Original Paper

Measurement and Control

2022, Vol. 55(5-6)
0(0) 1–8257–264
Progress in numerical simulation of casting © The Author(s) 2022
Article reuse guidelines:
process sagepub.com/journals-permissions
DOI: 10.1177/00202940221102656
journals.sagepub.com/home/mac

Zhengyang Chen1, Yueyang Li2, Fengrui Zhao2, Shi Li2  and Ji Zhang1

Abstract
The numerical simulation of casting process can calculate casting filling process, solidification process, and obtain change
and coupling information of temperature field, velocity field, pressure field, stress field and microstructure. At present,
numerical simulation of casting process has been quite mature at the macro level, and is developing towards the
direction of micro level. The coupling and integration of different fields between temperature, flow, stress, and mi-
crostructure, is the shape of things for numerical simulation research of casting process. This paper reviews and
summarizes the research history and current situation of numerical simulation of casting process. The progress in
numerical simulation from five aspects of casting solidification, casting filling, stress field, microstructure, commercial
software, is presented.

Keywords
casting process, numerical simulation, filling, solidification, microstructure

Date received: 22 December 2021

Introduction developing towards microstructure simulation, performance

optimization and service life prediction. The paper discusses
The numerical simulation of casting process numerically the status and trend of casting simulation from the following
calculates and simulates the filling flow and solidification five aspects: solidification, filling, stress, microstructure and
process of liquid metal in mold, so as to obtain the changes of simulation software.
temperature, flow, pressure, and microstructure of castings in
casting process. The numerical simulation of casting for-
mation can predict defects, including shrinkage cavity, Simulation of casting solidification
shrinkage porosity, hot crack. The simulation results can be After the high-temperature liquid metal is filled in the mold, it
used for optimization of casting process parameters, elimi- releases heat, and gradually cools and solidifies, finally forms
nating casting defects and improving product quality. The the casting product. Simulation of solidification mainly in-
qualities of different process schemes can be predicted and cludes the heat transfer between casting and mold and sur-
compared ahead of production through casting numerical rounding environment (heat conduction between metal and
simulation software, so as to reduce the cost and development mold, heat radiation between metal and environment, con-
time of new products. Therefore, the numerical simulation of vective heat transfer), latent heat releasing during solidifi-
casting process can shorten the process experiment cycle, cation (considered as a heat conduction problem with internal
ensure the casting quality, improve the process yield, reduce heat source), the temperature field simulation of liquid-solid
the production cost and waste loss, and has high practical state, and the formation of shrinkage cavity and porosity
value for enterprises.
Numerical simulation of casting process mainly consists
of temperature simulation, flow simulation, stress simulation, 1
Beijing Key Laboratory of Advanced High Temperature Materials, Central
and microstructure simulation.1 The casting solidification Iron and Steel Research Institute, Beijing, China
2
simulation began from the 1960s. The mold filling and stress- School of Electrical Engineering, University of Jinan, Jinan, China
strain evolution began in the 1980s. In the 1990s, it entered
Corresponding author:
the numerical simulation of casting microstructure. At Shi Li, School of Electrical Engineering, University of Jinan, Road
present, numerical simulation of casting process has pene- NanXinZhuangXiLu, Jinan 250022, China.
trated into all aspects of casting forming process, and is Email: [email protected]

Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License
(https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-
us/nam/open-access-at-sage).
2
258 Measurement andand
Measurement Control 55(5-6)
Control 0(0)

during solidification. Normally, a FDM (finite difference changes such as oxidation, heat transfer, impact and etc. The
method) algorithm is applied to simulate temperature field. casting defects, such as air entrainment, shrinkage cavity,
There are many methods to predict shrinkage cavity,2–9 in- shrinkage porosity, are related to filling process. The key
cluding isothermal curve method, isothermal gradient problem of filling simulation is the simultaneous solving
method, flow conductivity method, isosolid phase rate curve problem of flow and pressure in Navier-Stokes (N-S)
method, solid phase rate gradient method and liquid metal equation. Compared with numerical simulation of solidifi-
feeding distance method. These methodes can provide a cation process, the calculation of filling process involves
variety of criteria for shrinkage porosity and shrinkage cavity more complex theoretical equations. In addition to temper-
prediction. ature changes, N-S momentum equation needs to be solved.
In 1962, Forsund of Denmark firstly used computer and The iterative calculation of velocity field and pressure field is
FDM to calculate the heat transfer in casting solidification repeated, which requires a large amount of calculation and
process. In 1965, Campbell, Villen Weider of General the iterative results are easy to diverge. In addition, the
Electric (GE) applied FDM to simulate large thick castings position and shape of the free surface should be treated. And
and proposed a predictable solidification model. In the same simulation results are difficult to be verified by experimental
year, Henzel and keverian10 of GE applied the instantaneous methods. Therefore, it is difficult to calculate and simulate
heat transfer program to simulate automobile cylinder block casting filling.
castings and verified the results by actual measurement. In the In filling simulation, high-temperature liquid metal flowing
1970s, Sahm of Germany11 proposed the research direction is generally regarded as an unsteady flow of viscous incom-
of computer numerical simulation for casting solidification pressible Newton fluid. The turbulent model with k-ε equations
process. M.C. Flemings of Massachusetts Institute of tech- is used to solve mass, momentum, energy conservation
nology proposed to use simulation technology to study and equations. The volume function method is normally applied to
predict shrinkage cavity, porosity, hot crack and segregation track the changing volume function of fluid surface.
of castings. In 1973, Victor Davies of Norway applied FDM Mass conservation:
in sand casting, metal casting and low pressure casting of
aluminum products. In 1974, Los Alamos laboratory de- ∂u ∂v ∂w
veloped a color moving picture technology, which can de- þ þ ¼0 (1)
∂x ∂y ∂z
scribe strip or spot images in different temperature ranges,
and realize the mold profile visualization of solidification Navier-Stokes equation:
simulation technology.  
∂u ∂u ∂u ∂u ∂p
In the early 1980s, numerical simulation of solidification ρ þu þv þw ¼ � þ ρgx þ μ=2 u
process has gradually become a hot research direction, and ∂t ∂x ∂y ∂z ∂x
 
theoretical research and practical application have developed ∂v ∂v ∂v ∂v ∂p
rapidly. In 1982, Niyama proposed the criterion of shrinkage ρ þu þv þw ¼ � þ ρgy þ μ=2 v (2)
∂t ∂x ∂y ∂z ∂y
cavity and porosity defect, which is considered as an epoch-  
making research. He firstly proposed the G/R1/2 method to ∂w ∂w ∂w ∂w ∂p
ρ þu þv þw ¼ � þ ρgz þ μ=2 w
predict the shrinkage cavity and porosity of castings. The ∂t ∂x ∂y ∂z ∂w
criterion has good practicability for various alloys and dif- Energy conservation:
ferent castings. At the fifth International Foundry conference
       
in 1983, Pehlke and Berry has put forward the concept of ∂ ρcp T ∂ ρcp uT ∂ ρcp vT ∂ ρcp wT
casting process CAD, which is attributed to the organic þ þ þ
∂t ∂x ∂y ∂z
connection of calculated simulation, geometric modeling and      
(3)
database. Hansen proposed the concepts of pretreatment, ∂ k ∂T∂x
∂ k ∂T∂y
∂ k ∂T
∂z
posttreatment and intermediate calculation. Sahm discussed ¼ þ þ þS
the application of numerical simulation in microstructure and ∂x ∂y ∂z
casting properties of solidification process, summarized a The volume function:
series of relations based on the calculation, and envisaged
using these relations to organically connect geometric model, ∂F ∂Fu ∂Fv ∂Fw
þ þ þ ¼0 (4)
solidification parameters, alloy properties and microstructure ∂t ∂x ∂y ∂z
parameters.
At present, numerical simulation of casting solidification where
has been very mature at the macro simulation level. The
simulation of temperature field and the prediction of - u, v, w are velocities in x, y, z directions; gx, gy, gz are
shrinkage cavity and shrinkage porosity have been very close gravitational accelerations in x, y, z directions; =2 is the
to the actual production. Nowadays, the research of solidi- Laplace operator;
fication simulation is developing towards the microstructure - ρ is the density; P is the pressure; T is the temperature; μ
simulation at the micro level. is the kinematic viscosity; cp is the specific heat ca-
pacity at constant pressure;
- F is volume function, and 0 ≤ F ≤ 1.
Simulation of casting filling
During casting filling in the mold, the liquid metal is Several algorithms used in filling simulation are as
gradually cooled, accompanied by physical and chemical follows.
 al.
Chen et al 2593

· SIMPLE: the theoretical basis of this method is that the first combined the flow phenomenon and heat transfer in
pressure is not directly specified by mass conservation casting filling process. In 1985, Wang13 applied SOLA-VOF
equation. After substituting the pressure into mo- method to casting filling in his doctoral thesis, and verified it
mentum equation, the velocity satisfying mass con- with high-speed photography technology. In 1988, Lin and
servation equation can be obtained. Pressure and Hwang16 combined SOLA-VOF method with heat transfer in
velocity are coupled and need to be iterated simulta- fluid calculation, considering the computational coupling of
neously, hence the calculation speed is slow. Based on flow field and temperature field. In 1991, Jonson et al.17
this method, some researchers proposed improved simulated the filling results of liquid metal in the cavity under
SIMPLE methods to accelerate iteration speed and different states, and discussed the influence of liquid metal on
reduce calculation time. the filling results. In 1992, Yeh and Hwang18 used SOLA-
· MAC and SMAC: MAC method was proposed by VOF method to solve k-ε turbulence model with free surface.
Harlow and Welch in 1965.12 This method is based on Molten steel flowing is simulated and compared with hy-
finite difference grid, two ends of the momentum draulic model experiment.
equation are discretized to obtain Poisson equation for In 1995, Sirrell et al.19 of the University of Birmingham,
solving pressure field. The continuity equation is used UK, published the standard experimental results, which
as constraint conditions of the pressure to deform the choosed pure aluminum as the experimental alloy, a higher
Poisson equation. The corresponding pressure field sprue was designed to artificially cause turbulent filling ef-
and velocity field are solved by iterating the mo- fect, and the change of liquid metal filling state was recorded
mentum equation and continuity equation at the same by X-camera technology. The results showed that most of the
time. MAC method adds identification particles to the calculated filling states with time were similar to experi-
fluid to track and describe the fluid flow. The char- mental results. In 1997, Pan et al.20 used SOLA-VOF al-
acteristic of this method is to solve directly without gorithm to analyze gas-liquid flow in filling process,
changing the equations, the velocity boundary con- including velocity and bubble flow trajectory, and considered
ditions are easy to be given. It is suitable for solving the interaction between gas and liquid. In 1998, at the eighth
viscous incompressible fluid, and it is easy to realize international conference on Modeling of Casting, Welding
the transformation from two-dimensional to three- and Advanced Solidification Process, Jolly et al. of the
dimensional. However, the pressure field and veloc- University of Birmingham verified the simulation of
ity field need to be iterated simultaneously, and lots of MAGMAsoft and FLOW-3D software, and given the sim-
tracking particles need to be calculated to accurately ulation results of filling the cavity with liquid metal. Hirt et al.
reflect the shape of free surface, resulting in large of American fluid science company described the calculation
amount of calculation, slow calculation speed and low model, calculation method and defect prediction model of
efficiency. Based on MAC, SMAC algorithm intro- FLOW-3D software, which was successfully applied in
duces potential function to correct velocity field, which automotive castings. In 1998, Jurgen et al.21 improved the
can greatly improve operation speed and save com- VOF method. They used unstructured grids to control the unit
puting memory. volume, and converted the hyperbolic equation into a single
· SOLA-VOF: the SOLA-VOF method is proposed by parabolic equation to reduce the amount of calculation, which
Wang in his doctoral thesis in 1985.13 This method has has been used in the production of simple castings.
been the widely used numerical calculation method in In 1998, Stoehr and Wang22 improved the SOLF-VOF
casting filling simulation. The momentum equation method and applied it to the three-dimensional filling
and continuity equation are solved by SOLA method, problem. In 2001, Hong et al.23 proposed a mathematical
and the free surface is calculted by volume of fluid model of complex cavity filling process. Schwarze et al.24
(VOF) method. VOF method defines a volume func- proposed a model for the flow field and dispersed phase
tion F, which is used to represent the volume content of behavior in the melt of continuous casting of steel strip. In
liquid in a fluid unit. When the fluid unit is filled with 2003, Li and Zhou25 proposed a three-dimensional simula-
liquid, F = 1. When the unit has no liquid, F = 0. When tion program for die casting process, and the finite difference
0 < F < 1, the unit has a free surface. By solving the method was used to solve three-dimensional N-S equation. In
volume fraction F of each fluid element, the position 2004, Mirbagheri et al.26 proposed an improved SOLA-VOF
and shape of free surface can be determined. VOF method to analyze the impact and erosion degree of molten
method is verified sufficiently fast and stable to be used metal on different mold walls. Sakuragi and Takuya27 pro-
in free surface solving problem. With the development posed a filling process algorithm based on Cold Isostatic
of fluid equation from two-dimensional to three- Pressing (CIP) and considering surface tension for squeeze
dimensional, the difference scheme and computa- casting and die casting process simulation. The simulation
tional grid are more complex and diverse, various results were compared with experimentally measured pore
improved VOF methods have been proposed and used position and size.
in various filling numerical simulation software. In 2005, Kashiwai et al.28 simulated the vacuum suction
casting process of aluminum alloy castings. In 2007, Lee
The development of computational fluid dynamics for et al.29 simulated the filling process of semi-solid magnesium
filling simulation is reviewed as follows. In 1983, Hwang and alloy, analyzed the rheology and thixotropy of liquid metal at
Stoehr14 first combined computational fluid dynamics with different shear rates and cooling rates, and the effect of
casting filling problem, and simulated the filling flow process viscosity on the filling process. In 2011, Korti et al.30 sim-
of plate casting by two-dimensional method. In 1984, Desai15 ulated aluminum casting filling problem with high pressure.
4
260 Measurement andand
Measurement Control 55(5-6)
Control 0(0)

In 2013, Mcbride et al.31 developed a centrifugal casting various boundary conditions, and is suitable for casting
model, which can simulate the free surface details of complex simulation of stress field. Therefore, in the coupling simu-
centrifugal casting filling process, and verified by hydraulic lation of temperature field and stress field, FDM is often used
simulation experiment and high-speed camera. to calculate the flow field and temperature field, and FEM is
From the above litteratures, it can be seen that in filling often applied to calculate the stress field.42
simulation problems, MAC, SMAC and SOLA methods are
generally used to calculate N-S and continuity equations,
which are used to jointly solve the coupling calculation of
Simulation of casting microstructure
pressure and velocity. VOF method is generally applied for The macro simulation of filling and solidification has been
free surface treatment. Nowadays, casting filling simulation deeply studied, and many commercial simulation softwares
develops towards specialization and three-dimensional have been widely used and applied in casting production.
simulation. A specialized subdivision algorithm is devel- Numerical simulation of casting alloy microstructure is in the
oped for casting process of different types and materials.32,33 research stage and is a hot spot.43 In microstructure simu-
At the same time, the development of three-dimensional lation, the microstructure nucleation and growth model are
numerical simulation software is becoming more and more simulated to reproduce the formation and change process of
perfect, and many commercial simulation software have been grain, dendrite, eutectic, component segregation and micro
applied in practical production. pores. At present, the main numerical methods include de-
terministic method, stochastic method, phase field (PF)
method. Among these methods, Cellular Automaton (CA)
Simulation of stress field
method and PF method are the most concerned and studied.
The casting thermal stress generated in the solidification Nakajima et al.44 summarized the computer simulation
process will cause cracks and deformation in the processing and models of casting filling and solidification process, discussed
use of castings, which will seriously affect the properties and the development from macroscope-scale model to
performance of castings. Therefore, it is necessary to numer- microscope-scale model, especially the CA model and PF
ically simulate the generation and change of thermal stress. model. Zhang et al.45 summarized the development of tita-
The numerical simulation of stress field can be roughly nium alloy microstructure simulation, including the CA
divided into two parts, one is the simulation of solid-liquid model and PF model.
two-phase region (quasi solid region), and the other is the
simulation of the later stage of solidification.
The stress analysis of solid-liquid two-phase region is of great
Deterministic simulation method
significance to study and predict the casting defects, hot crack, In the early stage of solidification microstructure simulation,
residual stress and residual deformation formed in the solidifi- limited by the development level of computer and macro
cation zone, because many casting defects such as shrinkage transmission theory, the deterministic simulation method is
cavity, shrinkage porosity and hot crack are formed at this time. mostly used in the microstructure simulation of metal solid-
The simulation at this stage generally adopts the rheological ification process. At a given time, the nucleation density and
model. The rheological model adopts the combination of simple growth rate of grains in a certain volume of melt are deter-
fluid model and ideal mechanical models such as elastomer, mined functions. Deterministic simulation is based on volume
viscous body and plastic body to represent the complex flow and element to solve the continuity equation. Firstly, the calcu-
deformation law of materials, so as to more accurately reflect the lation space is divided into volume elements, and the tem-
change law of flow and deformation with time. Therefore, the perature of each volume element is uniform. Then, according
rheological model is more suitable to deal with the flow and to some nucleation laws, the volume element is further divided
deformation law of quasi solid zone during solidification.34–36 into micro volume elements, and each micro volume element
The stress calculation of post solidification stage often has a spherical grain growing at a certain speed.
uses some mechanical models,37–40 including the thermo- The deterministic method is based on classical solidifi-
elastic model, the thermoviscoelastic model, the thermo- cation kinetics and crystal growth characteristics. The de-
elastoplastic model, the thermoelastic viscoplastic model, the terministic functions of crystal nucleation density and growth
Perzyna model, the Heyn model and the unified variable rate in melt are derived, which can describe the changes of
model. nucleation density, dendrite tip radius and growth rate with
In the solidification process of castings, stress is closely solidification conditions. Several deterministic models have
related to cooling, and uneven temperature distribution will been proposed by researchers, including the continuous
produce thermal stress in castings. Therefore, coupling nucleation model,46 the instantaneous nucleation model,47
thermal analysis can improve the accuracy of stress analy- the Gaussian distribution nucleation model,48 the Rappaz-
sis.41 The mesh generation algorithm of the finite difference Theovz nucleation model,49,50 the dendrite growth model
method (FDM) is simple, the program design is simple, and LGK,51 KTG,52 the multi-scale multiphase model53 and the
the convergence is good. The FDM has unique advantages in eutectic growth model.54,55 However, the deterministic
predicting the dynamic boundary treatment in the process of method can not consider some random phenomena in the
shrinkage and porosity. Therefore, FDM is often used in solidification process, such as random nucleation distribu-
casting simulation of flow field and temperature field. The tion, random crystal orientation, etc., nor can describe the
mesh model of finite element method (FEM) is flexible, complex microstructure and explain some solidification
which can accurately simulate various complex curved phenomena, such as the transformation from columnar
surface or curve boundaries, realize the unified treatment of crystal to equiaxed crystal, the competitive elimination of
 al.
Chen et al 2615

dendrite growth, the fusing and bifurcation of dendrite arms, solid-liquid solute transport equation, solved the interface
etc. In recent years, there are few studies on deterministic moving velocity based on the solute conservation equation of
models, but its development and improvement have laid the interface element, and successfully simulated the growth of
foundation for the subsequent numerical simulation of free equiaxed and columnar crystals and CET transition in the
microstructure. undercooled melt.
After 2000, the microstructure simulation of solidification
process based on CA method has developed rapidly, more
Stochastic simulation method factors affecting dendrite growth, such as flow and alloy
The stochastic method uses the probability method to de- elements, have been considered, and many improved CA
scribe the nucleation and growth of grains. There are Monte methods have been proposed.62–66
Carlo (MC) method and CA method, of which CA method is
widely used.
MC method divides the calculation element into N micro
Phase field (PF) simulation method
elements on the grid, and the grain nucleation and growth on PF simulation is also called direct microstructure simulation.
each node are calculated according to the probability model. It introduces phase field variables to describe complex
CA method regards the grid element as a cell. The state growth process of the interface between new phase and
values of the cell include physical parameters such as tem- parent phase and the solid-liquid interface in equilibrium
perature, solubility, solid fraction, growth orientation and state. It can describe the changements of shape, curvature and
liquid phase flow rate. At each step, the state values of each movement of solid-liquid interface, and finely describe the
cell and the transition of the interface are realized by solving transformation of microstructure and the reproduction of
control equations such as temperature and solute field, complex solidification phenomena, such as coarsening of
coupling dendrite growth and capture algorithm. dendrite arm, fusing and bifurcation.67,68
In 1993, Rappaz and Gandin56 first used CA method to The important theoretical phase field model was devel-
simulate metal solidification microstructure, and predicted oped in the 1980s. The early work was dendrite solidification
the grain growth and get transformation of Al-Si alloy under simulation of pure materials and binary materials. Langer69
different alloy composition and cooling rate. Grandin and Fix70 first introduced the phase field model for first-order
et al.57–59 combined the finite element macro heat transfer phase transition. Collins and Levine71 independently de-
calculation with CA algorithm to simulate two-dimensional veloped a similar diffusion interface solidification model.
and three-dimensional grain structure of castings under non- Langer and Sekerka72 considered the model of diffusion
uniform temperature field, and applied to simulate three- interface motion in binary alloy system with miscible gap in
dimensional grain selection process of directional solidifi- solid solution phase. Caginalp et al.73–76 carried out im-
cation of superalloys and Al-Si alloys under continuous portant early works in the development and analysis of so-
casting conditions. In 1998, Dilthiy and Pavlik60 used CA lidification phase field model. Penrose and Fife77,78 provided
method to simulate the dendrite morphology of alloy. In a framework for deriving phase field equations from a single
1999, Nastac61 combined the macro temperature field and entropy function in a thermodynamically consistent manner,

Table 1. Main numerical simulation software of casting filling and solidification.

Software name Country Algorithm Analysis content Application field

Procast US.A. FEM Flow and heat transfer, stress Continuous, EPC, sand mold, shell mold, investment
deformation, microstructure, mold, pressure, metal mold, centrifugal, single
convective radiation crystal, rheological, electromagnetic castingetc.
MAGMA Soft Germany FDM/FEM Flow and heat transfer, stress Sand mold, shell mold, investment mold, pressure,
deformation, microstructure metal mold, low (high) die castingetc.
Flow3D US.A. FVM/FDM Flow and heat transfer, EPC, sand mold, pressure, metal mold, semi-solid,
microstructure inclined casting, centrifugal castingetc.
CastCAE Finland CVM/FDM Flow and heat transfer, Sand mold, shell mold, investment mold, pressure,
microstructure metal mold, semi-solid castingetc.
Simulor France FVM Flow and heat transfer Sand mold, precision, metal mold, low pressure
castingetc.
Novaflow&Solid Sweden FDM/ Heat transfer analysis Sand mold, shell mold, investment mold, pressure,
Modulus metal mold castingetc.
SOLSTAR U.K. FVM/ Heat transfer analysis Various casting processes
Modulus
AFS Solidification US.A. FDM Flow and heat transfer Sand mold, shell mold, investment mold, metal mold
system castingetc.
SOLIDA Japan FDM Flow and heat transfer Various casting processes
ADSTEFAN Japan FEM Flow and heat transfer Various casting processes
Any casting Korea FEM Flow and heat transfer, stress Various casting processes
deformation
6
262 Measurement andand
Measurement Control 55(5-6)
Control 0(0)

and the non-isothermal solidification process was modeled. metal matrix control. Micro level simulation technologies,
Kobayashi79,80 carried out pioneering simulation on the such as CA method and phase field method, as well as the
growth of two-dimensional and three-dimensional dendritic combination of micro level and macro level, make the
crystals in the early 1990s, the works proved that the phase coupling and integration of temperature field, flow field,
field method has the potential for complex microstructure stress field and microstructure, which makes the current
modeling, and promoted the rapid development of phase field development direction.
calculation in materials science. The combination of phase Casting numerical simulation technology, also known as
field and temperature field, flow field, solute field and ex- “flexible casting technology”, has been used not only in the
ternal action field can truly simulate the macro and micro- optimization and prediction of casting design and process
structure changes in the solidification process. In recent parameters, but also further used in the fields of digital
years, the adaptive variable grid and parallel algorithms81–84 manufacturing, virtual manufacturing, intelligent
were proposed, and PF method has been greatly improved, manufacturing.
showing strong development potential and application
prospect, and has become one of the most successful tech- Declaration of conflicting interests
niques for material microstructure simulation. The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Simulation software for casting filling and
Funding
solidification process
The author(s) disclosed receipt of the following financial support for
The first casting simulation software was MAGMA Soft the research, authorship, and/or publication of this article: This work
developed in 1989 under the auspices of Professor Sahm P.R. is partially supported by the National Natural Science Foundation of
of Aachen University in Germany. Since then, many com- China (Grant Nos. U1708253, 61973135, 91948201), and the
mercialized software have emerged, such as ProCAST of the National Key Research and Development Program of China (Grant
United States, CastCAE of Finland, Simulor of France, No. 2017YFB0702901).
Novaflow&Solid of Sweden, Solida of Japan, etc. Table 1
shows the main simulation softwares and its application ORCID iD
fields. These softwares can simulate the temperature, flow
and stress field of sand mold, shell mold, metal mold casting, Shi Li  https://orcid.org/0000-0002-8493-2344
and etc. Defects such as shrinkage porosity, shrinkage cavity,
hot crack, can be simulated and predicted. References
Development direction of current casting simulation soft- 1. Bastian KM. A look back at the 20th century casting process
ware is expanding from low-pressure casting, die casting and simulation. Mod Casting 2000; 12: 43–45.
investment casting on the one hand, and from macro simulation 2. Niyama E, Uchida T, Morikawa M, et al. A method of
to microstructure simulation on the other hand. At present, the shrinkage prediction and its application to steel casting prac-
development direction of numerical simulation is as follows: tice. Int Cast Met J 1982: 52–63.
3. Carlson KD and Beckermann C. Prediction of shrinkage pore
(1) Oxidation inclusion simulation; volume fraction using a dimensionless niyama criterion. Metall
(2) Complex multiphase flow simulation; Mater Trans A 2009: 163–175.
(3) Modeling/core making process simulation; 4. Song W and Xue X. Study of predicting model of shrinkage and
(4) Simulation of pattern generation process of gasified its parameters. Foundry 1986: 36–38.
mold casting; 5. Li H, Xue X and Zhang Y. Effect of molten metal static pressure
(5) Semi solid casting simulation; on shrinkage criterion. Foundry 2003: 426–428.
(6) Microscopic numerical simulation. 6. Reis A, Xu Z, Tol RV, et al. Modelling feeding flow related
shrinkage defects in aluminum castings. J Manuf Process 2012;
14: 1–7.
Conclusion and perspective
7. Li W and Cui J. A new method for quickly locating the hot
This paper reviews and arranges the research of numerical spots in solidification simulation for complicated castings. Int J
simulation of casting process in recent years, and discusses Cast Met Res 2002: 319–323.
and summarizes the solidification process, filling process, 8. Sutaria M, Gada VH, Sharma A, et al. Computation of feed-
stress field, microstructure simulation and the development of paths forcasting solidification using level-set-method. J Mater
commercial software. Process Tech 2012; 212: 1236–1249.
At present, the casting simulation technology has been 9. Zhou J, Wang M, Yin Y, et al. Feed paths and hot spots
very mature for the macro simulation of shape, size and computation based on a time gradient method in casting. Int J
contour, and the prediction of casting defects. The casting Adv Manuf Tech 2016: 1–12.
simulation research is developing towards the micro simu- 10. Henzel J and Keveran J. AFS Cast Met Res J 1965; 1: 19–23.
lation of predicting microstructure, involving crystallization, 11. Pehlke RD and Kirt MJ. Numerical simulation of solidification.
segregation, diffusion, gas precipitation, and phase trans- AFS Cast Met Res J 1973; 9: 49–52.
formation and so on. The micro simulation scale even reaches 12. Harlow HF and Welch EJ. Numerical calculation of time-
the nanometer level, involving crystal nucleation, crystal dependent viscous incompressible flow of fluid with free
growth, dendrite and columnar crystal transformation to surface. The Phys Fluids 1965; 12: 2182–2189.
 al.
Chen et al 2637

13. Wang CM, Computer simulation of fluid flow during the filling Sugeno fuzzy systems. IEEE Trans Fuzzy Syst 2022. DOI:
of casting molds, Pittsburgh, USA, Ph.D Thesis, University of 10.1109/TFUZZ.2022.3156735
Pittsburgh, 1985. 33. Li Y, Liu S, Zhao D, et al. Event-triggered fault estimation for
14. Hwant WS and Stoehr RA. Fluid flow modeling for computer- discrete time-varying systems subject to sector-bounded
aided design of castings. J Met 1983; 10: 22–29. nonlinearity: a Krein space based approach. Int J Robust
15. Desai PV. Computer simulation of forced and natural Nonlinear Control 2021; 31(11): 5360–5380.
convection during filling of a casting. AFS Trans 1984: 34. Ludwig O, Drezet JM, Ménésès P, et al. Rheological behavior
519–528. of a commercial AA5182 aluminum alloy during solidification.
16. Lin HJ and Hwang WS. Combined fluid flow and heat transfer Mater Sci Eng A 2005; s413–414(6): 174–179.
analysis for the filling of casting. AFS Trans 1988: 447–458. 35. Ludwig O, Drezet JM, Martin CL, et al. Rheological behavior
17. Jonson P, Sqluja N and Szekely J. Fluid flow phenomena in the of Al-Cu alloys during solidification constitutive modeling,
filling of cylindrical molds using Newtonian and non- experimental identification, and numerical study. Metallurgical
Newtonian fluids. AFS Trans 1991; 99: 291–297. Mater Trans A 2005; 36(6): 1525–1535.
18. Yeh JL and Hwang WS. Three-dimensional mathematical 36. Solek K and Dziarmagowski M. Tests of Steel Rheological
modeling of fluid flow in slab tundishes and its verification with Properties at Higher Temperatures for Computer Simulation in
water model experiments. J Materials Engineering Perfor- Procast Software. Arch Metall Mater 2014; 59(4): 1463–1466.
mance 1992; 1: 625–636. 37. Kozlowski PF, Thomas BG, Azzi JA, et al. Simple constitutive
19. Sirrel B, Holliday M and Campbell J, The benchmark test, equations for steel at high temperature. Metallurgical Trans A
Proceedings of international conference on modeling of cast- 1992; 23(3): 903–918.
ing, welding and advanced solidification process VII, London, 38. Ponthot JP. Unified stress update algorithms for the numerical
U.K., 1995, 915–1013. simulation of large deformation elasto-plastic and elasto-
20. Pan SM, Ho YH and Hwang WS. Three-dimensional fluid flow viscoplastic processes. Int J Plasticity 2002; 18(1): 91–126.
model for gas-stirred ladles. J Materials Engineering Perfor- 39. Garcı́a GC, Ribero VMS, Andı́a FS, et al. Numerical simu-
mances 1997; 6(3): 311–318. lation of finite strain viscoplastic problems. J Comput Appl
21. Jurgen N, Ingo S and Yves D. Modeling of free surfaces in Math 2013; 246(5): 174–184.
casting processes. Notes Numer Fluid Mechanics 1998; 66: 40. Careglio C, Canales C, Garino CG, et al. A numerical study of
168–186. hypoelastic and hyperelastic large strain viscoplastic Perzyna
22. Stoehr RA and Wang C. Coupled heat transfer and fluid flow in type models. Acta Mechanica 2016; 11: 1–14.
the filling of castings. AFS Trans 1998; 3: 733–740. 41. Fackeldey M, Ludwig A and Sahm PR. Coupled modeling of
23. Hong CP, Lee SY and Song K. Development of a new sim- the solidification process prediction temperatures, stresses and
ulation method of mold filling based on a body fitted coordinate microstructures. Comput Mater Sci 1996; 7: 194–199.
system. ISIJ Int 2001; 41: 999–1005. 42. Sih M and Cho C. A hybrid method for casting process simulation
24. Schwarze R, Obermeier F and Janke D. Numerical simulation by combining FDM and FEM with an efficient data conversion
of fluid flow and disperse phase behaviour in continuous algorithm. Mater Process Technol 2003; 133: 311–321.
casting tundishes. Model Simulation Materials Science Engi- 43. Asta M, Beckermann C, Karma A, et al. Solidification
neering 2001; 9(7): 279–287. microstructures and solid-state parallels: recent developments,
25. Li Y and Zhou W. Numerical simulation of filling process in die further directions. Acta Materialia 2009; 57(4): 941–971.
casting. Mater Technol 2003; 18(3): 36–41. 44. Nakajima K, Zhang H, Oikawa K, et al. Methodological
26. Mirbagheri SMH, Dadashzadeh M, Serajzadeh S, et al. progress for computer simulation of solidification and casting.
Modeling the effect on mould wall roughness on the melt flow ISIJ Int 2010; 50(12): 1724–1734.
simulation in casting process. Appl Mathematical Modelling 45. Zhang J, Li X, Xu D, et al. Recent progress in the simulation of
2004; 28(11): 933–956. microstructure evolution in titanium. Prog Natural Science:
27. Sakuragi Takuya. Prediction of gas hole by mold filling sim- Material International 2019; 29: 295–304.
ulation with consideration of surface tension. JSME Int Journal 46. Oldfield W. A quantitative approach to casting solidification-
, Ser B: Fluids Thermal Engineering 2004; 47(12): 699–708. freezing of cast iron. ASM Trans Q 1966; 59(4): 945–960.
28. Kashiwai S, Ohnaka I, Kimatsuka A, et al. Numerical simu- 47. Hunt JD. Steady state and columnar and equiaxed growth of
lation and X-ray direct observation of mould filling during dendrites and eutectic. Mater Sci Eng 1984; 65(1): 75–83.
vacuum suction casting. Int Journal Cast Met Research 2005; 48. Thevoz PH, Desbiolles JL and Rappaz M. Modeling of
18(3): 144–148. equiaxed microstructure formation in casting. Metallurgical
29. Lee DY, Moon JH, Seok HK, et al. Computer simulation for the Trans A 1989; 20(2): 311–322.
die filling behavior of semi-solid slurry of Mg alloy. J Korea 49. Rappaz M and Thevoz PH. Solute diffusion model for
Foundry Society 2007; 27(1): 31–35. equiaxed dendritic growth. Acta Metallurgical 1987; 35(7):
30. Korti AIN and Abboudi S. Numerical simulation of the in- 1487–1497.
terface molten metal air in the shot sleeve chambie and mold 50. Rappaz M and Thevoz PH. Solute diffusion model for
casting of a die casting machine. Heat Mass Transfer 2011; equiaxed dendritic growth: analytical solution. Acta Metal-
47(11): 1456–1476. lurgical 1987; 35(12): 2929–2933.
31. Mcbride D, Humphreys N, Croft TN, et al. Complex free 51. Lipton J, Glicksman M and Kurz W. Dendritic growth into
surface flows in centrifugal casting. Comput Modeling Vali- undercooled alloy metals. Mater Sci Eng 1984; 65: 57–63.
dation Experiments 2013; 62: 63–72. 52. Kurz W, Giovanola B and Trivedi R. Theory of microstructural
32. Li Y, Yuan M, Chadli M, et al. Unknown input functional development during rapid solidification. Metallurgical Mater
observer design for discrete time interval type-2 Takagi- Trans A 1993; 24(12): 2787–2802.
8
264 Measurement andand
Measurement Control 55(5-6)
Control 0(0)

53. Wang CY and Beckermann C. A multiphase solute diffusion 68. Wang ZJ, Li JJ, Wang JC, et al. Phase field modeling the
model for dendritic alloy solidification. Metallurgical Mater selection mechanisms of primary dendritic spacing in direc-
Trans A 1993; 24(12): 2787–2802. tional solidification. Acta Materialia 2012; 60(5): 1957–1964.
54. Jackson KA and Hunt JD. Lamellar and rod eutectic growth. 69. Langer JS, Grinstein G and Mazenko G. , Volume 1. World
Trans Metallurgical Soc AIME 1966; 236(8): 1129–1142. Scientific, 1986, pp. 165–186.Models of pattern formation in
55. Hunt JD and Jackson KA. Binary eutectic solidification. Trans first-order phase transitionsDir Condensed Matter Phys.
Metallurgical Soc AIME 1966; 236(8): 843–852. 70. Fix G. Phase Field Models for Free Boundary Problems.
56. Rappaz M and Gandin CA. Probabilistic modeling of micro- London: Longman, 1983.
structure formation in solidification processes. Acta Metal- 71. Collins JB and Levine H. Diffuse interface model of
lurgica et Materialia 1993; 41(2): 345–360. diffusion-limited crystal growth. Phys Rev B 1985; 31(9):
57. Gandin CA and Rappaz M. A coupled finite element-cellular 6119–6122.
automaton model for the prediction of dendritic grain structure 72. Langer JS and Sekerka RF. Theory of departure from local
in solidification processes. Acta Metallurgica et Materialia equilibrium at the interface of a two-phase diffusion couple.
1994; 42(7): 2233–2246. Scripta Metallurgica 1975; 9(6).
58. Gandin CA, Rappaz M and Tintillier R. Three-demensional 73. Caginalp G and Xie W. Phase-field and sharp-interface alloy
probabilistic simulation of solidification grain structures: ap- models. Phys Rev E 1993; 48(3): 1897–1909.
plication to superalloy precision casting. Metallurgical Trans A 74. Caginalp G and Socolovsky E. Computation of sharp phase
1993; 24(2): 467–479. boundaries by spreading: the planar and spherically symmetric
59. Gandin CA, Desbiolles JL, Rappaz M, et al. A three- cases. J Comput Phys 1991; 95(1): 85–100.
demensional cellular automaton finite element model for the 75. Caginalp G and Jones J. A derivation and analysis of phase
prediction of solidification grain structures. Metallurgical field models of thermal alloys. Ann Phys 1995; 237(1): 66–107.
Mater Trans A 1999; 30(12): 3153–3165. 76. Caginalp G and Fife P. Higher-order phase field models and
60. Dilthey U and Pavlik V. Numerical Simulation of Dendrite detailed anisotropy. Phys Rev B 1986; 34(7): 4940–4943.
Morphology and Grain Growth with Modified Cellular Au- 77. Penrose O and Fife PC. On the relation between the standard
tomation, Modeling of Casting. San Diego: welding and ad- phase-field model and a “thermodynamically consistent”
vanced solidification processes VII, 1998, pp. 589–596. phase-field model. Physica D: Nonlinear Phenomena 1993;
61. Nastac L. Numerical modeling of solidification morphologies 69(1): 107–113.
and segregation patterns in cast dendritic alloys. Acta Mate- 78. Penrose O and Fife PC. Thermodynamically consistent models
rialia 1999; 47(17): 4253–4262. of phase-field type for the kinetic of phase transitions. Physica
62. Wang W, Cee PD and Mclean C. A model of solidification mi- D: Nonlinear Phenomena 1990; 43(1): 44–62.
crostructures in nickel-based superalloys: predicting primary 79. Kobayashi R. A numerical approach to three-dimensional
dendrite spacing selection. Acta Materiallia 2003; 51: 2971–2987. dendritic solidification. Exp Math 1994; 3(1): 59–81.
63. Zhu MF, Lee SY and Hong CP. Modified cellular automation 80. Kobayashi R and Kai S. Pattern Formation in Complex
model for the prediction of dendritic growth with melt con- Dissipative Systems. Singapore: World Scientific, 1992.
vection. Phys Rev E 2004; 69: 1–12. 81. Chen Y, Qi XB, Li DZ, et al. A quantitative phase-field model
64. Yu J, Xu QY, Cui K, et al. Numerical simulation of micro- combining front-tracking method for polysrystalline solidifi-
structure evolution based on a modified CA method. Acta cation of alloys. Comput Mater Sci 2015; 104: 155–161.
Metallurgica Sinica 2007; 43: 731–738. 82. Takaki T, Schimokawabe T, Ohno M, et al. Unexpected se-
65. Yin H, Felicelli SD and Wang L. Simulation of a dendritic lection of growing dendrites by very-large-scale phase-field
microstructure with the lattice Boltzmann and cellular au- simulation. J Cryst Growth 2013; 382: 21–25.
tomaton methods. Acta Materialia 2011; 59: 3124–3136. 83. Takaki T, Ohno M, Shibuta H, et al. Two-dimensional phase-
66. Zhang H and Xu QY. Multi-scale simulation of directional field study of competitive grain growth during directional
dendrites growth in superalloys. J Mater Process Technol solidification of polycrystalline binary alloy. J Cryst Growth
2016; 238: 132–141. 2016; 442: 14–24.
67. Wesner E, Choudhurry A, August A, et al. A phase-field study 84. Tourret D and Karma A. Growth competition of columnar
of large-scale dendrite fragmentation in Al-Cu. J Cryst Growth dendritic grains: a phase-field study. Acta Materialia 2015; 82:
2012; 359: 107–121. 64–83.