Thermal Analysis and Process Control

for Compacted Graphite Iron

and Ductile Iron

©SinterCast 2014 www.sintercast.com

 Thermal Analysis and Process Control for Compacted Graphite Iron and Ductile Iron

S Dawson, P Popelar
SinterCast, Sweden

Abstract
Traditionally, cast iron microstructures have been controlled by measuring and adjusting the chemistry of
the liquid iron, the individual chemical elements being like the pieces of a jigsaw puzzle that the foundry can
assemble to infer the microstructure. However, spectographic analysis is not able to provide information about
the activity of the individual elements, to measure critical elements such as oxygen, or to indicate the efficiency
of the inoculant. Some pieces of the puzzle are always missing; the picture is incomplete. In contrast, thermal
analysis ‘listens’ to the graphite as it precipitates and grows. Regardless of the underlying chemical factors,
if a graphite particle grows as worm, it will show a worm’s heat. If it grows as a nodule, it will show a nodule’s
heat. Thermal analysis allows the foundry to see how the graphite is growing – similar to a micro but faster, and
with a larger sample size. While the spectrometer has for many years been the ultimate judge in the foundry,
the present paper proposes an evolution toward control by microstructure rather than by chemistry. Indeed, for
the determination of feedforward and feedback process control actions, it is proposed that chemistry should be
regarded as subordinate to microstructure determination by thermal analysis.

Introduction
As one of the largest consumers of castings, the automotive industry is undergoing a transformation.
Governments around the world are implementing increasingly stringent fuel economy and emissions standards,
with fuel economy in the US increasing from 27.5 mpg in 2010 to 54.5 mpg in 2025 and CO2 emissions in Europe
decreasing from 130 g/km (42 mpg) in 2010 to 95 g/km in 2020 (57 mpg). This legislation is motivating a wide
range of new technologies including lightweight cast components and body panels, downsized gasoline and
diesel engines, electric powertrains, improved aerodynamics and reduced rolling resistance.

The demand for more from less in the automotive industry places a parallel demand on the foundry industry.
In order to enable design engineers to reduce safety factors and trim component weight, microstructures must
be consistently controlled in a narrower range. And, in order to ensure cost effective production and global
competitiveness, castings must be produced right-first-time, satisfying the property requirements with a minimum
of alloying and feeding. One important element in achieving these goals is the continued transition from after-
the-fact quality control toward on-line process control.

Process Stability and Control Strategy

The need for process control is defined by the microstructure and property requirements of each casting. In
the case of Compacted Graphite Iron (CGI), complex components such as cylinder blocks and heads require
a narrow range of 0-20% nodularity, with no flake graphite, in order to provide the optimum combination of
castability, strength, machinability and thermal conductivity. For ductile iron, safety critical components must be
produced with more than 90% nodularity and a high graphite particle count, without resorting to excess alloying
that can cause casting defects and cost inefficiencies. In both cases, the microstructure requirements must be
maintained until the end of casting, even as the magnesium, inoculant and temperature fade.

Traditionally, the control of cast iron microstructures has been achieved through chemistry measurement.
However, chemical analyses provided by a spectrometer only show the total amount of each element and do
not provide any information about compounds that could serve as nuclei or the activity of the elements such as
the dissolved magnesium, oxygen and sulphur that influence the growth behaviour of the graphite. Therefore,
chemistry results provide only limited information about the expected solidification of the iron.

2
The reliance on chemical analysis is also limited by the changing conditions in the production environment, as
changes in the raw material mix and changes in melting and holding techniques influence the active oxygen and
sulphur content of the iron. The net result is that the stable microstructure range for both CGI and ductile iron
is a moving target. Ironically, ensuring that the chemistry is the same in every ladle will result in microstructure
variation. The most effective strategy to reliably remain within the specified microstructure range is to measure
the behaviour of the iron in each ladle and to dynamically adjust the magnesium and inoculant additions according
to the needs of the iron. Depending on the production conditions and process flow, this dynamic adjustment can
be applied by feedforward or feedback control. Chemical analysis remains important for the control of the base
iron chemistry, alloying, and tramp elements, but for microstructure control, chemistry should be regarded as
subordinate to the measured solidification behaviour.

The influence of the magnesium and inoculation on cast iron microstructures is illustrated in Figure 1, which
also serves to introduce the term “Modification”. The graphite morphology in cast iron is determined by the
combined influence of all nodularising elements, such as magnesium, rare earths, calcium, etc., and all
subversive elements, such as oxygen, sulphur, titanium, etc. Therefore, the horizontal axis of the microstructure
chessboard is referred to as “Modification” rather than simply as magnesium. In terms of thermal analysis, when
the cooling curves respond to the growth of graphite, it is as a result of the influence of all variables rather than
simply the magnesium, so the term “Modification” provides a more holistic representation of the graphite growth.
The microstructure chessboard shows how the foundry can influence the graphite morphology and particle count
by varying the Modification and Inoculation additions and therefore provides a quantitative basis for defining the
control limits of the stable windows for CGI and ductile iron production.

Figure 1: The microstructure chessboard provides a platform for defining

the control limits of the stable windows for CGI and ductile iron production.

Compacted Graphite Iron is stable within a four-sided window. When the Modification is too low, flake graphite
will form leading to local weakness and the risk of premature failure in service. When the Modification is too
high, excess nodularity will form, increasing the risk of porosity defects, impairing machinability and potentially
leading to thermal failures due to reduced thermal conductivity. Excess nodularity can also be caused by high
inoculation, as the increase in the number of nuclei favour the precipitation of nodular rather than compacted
graphite particles. The stable CGI window is also bounded at the lower inoculation limit by the onset of patchy
graphite distribution and thereafter by the potential for carbide formation. While many early attempts to produce
CGI focused on magnesium control, the perspective of the microstructure chessboard shows that the reliable
high volume production of CGI requires simultaneous control of the Modification and Inoculation, and that these
parameters must be given equal importance.

3
 In contrast to CGI, high quality ductile iron exists in the top-right corner of the chessboard and therefore affords
some opportunity for the overtreatment of both Modification and Inoculation to buffer the risk against low
nodularity, low nodule count, patchy graphite distribution and carbide formation. However, excess Modification
and Inoculation also increases the risk of casting defects and burdens the foundry with unnecessary ferroalloy
consumption and unnecessarily large feeders. Therefore, the objective of ductile iron production is to control
the iron within a narrow and low Modification and Inoculation window that reduces alloy consumption, defect
formation and porosity in order to ensure the consistency of the microstructure and the material properties, and
to maximise the economic efficiency related to mould yield and machinability.

Thermal Analysis
The mantra of process control is: you can’t control what you can’t measure. Accordingly, process control must
begin with an accurate measurement of the iron.

In order for a liquid to transform to a solid, it must release heat. The basis of thermal analysis is that this liberation
of heat, illustrated in cooling curve shown in Figure 2, can be measured and correlated to the solid phases that
are precipitating and growing. As the iron transforms from the liquid state to primary austenite, the liberated
heat causes an arrest in the natural cooling of the iron. The temperature at which this arrest occurs, shown at
Point A in Figure 2, provides a direct measurement of the Carbon Equivalent. Therefore, the first information
provided by thermal analysis of hypoeutectic irons is the Carbon Equivalent, which is important in determining
the shrinkage behaviour of the iron. The Carbon Equivalent also indicates the risk for graphite flotation and for
exploded graphite in ductile iron.

Figure 2: In hypereutectic irons, the heat released by

primary austenite results in a thermal plateau at Point A.
Thereafter, the onset of the eutectic solidification of graphite
and austenite defines the undercooling at Point B, and the
graphite growth beyond.

The Carbon Equivalent result determined by thermal analysis is different from the CE value derived from
chemical analysis (usually defined as C + ⅓Si) and from the definition of the eutectic CE value shown in the
iron-carbon phase diagram. This difference is because both the chemical value and the phase diagram are
defined under thermodynamic conditions, which means that the iron is under equilibrium, as could be achieved
by the controlled slow cooling of a thermal analysis sample over many hours. However, under normal production
conditions, the iron cools much faster than equilibrium. The CE result provided by thermal analysis therefore
represents the actual physical or kinetic solidification that occurs in the casting.

4
Physically, the eutectic is simply defined as the temperature at which the austenite and the graphite grow
simultaneously. Under equilibrium conditions, slow cooling provides enough time for the graphite to precipitate
out of solution, and therefore, the graphite can precipitate at a higher temperature: reaction kinetics play no role.
However, in series production, the carbon atoms require a driving force to squeeze them out of solution. This
force is provided by temperature (undercooling) and aided by the addition of inoculant – as more inoculant is
added and more nuclei are formed, less undercooling is needed to push the carbon out of solution.

With reference to Figure 2, as the Carbon Equivalent increases, the CE plateau (austenite plateau) at Point A
moves downward. The decrease is approximately 1°C for each 0.01 increase in CE. At the same time, as the
inoculation increases, the additional nuclei make it easier for graphite to precipitate, so the eutectic undercooling
defined by Point B moves upward. The point at which the CE plateau (A) merges with the undercooling (B) is
the eutectic, regardless of the CE value. Under production conditions, depending on the inoculation level of the
iron and the cooling conditions of the thermal analysis sample (or casting), CE values of up to approximately
4.55 can be measured before the eutectic is reached. In the example of Figure 2, it can be seen that the CE
is approximately 0.14 units below the eutectic. If eutectic solidification is targeted for this casting, the foundry
can adjust the carbon and the inoculant to bridge the 0.14 gap. Again, as with the Modification and Inoculation
window, it is seen that even the eutectic is a moving target and that chemical analysis cannot be relied upon to
determine the proximity of the iron to the eutectic point. Thermal analysis is required to define the true eutectic
point. Chemical analyses and phase diagrams should be regarded as subordinate to the measured solidification
behaviour.

Beyond the initial arrest in the cooling curve caused by the formation of primary austenite, the precipitation of carbon
atoms to form graphite particles triggers a second change in the cooling curve. As carbon atoms are consumed
from the liquid, the solidification temperature of the austenite increases. This constitutional undercooling effect
influences the speed of the solidification of the austenite. The sequence of events is (i) the speed of the graphite
growth determines the rate of consumption of carbon from the liquid, (ii) the rate of consumption of carbon from
the liquid determines the speed of the austenite growth, (iii) the speed of the austenite growth determines the
rate of heat released, and (iv) the released heat released determines the shape of the cooling curve. Working
backwards, the shape of the cooling curve can be used to determine the graphite morphology.

Figure 3: A typical 50 μm graphite nodule contains more

than 1020 carbon atoms. The rate of consumption of carbon
from the liquid, determined by the shape of the graphite,
controls the heat released during solidification and provides
the opportunity for measurement and control.

5
 The amount of heat liberated by the deposition of each carbon atom is the same, regardless of whether the atom
precipitates onto a flake, a ‘worm’, or a nodule. Further, the amount of heat liberated by the graphite, which only
represents approximately 3% by weight and 10% by volume, is dwarfed by the dominance of the heat released
by the austenite and the heat lost from the sample vessel to the atmosphere. But despite this, the graphite
growth still dictates the shape of the cooling curves - not directly by releasing heat, but indirectly by consuming
carbon from solution and controlling the austenite growth. The reason that the graphite is so influential is that the
number of carbon atoms that are precipitating is so immense. Each 50 μm graphite nodule, as shown in Figure
3, consumes more than 1020 carbon atoms and, in a simple thermal analysis sand cup, more than 1025 carbon
atoms will precipitate from solution in less than two minutes. Ultimately, it is the difference in the speed at which
the flakes, worms and nodules grow that determines the rate of carbon consumption from the liquid, the speed
of the austenite growth, and thus, the shape of the cooling curves.

Grey iron flakes grow in a eutectic cell, with many flakes branching from a single nucleation site. The flakes
grow in direct contact with the liquid iron, with the carbon atoms precipitating on the A-face (Prism plane) of the
basic graphite building block, in a coupled eutectic. The single nucleation event, combined with many points of
direct contact with the liquid, results in fast growth that allows the flakes to precipitate and grow with little or no
undercooling. In contrast, the precipitation of each graphite nodule in ductile iron requires a unique nucleation
event, followed by initial growth through the precipitation of carbon atoms onto the C-face (Basal plane) of the
basic graphite building block, in a divorced eutectic, and final growth by the diffusion of carbon atoms through a
solid austenite shell. The individual nucleation results in more undercooling than grey iron while the low surface
area to volume ratio and eventual diffusion-controlled growth through the austenite results in slower growth.

The growth mechanism of compacted graphite particles is intermediate between that of grey iron and ductile
iron. Similar to grey iron flakes, compacted graphite particles grow in branched eutectic cells, from single
nucleation sites, in direct contact with the liquid iron. The many branches and direct contact with the liquid
provide a high surface area for the deposition on new carbon atoms, resulting in fast growth. The predominant
deposition of carbon onto the A-face results in the characteristic elongated aspect of the compacted graphite
particles while the concurrent deposition of carbon atoms onto the C-face results in the irregular bumpy surface
of the particles. In contrast to ductile iron, where aggressive inoculation is preferred to maximise the nodule
count, the inoculation of CGI is lower in order to suppress the tendency to form nodular graphite. The net result
is that CGI generally shows more undercooling than ductile iron, followed by fast growth due to the high surface
contact with the liquid iron, providing a steeper recalescence in the cooling curve.

As illustrated in Figure 4, the differences in the nucleation and growth of grey iron, ductile iron and CGI provide
distinct differences in the thermal analysis cooling curves, and these differences can be used to determine
solidification behaviour and to decide upon control actions. Rather than measuring the chemistry to infer
the graphite microstructure, the philosophy of process control by thermal analysis is to directly measure the
solidification behaviour of the iron, as determined by the graphite growth. Regardless of the underlying chemical
factors, if the graphite particles grow as worms, the iron will show a worm’s heat. If the graphite particles grow
as nodules, the iron will show a nodule’s heat. It only remains to ‘listen’ to the graphite.

Figure 4: A The differences in the nucleation and growth

behaviour of graphite flakes, worms and nodules provide
distinct differences in the cooling curves for grey iron (A),
6 CGI (B) and ductile iron (C).
The three primary output values from the thermal analysis are the Carbon Equivalent, the Modification index
and the Inoculation index. The Modification and Inoculation index values are determined from the evaluation
of more than 100 different parameters extracted from the cooling curves, the derivatives and the heat evolution
(power) curves. The Modification and Inoculation index values are used to define the size and location of the
calibration window on the microstructure chessboard for each component during the calibration phase, and to
quantify the location of each ladle during series production. Together with the CE result, the Modification and
Inoculation values define the microstructure and the risk for casting defects such as porosity and carbides. The
use of dimensionless indices simplifies the process for the operators and allows the foundry engineers to focus
on production optimisation rather than allocating resources to the fundamental task of interpretation of cooling
curves.

Measurement Techniques and Devices

Thermal analysis is a heat balance. The ultimate shape – and thus the resolution – of the cooling curve is
determined by the balance between the heat liberated during solidification and the heat lost to the sampling
vessel and the atmosphere. It is evident that the amount of heat liberated by the solidification of a 200 gram
sample of cast iron is fixed. If the 200 gram sample is contained in a vessel that cools quickly, the heat liberated
by the solidification will be less able to prevail over the heat loss than it would be in a vessel that cools more
slowly. The result is that the faster cooling of the vessel will provide less resolution in the cooling curves. Fast
cooling caused by the vessel can also alter the true solidification behaviour of the iron by inducing chill or by
influencing the undercooling. In order to extract as much information as possible from the heat liberated by
the solidification, it is necessary to design the thermal analysis vessel such that it neither masks nor dilutes
the information provided by the solidification. The other major requirement of a thermal analysis vessel is that
it must ensure consistent sampling conditions. Because the differences in the liberated heat between a good
microstructure and an out-of-spec microstructure can be very small, it is critical that all variations measured are
due to differences in the iron and not due to differences in the sampling technique.

While conventional sand cups provide adequate resolution for the determination of the austenite arrest
temperature, the geometry and cooling conditions of sand cups introduce some limitations on the ability to
accurately determine the graphite microstructure. For example, the filling of sand cups is liable to variation in the
fill volume and differences in the starting temperature of the iron. The pour-in filling technique also introduces
a source of variation, as the exposure of the pouring stream to the atmosphere causes inconsistent amounts
of oxidation and air entrainment during the filling of the cup. Once the cup is filled, the high heat capacity of
the thick sand walls, shown in Figure 5 (a), causes the vessel to act as a heat sink, influencing the solidification
behaviour and effectively competing against the heat liberated by the solidification. The open surface of the
sand cups also results in large radiation heat losses from the surface and imbalanced heat losses from the top,
the exposed sides, and the thicker bottom, which is also influenced by the measurement stand. The net result
is that the design of conventional sand cups can contribute to measurement variation and limit the resolution of
the cooling curves.

(a) Conventional thermal analysis sand cup with open (b) SinterCast immersion sampling steel cup with
surface and variable wall thicknesses. regulated heat loss, showing the protective tube for
the containment of the reusable thermocouples.

Figure 5: Immersion sampling and pre-heating of the vessel provide more consistent sampling conditions and
higher resolution in the cooling curves.

7
 In order to improve upon the thermal analysis sampling technique, SinterCast has developed the immersion
sampling device shown in Figure 5 (b) for the thermal analysis of cast irons. The thermal analysis sample
is obtained by immersing the Sampling Cup into the iron for approximately three seconds. This sampling
technique minimises oxidation and simultaneously pre-heats the vessel to eliminate any influence on the natural
solidification potential of the iron. The immersion filling also simplifies the sampling procedure, as the overflow
provided by the upper annulus of the sample cup ensures consistent filling. In comparison to conventional sand
cups, the Sampling Cup is fabricated entirely from stamped and drawn steel sheet and has a predominantly
spheroidal containment area. The steel walls containing the molten iron provide a double-wall Dewar type
insulation of the sample. The void space between the walls varies symmetrically along the height of the vessel
to regulate heat loss, with the gap at the bottom being minimised to balance the heat loss from the surface of
the iron. The upper portion of the Sampling Cup also provides a barrier to reduce radiation heat losses from the
open surface. These design features promote a uniform spheroidal solidification behaviour. Beyond the design
of the Sampling Cup, six separate Go / No-Go checks are made on the cooling curves prior to the onset of
solidification to ensure that the sampling conditions are consistent and not influenced by the start temperature.
Together, the design of the thin-wall Sampling Cup and the immersion sampling technique ensure a constant
sample volume, prevent oxidation of the iron during pour-in filling, provide a uniform solidification profile and
yield a more accurate measurement of undercooling because of the elimination of chill-induced solidification.

Two different cooling curves are obtained from two thermocouples contained within the protective tube in the
Sampling Cup. The thermocouples are mounted in the sampling hardware and are reused up to 250 times as
the consumable Sampling Cups are positioned, used, and dispensed. One of the thermocouples is located in
the thermal centre of the Sampling Cup while the second is located at the bottom of the protective tube, thus
providing two different measurement conditions.

The walls of the Sampling Cup are coated with a reactive coating that consumes active magnesium in order to
simulate the fading of magnesium in the ladle. This patented Mg-fade simulation allows for the simultaneous
measurement of the solidification behaviour at the start of casting (through the centre thermocouple) and also
after a predetermined loss of magnesium (through the bottom thermocouple). Strategically, the strength of the
reactive coating is chosen for either CGI or ductile iron analyses to provide the desired difference between the
cooling curves of the centre and bottom thermocouples. In addition to the inherent differences in the cooling
curves of grey iron, CGI and ductile iron, as shown in Figure 4, this patented wall reaction provides additional
insight into the solidification behaviour of the iron to quantitatively determine the coordinates on the microstructure
chessboard and to decide upon control actions that ensure that the iron remains within the specification window
from the start until the end of casting.

Process Control
Thermal analysis provides the ability for the foundry to see inside the iron and, with the additional insight, to
decide upon control actions. Depending on the process flow, the control action can either be a feedforward
addition of magnesium and/or inoculant prior to casting, a feedback adjustment of magnesium and/or inoculant
in subsequent the base treatments, feedback control of furnace CE, or a combination of these actions.

Production experience with Compacted Graphite Iron has shown that the stable window for simultaneously
avoiding flake graphite and shrinkage defects in complex castings such as cylinder blocks and heads is too small
to reliably reach the start-cast coordinates directly with the base treatment. Even with advanced knowledge
of the parameters that determine the base treatment result, such as historical recovery, sulphur, ladle weight
and temperature, variation in the base treatment is too large to always remain within the narrow control range.
Figure 6 shows actual series production data for the variation of base treatment Modification (‘MGM’) results
during the production of heavy-duty CGI cylinder blocks from 281 ladles. The base treatment was conducted
by the addition of cored wire into a 2,300 kg ladle based on statistically optimised algorithms including input
from (i) MGM results from previous ladles; (ii) sulphur content of the base iron; (iii) ladle weight; and, (iv) ladle
temperature. The calibrated Modification limits required to achieve the specified microstructure, properties and
soundness in the heavy-duty blocks range from 38 to 46, but the results in Figure 6 (a) show that the actual base
treatment result varies from 26 to 44 – more than two times wider than the permissible casting range. The actual
Modification range after the corrective addition of magnesium cored wire is shown in Figure 6 (b), with all of the
281 production ladles falling within the specification range. The average corrective magnesium addition for this
series of 281 ladles was approximately 30 grams per tonne, showing that the final correction step is a precise
addition that manipulates the third and fourth decimal points in order to deliver consistency to the moulding line.

8
For the reliable high volume production of CGI, the control strategy must be to intentionally undertreat the iron during
base treatment to the extent that, even if all variables combine to produce the maximum recovery, the iron will still not be
overtreated. This strategy provides the opportunity to make bespoke additions of Mg and inoculant wire to adjust each
ladle prior to casting. In contrast, if the target for the base treatment result shown in Figure 6 (a) was shifted upward

Figure 6: Series production results for Modification after the base treatment
and correction of 281 ladles show the need for a two-step casting process
to simultaneously eliminate flake graphite and shrinkage defects in complex
CGI components.

in an attempt to target the centre of the MGM casting range of 38 to 46, many of the ladles would be overtreated,
resulting in the formation of shrinkage defects, ultimately causing leakage scrap after machining. The only recourse
to avoid such scrap would be to resort to excessive feeding, but the cost of such feeding is generally more than the
cost of the process control, especially when exothermic feeders are required. It must also be considered that not all
areas of complex castings can be reached by feeders, so that not all defects can be eliminated. But beyond these
considerations, shifting the base treatment target to the centre of the MGM casting range is not sufficient, because this
would also result in some ladles being undertreated, causing flake graphite. In order to avoid flake graphite, the base
treatment target would need to be increased to the point where no ladles fell below the minimum MGM limit. In turn, this
would necessitate more feeding, further increasing the cost and increasing the risk for more scrap.

In the case of ductile iron, the analysis of the cooling curves can be used to indicate microstructure results and the
potential for casting defects. Following a calibration period to correlate the cooling curves to the casting results for
each new component, the analysis can predict results for nodularity (%), nodule count (No/mm2), carbide risk (1-5
Severity Index), shrinkage risk (1-5 Severity Index) and, if the thermal analysis is allowed to continue until the eutectoid
reaction, pearlite content. These results can be used to augment quality control and to enable the foundry to implement
feedforward and feedback control actions to improve the microstructure and the production efficiency.

As shown in Figure 7, the CE, Modification and Inoculation results are also provided to serve as the basis for on-
line process control. These results are displayed as running bar charts with the current analysis on the far
right and the previous results trailing to the left. The absolute CE value is displayed in the upper run chart while
additional CE information indicating the proximity of the CE result to the true eutectic is reported in the “Δ-Eutectic”
field provided in the final result window at the bottom of the display. With reference to the previous explanation of the
eutectic in Figure 2, the Δ-Eutectic value shows the degree of separation between the austenite arrest (Point A) and the
minimum undercooling (Point B). In the example of Figure 7, the Δ-Eutectic result shows that the CE is 0.18 units below
the eutectic, allowing the foundry to make carbon, silicon and/or inoculant additions to adjust the iron toward the eutectic
composition to optimise the castability. Likewise, the Modification and Inoculation results are provided as simple index
values corresponding to the pre-calibrated target window for each product on the microstructure chessboard. These
index values provide the basis for automatic or manual feedforward control of the in-stream inoculation and/or feedback
control of subsequent base treatments to iteratively adjust the iron into the target window.

9
 Figure 7: The primary output values of the thermal analysis are the Carbon
Equivalent, the Modification index and the Inoculation index. Together,
these parameters define the size and location of the calibration window for
each individual component and can be used to predict the microstructure
and the risk for casting defects.

In a ductile iron pilot application that has been running for more than a year, the process has evolved such that the
foundry obtains a thermal analysis sample from the first ladle of each furnace campaign and from every other ladle
thereafter, until the end of the 16 ladle campaign. The CE results have been adopted to trim the carbon content in the
furnace throughout the campaign, while the Modification and Inoculation results are used for dynamic feedback control
of the tundish cover base treatment process. The optimisation of the base treatment has resulted in a 7.5% reduction
in the FeSiMg alloy consumption and, although the average inoculant addition has not changed, the dynamic control
has resulted in more consistency in the graphite distribution and the particle count. The foundry is also contemplating
the acquisition of an in-stream inoculant dispenser to allow for the dynamic feedforward control of the in-stream addition
based on the Inoculant index results. In addition to the improved microstructure consistency, the external machining
company has stated that the tool life has improved as a result of the process control.

The potential to reduce the magnesium and inoculant consumption provides a direct and obvious cost reduction, but the
most significant opportunity for cost reduction is the ability to reduce scrap levels and feeding. Incremental magnesium
and inoculant additions beyond the minimum amount needed to satisfy the microstructure specification increase the risk
of scrap and the shrinkage sensitivity of the iron. For each component, the foundry must determine the safety margin
in the size of the feeders to compensate for the worst case condition, which occurs when the variation in the process
results in the highest Modification. Process control provides the opportunity both to reduce the variation and to shift
the location of the target window toward lower levels, thus providing a double-opportunity for reducing the worst case
condition.

Depending on the sales price of the casting and the mould yield, every 1% reduction in scrap or improvement in mould
yield can provide a cost benefit of $10~20 per tonne of castings. Accordingly, the application of additional process
control for ductile iron production should be approached as a cost-benefit opportunity rather than as an absolute cost.

Conclusion
The global foundry industry continues to become more polarised. Commodity components are under pressure to be
made less expensively while high tech components are under pressure to be made with better and more consistent
properties. The challenge for the high tech foundry industry is that the stable microstructure ranges for engineered
castings such as Compacted Graphite Iron and ductile iron are moving targets. Despite all efforts and discipline,
experience has shown that process consistency can’t be ensured by reproducing the same chemistry in every ladle.
This realisation leads to the concept of dynamic process control; to measure the solidification behaviour of the ladle and
to adjust the iron prior to casting. As the starting point for process control, thermal analysis provides the opportunity for
the foundry engineers to see into the iron and to understand the true solidification behaviour. With this insight, bespoke
feedforward and/or feedback control actions can be determined and implemented before the castings are poured to
improve consistency and cost efficiency.

10
Head Office China
SinterCast Ltd SinterCast Trading (Beijing) Co., Ltd
Kingswick House Room 1131, 11/F, Block A, Gateway
Kingswick Drive No. 18 Xiaguangli
Sunninghill Berkshire North Road East Third Ring
SL5 7BH Chaoyang District, Beijing
United Kingdom China, 100027
Tel: +46 150 794 40 Tel: +86 10 5923 1163

Technical Centre Korea

SinterCast AB (publ) SinterCast Korea Co., Ltd
Kungsgatan 2 Rosavallcity No. 310
641 30 Katrineholm HyoJa-Ro 194, WanSan-Gu
Sweden Jeon Ju City, 560-900
Tel: +46 150 794 40 South Korea
Tel: +82 10 9228 1764

USA
SinterCast Inc
1755 Park Street
Suite 200
Naperville
IL 60563
USA
Tel: +1 630 778 3466

www.sintercast.com
[email protected]