Study of the Distribution of Bolt Forces in a Multi-Bolted

System under Operational Normal Loads

Rafał Grzejda1, a)
1
West Pomeranian University of Technology, Szczecin
Faculty of Mechanical Engineering and Mechatronics, 19 Piastów Ave., 70-310 Szczecin, Poland.
a)
Corresponding author: rafal.grzejda@zut.edu.pl

Abstract. The paper is related to modeling of multi-bolted joints treated as a system. The model of the joint with various
substitute models of bolts is presented. The system was pre-tensioned and externally loaded with normal forces.
Calculations were made for an exemplary asymmetric joint. As a result of these calculations, distributions of operational
forces in the bolts are shown.

INTRODUCTION

In many engineering applications to connect components bolted joints are used. Most often they are designed as
multi-bolted joints [1-3]. In order to provide such joints sufficient load-carrying ability in operational conditions,
they are pre-tensioned according to different techniques [4-6].
Various analyzes of multi-bolted joints in operational conditions have so far been presented in several papers.
They mostly concerned traditional types of bolted connections: flange joints [7-9], beam-to-column joints [10-12]
and lap joints [13-15]. Abid [7] has proposed safe operational conditions for a single bolted connection separated
from a flange joint and loaded with a combination of internal pressure and temperature. Murali Krishna et al. [8]
have published a study on the sealing performance of flange joints with three types of gaskets. Gorla and Tanawade
[9] have described results of a probabilistic structural and thermal analysis of a gasketed flange. Liu et al. [10] have
investigated the behaviour of beam-to-tubular column angle joints under shear loads. Li et al. [11] have studied the
performance of semi-rigid joints between rectangular tubular columns and steel beams. Saberi et al. [12] have
compared sensitivity to the bolt diameter on the cyclic behaviour in the case of two different beam-to-column joints.
Ju et al. [13] have studied the structural behaviour of three-bolted lap joints. Salih et al. [14] have carried out a study
on the bearing behaviour of lap joints between thick and thin plates. Oinonen and Marquis [15] have simulated the
shear damage of adhesive reinforced bolted lap joint interfaces. All these publications are based on the finite
element method (FEM) and are related to FE-models of steel pre-tensioned and externally loaded joints, however in
none of them a systemic approach to modeling of bolted joints is not taken into account.
By treating the multi-bolted joint as a system, it is possible to separate subsystems from this system, which can
be modeled in various way. These subsystems include: a pair of connected elements, a contact layer between them
and a set of bolts. Assuming the normal direction of load of the joint, the contact layer can be represented by the
nonlinear [16] or linear Winkler model [17]. The set of bolts, by contrast, can be replaced with very different
models, which are listed e.g. in [18]. In the paper, the following two of these models are used:
• rigid body bolt models (RBB models) – models consisting of a flexible shank of the bolt modeled as a beam
and a rigid head of the bolt,
• spider bolt models (SB models) – models consisting of a flexible shank of the bolt modeled as a beam and
a flexible head of the bolt modeled as a set of beams with a total volume equal to the volume of the head in
the actual bolt.
 The subject of modeling nonlinear pre-tensioned multi-bolted systems on the operational state has been initiated
in [19], and the current paper is an extension of this subject. The aim of the study is to compare the values of
operational bolt forces determined for the following four FE-models of a steel multi-bolted system:
• RBB-NL – nonlinear system with RBB models and the nonlinear contact layer between the connected
elements,
• RBB-L – linear system with RBB models and the linear contact layer between the connected elements,
• SB-NL – nonlinear system with SB models and the nonlinear contact layer between the connected elements,
• SB-L – linear system with SB models and the linear contact layer between the connected elements.

METHODOLOGY FOR CREATING A MULTI-BOLTED SYSTEM

The structure of the multi-bolted system is based on the model of subsystems shown schematically in Fig. 1.
These subsystems include: a pair of connected elements (subsystems F and S), a contact layer between them
(subsystem C) and a set of bolts (subsystem B).

2 RBB model

SB model
1

FIGURE 1. Multi-bolted system (1 – subsystem B, a set of the bolts, 2 – subsystem F, the flexible flange element,
3 – subsystem C, the contact layer, 4 – subsystem S, the flexible support)

Between the above-mentioned subsystems occur dependencies that can be visualized using the stiffness matrix of
the system:

KBB KBF 0 KBS

K KFF KFC 0
K= [ FB ] (1)
0 KCF KCC KCS
KSB 0 KSC KSS

where: Kii – stiffness matrices of individual subsystems, Kij – matrices of elastic couplings between separated
subsystems, i  {B, F, C, S}, j  {B, F, C, S} (for comparison, see Fig. 1).
The equation of the equilibrium state of the multi-bolted system shown in Fig. 1 can be written in the form:

K∙q = p (2)

where: q – displacements vector of the system, p – loads vector of the system.

RESULTS OF CALCULATIONS

Exemplary calculations were made for a selected asymmetrical multi-bolted system shown in Fig. 2a. The shape
of the connected elements is irregular in order to give the system the features of generality. The thickness of the
connected elements is equal to 20 mm. The model of the joint has been fastened by:
• receiving all degrees of freedom on the bottom surface of the support,
• receiving degrees of freedom in a plane perpendicular to the axis of bolts in five nodes on the side surface of
each of the connected elements.
 1.06
0.98
0.9
0.82
0.74
1 2 3 4 5 6 7
SB-L SB-NL RBB-L RBB-NL
1.06
4 5
FI 3 FII 0.98

Fe /Fm
0.9
2 FIII
6
0.82
7
1 0.74
1 2 3 4 5 6 7
Number of the bolt
(a) (b)

1.06 1.06

0.98 0.98

Fe /Fm
Fe /Fm

0.9 0.9

0.82 0.82

0.74 0.74
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Number of the bolt Number of the bolt

(c) (d)

FIGURE 2. Example of calculations: a) FEM-based model of the multi-bolted system and numbering of the bolts, b) distribution
of operational bolt forces for the system loaded by the force FI, c) distribution of operational bolt forces for the system loaded by
the force FII, d) distribution of operational bolt forces for the system loaded by the force FIII

The joint was assembled by means of seven M10 bolts with the preload Fm equal to 20 kN. In this case, all the
bolts were pre-tensioned simultaneously, but it is also possible to include in the calculations the bolt tensioning
conducted according to a specific sequence. Two types of bolt models were adopted for modeling of the subsystem
of bolts: RBB models and SB models. In the current models of the system threaded connections between the bolts
and the support are not included. Beams modeling the shanks of individual bolts are connected to the support in one
node.
The pre-tensioned multi-bolted system was subjected to an external normal force Fa equal to 30 kN and applied
in three positions as shown in Fig. 2a (for a  {I, II, III}). The distribution of operational forces in the bolts Fe,
related to the preload Fm, for the multi-bolted system loaded externally by the force Fa is illustrated in Fig. 2b–d.

TABLE 1. List of calculation results for the Z1 index

Model External load Bolt number Z1, %
RBB model FI 2 1.9
RBB model FI 3 1.8
RBB model FII 5 2.0
RBB model FII 6 1.8
RBB model FIII 1 1.6
RBB model FIII 7 1.8
SB model FI 2 -1.2
SB model FI 3 -1.1
SB model FII 5 -2.0
SB model FII 6 -1.5
SB model FIII 1 -1.0
SB model FIII 7 -1.3
 The assessment of the usefulness of linear models of the multi-bolted system can be made on the basis of the Z1
index:

FeFEM-L - FeFEM-NL
Z1 = ∙100 (3)
FeFEM-NL

where: FeFEM-L – operational force in the bolt according to the linear model of the multi-bolted system, FeFEM-NL –
operational force in the bolt according to the nonlinear model of the multi-bolted system.
The values of the Z1 index obtained for the bolts with the largest decrease in the operational force in relation to
their preload are collected in Table 1.
Based on the results listed in Table 1 it can be noticed that the use of linear models may cause the lowering of
the values of forces in the most loaded bolts by 2 % in the case of the system with spider bolt models and the
overestimation of the values of forces in the most loaded bolts by 2 % in the case of the system with rigid body bolt
models.
The assessment of the usefulness of adopted bolt models can be made on the basis of the Z2 index:

FeRBB - FeSB
Z2 = ∙100 (4)
FeSB

where: FeRBB – operational force in the bolt according to the multi-bolted system with rigid body bolt models, FeSB –
operational force in the bolt according to the multi-bolted system with spider bolt models.
The values of the Z2 index obtained for the bolts with the largest decrease in the operational force in relation to
their preload are collected in Table 2.

TABLE 2. List of calculation results for the Z2 index

Model External load Bolt number Z2, %
NL model FI 2 15.3
NL model FI 3 13.9
NL model FII 5 11.6
NL model FII 6 9.4
NL model FIII 1 7.2
NL model FIII 7 12.0
L model FI 2 18.9
L model FI 3 17.2
L model FII 5 16.0
L model FII 6 13.0
L model FIII 1 10.1
L model FIII 7 15.5

Based on the results listed in Table 2 it can be noticed that the use of partially rigid bolt models instead of
flexible bolt models may cause the overestimation of the values of forces in the most loaded bolts by about 16 % in
the case of the nonlinear multi-bolted system model and by about 19 % in the case of the linear multi-bolted system
model. Further recognition of the suitability of the considered bolt models in the modeling of multi-bolted systems
should be continued based on experimental research.

CONCLUSIONS

In the paper a general systemic approach to modeling and calculations of arbitrary multi-bolted systems is
presented. It can be implemented in the operational state of the system for different bolt models and different models
of the contact layer between the connected elements. It is planned to carry out experimental research to determine
the usefulness of substitute bolt models in the modeling of multi-bolted systems.
Due to the fact that calculations were prepared for a small seven-bolt system, it was considered that providing
results for three external forces positions will be sufficient to show the specificity of operational force changes in the
bolts. Calculations for other load cases will give a similar effect: the largest decreases in operational forces will
occur in the bolts nearest to the location of the external forces.

REFERENCES
1. F. Hu, G. Shi, Y. Bai, and Y. Shi, J. Constr. Steel Res. 102, 204–216 (2014).
2. H. Augusto, L. S. da Silva, C. Rebelo, and J. M. Castro, J. Constr. Steel Res. 116, 271–293 (2016).
3. H. Ma, S. Ren, and F. Fan, Eng. Struct. 126, 725–738 (2016).
4. I. Coria, M. Abasolo, I. Olaskoaga, A. Etxezarreta, and J. Aguirrebeitia, Ocean Eng. 129, 441–450 (2017).
5. R. Grzejda, Procedia Engineer. 206, 1808–1812 (2017).
6. R. Grzejda, “Impact of nonlinearity on bolt forces in multi-bolted joints: A case of the assembly stage,” Sci.
Iran. Trans. B (in press).
7. M. Abid, Int. J. Pres. Ves. Pip. 83, 433–441 (2006).
8. M. Murali Krishna, M. S. Shunmugam, and N. Siva Prasad, Int. J. Pres. Ves. Pip. 84, 349–357 (2007).
9. R. S. R. Gorla and A. Tanawade, Therm. Eng. 59, 535–541 (2013).
10. Y. Liu, C. Málaga-Chuquitaype, and A. Y. Elghazouli, Eng. Struct. 42, 434–456 (2012).
11. G. Li, H. Yu, and C. Fang, Front. Struct. Civ. Eng. 7, 296–303 (2013).
12. H. Saberi, A. Kheyroddin, and M. Gerami, Int. J. Steel Struct. 14, 633–647 (2014).
13. S.-H. Ju, C.-Y. Fan, and G. H. Wu, Eng. Struct. 26, 403–413 (2004).
14. E. L. Salih, L. Gardner, and D. A. Nethercot, Eng. Struct. 33, 549–562 (2011).
15. A. Oinonen and G. Marquis, Eng. Fract. Mech. 109, 341–352 (2013).
16. R. Grzejda, “Modelling nonlinear multi-bolted connections: A case of operational condition, ” in Proceedings of
the 15th International Scientific Conference “Engineering for Rural Development 2016“, (Jelgava, Latvia,
2016), pp. 336–341.
17. R. Grzejda, “Impact of nonlinearity on bolt forces in multi-bolted systems: a case of the assembly state,” in
Proceedings of the 22nd International Scientific Conference “MECHANIKA 2017”, (Birštonas, Lithuania,
2017), pp. 110–113.
18. R. Grzejda, “New method of modelling nonlinear multi-bolted systems,” in Advances in Mechanics:
Theoretical, Computational and Interdisciplinary Issues, edited by M. Kleiber et al. (CRC Press/Balkema,
Leiden, 2016), pp. 213–216.
19. R. Grzejda, Engng. Trans. 64, 525–531 (2016).