id=180&L=1, 02-02-2009.
13 Bousseau M. and Millot T., Fatigue Life Improvement of Welded Structures by
Ultrasonic Needle Peening Compared to TIG Dressing, IIW-Doc. XIII-2125-06,
2006.
50

14 Lihavainen V.M., A Novel Approach for Assessing the Fatigue Strength of Ultra-
sonic Impact Treated Welded Structures, PhD thesis, Lappeenranta University of
Technology, Finland, 2006.
15 Haagensen P.J. IIW’s Round Robin and Design Recommendations for Improve-
ment Methods, Proceedings of the IIW 50th Annual Assembly Conference, San
Francisco, Welding Research Council Inc., New York, USA, 1997.
16 Statnikov E.S., Muktepavel V.O. and Blomqvist A. Comparison of Ultrasonic
Impact Treatment (UIT) and Other Fatigue Life Improvement Methods, Welding
in the World 46, pp. 28- 39, 2002.
17 Trufiakov V.I., Statnikov E.S., Mikheev P.P. and Kuzmenko A.Z. The Efficiency
of Ultrasonic Impact Treatment for Improving the Fatigue Strength, IIW Doc.
XIII-1745-98, 1998.
18 Jonsson B. and Samuelsson J., A New Weld Quality System, IIW Doc. XIII-2235-
08, 2008.
19 Lopez-Martinez L. and Haagensen P.J. Life Extenion of Class F and Class F2
Details Using Ultrasonic PeeningIIW Doc. XIII-2143-07, 2007.
 B
Re-analysis of Fatigue Data for
Welded Joints Using the Notch
Stress Approach

M.M. Pedersen, O.Ø. Mouritsen, M.R. Hansen, J.G. Andersen, J. Wenderby

International Journal of Fatigue, 32, pp.1620-1626, 2010.

Abstract. Experimental fatigue data for welded joints have been collected and subjected
to re-analysis using the notch stress approach according to IIW recommendations. This
leads to an overview regarding the reliability of the approach, based on a large number of
results (767 specimens). Evidently, there are some limitations in the approach regarding
mild notch joints, such as butt joints, which can be assessed non-conservatively. In order
to alleviate this problem, an increased minimum notch factor of Kw > 2.0 is suggested
instead of the current recommendation of Kw > 1.6. The data for most fillet welded joints
agree quite well with the FAT225 curve; however a reduction to FAT200 is suggested
in order to achieve approximately the same safety as observed in the nominal stress
approach.

Introduction
The notch stress approach for fatigue assessment of welded joints correlates the stress
range in a fictitious rounding in the weld toe or root to the fatigue life using a single
S-N curve. The notch stress is typically obtained using finite element models with the
reference radius of 1mm in order to avoid the stress singularities in sharp notches.

The approach has received much attention lately, due to the increasing available
computational power. The approach is very flexible in the sense that all types of welded
joints can be assessed using a single S-N curve. It does, however, require more modelling
and analysis work than e.g. the nominal or structural stress approaches.

Radaj et al. [1] presents a thorough review of the history of the approach. Fricke [2] gives
practical guidelines for the notch modelling and stress analysis and Sonsino [3] proposes
S-N curves to be used under different conditions. The approach is included in the IIW
fatigue design recommendations by Hobbacher [4].
52

In this paper, we consider the notch stress approach according to the IIW [2]. The
approach is based on the work by Radaj [5] and modified by Seeger and co-workers,
see Olivier et al. [6,7]. The reference radius of R1 is determined as a mean value and the
fatigue strength (FAT225) is derived from experiments.

While many recent publications on fatigue experiments also discuss the results in terms
of notch stresses [8-11], the amount of published experimental evidence of the reliability
of the approach is very limited.

This investigation therefore presents a systematic re-analysis of fatigue data extracted

from the literature and converted to the notch stress system. This provides an overview
of the reliability of the approach and a basis for discussion of the observed limitations.

Extraction of Fatigue Data

Constant amplitude fatigue testing results have been extracted from the literature [12-
35]. Tab. B.1 presents an overview of the different test series, for which data have been
extracted. All data are plotted in Fig. B.1.

The data have been limited to small scale specimens from recent investigations. Only
investigations with positive stress ratios are considered and only papers with thorough
description of specimen geometry. The only difference in the data considered here is
the specimen geometry, since this is the only parameter considered in the notch stress
approach.

Only specimens failing from the weld toe is considered and only welded specimens of
comparable quality in as-welded condition. The steel grades vary from S235 to S1100,
specimen thickness varies from 5 − 25mm and stress ratios vary from 0 to 0.5+. Run-outs
are included as well, but the main focus of this study is on the finite life region.

In many cases, the fatigue data were listed in the references; otherwise software assisted
extraction from the SN diagrams has been performed. The presentation of fatigue
data in Fig. B.1 is divided according to specimen type, in the nominal stress system.
Only the four most popular specimen types are considered; T-joints, double sided
transversal attachments (non-load carrying cruciform joints), butt joints and double sided
longitudinal attachments.

The relatively large scatter in the results is explained by different thickness, weld quality,
misalignment, stress ratio and so forth. The large scatter is considered positive in this
investigation, since a more general overview can be achieved. Fig. B.2 shows all specimens.

All data agree quite well with the FAT classes suggested by the IIW, and only very few
data points fall below the design S-N curves. It is clear that the T-joints show much
better results than suggested by the FAT80 curve. This is expected though, since they
are tested in bending, considering the positive effect of the steep stress gradient and little
to none negative effect from misalignment.
Paper B. Re-analysis of Fatigue Data for Welded Joints Using the Notch Stress
Approach 53

ID Ref Sy [MP a] t[mm] Kt R Loading Process

T-joints Haa1 [12] 420 20 2.80 0.1 bending SAW
Bud1 [16] 550-690 16 2.64 0.0-0.5 bending SAW
Sta1 [13] 420 20 2.91 0.1 bending SAW
Gal1 [23] 700 5 1.99 0.1 bending ?
Gal2 [23] 355 6 2.03 0.1 bending ?
Ped1 [25] 700 6 2.03 0.1 bending MAG
Tru1 [31] 420 20 2.73 0.1 bending ?
Transversal Lif1 [17] 700-900-1100 8 2.35 0.2 tension MAG
attachments Fos1 [18] 355-460-690 12 2.49 0.1-0.5 tension MAG
Fos2 [18] 355-460-690 12 2.57 0.1 tension MAG
Fos3 [18] 690 25 2.69 0.1 tension MAG
Son1 [26] 1100 8 2.20 0.0 tension MAG
Man1 [15] 355-700 12.5 2.72 0.1 tension MMA
Oht2 [30] 570 20 3.01 0.0-0.5 tension MMA
Muc1 [32] 460 13 2.51 0.1 tension MAG
Hut2 [34] 235-355 8 2.32 0.5 tension ?
Kud1 [35] 260 20 3.10 0.0 tension ?
Butt joints Lif2 [17] 700-900 8 1.60 0.2 tension MAG
Wan2 [20] 235-390-700 8 1.60 0.0-0.1 tension ?
Hut3 [34] 235-355 8 1.60 0.5 tension ?
Dic1 [21] 318 24 2.05 0.0 tension MMA
Oht1 [27] 284-579 20 1.95 0.0-0.5 tension SAW
Nak1 [28] 431 20 1.95 0.0-Sy tension MMA
Longitudinal Lif3 [17] 690-900-1100 8 3.42 0.2 tension MAG
attachments Lif4 [17] 690-900-1100 8 3.85 0.2 tension MAG
Haa2 [19] 355-700 8 3.73 0.1 tension ?
Hut1 [22] 700 8 3.73 0.1-0.5 tension ?
Wan1 [20] 235-390-700 8 2.69 0.0-0.1 tension ?
Mad1 [24] 355 13 3.32 0.1 tension MMA
Lop1 [14] 355-590 12 3.82 0.0 tension MAG
Dim1 [33] 333 12.7 3.62 0.1 tension MAG
Mor1 [29] 417 12 4.01 0.0 tension MAG

Table B.1: Extracted experimental fatigue data series. Kt is determined according

to the IIW recommendations for fatigue assessment using the notch stress approach
[2]

Conversion to the Notch Stress System

Using the notch stress approach, the stress concentration factor (SCF) of an arbitrary
welded joint can be determined using finite element analysis. Radaj et al. [1] explain how
the geometric SCF Kt corresponds to the fatigue effective SCF Kf , due to the fictitious
rounding of the notch.

Kf = Kt (Rref = 1mm). (B.1)

The notch approach thereby uses an idealized geometry, which makes the above statement
true. Conversion of the extracted fatigue data in the nominal stress system ∆σn to the
notch stress system ∆σk for a given specimen can thus be accomplished as follows.

∆σk = Kf · ∆σn . (B.2)

54

Figure B.1: Extracted fatigue data in the nominal stress system.

The stress concentration factors of all specimens are therefore determined using complete
FE models, according to IIW recommendations [2,4]. Here, the procedure is exemplified
using the specimens from Lagerqvist et al. [17], (Lif1). The FE analysis was performed
using mesh refinement in the area around the weld toe, as shown in Fig. B.3.

A reference radius of 1mm was used and flank angles of 45◦ for fillet welds and 30◦ for
butt welds, as recommended in [4]. A tensile nominal stress of 1MPa was applied to the
specimen, such that the maximum principal stress observed in the notch corresponds to
the SCF, see Fig. B.4. The principal stress hypothesis is used for all notch stress analysis.

No misalignment is considered directly in the FE analysis for the determination of the

SCFs. However, the SCF for the butt joints is multiplied by a stress magnification factor
km, since these specimens are very prone to misalignment. Hobbacher [4] suggests km =
1.10 for butt joints made in flat position in shop. This value has been applied here.

Extracted Fatigue Data in the Notch Stress System

The converted fatigue data are plotted in Fig. B.5. The immediate conclusion is that
the results agree quite well with the FAT225 curve for all fillet welded joints, but not so
Paper B. Re-analysis of Fatigue Data for Welded Joints Using the Notch Stress
Approach 55

Figure B.2: Specimen geometry and associated ID.

Figure B.3: Lif1-specimen: element size in the notch is 0.1mm. Only the weld toe
is considered here.

well for the butt joints. As expected, the results for the T-joints are somewhat above the
FAT225 curve, which can be explained by them being tested in bending.

For the double sided transversal and longitudinal attachments, the results agree
reasonably well with the FAT225 curve, however, some data points fall below the curve.
The reason for this is unclear, but the few specimens falling below the FAT225 curve
is assumed to suffer from some unfortunate conditions, e.g. misalignment or poor local
weld toe profile. It is noted, that the FAT225 curve is derived for ’welds with relatively
good quality toe profiles’, according to Fricke [2, p.13].

The results for many of the butt joints, on the other hand, lie significantly below the
FAT225 curve. This is expected to be due to the relatively mild notch present in butt
joints with little overfill. The stress concentration factor determined for these joints were
calculated to 1.6 − 2.0, whereas the stress concentration factor for the fillet welded joints
were in the range of 2.0 − 4.0. The problem is especially pronounced for thin butt joints,
e.g. in 8mm plate, which has a stress concentration factor of approximately 1.6.

It is well known that some low-SCF joints, such as butt joints, can be assessed in a
non-conservative manner using the notch stress approach. Fricke [2] therefore suggests a
correction for mild notches, i.e. assuming a notch factor Kw = σk /σhs of at least 1.6,
56

Figure B.4: The SCF of the specimen is determined by applying a nominal stress
of 1MPa to the FE model.

where σhs is the structural hot spot stress. However, in this investigation, the notch factor
was above 1.6 for all butt joints, and this correction was thus not applied.

Fatigue assessment of butt joints by the notch stress approach is investigated further in
the following section.

Notch Stress Assessment of Butt Joints

In order to obtain a more reliable basis for evaluating the notch stress approach for butt
joints, additional test results from many investigations of butt joints were extracted from
the S-N curve catalogue by Olivier and Ritter [36]. Only fully penetrated butt joints (I,
V and X joints) and only toe-failures are considered. Extreme values are ignored, i.e.
only series with a probability of occurrence in the interval 10 − 90% are considered here.
Series outside of this interval are presumably affected by severe weld defects on the one
side or a significant crack-initiation period on the other side. Run-outs are excluded as
well.

In Fig. B.6 the fatigue data for the butt joints are compared in the nominal (left) and
notch stress system (right). The fatigue data is converted using the formula for Kt in
Fig. B.7 (left) and km = 1.10 for misalignment. It is seen that the fatigue strength of
all butt joints are approximately identical in the nominal stress system, regardless of
the different specimen thickness. In the notch stress system, however, the thin joints
are assessed non-conservatively because of the very low SCF determined for these joints
(Kt < 2.0).

Sonsino et al. [37] also reports problems with fatigue assessment using the notch stress
approach considering thin/flexible welded joints, e.g. butt joints. They observed shallower
slopes for these particular joints and therefore suggest the use of m = 5 while maintaining
the FAT225 value. As it is seen in Fig. B.6 (right), this approach seems promising in the
high cycle area, but too conservative in the medium-to-low cycle area.

In Fig. B.7 (left), it is seen how the stress concentration factor decreases rapidly for
thin butt joints. The stress concentration factor determined using FEM is compared to
a formula by Anthes et al. [38], which shows a similar tendency. Accordingly, the notch
stress approach will estimate very high fatigue strength for thin butt joints.
Paper B. Re-analysis of Fatigue Data for Welded Joints Using the Notch Stress
Approach 57

Figure B.5: Fatigue data converted to the notch stress system.

The fatigue capacity of all butt joint specimens from this investigation and [36] are
plotted as a function of the thickness in Fig. B.7 (right). The fatigue capacity is plotted
instead of e.g. the characteristic fatigue strength, in order to avoid uncertainties due to
statistical treatment. The fatigue capacity is calculated assuming m = 3.0. For the notch
stress assessment, misalignment is considered using km = 1.10.

It is seen, that the fatigue capacity estimated by the notch stress approach can become
very non-conservative especially for thin butt joints. The experimental data in Fig. B.7
(right) does not show a significantly higher fatigue capacity for thin joints. There do not
seem to be any clear thickness dependency at all in this range. Or at least, there do not
seem to be any support for the tendency suggested by the notch stress approach for thin
butt joints.

Fig. B.7 furthermore shows the effects of a minimum notch factor Kw limit of 1.6 and 2.0,
respectively. It is seen, that the current recommendation of Kw = 1.6 only affects butt
joints thinner than approximately 7mm and has relatively little effect. If the minimum
notch factor is increased to Kw = 2.0 on the other hand, a much more conservative result
is obtained for thin butt joints and other low SCF joints.
58

Figure B.6: In the nominal stress system (left) the fatigue strength seems
independent of thickness. In the notch stress system (right), on the other hand,
the low SCF of thin specimens cause them to be assessed non-conservatively.

In most standards and e.g. the IIW recommendations [4], the thickness effect is only
considered for welded joints in plate thicker than 25mm. However, recent work e.g.
Gustafsson [39], reports higher fatigue strength in plates thinner than 25mm. He reports
a significant improvement in the fatigue strength for longitudinal attachments down to
3mm.

Ohta et al. [40] have studied the thickness effect of butt joints. They report that the
fatigue strength of butt joints in 9 and 40mm plate was similar when testing by cycling
down from the yield strength (σmax = σy ). On the other hand, if tested at R = 0, the
9mm specimens showed significantly higher fatigue strength. They suggest that thinner
butt joints have an apparent higher fatigue strength (when tested at low R-ratios), than
thick butt joints. This is only indirectly due to the thickness, but is because butt joints
in thicker plates better hold a high level of tensile residual stresses.

Conclusively, it seems that the fatigue strength for thin butt joints estimated by the
notch stress approach is too optimistic and there is a need for guidance in order to
alleviate this problem, e.g. by requiring an increased minimum notch factor of Kw = 2.0
or alternatively use a shallower slope of m = 5, as suggested by Sonsino et al. [37].

Further Observations
In the following, a minimum notch factor of Kw = 2.0 is assumed for thin butt joints. By
considering the fatigue data in the notch stress system, the difference in the geometry of
the specimens can be disregarded to some extend. At least the effect of different stress
concentration factors of the specimens can be disregarded; however, some specimens hold
Paper B. Re-analysis of Fatigue Data for Welded Joints Using the Notch Stress
Approach 59

Figure B.7: The thickness has a large influence on the stress concentration factor
of the idealized butt joint and thus the estimated fatigue strength. However, this
estimation does not correspond well with experimental data.

tensile residual stresses better than others and this effect cannot be disregarded. Still,
observations can be made based on a larger amount of fatigue data than usual.

It is clear from Fig. B.8, that the cutting off by the parent material curve FAT160·Kw ,
m = 5 seems unnecessary. However, high strength steel must be applied for high stress
ranges and the specimen edge roughness must be sufficiently fine. Sperle [41] showed that
the fatigue strength even for thermally cut edges in high strength steel can be significantly
higher, than the current IIW recommendation for parent material.

If excluding run-outs and using a slope of m = 3.0, the mean fatigue strength (Ps = 50%)
is FAT305. Assuming a log-normal distribution, the standard deviation of log(C) is 0.28,
which yields a design curve (Ps = 97.7%) of FAT199 (mean - 2 standard deviations).

Comparing the plots of Fig. B.1 and Fig. B.5, it seems that the nominal stress approach
is more conservative than the notch stress approach, since less data points fall below
the respective S-N curves. The reason for this is not clear, however, it seems that if the
FAT225 curve is reduced to FAT200, it gives approximately the same safety as observed
in the nominal stress system, see Fig. B.8.

Conclusions
The following conclusions are drawn based on re-analysis of a large amount of recent
fatigue data in the notch stress system using the principal stress hypothesis.

1. For most fillet-welded joints, the experimental fatigue data agrees reasonably well
with the current IIW guidance, i.e. using FAT225 S-N curve, except for few data
points.
60

Figure B.8: Fatigue data in notch stresses for all series pooled together.

2. FAT200 is proposed instead though, as it seems to give the same safety as observed
in the nominal stress system.
3. The current IIW recommendations can cause non-conservative assessment for thin
butt joints. Increasing the minimum notch factor from Kw = 1.6 to 2.0 is therefore
proposed to alleviate this.
4. The parent material limit of FAT160·Kw seems unnecessary if high strength steel
is applied.

Acknowledgements
The research work is funded by an Industrial PhD-grant from the Danish Agency for
Science, Technology and Innovation under the Danish Ministry for Science, Technology
and Innovation. IIW commission XIII members are gratefully acknowledged for their
comments and suggestions for this work.

References
1 D. Radaj, C.M. Sonsino, W. Fricke, Fatigue Assessment of Welded Joints by Local
Approaches, second ed., Woodhead Publishing Ltd., Cambridge, 2006.
2 W. Fricke, Guideline for Fatigue Assessment by Notch Stress Analysis for Welded
Structures, IIW Doc. XIII-2240r1-08/XV-1289r1-08, 2008.
Paper B. Re-analysis of Fatigue Data for Welded Joints Using the Notch Stress
Approach 61

3 C.M. Sonsino, Suggested Allowable Equivalent Stresses for Fatigue Design of

Welded Joints According to the Notch Stress Concept with the Reference Radii
rref=1.00 and 0.05mm, IIW Doc. XIII-2216-08/XV1285-08, 2008.
4 A. Hobbacher, Recommendations for Fatigue Design of Welded Joints and
Components, IIW Doc. IIW-1823-07, 2008.
5 D. Radaj, Design and Analysis of Fatigue Resistant Welded Structures, Woodhead
Publishing Ltd., Cambridge, 1990.
6 R. Olivier, V.B. Köttgen, T. Seeger, Schweissverbindung I, FKM Forschungshefte
143, Frankfurt/M, 1989.
7 R. Olivier, V.B. Köttgen, T. Seeger, Schweissverbindung II, FKM Forschungshefte
180, Frankfurt/M, 1994.
8 W. Fricke, A. Kahl, Local Stress Analysis and Fatigue Assessment of Bracket Toes
Based on Measured Weld Profile, IIW-Doc. XIII-2166-07/XV-1253-07, 2007.
9 Z. Barsoum, Residual Stress Analysis and Fatigue Assessment of Welded Steel
Structures, Doctoral Thesis, KTH, Stockholm, 2008.
10 B. Kranz, C.M. Sonsino, Verification of the Notch Stress Concept for the Reference
Radii of rref=1.00 and 0.05mm, IIW-Doc. XIII-2274-09/XV-1312-09, 2009.
11 W. Fricke, H. Paetzold, Full-scale Fatigue Tests of Ship Structures to Validate the
SN-approaches for Fatigue Strength Assessment, IIW-Doc. XIII-2279-09/XV-1321-
09, 2009.
12 P.J. Haagensen, IIW’s Round Robin and Design Recommendations for Improve-
ment Methods, Proc. IIW 50th Annual Assembly Conference, San Francisco,
Welding Research Council Inc., New York, 1997.
13 E.S. Statnikov, V.O. Muktepavel, A. Blomqvist, Comparison of Ultrasonic Impact
Treatment (UIT) and Other Fatigue Life Improvement Methods, Welding in the
World 46, pp.28-39, 2002.
14 L. Lopez Martinez, A. Blom, Influence of Life Improvement Methods on Different
Steel Grades under Fatigue Loading, Fatigue Design 1995. Proc. VTT Symposium
157, Helsinki, 5-8. Sept., 1995.
15 S. Manteghi, S.J. Maddox, Methods for Fatigue Life Improvement of Welded Joints
in Medium and High Strength Steels, IIW-Doc. XIII-2006-04, 2004.
16 S. Budano, M. Kuppers, H. Kaufmann, A.M. Meisozo, C. Davis, Application of
High Strength Steel Plates to Welded Deck Components for Ships and Bridges
Subjected to Medium/High Service Loads, EUR22571EN, European Commission,
Brussels, 2007.
17 O. Lagerqvist, M. Clarin, J. Gozzi, B. Völling, D. Pak, J. Stötzl, H.P. Lieurade,
B. Depale, I. Huther, S. Herion, J. Bergers, R.M. Martsch, M. Carlsson, A.
Samuelsson, C. Sonander, LiftHigh - Efficient Lifting Equipment with Extra High-
Strength Steel, European Commission, EUR22569EN, Brussels, 2007.
18 U. Kuhlmann, A. Dürr, J. Bergmann, R. Thumser, Effizienter Stahlbau aus
höherfesten Stählen unter Ermüdungsbeanspruchung, Forschungsvorhaben P620,
FOSTA, Verlag und Vertriebsgesellschaft GmbH, Düsseldorf, 2006.
62

19 P.J. Haagensen, Ø. Alnes, Progress Report on IIW WG2 Round Robin Fatigue
Testing Program on 700MPa and 350MPa YS Steels, IIW Doc. XIII-2081-05, 2005.
20 T. Wang, D. Wang, L. Huo, Y. Zhang, Discussion on Fatigue Design of Welded
Joints Enhanced by Ultrasonic Peening Treatment (UPT), International Journal
of Fatigue, 31, pp. 644-650, 2009.
21 T. Dickerson, C. Moura Branco, Weld Improvement Methods for Low Cycle Fatigue
Applications, EUR17823, European Commission, Brussels, 1997.
22 I. Huther, Y. Suchier, H.P. Lieurade, Fatigue Behaviour of Longitudinal Non-load
Carrying Joints Improved by Burr Grinding, TIG dressing, IIW Doc. XIII-2108-06,
2006.
23 A. Galtier, E.S. Statnikov, The Influence of Ultrasonic Impact Treatment on
Fatigue Behaviour of Welded Joints in High-Strength Steel, Welding in the World,
48 (5/6), 2004.
24 S.J. Maddox, Improving the Fatigue Strength of Toe Ground Welds at the End of
Longitudinal Stiffeners, IIW-Doc. XIII-2156-07, 2007.
25 M.M. Pedersen, O.Ø. Mouritsen, M.R. Hansen, J.G. Andersen, J. Wenderby,
Comparison of Post Weld Treatment of High Strength Steel Welded Joints in
Medium Cycle Fatigue, IIW Doc. XIII-2272-09, 2009.
26 C. Sonander, Ermüdung von geschweissten Kreuzstössen aus Weldox 1100, Stahlbau,
69 (4), 2000.
27 A. Ohta, N. Suzuki, Y. Maeda, Shift of S N Curves with Stress Ratio, Welding in
the World, 47 (1/2), 2003.
28 Nakamura, Nishijima and A. Ohta, A Method for Obtaining Conservative S- N
Data for Weld Structures, Journal of Testing and Evaluation, JTEVA, 16 (3),
pp.280-285, 1988.
29 T. Mori, T. Inomata, Influence of Grinding on Fatigue Strength of Out of Plane
Gusset Welded Joint, IIW Doc. XIII-1970-03, 2003.
30 A. Ohta, Y. Maeda, N. Suzuki, Residual Stress Effect on Fatigue Strength of Non-
Load-Carrying Cruciform Welded Joints of SM570Q Steel for Welded Structures,
IIW Doc. XIII-1921-02, 2002.
31 V.I. Trufiakov, E.S. Statnikov, P.P. Mikheev, A.Z. Kuzmenko, The Efficiency of
Ultrasonic Impact Treatment for Improving the Fatigue Strength, IIW Doc. XIII-
1745-98, 1998.
32 A. Muck, Ertüchtigung von ermüdungsbeanspruchten Schweißverbindungen durch
Anwendung von Ultrasonic Impact Treatment (UIT), Diplomarbeit, Universität
Stuttgart, 2005.
33 S.D. Dimitrakis, F.V. Lawrence, Improving the Fatigue Performance of Fillet Weld
Terminations, Fatigue Fract Engng Mater Struct 24, 429-438, 2001.
34 I. Huther, V. Minard, Y. Royer, H.P. Lieurade, Burr Grinding Effect on the Fatigue
Strength as Regard to Initial Weld Quality, IIW Doc. XIII-2038-04, 2004.
35 Y. Kudryavtsev, J. Kleiman, A. Lugovskoy, L. Lobanov, V. Knysh, O. Voitenko,
G. Prokopenko, Rehabilitation and Repair of Welded Elements and Structures by
Ultrasonic Peening, Welding in the World, 51 (7/8), 2007.
Paper B. Re-analysis of Fatigue Data for Welded Joints Using the Notch Stress
Approach 63

36 R. Olivier, W. Ritter, Wöhlerlinienkatalog für Schweissverbindungen aus Baustählen,

DVS GmbH, Düsseldorf, 1979.
37 C.M. Sonsino, T. Bruder, J. Baumgartner, SN-curves for Welded Thin Joints -
Suggested Slopes and FAT-values for Applying the Notch Stress Concept with
Various Reference Radii, IIW-Doc. XIII-2280-09, 2009.
38 R.J. Anthes, V.B. Köttgen, T. Seeger, Kerbformzahlen von Stumpfstössen und
Doppel-T-Stössen Schweissen und Schneiden, 45 (12), pp.685-688, 1993.
39 M. Gustaffsson, A Study on Thickness Effect on Fatigue in Thin Welded High
Strength Steel Joints, Steel Research International, 77 (12), 2006.
40 A. Ohta, T. Mawari, N. Suzuki, Evaluation of Effect of Plate Thickness on Fatigue
Strength of Butt Welded Joints by a Test Maintaining Maximum Stress at Yield
Strength, Engineering Fracture Mechanics, 37 (5), pp.987-993, 1990.
41 J.O. Sperle, Influence of Parent Metal Strength on the Fatigue Strength of Parent
Material with Machined and Thermally Cut Edges, IIW-Doc. XIII-2174-07, 2007.
64
 C
Experience with the Notch Stress
Approach for Fatigue Assessment
of Welded Joints

M.M. Pedersen, O.Ø. Mouritsen, M.R. Hansen, J.G. Andersen

Proceedings of the Swedish Conference on Light Weight Optimized Welded Structures,
Borlänge, Sweden, pp. 122-133, March 24-25, 2010.

Abstract. In this paper, fatigue assessment using the notch stress approach is discussed
based on re-analysis of many fatigue test results and experience from practical applica-
tion. Three topics are treated; evaluation of the fatigue strength for as-welded details
(FAT225) in the notch stress system, problems regarding assessment of mild-SCF details
and a novel proposal for extension of the notch stress approach for use with post-weld
treated details.

Introduction
The notch stress approach has received much attention lately due to the increasing
available computational power and the need for assessing more increasingly complex
geometries. The approach is very flexible in the sense that both the toe and the root of
all types of welded joints can be assessed using a single S-N curve.

Radaj et al. [1] presents a thorough review of the history of the approach. Fricke [2] gives
practical guidelines for the notch modelling and stress analysis and Sonsino [3] proposes
S-N curves to be used under different conditions. The approach is included in the IIW
fatigue design recommendations by Hobbacher [4].

The notch stress approach correlates the stress range in a fictitious rounding in the weld
toe or root to the fatigue life using a single S-N curve. The notch stress is typically
obtained using finite element models with the reference radius of 1mm in order to avoid
the stress singularities in sharp notches. The approach is schematically illustrated in
Fig. C.1. In this paper, notch stresses are calculated using the first principal stress
(denoted by index PS in diagrams).
66

Reference Elements Principal ∆σ

radius R1 <0.25mm stresses

FAT225

Nf

Figure C.1: Schematic principles of fatigue assessment using the notch stress
approach.

The approach is based on the work by Radaj [5] and modified by Seeger and co-workers,
see Olivier et al. [6,7]. Here, the reference radius of R1 is determined as a mean value
and the design fatigue strength (FAT225) is derived from experiments.

Fatigue strength in the notch stress system

The IIW recommends the use of FAT225 when considering the fatigue strength of welded
steel joints using the notch stress approach. This value was derived based on the results
of a large experimental investigation of different T- and Y-joints [6,7] and further re-
analysis of the results assembled by Olivier and Ritter [8]. Kranz and Sonsino [9] explain
the assumptions that led to the derivation of the FAT225 fatigue class.

Pedersen et al. [10] carried out a re-analysis using the notch stress approach, and found
the FAT225 to be slightly non-conservative, as shown in Fig. 2. The re-analysis was
carried out to evaluate the notch stress approach according to the IIW and provide
further experimental evidence to the approach.

The fatigue data was extracted from recent publications and converted to the notch
stress system by scaling the nominal stress range with the stress concentration factor
determined by FE analysis. Only tests carried out in the as-welded condition and under
positive stress ratio (R ≥ 0) were considered. The steel grade varied from S235-S1100
and the thickness was 5 − 25mm.

Four different specimen types were considered; T-joints, double-sided transversal attach-
ments, butt joints and double-sided longitudinal stiffeners. The T-joints are assessed
very conservatively because they are tested in bending. In the nominal stress system, the
fatigue data agree quite well with the FAT classes suggested by the IIW [4], therefore,
the quality of the specimens is considered to be representative of normal quality.

If excluding run-outs and using a slope of m = 3.0, the mean fatigue strength (PS =
50%) is FAT305. The standard deviation of log(C) is 0.28, and thus the design curve
can be calculated (PS = 97.7%) to FAT199 (mean - 2 standard deviations). It was
therefore suggested to reduce the fatigue strength from FAT225 to FAT200 to achieve
approximately the same safety as observed in the nominal stress system.
Paper C. Experience with the Notch Stress Approach for Fatigue Assessment of
Welded Joints 67

Notch stresses
Haa1 Lif1
Lif2
1500 R≥R0,0,n=767
n=767
Lif3
stdvlog(C) =0.28
Fos1 Wan 2
Bud 1
Lif4
1000
Hut3
Fos2
700 m=3 Sta1 Haa2
∆σk,PS [MPa]

r ref=1mm
→ → Fos3 Dic1
K w
2.0
500 → → Gal1
Hut1

→→ → →
400 →→ → → Oht1 Wan 1
→ → → → Gal2
Son1
→
→ →→
→305 → →
→
300 → →
→ → → Mad1
→
→
Ped 1
Man1
Nak1
225
200 199
→
→→ Lop1
Tru1 Oht2
→
→ → Dim1
→
Muc1
100 4 5 6 7 8
10 10 10 10 10 Mor1
N
Nf [cycles]
f
Hut2

Figure C.2: FAT225 seems to be too optimistic and the reduced FAT200 is therefore
proposed in order to achieve approximately the same safety as in the nominal stress
system [10].

Mild SCF joints

It is well known that the notch stress approach can lead to non-conservative assessment
of mild SCF joints, e.g. thin butt joints, thus there is a need for further insight into the
problem. The subject is discussed in terms of thin butt joints and a mild SCF crane
detail.

Butt Joints
For butt joints, the IIW suggests the idealized geometry shown in Fig. C.3 for notch
stress analysis. The problem with this geometry is that the resulting SCF will be very,
low especially for thin joints, e.g. Kt ≈ 1.6 for 8mm thickness.

R1 30°

F t F

Figure C.3: Idealized geometry for notch stress assessment of butt joints.

In Fig. C.4 fatigue test data for many butt joints are compared in the nominal (left) and
notch stress system (right). The fatigue data is converted using the formula Kt (t) = 1.055·
68

t0.216 derived in [10] and km = 1.10 to consider a stress increase due to misalignment, as
suggested by Hobbacher [4].

Figure C.4: Thin butt joints can be assessed non-conservatively by the notch stress
approach [10].

It is seen that the fatigue strength of all butt joints are approximately identical in the
nominal stress system, regardless of the different specimen thickness. In the notch stress
system, however, many thin joints are assessed non-conservatively because of the very low
SCF determined for these joints. A remedial action was therefore suggested, i.e. requiring
a minimum notch factor of Kw ≥ 2.0, instead of the current Kw ≥ 1.6 given by the IIW
[2].

Sonsino et al. [11] also reports problems with fatigue assessment using the notch stress
approach considering thin/flexible welded joints, e.g. butt joints. They observed shallower
slopes for these particular joints and therefore suggest the use of a shallower slope, m =
5.0, while maintaining the FAT225 value. As it is seen in Fig. C.4 (right), this approach
seems promising in the high cycle area, but too conservative in the medium-to-low cycle
area.

Crane Detail
Another type of mild SCF joint where the notch stress approach can lead to non-
conservative results is the crane detail shown in Fig. C.5. The fatigue critical location
is the weld toe in front of the termination of the weld seam around the reinforcement
plate. This type of reinforcement is often used in high strength steel structures, where
a concentrated load is distributed into the main plate, e.g. in a revolute joint, since
the bearing load will otherwise be too high. There are many examples of complicated
geometries used in order to 1) reduce the stress concentration factor by softening/tapering
out the reinforcement and 2) move the fatigue critical location to an area with reduced
loading, e.g. near the centreline of beams.
Paper C. Experience with the Notch Stress Approach for Fatigue Assessment of
Welded Joints 69

Figure C.5: a) Crane with reinforcement detail, b) actual weld, c) idealized weld.

Rasmussen [12] investigated the considered detail experimentally and numerically, Tab.
C.1 and Fig. C.6. This detail is a good example of a weld, where the geometry of the weld
itself has a significant influence on the fatigue strength and therefore has to be taken into
account, e.g. by using the notch stress approach.

The welded specimens were manufactured by the Danish crane manufacturer HMF A/S
using 5mm thick S700 and normal quality MAG welding. Fatigue testing were carried
out at Aalborg University at a stress ratio of R = 0.1 and with special attention to the
medium cycle area.

∆F [kN ] ∆σn [M P a] Nf [cycles]

168 509 18.414
F
168 509 12.966
88 267 183.228
86 261 131.935
115 348 48.740
63 191 306.932
164 497 12.805
164 497 11.148
87 264 346.114
F
115 348 64.980
64 194 242.445

Figure C.6: Specimen geometry [12]. Table C.1: Fatigue testing results [12].

For such a detail, the question arises, how to idealize the local weld geometry? Assuming
a reference radius of R1 seems logic, but what about the remainder of the weld, especially
the flank angle?
70

The results are presented in Fig. C.7 in the notch stress system using two different
modelling techniques - real flank angle (15◦ ) and assumed flank angle (45◦ ). Using the
real flank (15◦ ), the SCF is determined to Kt = 1.72 and using the assumed 45◦ flank
angle, the SCF is Kt = 2.53. It is clear from Fig. C.7, that the results fit the FAT225 curve
best, if the assumed 45◦ flank angle is applied. It is however a quick-and-dirty solution for
achieving conservative results using the notch stress approach for this particular detail.

Notch
Notch stress stress
system: system
R=0.1, S700, t=5mm
1500
Kt=1.72
R1
1000 15°
900
800
700 Kt=2.53
600 R1
45°
500
'V k,PS [MPa]

400 A-A

300

225
200 203

Kt 138

100 A A

4 5 6 7
10 10 10 10
Nf [cycles]

Figure C.7: Fatigue testing results for crane detail, incl. design curves (PS = 97.7%)
for two different modelling techniques compared with the FAT225 curve.

The example shows how important the idealization of the weld geometry is to the final
result of the fatigue assessment using the notch approach. It is clear, that experience
with the approach is needed, and it is highly recommended to compare results with other
(local) approaches.

Extension to post weld treatment

Using the notch stress approach for fatigue assessment of post weld treated details is
a subject still under debate. The IIW recommendations by Fricke [2] suggest using a
model of the real toe radius +1mm and evaluate the principal stress range here against
the FAT200 SN curve. However, it is also stated, that this approach is only valid for
relatively sharp notches, i.e. with a toe radius of R1 − 3mm, and that the approach has
not been verified.

Fig. C.8 shows examples of increased weld toe radii due to different post weld treatments.
Typically, when performing post weld treatment, a large toe radius is desired and e.g.
grinding and TIG dressing will generally leave a radius larger than R3.
Paper C. Experience with the Notch Stress Approach for Fatigue Assessment of
Welded Joints 71

Figure C.8: Typical toe profiles for post-weld treated welds compared to as welded
condition [23].

Fig. C.9 shows the results of measurements of weld toe profiles published in the literature.
It is clear, that the toe radius will vary a lot in production and will be difficult to
estimate in a design situation. The question therefore arises, which ’real’ radius to use
- the mean, or some lower value? Since the value of the toe radius typically follows a
Gaussian distribution, the minimum value is not well defined.

As welded Burr ground TIG dressed HFP treated

2
Probability density

[15] [13]
[14][20] [14]
[19] [13]
1 [14]
[18] [18] [17] [13]
[21][22] [17] [20] [21]
[18][19]
[16][17]
0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Toe radius, R [mm]

Figure C.9: Measurements of toe radius for as welded and post weld treated welds
toes [13-22].

For welds in the as-welded condition, the average toe radius is approximately R1 − R1.5,
depending on the quality level. For burr ground welds, the average is in the order of R4,
depending on the chosen burr and the skill of the operator. The radius of TIG dressed
welds varies the most - from R0 to R12+. This radius depends heavily on the skill of the
operator and position of the weld relative to gravity. The average is approximately R6.
Welds treated by high frequency peening (HFP) also achieves varying radii, but not as
random as TIG - here the average is around R2.5.

Conclusively, if the approach suggested by IIW [2], should be applied with the average
radius, it will only be applicable for HFP post weld treatment, since BG and TIG typically
results in radii larger than R3.

Weich [24] recently proposed an approach for assessing HFP treated welded joints using
the notch stress approach. The idea is to consider an effective stress ratio and hereby
′
apply an improvement factor of up to f (Ref f ) = 1.6 for effective stress ratios of
′
Ref f < −1. The effective stress ratio is based on a superposition of the local stresses and
compressive residual stresses introduced by the treatment. This approach is presumably
72

quite accurate, but difficult to apply in a design situation, because the level of compressive
residual stresses introduced by the treatment will be unknown.

A New Approach
An alternative procedure for assessing post weld treated joints using the notch stress
approach is investigated in the following. The basic idea is to maintain the reference
radius at R1 and use a higher FAT class for fatigue assessment of post weld treated joints.
This is chosen such that an engineer can easily assess whether post weld treatment is
necessary and sufficient, using the same FE model as used for assessing the joint in as-
welded state. This approach seems very practical, but also has the obvious drawback, that
the computed notch stress becomes even more of a model-number, than in the as-welded
case, because the model geometry differs more from the actual geometry.

The higher FAT classes are derived based on a large collection of experimental results for
post weld treated details. Only burr grinding, TIG dressing and high frequency peening
(HFP), e.g. UIT, are considered and only fillet welded joints. Post weld treated butt
joints are considered unsuitable for assessment using the notch stress approach, since the
notch stress concentration may be completely removed, as in the case of disc grinding.

Most test results considered are for high strength steel, S355-S1100, except some in the
series denoted Hut2, Kud1 and Wan1 which also considers mild steel. Only fatigue tests
carried out under positive stress ratios are considered and only publications with thorough
description of the specimen geometry.

By considering the fatigue data for post weld treatment in the notch stress system, the
difference in the geometry of the test specimens can be disregarded to some extend.
At least the effect of different stress concentration factors of the specimens can be
disregarded. However, some specimens can contain higher levels of tensile residual stresses
than others and this effect cannot be disregarded. Still, suggestions for design curves can
be given based on a larger volume of fatigue data than usual.

Traditional statistical treatment of the experimental results yields poor results because
of very large scatter in the data. This is due to different testing conditions, quality of the
post weld treatment, thickness etc. The suggested FAT classes are therefore not directly
calculated but simply given based on practical estimation.

T-joints Transversal attachments Longitudinal attachments

ID Ref Sy t Kt ID Ref Sy t Kt ID Ref Sy t Kt
Haa1[25] 420 20 2.80 Lif1 [30]700-1100 8 2.35 Lif3 [30]690-1100 8 3.42
Bud1[26]550-690 16 2.64 Fos1 [14] 355-690 12 2.49 Lif4 [30]690-1100 8 3.85
Sta1 [27] 420 20 2.91 Fos2 [14] 355-690 12 2.57 Haa2 [34] 355-700 8 3.73
Gal1 [28] 700 5 1.99 Fos3 [14] 690 25 2.69 Hut1 [18] 700 8 3.73
Gal2 [28] 355 6 2.03 Son1 [31] 1100 8 2.20 Wan1[35] 235-700 8 2.69
Ped1 [23] 700 6 2.03 Man1[32] 355-700 12.5 2.72 Wei2 [36] 690 16 2.40
Tru1 [29] 420 20 2.73 Hut2 [19] 235-355 8 2.32 Mad1[37] 355 13 3.32
Kud1[33] 260 20 3.10 Lop1 [38] 355-590 12 3.82
Dah1 [39] 355-590 12 3.82
Mor1 [20] 417 12 4.01

Table C.2: Extracted fatigue data series. Kt is determined according to [2].

Paper C. Experience with the Notch Stress Approach for Fatigue Assessment of
Welded Joints 73

Burr Grinding and TIG Dressing

Grinding reduces the stress concentration factor of welded joints and removes included
defects in the ground area. The experimental data suggests parallel shifting of the S-N
curve upwards and FAT300 is suggested, as shown in Fig. C.10. This corresponds to an
increase of approximately a factor of 1.3 on the fatigue strength. This is the same factor
as suggested by the IIW [40] for mild steel joints in the nominal stress system.

Figure C.10: Fatigue data for burr grinding (left) and TIG dressing (right)
compared to the as-welded design curve in the notch stress system.

TIG dressing improves the welded joint in much the same way as grinding, i.e. the stress
concentration is reduced and defects are removed. The experimental data also supports
the use of the same S-N curve for both ground and TIG dressed welded joints, i.e. FAT300.

High Frequency Peening (HFP) Treatment

The following different types of high frequency peening treatments are described in the
literature; ultrasonic impact treatment (UIT) [14], ultrasonic peening (UP) [29], high
frequency impact treatment (HiFIT) [36] and ultrasonic needle peening (UNP) [41].
UIT is based on magnetostriction, UP uses a piezoelectric transducer, whereas HiFIT
is pneumatically actuated.

Although they are different processes, their properties and resulting improvement in
fatigue strength appear to be similar. Here, the treatments are considered equal and
denoted HFP treatments.

The HFP treatments improve the fatigue strength of the welded joint in several ways
by plastic deformation of the weld toe. Firstly, the tensile residual stress state present
in the weld seam is relieved and beneficial compressive residual stresses are introduced.
Secondly, the sharp notch in the weld toe is blunted and the treatment leaves behind
74

a smooth trace with a radius of 2 − 3mm, see Fig. C.8. Finally, the surface material is
mechanically hardened, which locally increases the fatigue strength of the material in the
notch [42].

Since peening treatments improve the fatigue strength of welded joints primarily based
on the introduction of compressive residual stresses, a flatter S-N curve is expected. As
evident from Fig. C.11, a rotated S-N curve using FAT360 with the slope of m = 5.0, fits
the experimental data well.

Figure C.11: Fatigue data for HFP treatment compared to the as-welded design
curves in the notch stress system.

Due to the flatter slope, the HFP design curve is lower than the as-welded design curve for
N < 50, 000. However, it is clear that the lack of further experimental data, especially in
the low cycle area, limits the reliability of the suggested design curve here. It is not likely,
that the HFP treatment decreases the fatigue strength in the low cycle area; therefore,
the as-welded design curve could be applied here.

Conclusions
The following conclusions are drawn based on re-analysis of fatigue data and practical
experience with the notch stress approach.

1. For fatigue assessment using the notch stress approach, it is proposed to reduce the
fatigue strength to FAT200 for as-welded normal quality, based on a large amount
of experimental data.
2. For mild SCF joints, it is suggested to use a minimum notch factor of Kw ≥ 2.0
instead of the current Kw ≥ 1.6. Alternatively, conservative results can be achieved
by modelling the flank angle steeper than the real flank angle.
Paper C. Experience with the Notch Stress Approach for Fatigue Assessment of
Welded Joints 75

3. For fatigue assessment of post-weld treated details, an approach is suggested where

the stress analysis is carried out identically to as-welded details, i.e. the reference
radius of R1 is maintained, but the FAT class is increased.
(a) For both burr ground and TIG dressed details, a FAT class of FAT300 using
m = 3.0 is suggested.
(b) For HFP treated details, e.g. UIT, FAT360 is suggested with a flatter slope
of m = 5.0.

Acknowledgements
The research work is funded by HMF A/S and an industrial PhD-grant from the Danish
Ministry for Science, Technology and Innovation.

References
1 Radaj D., Sonsino C.M., Fricke W., Fatigue Assessment of Welded Joints by Local
Approaches, 2nd ed., Woodhead Publishing Ltd., Cambridge, 2006.
2 Fricke W., Guideline for Fatigue Assessment by Notch Stress Analysis for Welded
Structures, IIW Doc. XIII-2240r1-08/XV-1289r1-08, 2008.
3 Sonsino C.M., Suggested Allowable Equivalent Stresses for Fatigue Design of
Welded Joints According to the Notch Stress Concept with the Reference Radii
rref=1.00 and 0.05mm, IIW Doc. XIII-2216-08/XV1285-08, 2008.
4 Hobbacher A., Recommendations for Fatigue Design of Welded Joints and
Components, IIW Doc. IIW-1823-07, 2008.
5 Radaj D., Design and Analysis of Fatigue Resistant Welded Structures, Woodhead
Publishing Ltd., Cambridge, 1990.
6 Olivier R., Köttgen V.B., Seeger T., Schweissverbindung I, FKM Forschungshefte
143, Frankfurt/M, 1989.
7 Olivier R., Köttgen V.B., Seeger T., Schweissverbindung II, FKM Forschungshefte
180, Frankfurt/M, 1994.
8 Olivier R., Ritter W., Wöhlerlinienkatalog für Schweissverbindungen aus Baustählen,
DVS GmbH, Düsseldorf, 1979.
9 Kranz B., Sonsino C.M., Verification of the Notch Stress Concept for the Reference
Radii of Rref = 1.00 and 0.05mm, IIW Doc. XIII-2274-09, 2009.
10 Pedersen M.M., Mouritsen O.Ø., Hansen M.R., Andersen J.G., Wenderby J., Re-
analysis of Fatigue Data for Welded Joints Using the Notch Stress Approach,
accepted for publication in International Journal of Fatigue, 2010.
11 Sonsino C.M., Bruder T., Baumgartner J., SN-curves for Welded Thin Joints -
Suggested Slopes and FAT-values for Applying the Notch Stress Concept with
Various Reference Radii, IIW-Doc. XIII-2280-09, 2009.
12 Rasmussen L.V., Levetidsbestemmelse og analyse af typisk svejst krandetalje i
højstyrkestål (Danish), M.Sc. thesis, Inst. of Mechanical Engineering, Aalborg
University, 2008.
76

13 Lieurade H.P., Huther I., Lefebvre F., Effect of Weld Quality and Post Weld
Improvement Techniques on the Fatigue Resistance of Extra High Strength Steels,
IIW Doc. XIII-2184-07, 2007.
14 Kuhlmann U., Dürr A., Bergmann J., Thumser R., Effizienter Stahlbau aus
höherfesten Stählen unter Ermüdungsbeanspruchung, Forschungsvorhaben P620,
FOSTA, Verlag und Vertriebsgesellschaft GmbH, Düsseldorf, 2006.
15 Jonsson B.J., Barsoum Z., Ghavi Bazou A., Influence From Weld Position on Fillet
Weld Quality, IIW Doc. XIII-2273-09, 2009.
16 Barsoum Z., Residual Stress Analysis and Fatigue Assessment of Welded Steel
Structures, Ph.D. Thesis, KTH, Stockholm, 2008.
17 Moura Branco C., Gomes E.C., Development of Fatigue Design Curves for Weld
Improved Joints, Fatigue Design 1995. Proceedings VTT Symposium 157, Helsinki,
5-8th Sept.1995.
18 Huther I., Suchier Y., Lieurade H.P., Fatigue Behaviour of Longitudinal Non-load
Carrying Joints Improved by Burr Grinding, TIG dressing, IIW Doc. XIII-2108-06,
2006.
19 Huther I., Minard V., Royer Y., Lieurade H.P., Burr Grinding Effect on the Fatigue
Strength as Regard to Initial Weld Quality, IIW Doc. XIII-2038-04, 2004.
20 Mori T., Inomata T., Influence of Grinding Method on Fatigue Strength of Out-
of-Plane Gusset Welded Joints, IIW Doc XIII-1970-03, 2003.
21 Tominaga T., Matsuoka K., Sato Y., Suzuki T., Fatigue Improvement of Weld
Repaired Crane Runway Girder by Ultrasonic Impact Treatment, IIW Doc. XIII-
2170-07, 2007.
22 Lee C-H., Chang K-H., Jang G-C, C-Y. Lee, Effect of Weld Geometry on the
Fatigue Life of Non-load-Carrying Fillet Welded Cruciform Joint, Engineering
Failure Analysis, 16, pp.849-855, 2009.
23 Pedersen M.M., Mouritsen O.Ø., Hansen M.R., Andersen J.G., Wenderby J.,
Comparison of Post Weld Treatment of High Strength Steel Welded Joints in
Medium Cycle Fatigue, IIW Doc. XIII-2272-09, 2009.
24 Weich I., Edge Layer Condition and Fatigue Strength of Welds Improved by
Mechanical Post Weld Treatment, IIW Doc. XIII-2265-09, 2009.
25 Haagensen P.J., IIW’s Round Robin and Design Recommendations for Improve-
ment Methods, Proc. IIW 50th Annual Assembly Conference, San Francisco,
Welding Research Council Inc., New York, 1997.
26 Budano S., Kuppers M., Kaufmann H., Meisozo A.M., Davis C., Application of
High Strength Steel Plates to Welded Deck Components for Ships and Bridges
Subjected to Medium/High Service Loads, EUR22571EN, European Commission,
Brussels, 2007.
27 Statnikov E.S., Muktepavel V.O., Blomqvist A., Comparison of Ultrasonic Impact
Treatment (UIT) and Other Fatigue Life Improvement Methods, Welding in the
World 46, pp.28-39, 2002.
28 Galtier A., Statnikov E.S., The Influence of Ultrasonic Impact Treatment on
Fatigue Behaviour of Welded Joints in High-Strength Steel, Welding in the World,
48 (5/6), 2004.
Paper C. Experience with the Notch Stress Approach for Fatigue Assessment of
Welded Joints 77

29 Trufiakov V.I., Statnikov E.S., Mikheev P.P., Kuzmenko A.Z., The Efficiency of
Ultrasonic Impact Treatment for Improving the Fatigue Strength, IIW Doc. XIII-
1745-98, 1998.
30 Lagerqvist O., Clarin M., Gozzi J., Völling B., Pak D., Stötzl J., Lieurade H.P.,
Depale B., Huther I., Herion S., Bergers J., Martsch R.M., Carlsson M., Samuelsson
A., Sonander C., LiftHigh - Efficient Lifting Equipment with Extra High-Strength
Steel, European Commission, EUR22569EN, Brussels, 2007.
31 Sonander C., Ermüdung von geschweissten Kreuzstössen aus Weldox 1100, Stahlbau,
69 (4), 2000.
32 Manteghi S., Maddox S.J., Methods for Fatigue Life Improvement of Welded Joints
in Medium and High Strength Steels, IIW-Doc. XIII-2006-04, 2004.
33 Kudryavtsev Y., Kleiman J., Lugovskoy A., Lobanov L., Knysh V., Voitenko O.,
Prokopenko G., Rehabilitation and Repair of Welded Elements and Structures by
Ultrasonic Peening, Welding in the World, 51 (7/8), 2007.
34 Haagensen P.J., Alnes Ø., Progress Report on IIW WG2 Round Robin Fatigue
Testing Program on 700MPa and 350MPa YS Steels, IIW Doc. XIII-2081-05, 2005.
35 Wang T., Wang D., Huo L., Zhang Y., Discussion on Fatigue Design of Welded
Joints Enhanced by Ultrasonic Peening Treatment (UPT), International Journal
of Fatigue, 31, pp. 644-650, 2009.
36 Weich I., Ermüdungsverhalten mechanisch nachbehandelter Schweißverbindungen
in Abhängigkeit des Randschichtzustands, Ph.D. Thesis, Technischen Universität
Brauschweig, 2008.
37 Maddox S.J., Improving the Fatigue Strength of Toe Ground Welds at the End of
Longitudinal Stiffeners, IIW-Doc. XIII-2156-07, 2007.
38 Lopez Martinez L., Blom A., Influence of Life Improvement Methods on Different
Steel Grades under Fatigue Loading, Fatigue Design 1995. Proc. VTT Symposium
157, Helsinki, 5-8. Sept., 1995.
39 Dahle T., Design Fatigue Strength of TIG-dressed Welded Joints in High-Strength
Steels Subjected to Spectrum Loading, ABB Corporate Research, S-721, 78,
Sweden, 1998.
40 Haagensen P.J., Maddox S.J., IIW Recommendations on Post Weld Improvement
of Steel and Aluminium Structures, IIW Doc. XIII-2200-07, 2007.
41 Bousseau M. and Millot T., Fatigue Life Improvement of Welded Structures by
Ultrasonic Needle Peening Compared to TIG Dressing, IIW-Doc. XIII-2125-06,
2006.
42 Ummenhofer T., Weich I., Fatigue Design Concepts for Welds Improved by High
Frequency Hammer Peening Methods, Proceedings of the Fatigue Design 2007
conference, CETIM 21-23. November, 2007.
78
 D
Developing a Tool Point Control
Scheme for a Hydraulic Crane
Using Interactive Real-time
Dynamic Simulation

M.M. Pedersen, M.R. Hansen, M. Ballebye

Modelling, Identification and Control 31, Vol. 4, pp. 133-143, 2010.

Abstract. This paper describes the implementation of an interactive real-time dynamic

simulation model of a hydraulic crane. The user input to the model is given continuously
via joystick and output is presented continuously in a 3D animation. Using this simulation
model, a tool point control scheme is developed for the specific crane, considering the
saturation phenomena of the system and practical implementation.

Introduction
Current mobile hydraulic manipulators, such as loader cranes, forestry machines and
aerial lifts are typically controlled by electro-hydraulic actuation via joystick or remote
control. One axis on the joystick commands one hydraulic cylinder. In order to move
the tool point (end effector) in a controlled way, it is normally necessary to use several
DOF of the manipulator simultaneously. This is often a highly nonlinear task, which
only experienced operators can do accurately. Therefore one of the main advantages of
automated or semi-automated tool point control is the reduced skill level required or,
alternatively, the increased productivity.

Tool point control of mobile hydraulic manipulators has been subjected to research for
several years. Such control is different from control of typical industrial robots, due to
the non-ideal behavior of mobile hydraulic directional control valves, complex dynamics
of the hydraulic and mechanical system, continuous online control by an operator,
significant structural flexibility, vehicle mounting flexibility, highly varying loads and
eigenfrequencies. Furthermore, there are several saturation phenomena associated with
the hydraulic actuation.
80

Thus, for practical implementation of tool point control in mobile hydraulic manipulators,
there are many issues to be addressed. Model based development and testing is preferred,
since practical testing of such machinery is unsafe and costly in the initial design stages.

Krus and Palmberg (1998) presented a simple vector control strategy for a 2 DOF
hydraulic crane in the early nineties. Mattila and Virvalo (2000) describe a more advanced
control scheme for a similar crane; where an online model is used to reduce the pressure
levels and thus the energy consumption. Beiner (1997) solves the redundancy in a
3 DOF hydraulic crane using the minimum norm of the actuator forces. Recently,
Yuan et al. (2009) presented a motion control for a 4 DOF non-redundant aerial lift
including vibration suppression and static deflection compensation. An interactive real
time simulation was achieved by Esque et al. (2003) for a simplified 2 DOF crane.

Adapting robot control principles in motion control of mobile hydraulic manipulators

has been a key area of research at Aalborg University for several years, see Pedersen and
Nielsen (2002), Münzer (2003), Hansen and Andersen (2005), Ebbesen (2007) and Kabus
and Hastrup (2008) where the Kuhn-Tucker optimality criteria, are applied continuously
in directly calculating optimal actuator velocities, while simultaneously minimizing power
consumption.

Typically, a tool point control scheme for such systems is based on velocity control of
the manipulator, and the premise, that the operator closes the loop regarding the tool
point position. In order to verify this concept, an interactive real-time simulation model
is developed, with actual operator-in-the-loop capabilities.

Recent advances in computational power has made it possible to make interactive

dynamic simulations run in real-time on an ordinary laptop computer for limited systems.
Here, the implementation of such a simulation model is presented in 3D for a 4 DOF
crane including simple flexibility.

This paper describes the development of a tool point control scheme for commercial loader
cranes, considering the practical implementation. This means that the costs of sensors
and control valves must be kept at a minimum. The developed tool point control scheme
is purely feed forward, based on a pseudo inverse approach for handling the redundancy.
Flow sharing, deflection compensation, configuration control and joint limit avoidance is
also implemented.

Considered system
An HMF 2020K4 mobile loader crane is used as case study, Figure D.1. The crane consists
of the following 5 bodies; base, column, main boom, jib and extension boom system. The
base is considered stationary and neglected from the dynamic analysis.

The crane has 4 controlled joints; slew, main, jib, and extension. The first 3 are revolute
joints and the 4th is prismatic. Each joint is actuated by one or more hydraulic cylinders.
The slew joint connects through a rack-and-pinion. Both the main and jib joints connects
through a linkage system. The extension joint consists of beam segments sliding within
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 81

main joint
jib
jib joint extension system,
main boom
extension joint

column
slew joint

base

Figure D.1: HMF 2020K4 loader crane.

each other. It is actuated by 4 cylinders coupled in parallel and the sequence is controlled
hydraulically.

For the simulation task at hand, the crane is modelled in 3 subsystems; a control system
model, a hydraulic model and a mechanical model.

The control model establishes the reference input for the hydraulic model based on
user input via an ordinary PC joystick. The hydraulic model then determines the
actuator forces based on the states of both the mechanical system and hydraulic system
and the current reference input. The actuator forces are then used as input in the
mechanical model, which solves the forward dynamics of the system and returns the
joint accelerations for integration, see Figure D.2.

u u Virtual crane
θ Fc Fc
θ θ θ ∫ ∫ θ
Joystick Hydraulic θ
input model Mechanical 3D
model animation

Figure D.2: Overview of the simulation model.

The modelling is as detailed as possible without violating the demand for real time
simulation.

This paper is organized as follows; chapter 3 and 4 present the mechanical and hydraulic
models. Chapter 5 discusses the implementation of these, and chapter 6 presents the
developed tool point control algorithm. Results are presented in chapter 7.
82

Mechanical model
A simplified 6 DOF model of the crane is considered in order to limit the computational
effort. All flexibility is lumped in a passive universal joint within the jib body, modelled
by rotational springs around the vertical and lateral axes, see Figure D.3 and D.4.

θ4 θ5 θ6
θ3 (5)
(4)
(3)
extension system
θ2 (2) jib

main boom

(1) Y
column

θ1 X
(0) base Z

Figure D.3: Mechanical model.

The dynamics of the mechanical system is implemented using the joint coordinate
formulation and non-centroidal coordinate systems, as described by Nikravesh (1990 and
1991).

θi body i

yi ρi
zi xi
ei
ri Y

X
Z

Figure D.4: Joint and body definitions.

The body-fixed coordinate system is located in the joint connecting the given body to
the previous in the chain by the vector r. The vector e describes the axis of the joint.

Equations of motion
Each body is defined by its mass m, inertia matrix J and location of centre of mass ρ.
The equations of motion is given by

Mi v̇i = gi . (D.1)
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 83

Here, the ith mass matrix and force vector are

   
mI −mρ̃ f −m ω̃ ω̃ρ
Mi = and gi = , (D.2)
mρ̃ J i n −ω̃Jω i

where f and n are the forces and moments, respectively, acting on the body. The vector
of Cartesian velocities v is defined as
 
 T ṙi
v = v1 v2 v3 v4 v5 and vi = . (D.3)
ωi
For a system of constrained bodies, the equations of motion are combined with the
constraints, yielding
Mv̇ − DT λ = g. (D.4)
Here, D and λ are the Jacobian of the constraints and a vector of Lagrange multipliers.

This equation system can be solved for the accelerations and Lagrange multipliers by
inverting a system matrix of size 54x54 (6x number of bodies + 24 constraints). However,
in order to increase computational efficiency, the system is reduced to a minimum set
using velocity transformations.

Reduction to a minimum set

The vector of joint velocities is defined as
 T
θ̇ = θ̇1 θ̇2 θ̇3 θ̇4 θ̇5 θ̇6 . (D.5)

Transformation between the Cartesian and joint space is handled by the transformation
matrix B as follows
v = Bθ̇. (D.6)
Accelerations are then transformed by

v̇ = Bθ̈ + Ḃθ̇. (D.7)

It can be shown that the B matrix is orthogonal to the constraint Jacobian D, such that

DB = 0. (D.8)
T
Thus, by inserting eq. (D.7) into eq. (D.1) and pre-multiplying with B we get

BT MBθ̈ = BT (g − MḂθ̇). (D.9)

Using the inertia matrix and force vector in joint space, M and f , then leads to the
following short notation
Mθ̈ = f , (D.10)
where
M =BT MB
(D.11)
f =BT (g − MḂθ̇).
Here, the constraints will be implicitly enforced and the constraint Jacobian as well as
the Lagrange multipliers need not be determined. The system matrix, which needs to be
inverted in this formulation, is only 6x6, corresponding to the number of DOFs.
84

Velocity transformation
The velocity transformation matrix B and its time derivative can be established from
joint specific block matrices. For the ith body and the jth revolute (R) or prismatic
joints (P), these are    
−d̃ij ej ej
Rij = and Pij = . (D.12)
ej 0
The distance vectors dij = ri − rj are illustrated in Figure D.5, where it should be noted
that dii = 0. Combining the block matrices of two revolute joints yields a Universal joint
(U).

(4) d54 (5)

d43
(3) d 53
d32
d42 d52
(2)

d31
d41
(1) d d51
21

(0)

Figure D.5: Distance vectors.

The B matrix is thus built by inserting block matrices in the following way, such that
Rij relates the joint velocity contribution of joint j to the Cartesian velocity of body i.
 
R11 0 0 0 0
R21 R22 0 0 0 
 
B= R
 31 R 32 R 33 0 0 
 (D.13)
R41 R42 R43 U44 0 
R51 R52 R53 U54 P55

The derivative matrix Ḃ is built in the same manner as eq. (D.13), but using the derivative
block matrices " #
˜ 
ω̃j ej

Ṙij = −( ḋij + d̃ij ωj )ej
and Ṗij = (D.14)
ω̃j ej 0

where ḋij = ṙi − ṙj .

Flexibility
Mobile manipulators are typically highly weight optimized and therefore, the influence of
the structural flexibility cannot be ignored. The influence of flexibility is investigated by
Linjama and Virvalo (1999) and Mikkola (1997). The flexibility significantly influences
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 85

the eigenfrequency of the structure and thus needs to be included in the mechanical
model.

Here, structural flexibility is included in the mechanical model, by dividing the jib-body
in two and connecting the halves by a universal joint equipped with rotational springs,
as shown in Figure D.3. This enables the crane model to flex both in its working plane
and in the horizontal plane.

The applied spring torques Ts are calculated by

Ts = (Ks,i · δθi − Kd,i · θ̇i )ei . (D.15)

The spring constant Ks of the rotational springs are determined by physical experiments,
such that the deflection of the model resembles the deflection of the real crane. Damping is
included in order to stabilize the model and the damping coefficient Kd is also determined
to match experiments.

This simple flexibility model lowers the eigenfrequency and gives a deviation in the tool
point position, and is considered sufficient for the purpose of this model.

Extension system modelling

The extension system is modelled as a single body with variable geometry and inertia
properties. The dynamics of the relative motion between segments is thus ignored. This
is justified, since this motion is highly damped due to friction and its dynamics has only
negligible influence.

ρ5

θ6

Figure D.6: Variable location of extension system Centre of Mass (COM).

An expression for the position of the centre of mass is obtained by determining it for
different values of θ6 using CAD and fitting a curve through the results, as shown in
Figure D.6.  
ρ5 = ρx (θ6 ) ρy ρz (D.16)
The same approach has been applied for the inertia matrix, which is approximately
constant, except for the Jyy ≈ Jzz terms, which vary notably with θ6 ,
 
Jxx Jxy Jxz
J5 = Jyx Jyy (θ6 ) Jyz  (D.17)
Jzx Jzy Jzz (θ6 )
86

The results of the fitting procedure are shown in Figure D.7.

-1 2 4
x10

Jyy,Jzz [kg·m2]
-2 1.5
ρ [m]

-3 1
x

-4 CAD 0.5
fit
-5 0
0 2 4 6 8 10 0 2 4 6 8 10
θ6 [m] θ6 [m]

Figure D.7: Fitted location of COM and moments of inertia for the extension
system.

Hydraulic model
The hydraulic system is modelled with the highest possible detail level without
compromising real time simulation. Pressure losses in hoses and fittings are largely
ignored and the dynamics of the directional control valves is modelled as a first order
system.

All cylinders on the crane are mounted with over-center valves (OCVs) due to practical
and legislative reasons. OVCs serve a number of purposes, including leak tight load
holding, shock absorbtion, cavitation protection during lowering and load holding at
pipe/hose bursts. Their influence on the system is significant, and their effect is included
in a simple and efficient manner.

cmax
Vp0 c Vr0
Fc
Fc pp pr
Qp Ap Ar Qr
pp
pcr
Μ pr

Qin Qout

Figure D.8: Simplified hydraulic circuit; cylinder, over-center valve and directional
control valve.
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 87

The system in Figure D.8 illustrates the main lift cylinder, which has only one OCV
controlling the lowering motion. All other joints have double acting OCVs mounted,
since the gravity load for these joints act in both directions.

The dynamics of the hydraulic fluid is introduced via the pressure gradient equation for
a volume of Newtonian fluid, Merrit (1967)
β
ṗ = (Q − V̇ ), (D.18)
V
where β is the effective bulk modulus of the fluid, Q is the flow into the volume V and
V̇ is the time derivative of the volume expansion. From this, the piston and rod side
pressure gradients of the cylinders are calculated
β
ṗp,i = (Qp,i − ċi · Ap,i ) and (D.19)
Vp0,i + ci · Ap,i
β
ṗr,i = (Qr,i + ċi · Ar,i ). (D.20)
Vr0,i + (cmax,i − ci )Ar,i
These are then integrated to get the pressures on both sides of the piston.
Z Z
pp,i = ṗp,i dt and pr,i = ṗr,i dt (D.21)

The flow into the circuit Qin is determined from the valve rated flow and the control
signal ui and saturates at maximum valve flow Qv,max

ui · Qv,max,i f or Qin,i < Qv,max,i
Qin,i = (D.22)
Qv,max,i f or Qin,i ≥ Qv,max,i

The flow out of the circuit is governed by the OCV.

Qout,i = −QOCV,i (D.23)

Finally, the cylinder force can be determined from the pressure and area differences and
friction.
Fc,i = pp,i · Ap,i − pr,i · Ar,i − Fµ,i (D.24)
Nielsen (2005) obtained good results using the following simple cylinder friction model,
which has also been implemented here. The tanh is continous and crosses through the
origin, thus ignoring stick-slip effects.
ċi
Fµ,i = tanh( ) · |pp,i · Ap,i − pr,i · Ar,i | · (1 − η) (D.25)
ċη
A mechanical efficiency of η = 0.92 is assumed and the normalization velocity ċη is set
at 0.02m/s.
88

Over-center valves
The OCV model determines the flow out of the cylinders in most cases. The dynamics of
the over-center valves is disregarded, and the following static approximation is applied.

pcr,i = pp,i + ψi · pr,i (D.26)

The crack pressure pcr must be balanced by the pressures in the cylinder chambers
(affected by the pilot ratio ψ, typically ≈ 4 for this application) for the over-center valve
to open. Assuming flow out of the piston side chamber, we have
p
QOCV,i = KOCV,i pp,i − pt (D.27)

The flow out of an OCV is determined by the orifice equation, assuming a linear discharge
characteristic, as shown in Figure D.9. The tank pressure pt is assumed to be zero.

Kocv
Kocv,max

pp+ψpr
pcr pcr+p

Figure D.9: Assumed OCV discharge characteristic.

The maximum value of the flow coefficient KOCV,max can be calculated from the OCV
manufacturer datasheets. The pressure increment ∆p at which the OCV is fully opened,
is set to 25bar. The OCV flow coefficient is given as
KOCV,max,i KOCV,max,i
KOCV,i = (pp,i + ψi · pr,i ) − pcr,i . (D.28)
∆p ∆p

Connecting the mechanical and hydraulic models

In the hydraulic model, the crane state is expressed by cylinder stroke and velocity,
i.e. actuator coordinates. The mechanical model, however, returns joint coordinates. Thus
a transformation is necessary from the joint space to the actuator space.

ċ = Aθ̇ (D.29)

A is the Jacobian matrix of the joint to actuator transformation, i.e. the instantaneous
effective torque arms of the actuators, given by
 
a1 0 0 0
 0 a2 (θ2 ) 0 0
A= 0
 (D.30)
0 a3 (θ3 ) 0 
0 0 0 a4

Here, a1 = 0.105m is the constant radius of the rack and pinion slew mechanism, a2 and
a3 vary significantly due to the linkage systems, see Figures D.10 and D.11. For the linear
extension joint, a4 = 1.
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 89

T2

T2
c2 m
șm h
n
șl
l
șc
c2

y1 k c2

x1

Figure D.10: Cylinder coupling through linkage system.

Given the constant lengths k, l, m, n and h and the angle θ2 , its possible to determine
the angles θl , θm , θc and the cylinder stroke c2 . The instantaneous effective relationship
between the cylinder velocity and joint velocity is then given by

ċ2 = a2 (θ2 ) · θ̇2 (D.31)

where
m · sin θm sin θc
a2 (θ2 ) = (D.32)
sin θl
The same holds for the jib joint. In order to reduce the computational burden, curve fits
are used to determine a2 and a3 as a function of θ2 and θ3 , as shown in Figure D.11.

0.4 0.4

0.3 0.3
a2 [m]

a [m]
3

0.2 0.2 calc.

fit
0.1 0.1
-1 0 1 2 -2 -1 0
θ [rad] θ [rad]
2 3

Figure D.11: Effective torque arms of cylinder 2 and 3.

The A matrix furthermore gives the relationship between the cylinder forces and joint
torques
Tc = AFc (D.33)
90

It is clear, that the maximum lifting capacity of joint 2 (main) and 3 (jib) will be where
the effective torque arms have their maximum. The typical pressure layout of such a
crane is that the jib joint is set to be stronger than the main joint. Therefore, for general
maximum lifting capacity, the torque arm of the main joint should be maximized. The
torque arm of the jib joint is approximately constant for θ3 > −π/2.

Model implementation
The simulation model is implemented in Simulink using the 3D animation toolbox
and blocks of Embedded Matlab, Figure D.12. The Embedded Matlab blocks execute
very fast, since their content is automatically pre-compiled in the background prior to
execution of the Simulink model.

Using the fixed step Runge-Kutta integrator (ODE4), real time performance is achieved
with a time step of 2.5ms on a standard 2GHz laptop computer.

The Simulink 3D animation toolbox includes a source block for joystick input. This is
used as interactive reference input to the model, and resembles the input device typically
used for controlling the real crane.

Figure D.12: Example of the 3D animation.

The 3D animation sink block featured in the same toolbox is easily applied together with
CAD files for the given crane. The CAD files are converted to VRML format, and a
VR-world hierarchy files is written, defining the relationship between the CAD parts.

The interactive real time model provides a better feeling with the control system during
development, which can otherwise be difficult to obtain, using non-real time models with
preset trajectories. It provides an excellent target for developing and testing new control
strategies.
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 91

Tool point control

A feedforward control scheme is implemented, where the operator controls the velocity,
i.e. direction and speed, of the tool point (hook) of the crane online. The operator will
then ’close the loop’ and ensure sufficient positional accuracy.

The purpose of creating the simulation model was to aid in the development of a new
tool point control scheme for loader cranes. Thus a control block is inserted to transform
the joystick input to a suitable reference signal for the virtual crane.

Input Control Virtual crane 3D animation

rTP rTP
u u θ θ
y

Feedback

Figure D.13: Overview of the simulation.

The tool point control block must take the Cartesian velocity reference input ṙT P and,
based on some feedback, return the valve flow reference u to the virtual crane, as shown
in Figure D.13.

Joint controller
Low cost, low resolution sensors are used for the practical implementation in order to
keep it commercially viable. A pure feed forward velocity controller is chosen, in order
not to rely too much on the sensors for accurate feedback in a closed loop control scheme.

The Jacobian matrix of the Cartesian to joint transformation is easily established

considering that each column represents the associated joint’s velocity contribution to
the tool point velocity.
ṙT P = Jθ̇ (D.34)
where  
J(θ) = ẽ1 d51 ẽ2 d52 ẽ3 d53 e6 (D.35)
Due to the redundancy, the Jacobian is not square, and the Moore-Penrose pseudo-inverse
is used
θ̇ = J+ ṙT P (D.36)
This corresponds to finding the minimum velocity norm in the joint space, Beiner and
Mattila (1999). The requested cylinder velocity is then calculated by eq. (D.29) which is
subsequently used for calculating the flow set-point.

Joint limit avoidance

Joint limits are avoided by the weighted least norm (WLN) approach, Chan and Dubey
(1993), using the weighted pseudo inverse, which penalizes the motion of joints close to
their limits
−1 T
J+W = W J (JW−1 JT )−1 (D.37)
92

Here a W is a symmetric positive definite weighing matrix

 
w1 0 0 0
 0 w2 0 0
W=  . (D.38)
0 0 w3 0
0 0 0 w4

The elements of W is given as the partial derivative of a performance criterion, e.g.

4
X 1 (θmax,i − θmin,i )2
H(θ) = . (D.39)
i=1
4 (θmax,i − θi )(θi − θmin,i )

The value of H increases exponentially when a joint approaches its limit. The partial
derivative of H is then used as weights in W.

∂H (θmax,i − θmin,i )2 (2θi − θmax,i − θmin,i )

= . (D.40)
∂θi 4(θmax,i − θi )2 (θi − θmin,i )2
(
∂H ∂H
1 + | ∂θ | f or ∆| ∂θ |≥0
wi = i
∂H
i (D.41)
1 f or ∆| ∂θi
| < 0.
The ith weight will go towards infinity, when the ith joint approaches its limit, thus
effectively stopping the joint motion. On the other hand when moving away from a limit,
the joint is not penalized. Here, ∆| ∂H
∂θi
| is the rate of change of the partial derivative. It
∂H
can be shown, that the weights are continuous, also when ∆| ∂θ i
| changes sign, Chan and
Dubey (1993).

Configuration control
Experience shows that it is feasible to maximize the effective joint torque arm of the main
joint, i.e. a2 in Figure D.11. This can be achieved by modifying the associated weight w2
by adding the following penalization, i.e. when moving away from the maximum.
(
Ka | da
dθ2
2
| f or θ̇2 · | da
dθ2
2
|<0
w2 = da2 (D.42)
0 f or θ̇2 · | dθ2 | ≥ 0.

In some cases it might also be necessary for the operator to control the configuration of
the crane, as shown in Figure D.14. This will override the above maximization of the main
joint torque arm and can be used to avoid obstacles or fine-tuning the lifting capacity.

ϕ s

Figure D.14: Configuration control (self motion).

Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 93

For a redundant system, a set of joint velocities that do not cause any tool point motion
can be determined according to so-called self-motion, see e.g. Chan and Dubey (1993)
or Beiner and Mattila (1999). This is typically used in some sort of optimization, e.g. in
the gradient projection method, but here it is controlled directly by the operator.

θ̇s = (I − J+ J)z. (D.43)

The first term is the null space projection matrix of the Jacobian and z is an arbitrary
vector in the null space. Then θ̇s can be added to eq. (D.36) without interfering with the
tool point motion. Since there is only on degree of redundancy, choosing
 
z = 0 ϕ̇s 0 0 , (D.44)

and taking ϕ̇s as input from joystick, the operator can control the configuration online.
This is possible, even while simultaneously controlling the tool point motion.

Deflection compensation
The total deflection of the tool point is due to; 1) structural flexibility, 2) deformation
of sliding pads between segments in the extension system, 3) mechanical slack between
segments and 4) compression of hydraulic fluid in cylinders.

In a practical implementation, the joint angles of the main and jib joints would be
measured using inclinometers, i.e. relative to horizontal. This means that these joint
angles would be measured in the deflected state.

Stiff
Actual

6
ref
rTP
d 
w
real
rTP
L
P

Figure D.15: Tool point velocity deviation due to deflection.

Thus, only the deflection of the extension system needs to be compensated for actively.
In practise, only the in-plane vertical deflection is significant, i.e. up to ≈ 0.5m. The
deflection is especially problematic for horizontal straight-line tool point motion.

Figure D.15 illustrates how the real tool point velocity deviates from the reference due
to the deflection, which can be expressed as

ṙreal ref
T P = ṙT P + ẇ. (D.45)
94

In order for the real tool point velocity to match the reference, the reference must be
compensated as
ṙcomp
TP = ṙref
T P − ẇ. (D.46)
Determining ẇ is difficult in practise, since the deflection contribution caused by 2), 3)
and 4) is difficult to isolate and quantify. Hence, a simplified approach is used here;
 
0
ẇ = −Kc sin(ϕd )θ̇6  . (D.47)
0

Only the vertical deflection is considered and the total deflection velocity is assumed to
be proportional to the telescope extension velocity θ̇6 . The deflection angle of an inclined
cantilever beam with a load P at the free end can estimated as follows
cos(ϕ6 )P L2
ϕd = (D.48)
2EI
where ϕ6 = θ2 + θ3 is the inclination angle of the extension system relative to horizontal.
Thus, only the scaling factor Kc in eq. (D.47) needs to be determined from experiments.

Practically, this deflection compensator adds a small upwards velocity to the tool point
reference velocity, when extending, which cancels out due to the deflection.

Flow setpoint calculation

The requested cylinder velocities are calculated using eq. (D.29) and the flow request
Qreq is determined by the piston or rod side area, depending on the direction of the
requested cylinder velocity

Ap,i · ċi f or ċi ≥ 0
Qreq,i = (D.49)
Ar,i · ċi f or ċi < 0

Two saturation phenomena need to be included; firstly the valve flow limit and secondly
the pump flow limit. The set flow is different from the requested flow if any of the two
limits are exceeded in order to maintain the tool point velocity and only violate the speed.

The requested set point signal ureq,i is determined from the requested flow normalized
by the maximum valve flow
Qreq,i
ureq,i = (D.50)
Qv,max,i

Flow sharing
The requested flow is limited to the valve flow limit, typically 40-100 l/min. If one valve
limit is exceeded, all requested flows are scaled down by kv .
 
ureq = u1 u2 u3 u4 (D.51)

1
kv = (D.52)
max(|ureq |)
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 95


ureq f or kv > 1
ureq,v = (D.53)
kv ureq f or kv < 1
The total requested flow is determined by summation
 P
Qtot = P|Qreq,i | f or kv > 1
(D.54)
|kv · Qreq,i | f or kv < 1

Flow sharing is applied again if the total requested flow Qtot exceeds the maximum pump
flow Qpump :
Qpump
kp = (D.55)
Qtot

kp ureq,v f or Qtot ≥ Qpump
u= (D.56)
ureq,v f or Qtot < Qpump
It is then ensured, that neither the valve flow limit nor the pump flow limit are exceeded.

Results
Figure D.16 shows the overall control scheme developed. All motion is directly propor-
tional to the input given by the operator, i.e. no closed loop control has been applied, in
order to leave the operator in complete control.

Deflection
compensation
w
rTP c u
+- comp JW+ ++ A Flow-
sharing
r&TP θ&
Maximize joint Joint limit θ&s
2 torque arm avoidance

ϕs Configuration
control

Figure D.16: Overview of the control scheme.

The following working cycle is established, in order to illustrate the advantages of real-
time versus non-real-time simulation, see Figure D.17.

Traditionally, in a non-real-time simulation, the operator will be modelled, e.g. as a set

of predefined trajectories which should follow a reference. This is exemplified as the
“predefined operator”. Using the simple feedforward joint controller, the tool point speed
will generally be slightly lower than the reference, which accumulates to a position error.
This is due to pressure build-up time in the hydraulics and acceleration of the mechanics.

In real-time simulation, an actual operator controls the crane tool point (hook) such that
it follows the same trajectory. This is shown as the “active operator”. In this case, the
operator uses visual feedback to close the loop and adjust the input through the joystick,
until he/she finds that the reference position is sufficiently achieved. Since the direction
of the tool point is always maintained, and only the speed is violated, it is not noticed
by the operator.
96

Tool point trajectory

6
Reference
5.5 Predefined operator
5 Active operator

4.5

4
y [m]

3.5

2.5

1.5
4 5 6 7 8
x [m]

Figure D.17: Reference and actual tool point position.

Conclusions
Under normal evaluation conditions (i.e. using a predefined operator), such a control
scheme would probably be deemed unacceptable. However, when the operator capabilities
are included in real-time simulation, the controller performance is excellent.

Tool point control of a hydraulic loader crane seems very advantageous for many tasks,
however, the operator needs the ability to control the crane traditionally as well, e.g. when
folding the crane, or operating close to obstacles. Furthermore, some kind of configuration
control will be needed in order to avoid obstacles in the workspace.

Interactive real-time dynamic simulation is a powerful tool for developing control

strategies. It provides valuable insight into the practical application of a given control
scheme. Future works could include more detailed modelling of the operator as part of a
simulation. However, the only authentic way to include operator abilities is to let him/her
be part of the loop.

Acknowledgements
This research work is funded by HMF A/S and an Industrial PhD-grant from the Danish
Ministry for Science, Technology and Innovation.
Paper D. Developing a Tool Point Control Scheme for a Hydraulic Crane Using
Interactive Real-time Dynamic Simulation 97

References
Beiner L. (1997), Minimum force redundancy control of hydraulic cranes. Mechatronics
7(6), pp.537-547.

Beiner L. and Mattila J. (1999), An improved pseudoinverse solution for redundant

hydraulic manipulators. Robotica, 17, pp.173-179, 1999.

Chan T.F. and Dubey R.V. (1993), A weighted least-norm solution based scheme for
avoiding joint limits for redundant manipulators. Robotics and Automation, 3, pp. 395-
402, 1993.

Ebbesen M.K. (2007), Optimal Design of Flexible Multibody Systems. PhD Thesis,
Aalborg University, Denmark, 2007.

Esqué S., Raneda A., and Ellman A. (2003), Techniques for studying a mobile hydraulic
crane in virtual reality. Int J Fluid Power, 4(2) pp.25–34, 2003.

Hansen M.R. and Andersen T.O. (2005), A method for deriving the optimal operation of
mobile hydraulic manipulators. In Proc. 9th Scandinavian International Conference on
Fluid Power, SICFP’05. Linköping, Sweden, 2005.

Kabus S. and Haastrup, M. (2008), Simulations of Flexible Loader Crane with Designed
Tool Point Control. Masters Thesis, Aalborg University, Denmark, 2008.

Krus G. and Palmberg J.O. (1998), Vector control of a hydraulic crane. In Proc
International Off-Highway and Power plant Congress and Exposition. Milwaukee, USA,
1992.

Linjama M. and Virvalo T. (1999), State-space model for control design of multi-link
flexible hydraulic cranes. In 6th International Conference on Fluid Power. Tampere,
Finland, 1999.

Mattila J. and Virvalo T. (2000), Energy efficient motion control of a hydraulic

manipulator. In Int Proc Robotics and Automation. San Francisco, USA, 2000.

Merrit H.E. (1967), Hydraulic Control Systems. John Wiley and Sons, Inc. New York,
USA, 1967.

Mikkola A.M. (1997), Studies of Fatigue Damage in a Hydraulic Driven Boom System Us-
ing Virtual Prototype Simulations. PhD Thesis, Lappeenranta University of Technology,
Finland, 1997.

Münzer M.E. (2003), Resolved Motion Control of Mobile Hydraulic Cranes. PhD Thesis,
Aalborg University, Denmark, 2003.

Nielsen B.K. (2005), Controller Development for a Separate Meter-in Separate Meter-out
Fluid Power Valve for Mobile Applications. PhD Thesis, Aalborg University, Denmark,
2005.
98

Nikravesh P.E. (1990), Systematic reduction of multibody equations of motion to a

minimal set. Int J Nonlinear Mechanics, 25(2/3) pp.143-151, 1990.

Nikravesh P.E. (1991), Computational Methods in Multi-body Systems. Notes, COMETT,

DTH, Lyngby, Denmark, 1991.

Pedersen H.C. and Nielsen B. (2002), Resolved Motion Control of Flexible Hydraulic
Manipulators. Master Thesis, Aalborg University, Denmark, 2002.

Yuan Q., Lew J., and Piyabongkarn D. (2009), Motion control of an aerial work
platform.In Proc. American Control Conference. St. Louis, MO, USA, 2009.
 E
A Cost-effective Approach to
Hardware-in-the-loop Simulation

M.M. Pedersen, M.R. Hansen, M. Ballebye

Submitted to the 9th International Conference: Mechatronics 2011, September 21–24,
Warsaw, Poland, 2011.

Abstract. This paper presents an approach for developing cost effective hardware-in-the-
loop (HIL) simulation platforms for the use in controller software test and development.
The approach is aimed at the many smaller manufacturers of e.g. mobile hydraulic
machinery, which often do not have very advanced testing facilities at their disposal.
A case study is presented where a HIL simulation platform is developed for the controller
of a truck mounted loader crane. The total expenses in hardware and software is less
than 10.000$.

Introduction
The safety demands for machinery are ever increasing, as described in e.g. the new
Machinery Directive (2006) or EN13849 (2006). More and more electronic safety and
control systems are therefore implemented in order to avoid potential danger, both from
conventional use, but also from foreseeable misuse.

As an example from the loader crane industry, the new revision of EN12999 (2010)
requires the stability of the truck to be monitored and secured via the crane safety
system, in order to avoid overturning of the truck. In the case of using the loader crane
for personal lifts, the safety requirements are further increased due to the obvious risk to
the person(s) being lifted.

The safety system for an advanced loader crane as shown in Fig. E.1 may include in the
order of 50 IOs and up to some 20 CAN modules connected to a main controller. The
trend is towards more distributed IO systems communicating via CANbus networks, but
also using sensors with integrated CANbus interface.

The complexity of the controller software increases in the same rate as the safety demands,
and is further increased by a high level of product configurability. It thus becomes
100

increasingly difficult to perform end-to-end tests for such systems, using a breakout board
with switches, diodes and potentiometers.

In other industries with complex safety systems, e.g. the automotive, aerospace or offshore
industries, hardware-in-the-loop simulation have long been the standard platform for
testing controller software. In the crane industry and other similar industries, controller
software testing have typically been performed using the real machine, though this is
neither safe nor efficient.

It is possible, however, to implement very cost-effective HIL simulation platforms for such
machinery, and thus significantly improving testing capabilities.

Figure E.1: Truck with loader crane used for personal lift.

Earlier HIL simulation (HILS) platforms typically considered a single component or minor
isolated systems for testing controller software with limited visual output. However,
advances in computational power have propelled a rapid evolution of HILS platforms
which now considers entire systems, such as heavy construction machinery or offshore
structures and offer rich 3D animation output.

A HILS system for a hydraulic excavator with both audio and visual simulation output
integrated in the operator cabin of a real excavator, was presented by Elton et al. (2009).
The system included a remote hydraulic load emulator for a variable displacement pump,
connected via Internet. The simulation are thus feed accurate (non-modeled) information
from the pump.

Ghabcheloo and Hyvonen (2009) discusses development of a motion control system

for a hydraulic wheel loader in a HILS platform called GIMsim. It consists of 5 PCs
communicating via LAN to accommodate animation, dynamics, tire contact, IO + CAN
Paper E. A Cost-effective Approach to Hardware-in-the-loop Simulation 101

and simulation control. The simulation is developed in Matlab/Simulink, compiled and

transfered to an xPCtarget for hard real-time operational speed.

Recently, Trydal (2010) presented a very advanced proprietary HILS platform from
National Oilwell Varco. It concerns an offshore drill-rig including the entire drilling
process equipment and displays a 3D animation in a dome in order to completely immerse
the operators. Instead of physical IOs, each controller are connected to a similar controller
which simulates the IO functionality.

The general trend is to develop the simulation in Matlab/Simulink, use several PCs
connected via LAN and use hardware IO connectivity from dSpace, National Instruments
or others. Simulation input are given from real operator interfaces and results are
presented using more and more realistic 3D animation. The animation software is
generally developed specifically for the considered system using OpenGL.

In this paper, a cost effective approach to HIL simulation is demonstrated for a non
time-critical application using only standard high-level engineering software. It should
thus be possible for it to be implemented and maintained by engineers, without in-depth
software development background.

Considered System
We consider a truck with an advanced folding loader crane attached to it, Fig. E.1. The
crane is equipped with a so-called flyjib, and has 6 degrees of freedom; slew, main boom,
jib boom, telescope extension, fly-jib, and fly-jib telescope extension. In the tool point,
a variety of accessories can be mounted, such as a rotator and grab or a personal lift
basket. The crane is furthermore equipped with a winch mounted in a revolving console
for parking during transport.

The crane is mounted with a stabilizer system containing 4 extensible booms each with a
stabilizer leg, and an additional fixed front stabilizer leg. In total we have 19 hydraulically
actuated DOFs controlled by 12 directional valves and several shift valves. The truck
furthermore has two passive DOFs, enabling inclination around the roll and pitch axes in
order to simulate stability issues (i.e. overturning of the truck). Fig. E.2 shows the DOFs
included in the simulation.

The safety and control system, Fig. E.3, of such a crane primarily covers a number of
capacity limitation, stability monitoring and collision avoidance features, but also several
secondary safety features, such as smart speed reduction and selective stopping of load-
increasing motion. All safety features are supported by visual and audible warning signals.

The electronic hardware consists of a CAN network with a main controller, 3 I/O
modules (flyjib, winch and stabilizers), a radio remote control, a slew sensor, 2 dual
axis inclinometers, and up to 12 electro-hydraulically actuated directional control valves.
Inputs and outputs are thus distributed to the corresponding function via the CAN
network.
102

x9 x10 x4

x3
x2 x12 x5 x6
x1 x7
x14

x19 x8

x11
x13
x18
x16

x20
x21
x15
x17

Figure E.2: Degrees of freedom in the simulation model.

The critical components are redundant, e.g. pressure transducers and inclinometers. Also
the I/O modules and controller consist of two separate processors, each of which monitors
the operation of the other, in order to achieve a high level of safety.

All IOs are monitored to comply with defined limits according to their use. Operating
status and error reports are communicated via 12 LED indicators and a 3x7 segment
display. The operator interface is typically a radio based remote control with up to 8
analog input channels, 12-20 buttons and a LCD display.

Slew Sensor Inclinometer Valve actuation

Inclinometer

CANbus

RC Remote Fly-jib I/O Winch I/O Stabilizer I/O Controller

IO IO IO IO IO

Figure E.3: Electronic crane safety and control system.

Paper E. A Cost-effective Approach to Hardware-in-the-loop Simulation 103

HIL Simulation
The developed HIL simulator consists of three subsections; 1) a simulation PC, 2) a
hardware-gateway PC and 3) the real controller hardware, as shown in Fig. E.4.

The simulation is developed in Matlab using the Simulink 3D animation toolbox

(which doesn’t actually require Simulink) for visualization. The Matlab and Simulink
programming environments are well known for their simple syntax and high productivity.

Figure E.4: Physical setup, from left: 1) controller hardware, 2) Gateway PC, 3)
Simulation PC and operator interface.

The hardware-gateway is developed in LabView using low-cost USB I/O and CAN devices
from National Instruments for low level communication with the controller hardware. The
link between the simulation running in Matlab and the hardware controlled by LabView
is achieved over the local area network (LAN), using the IP/UDP protocol.

Co-simulation via LAN/UDP

LAN communication using the UDP protocol is an effective approach to achieve co-
simulation with two different programs, e.g. Matlab and LabView or others. The programs
can run on either two different PCs connected by LAN, or on a single PC sending data
to itself - between programs. Using several PCs gives a performance advantage though
and enables additional monitoring possibilities and distributed location of equipment.

The UDP protocol is typically used for streaming media over the Internet and other
applications where it is preferable to drop packets rather than wait for them. In many
applications, this is not a problem, if the same data are transmitted repeatedly. Sending
data with UDP requires specification of the receiver IP address and port number,
receiving requires only specification of the local port number and the size of the packet
in bytes.
104

In this case study, the UDP protocol is used to communicate between the simulation
running in Matlab on the Simulation PC and the Hardware Gateway PC running
LabView, see Fig. E.5. An UDP packet are sent both ways at every time-step in the
simulation, all necessary data are collected and transfered in these packets.

Hardware Hardware Gateway PC Simulation PC

Simulated CAN hardware LabView Matlab
CAN loop Initialization

Simulation loop
Get inputs
Slew Sensor Inclinometer Valve actuation LAN loop (UDP) & Hydraulics model
Simulation control Kinematic model
Animation data
CANbus Time integration
Send outputs

Animation
Remote Fly-jib I/O Winch I/O Stab. I/O Controller
IO IO IO IO IO
A/D loop
A/D IO devices

Figure E.5: Overview of the HIL simulation.

The simulation model is thereby updated with the valve flow command sent from the
controller to the virtual valves, and the state of all outputs from the controller and
modules, e.g. to activate shift valves or perform engine control. In the other direction, the
controller hardware is updated with the state of the virtual sensors in the simulation, e.g.
hydraulic pressure, truck inclination, slew angle and estimated flow from the simulated
valves, etc.

Hardware Gateway
Using a separate PC with LabView to facilitate communication between the controller
hardware and the simulation has several advantages. Firstly, the price is low compared
to dedicated systems. Secondly, the system is very flexible and scalable. The gateway PC
can furthermore be used to control and monitor values in the HIL simulation while it
runs. Additional filtering or noise can also be added to the signals here.

As shown in Fig. E.6 and E.5 the Labview program is divided in three separate loops;
a LAN loop, CAN loop and A/D loop. Each loop covers communication between the
gateway and LAN network, CAN network or analog/digital hardware. Separate loops
in LabView runs in parallel for efficient execution and at different rates of 200Hz, 50Hz
and 20Hz respectively. The hardware gateway additionally simulates a number of CAN
devices, i.e. sensors and valve actuation.
Paper E. A Cost-effective Approach to Hardware-in-the-loop Simulation 105

Figure E.6: Screenshot from the Hardware Gateway (LabView).

Simulating CANopen Devices

The communication profile of most CAN devices such as sensors and valves is relatively
simple and these can therefore be simulated quite easily. For a thorough overview of
CAN and CANopen, see Pfeiffer et al. (2003). A first approach is to let LabView send
CAN frames corresponding to the the necessary transmit PDO(s) of the devices to the
controller.

If needed, it is also possible to implement a limited object dictionary and SDO server
in the LabView program containing the entries the controller (or other devices, e.g. a
service terminal) will inquire on. Typically, to simulate e.g. a generic CAN I/O module,
it will only be necessary to include in the order of 10 entries in the object dictionary.

The SDO protocol is straightforward for the expedited transfers (up to 4bytes), however
if segmented or block transfers are required, this will take some more work, see CiA 301
(2002).

For the application at hand, strict real-time performance is not critical. Most of the CAN
devices sample at 10-20Hz only and a small delay in the response of the simulation is
acceptable.

Simulation Model
Since the simulation model must run in real time (or close to) in order to communicate
with the controller, it is essential to limit the computational burden to what can be
achieved with the available hardware.
106

Most dynamics of the system is omitted, which is acceptable since the controller mainly
considers steady state values of the input. This is necessary in order to achieve near
real-time performance in Matlab.

Referring to Fig. E.5, the simulation is initiated by loading the crane parameters, such
geometric and inertial properties, joint states and limits, hydraulic system properties,
etc.. The main simulation loop is then started which is executed as fast as possible
(around 200Hz) and updates all simulation parameters and controls the animation.

All required data is retrieved from the UDP packet; flow command to the valves, state of
the outputs on the controller and modules, and necessary simulation control parameters
which are set in the Hardware Gateway PC, i.e. mass of payload, mass of truck, crane
setup, accessories mounted, etc.

Based on the flow commands and output states, the reference velocity of all DOFs are
then calculated. Since several outputs controls shift valves, in order to use the directional
control valves for multiple functions, the state of these leads to different modes of
operation; i.e. crane mode, stabilizer mode, winch mode, etc.

Then a complete kinematic analysis is performed in order to determine position,

orientation and velocity of all parts and the resulting tool point motion. The kinematic
parameters are needed in order to drive the animation.

A simple kinetic analysis is also conducted, which determines the needed parameters;
crane lifting moment, truck inclination due to crane load, hydraulic pressures in the main
cylinder, flyjib cylinder and winch. Additionally, the ground projection of the total Center
of Mass (COM), are calculated and displayed in the animation, in order to visualize the
stability margin.

The only dynamics included is for a suspended payload, which is modeled as a 2 DOF
pendulum. This gives the user the impression of operating a dynamic system and thus
experience the consequences of abrupt motion.

The state of all virtual sensors are then determined. These include digital sensors,
monitoring whether e.g. stabilizers are fully deployed, the jib is above horizontal, the
extension system is fully retracted, etc. Several analog sensors are also simulated, e.g.
pressure transducers for measuring pressure in selected cylinders, crane slew angle, truck
inclination and directional control estimated valve flow, etc.

Time integration of the calculated velocities for all DOF is then performed using the
trapezoidal rule and the animation is updated. All the simulated devices are modeled as
ideal, i.e. they respond instantaneously and with with perfect precision. In the Hardware
Gateway, however, noise is added to selected signals, in order to test the controller filtering
capabilities.

The necessary values are then assembled in a UDP packet and feed back to the controller
via the Hardware gateway, and the simulation loop starts over.
Paper E. A Cost-effective Approach to Hardware-in-the-loop Simulation 107

3D Animation
Using the Simulink 3D Animation Toolbox, a Virtual Reality (VR) world can be
manipulated from Matlab by translating, rotating or scaling the 3D parts at each time-
step, to make it appear that the crane is moving. The VR world is defined using the
VRML markup language, i.e. the relation between the parts and the surroundings.

Fig. E.7 shows a screenshot from the implemented 3D animation. Some CAD programs
can export parts to VRML format, otherwise, simple geometries, e.g. boxes and cylinders,
can be directly defined in the VRML language.

Using the Simulink 3D Animation Toolbox makes it simple to create an advanced 3D

environment for simulation output. Its flexible and scalable and based on a simple markup
language, unlike the typical custom-made OpenGL applications used for HILS animation.
Output from the animation can either be displayed directly or recorded for later use.

Figure E.7: Screenshot from the 3D animation.

Conclusion
The HIL simulator described in this paper have proven to posses the necessary level of
detail to test all the features of the controller concerned. Indeed, its possible to create
much more precise testing conditions, compared to using the real machine or a board
with switches and potentiometers.

Especially when testing dangerous conditions, such as truck overturning or faults in the
control system, simulation is preferred, since it is both expensive and time consuming to
perform such test using the real machine.

Table E.1 gives an estimate of the total costs of the HILS platform, excluding the
development of the software, which took approximately three months.
108

Matlab 2800 $
3D animation toolbox 1400 $
LabView 1700 $
IO modules 1000 $
2 PCs 3000 $
Total 9900 $

Table E.1: Approximate HILS platform costs.

The HIL simulator currently only comprise ad-hoc testing capability, however, it could
be extended to included fully automated end-to-end testing of the controller, which will
lead to additional savings on the testing expenses.

As a replacement for ad-hoc testing of controller software using the real machine or a
breakout board, HIL simulation is a very effective tool. With the additional control over
the testing conditions and increased monitoring possibilities, it is possible to create more
repeatable and precise testing conditions. The effort required and risk of testing controller
software can thus be significantly reduced.

Using only standard engineering software, such as Matlab and LabView, it is should be
possible for the engineers who possess the domain knowledge to implement and maintain
the HILS system, without requiring assistance from software developers.

It’s difficult to determine the payback period on the investment of such HIL simulator,
but a qualified estimate is 1-2 years. This is due to saved engineer-hours for testing, when
comparing to testing using the real machine.

Acknowledgments
This work is funded by an Industrial PhD grant from the Danish Ministry of Science,
Technology and Innovation and HMF A/S.

References
M.D. Elton, A.R. Enes and W.J. Book A Virtual Reality Operator Interface Station
with Hydraulic Hardware-in- the-Loop Simulation for Prototyping Excavator Control
Systems. IEEE/ASME International Conference on Advanced Intelligent Mechatronics,
Singapore, July 14-17, 2009.

R. Ghabcheloo and M. Hyvnen Modeling and motion control of an articulated-frame-

steering hydraulic mobile machine. 17th Mediterranean Conference on Control and
Automation, Thessaloniki, Greece, June 24-26, 2009.
Paper E. A Cost-effective Approach to Hardware-in-the-loop Simulation 109

S. Trydal HIL Simulator. Hardware In Loop Simulator med fokus paa integrerte prosesser.
(Norwegian) NODE seminar, Grimstad, Norway, 29 September, 2010.
http://www.nodeproject.no/index.php?articleid=393 (09-11-2010)

EU Machinery Directive, 2006/42/EG European Union, Brussels, 2006.

EN ISO 13849-1:2006 Safety of machinery - Safety-related parts of control systems - Part

1: General principles for design. CEN, Brussels, 2006.

FprEN12999:2010 (Final Draft) Cranes - Loader cranes. CEN, Brussels, 2010.

O. Pfeiffer, A. Ayre and C. Keydel Embedded Networking with CAN and CANopen
Copperhill Media, MA, USA, 2003.

CAN in Automation (CiA), e.V. CANopen, Application Layer and Communication

Profile. CiA 301, 2002.