Fluctuations in Magnesium Treatment of Ductile Iron

Some Reasons Some Solutions

Excerpts from PowerPoint presentation by Mr. Pierre-Marie Cabanne, Manager, Sorelmetal Technical Service, Rio Tinto
Iron & Titanium Inc., Canada, at the 1-day S. G. Iron Conclave conducted jointly by Rio Tinto Iron & Titanium, and Kastwel
Foundries, Ahmedabad in Mumbai on 16-9-11.

(A) Introductory Remarks

This Subject is often discussed during our visits to various foundries. Recommendations are easy to give if the foundry has
sound process control practice (in terms of weighing practice, metal temperature, sulphur content in molten metal etc.) However,
different process parameters need to be analyzed.
Please Remember : FeSiMg is a costly alloy. Any saving in its use and consistency in process saves you money, besides
improved consistency in Quality of Ductile Iron (DI) castings!

(B) General Formula

General Formula for Calculating the Weight of FeSiMg Alloy needed for Mg Treatment
WFeSiMg = Wladle x

(Mgfinal + 0.76 (Si - Sf) + 10-3 t)

x (T C/1450)2

% MgFeSiMg x R
Where:
W FeSiMg = Weight of FeSiMg (kg)

Si = initial (before Mg treatment) Sulfur in the iron (%)

W Ladle = Weight of molten metal in ladle (kg)

Mg final = Mg in molten iron after treatment (%)

t = time between treatment and beginning of pouring (minute)

% Mg FeSiMg = Mg content in FeSiMg alloy (%)

R = Yield of Mg in Mg treatment (%)

Sf = final (after Mg treatment) Sulfur in the iron (%)

T = temperature of iron (C)

(C) Effect of Variation in Each Parameter

1. W Ladle = weight of molten metal in ladle (kg)

2. W FeSiMg = Weight of FeSiMg (kg)

Only the cored wire process takes into account the weight of
molten Dl in the ladle. Other processes are influenced by the
fluctuation of weight of molten metal in the ladle.

All Mg treatment processes present a possible variation

in the weight of FeSiMg
A fluctuation of 1% is frequent (due to inaccuracy of the
balance).

A fluctuation of 5% is normal, and automatically generates

variations in the final Mg content of treated iron.
Parameter

(kg)
(%)
(%)
Mg final
(%)
Si
(%)
(%)
Sf
t
(minutes)
T
(C)
W FeSiMg
(kg)

Target

W ladle

Actual

(kg)
(%)
R
(%)
Mg final
(%)
(%)
Si
(%)
Sf
t
(minutes)
T
(C)
W FeSiMg
(kg)

W ladle
Mg FeSiMg

Change (%)

1000

1050

5.0

Nil

60

60

Nil

Mg FeSiMg
R

Parameter

0.0350

0.0329

-6

0.01

0.01

Nil

0.005

0.005

Nil

Nil

1450

1450

Nil

12.17

12.17

Nil

Target

Actual

Change (%)

1000

1000

Nil

Nil

60

60

Nil

0.0350

0.0355

1.4

0.01

0.01

Nil

0.005

0.005

Nil

Nil

1450

1450

Nil

12.17

12.30

1.1

Actions to minimize this variation

Actions to minimize this variation

Use of clean balance with calibration

Use of a good balance

Frequent audits of the automatic distributor system

Good hydraulic lay-out on the furnace, low speed of

tapping towards the end of pouring

Capability of the balance

Training to Operators

Good design of the furnace neck and ladle

********

Good exhaust system

(continued on next page)

Training to Operators

********
FOUNDRY

33

SEPTEMBER/OCTOBER, 2011

3. Si = initial Sulfur (before Mg treatment)

5. Mg final = final Mg before pouring

Depending on the melting process and charge composition

Si may fluctuate between 0.08% and 0.01 %.

Final Magnesium before pouring depends on :

Parameter

(kg)
(%)
(%)
Mg final
(%)
Si
(%)
(%)
Sf
t
(minutes)
T
(C)
W FeSiMg
(kg)
W ladle

Target

Actual

1000

1000

Nil

Nil

60

60

Nil

0.0350

0.0312

-11

0.01

0.015

50

0.005

0.005

Nil

Nil

1450

1450

Nil

12.17

12.17

Nil

Mg FeSiMg
R

casting thickness,
level of S, Zn and free oxygen in the iron (i.e. quality of
charge material, melting process etc.),

Change (%)

time between pouring of first mould and last one (in that
case the time t will include this period also).
Final Magnesium before pouring depends on the level of S,
Zn and free oxygen in the iron (i.e. quality of charge material,
melting process etc.):
%Mg* final = MgS + MgO + MgZnO3 + MgX + free Mg
(* Spectro results always show total Mg, i.e. %Mg final.)

Only the free Mg promotes nodularization, and a very

small part of MgS and MgO will create some nuclei !
Many defects (like poor graphite shape) come from too
low free Mg.

Actions to minimize this variation

Good Raw Materials & Charge Composition

Excessive Magnesium content also promotes defects

such as carbides, dross, micro-shrinkage!!

Good Recarburiser and Ferro-Alloys

Good Process Control (Spectro, Leco)

Preventive Action : Thus we ought to look at the Pl ratio,

like our HPl, to minimize the presence of free O, S and Zn
from steel scraps!

Use of Sorelmetal !
The best advantage of Sorelmetal versus other metallic
charges is illustrated in the table below.

********

Sorelmetal
RF10
Other Charges
W ladle

(kg)

1000

(%)
(%)
Mg final
(%)
Si
(%)
Sf
(%)
(minutes)
t
T
(C)
(kg)
W FeSiMg
Mg FeSiMg

60

60

0.0350
0.0073
0.005
5
1450
11.60

0.0350
0.0153
0.005
5
1450
13.29

Rs. 1450

Rs. 1660

Parameter

(kg)
Mg FeSiMg
(%)
(%)
R
(%)
Mg final
Si
(%)
Sf
(%)
t
(minutes)
(C)
T
W FeSiMg
(kg)

i.e. Use of Sorelmetal saved Rs. 210 in treatment of 1 tonne.

Target

Actual

Change (%)

1000

1000

Nil

5.8

-3.3

60

60

Nil

0.0350

0.0335

-4.2

0.01

0.01

Nil

0.005

0.005

Nil

Nil

1450

1450

Nil

12.17

12.17

Nil

Actions to minimize this variation

********

Selection of a good supplier of FeSiMg alloy

4. Sf = final Sulfur after Mg Treatment

Frequent audits of the melting platform

Final sulfur is more a result of the Mg treatment rather

than other influencing parameters. However, depending
on the quality of the FeSiMg (e.g. presence of Ce & Ca
and their level ), its content can fluctuate.

Capability on the product

Training to operators

********

The higher the initial sulfur, the higher the final sulfur, and
the higher the risk of getting dross in the casting.

7. t = time between end of treatment and beginning

of pouring

We recommend final Sulfur < 0.005% for heavy castings,

or between 0.005% & 0.015% for usual castings.

Only a fluctuation of 2 minutes in the time t between end of

treatment and pouring decreases the level of Mg by 0.002%.
Causes are mainly due to transport and slag removing.
Excesses of slag comes from the quality of charge material:
sand, rust, sterile, steel coating etc.

Many foundries have decided to switch from the use of

normal Pl to our UHPI to minimize the dross defect: heavy
castings or big thickness, castings submitted to impact
& fatigue,. . .

********

% of Mg in FeSiMg Alloy

W ladle

* Assuming FeSiMg Alloy Price at Rs. 125/kg

FOUNDRY

FeSiMg =

Often, this ratio is overlooked. But FeSiMg containing 5-7%

Mg presents a natural variation close to 0.5 (0.25)% due to
different lots, grain size segregation, grain size, origin etc.

1000

Alloy Cost* of Mg treatment

6. %Mg

34

SEPTEMBER/OCTOBER, 2011

Parameters

(kg)
Mg FeSiMg
(%)
(%)
R
Mg final
(%)
Si
(%)
Sf
(%)
(minutes)
t
T
(C)
W FeSiMg
(kg)

An increase of pouring temperature by 50 C decreases the

final magnesium by 0.003%.

Target

Actual

Change (%)

1000

1000

Nil

Nil

Parameter

(kg)
(%)
(%)
Mg final
(%)
(%)
Si
Sf
(%)
t
(minutes)
(C)
T
(kg)
W FeSiMg

W ladle

60

60

Nil

0.0350

0.0330

-5.71

0.01

0.01

Nil

0.005

0.005

Nil

40

1450

1450

Nil

12.17

12.17

Nil

W ladle

Mg FeSiMg
R

Actions to minimize this variation

Good maintenance and transportation machine

Target

Actual

Change (%)

1000

1000

Nil

Nil

60

60

Nil

0.0350

0.0321

-8.3

0.01

0.01

Nil

0.005

0.005

Nil

Nil

1450

1500

3.4

12.17

12.17

Nil

Training to Operators

Actions to minimize this variation

Good Raw Materials to minimize dirt & slag

Use of insulating refractory + cover for ladle(s)

Small addition in the ladle

********

Quick transfer(s) and short pouring period

8. t : In case of pouring of large number of castings

from the same ladle, time t needs to be considered
as time between Mg treatment and pouring of last
mould.

********
10. R = yield of Mg treatment
From the mathematical formula, we automatically get the
yield, and all the past parameters could have an influence!

For a final Mg of 0.035%:

If total pouring period is 15 minutes, FeSiMg needed is 14.94
kg.

WFeSiMg = Wladle x

If this time becomes only 10 minutes, the need of FeSiMg

decreases to 13.56 kg.
Parameter

(kg)
(%)
Mg FeSiMg
R
(%)
Mg final
(%)
(%)
Si
Sf
(%)
t
(minutes)
(C)
T
(kg)
W FeSiMg

(Mgfinal + 0.76 (Si - Sf) + 10-3 t)

x (T C/1450)2

% MgFeSiMg x R

However, yield is a typical parameter like the previous ones,

And, it is more a sum of subjective items than purely a
mathematical list.

Target

Actual

Change (%)

1000

1000

Nil

Actions to minimize this variation

Nil

60

60

Nil

0.0350

0.0350

Nil

Quality of the spectrometer and spectro samples, how and

where the spectro sample is taken, its design, cleanliness
of the ladle (body and pocket) and quality of its lining

0.01

0.01

Nil

Quality and cleanliness of the steel cover

0.005

0.005

Nil

Output of the furnace during tapping

10

15

50

Time between input of FeSiMg and beginning of tapping

1450

1450

Nil

13.56

14.94

10.2

Level of oxidation of molten iron and practice of preconditioning

W ladle

Exhaust system directly attached on the ladle as in case

of cored wire process etc. ..

Actions to minimize this variation

Good maintenance and transportation machine

If all the previous parameters are under control, we

recommend to recalculate the yield for each treatment.

Training to Operators
Ladle with Cover
Automatic pouring line

In case of major deviation, the foundrymen will be alerted

and will check all items in the precedent list.

********
9. T = treatment temperature (C)

For a similar design of ladle and process (melting

treatment,) the yield can vary from 60 to 40% from one
foundry to another!!!

Treatment temperature is the sum of:

Pouring temperature (function of casting thickness and type
of risering) + Loss of temperature during transfer and
pouring* + Loss of temperature during Mg treatment**.

Our job is to give information to our customers or

potential customers.

********

(* between 5 to 12 C per minute)

(** 10 kg of cold ferro-alloy in a ladle of 1 tonne at 1500 C
decreases the temperature of treated iron by 15 C.)

FOUNDRY

(continued on next page)

35

SEPTEMBER/OCTOBER, 2011

11. Summary of Calculations

(a) For the same final Mg = 0.035 %, if we consider all the
possible fluctuations (on industrial level), Mg alloy cost of
the treatment will become:
Parameter

Target with
RF10 & Well
Process

(kg)
Mg FeSiMg (%)
R
(%)
Mg final
(%)
(%)
Si
Sf
(%)
t
(minutes)
T
(C)
W FeSiMg (kg)

W ladle

Cost of (Rs.)
FeSiMg Alloy

Actual Change (%)

or Others

1000

1050

5.7

-5

60

57

-5

0.0350

0.0350

Nil

0.0073

0.0153

110

0.005

0.005

Nil

20

1450

1500

3.4

11.60

16.89

46

1,450

2,110

46
(i.e. Rs.660)

(b) But for the same weight of FeSiMg, if we consider all the
negative fluctuations (on industrial level), the final Mg will
become:
Parameter

(kg)
(%)
(%)
(%)
Mg final
Si
(%)
Sf
(%)
t
(minutes)
(C)
T
W FeSiMg
(kg)

W ladle

Mg FeSiMg
R

Target

Actual

Change (%)

1000

1050

5.7

-5

60

57

-5

0.0350

0.0197

-44

0.0073

0.0153

110

0.005

0.005

Nil

20

1450

1500

3.45

11.60

11.60

Nil

(D) Concluding Remarks

The final Magnesium content is one of the most important
parameters to obtain spheroidal graphite. Possible causes
for its fluctuation are multiple.
The quality of the process, including charge materials, is
the major way to get good castings in Dl.
Rio Tinto Iron & Titanium, our Agents, our Literature, our
Technical Services, and our famous Sorelmetal can help
foundries to maintain magnesium treatment under control.
123456789012345678
123456789012345678

Ductile Iron Foundrymen!

This presentation is valuable. You will find its content
highly educative. Please read it carefully and study
the calculations shown here. It will act as a Road Map
to improve Process Control in your foundry,
particularly for producing high value and/or heavy
ductile iron castings for stringent and/or sub-zero
temperature applications.
Editor

FOUNDRY

36

SEPTEMBER/OCTOBER, 2011