The 8th Pacific Rim Conference on Modeling of Casting and Solidification Processes (MCSP8) was organised by the Korean Institute of Industrial Technology at the Sheraton Hotel, Incheon, Korea. During the three days of technical sessions from 12 to 15 April 2010, 52 papers were presented in seven sessions: solidification, microstructure evolution, continuous casting, mould filling, deformation, new application of CAD/CAE, and casting process.
More than 90 participants from Korea, Japan, China, Taiwan, Germany, UK and USA exchanged their theoretical and technical knowledge of computer simulation and casting in earnest during the presentation. Written versions of 24 of the presentations at the conference have been selected by the organisation for publication, following peer review, in this special issue of International Journal of Cast Metals Research.
The main directions in which progress in computer simulation of casting processes is sought may be summarised briefly under three categories. First, more accurate, rapid computation is indispensable in realising satisfactory simulation in the foundry shop. Second, extended applicability of modelling to handle overall casting processes and a wider range of materials should be established. Third, measures must be taken for the distribution of simulation packages to all foundries, even small facilities.
It is obvious that computer simulation of solidification and fluid flow is in practical use worldwide and is making a great contribution to improved productivity and defect prevention. For further successful application, in depth research and numerical formulation on the complex thermal, physical and metallurgical micro-phenomena occurring in the mould cavity during various casting process are demanded. Based on these fundamentals, the quality of numerical simulation will improve progressively.
As high speed internet networks are being built up everywhere, the chance to perform casting simulation on a distributed basis is just around the corner. With these new circumstances, powerful strategies of software distribution and access have to be considered.
In the first paper of this issue, Ohnaka et al. review the problems faced most frequently in the computer simulation of casting. Although many commercial codes have progressed rapidly and are widely used in the casting industry, there remain many challenges to overcome: more accurate simulation of defects, optimisation of casting processes, etc. To solve these problems, improved understanding of the real mechanisms of defect formation and the development of ways to handle the defects are strongly demanded.
Ohnaka et al. consider in detail the overall issues on which effort should concentrate in the near future. These include, among many others: estimation of material properties and boundary conditions, simulation of mould filling and solidification, prediction of porosities, and optimisation of casting. They also emphasise the importance of estimating the materials properties of sand moulds and the boundary conditions when considering the behaviour of oxide films on the molten metal, and how best to compare the results of simulation with real casting processes. The accurate prediction of inclusion entrapment, deformation and residual stress, and simulation of new casting processes are also very important. The overview surely provides an important map to anyone considering an overall master plan for development of casting simulation.
Even though many computational models have been proposed to predict microporosity formation in cast alloys, it is clear that stumbling blocks remain to realising optimal prediction of defects. This is mainly due to the complexity of a problem that involves many materials and process variables intertwined in complex physics. Nevertheless, theoretical efforts to solve these complicated problems should continue as a means to improve prevention of microdefects.
Stefanescu et al. propose the use of two models, based on physical concepts, for the growth of a gas pore in the mushy zone at the end of solidification of aluminium alloys. In this concept, several process variables – initial gas content, temperature and fraction solid, cooling rate, temperature gradient, applied pressure, melt purity (which affects pore nucleation) – are interconnected.
The first model is based on the growth of gas pores by diffusion, and the second on the equilibrium mass balance. Based on these two models, it is possible to calculate the final pore radius and total porosity volume depending on initial hydrogen concentration in the melt, applied pressure and impurities. By minimising the condition of initial hydrogen content and impurities, the conditions for minimal porosity in the casting can be inferred by calculation. These models have provided an effective theoretical basis on which modelling microporosity of aluminium alloy castings can advance.
Several approaches to the fluid flow simulation of molten metal, to improve accuracy and speed in calculation and to extend the application with respect to various casting processes, were reported, covering, for example, development of two-phase flow simulation in gas and molten metal, a pre-/post-processor for flexible treatment of curved surface using a cut-cell method, and lost foam casting simulation. Notable is the suggestion of Mouhamadou Diop et al. for a basic model to simulate the solidification and flow of aluminium foam during casting using the lattice Boltzmann method (LBM), which is a kinetic approach that considers flows to be composed of a collection of pseudo-particles represented by a distribution function.
In the field of microstructure analysis, grain size and microsegregation for practical applications, aluminium alloys were studied on the theoretical basis of cellular automaton (CA) prediction of dendritic growth. Dong-Ke Sun et al. have coupled CA to a kinetics based LBM solution, which has higher numerical convergence and stability than the Navier–Stokes (NS) equation, to model forced melt flow and solute transport in the liquid. CA was also coupled with PanEngine, the thermodynamic phase diagram software, to simulate multi-dendritic growth of the ternary alloy Al–4Cu–1Mg during solidification in the presence of melt convection. Validation showed good results when comparing simulated composition profiles of two solutes in solidified dendrites with the predictions of the Scheil model. This newly applied model is expected to be implemented effectively for the simulation of dendritic growth of ternary alloys in the presence of melt convection.
Recently, as the possibilities of cloud computing emerge rapidly, high performance computing assisted by parallel computing on the internet has the potential for powerful application to advanced casting design. In other words, a ubiquitous approach in which everyone uses high speed simulation on the internet easily and rapidly, at every place and at any time, has been created. By combining casting simulation technology with these new circumstances, it may be possible even for small foundries that cannot afford to use expensive commercial software to implement almost all kinds of new simulation software easily.
Sang-Hyun Cho et al. attempted to realise a cloud computing network in the casting field, which constructs parallel programmed solidification and fluid flow simulation on the high speed internet with cluster super-computing. They have established an Internet Simulation Center (ISC) on a trial basis, through which foundry engineers can conveniently use various simulation packages on the internet. In the near future, it is envisaged that such ubiquitous applications will make big differences at the foundry. For example, users will be able easily to select new software, whatever they want, from the huge sea of software on the internet without purchasing expensive packages – and even use them on their mobiles on a high speed train. ISC keeps running while the user’s notebook computer is turned off. After an hour or minutes, he or she can get the simulation results on a mobile device, or after getting off the train at his or her own computers. The ubiquitous distribution of the simulation technology will maximise the diffusion and impact of high technology at the foundry. The key to success will be to optimise supply and the quality of the contents of the service, and to guarantee security.
Many case studies to find optimum conditions for various casting processes – die casting, sand casting, electromagnetic casting, directional solidification casting, for example – were reported using commercial software and hardware for factors such as fluid flow, stress analysis and 3D computer tomography. Practical experiments combined with simulation results were reported and optimum parameters were suggested to minimise hot cracking, distortion, surface defects, porosity, blistering, fracture and microsegregation. Experiments were also reported to find accurate heat transfer coefficients during the various stages during solidification.
As artificial intelligence is being increasingly adopted in high technology industries, interest in applying it to casting simulation is growing. If this dream comes true, someday casting design experts may no longer be necessary! Artificial design experts will trace the most effective path to the final solution once initial conditions such as the modelling data, mould and casting condition are given. Trial and error will be minimised and optimum design can be obtained within a short time.
Dun-Ming Liao et al. present a study of the very first stage of automatic optimal design. They attempted to develop an automatic initial riser design system that couples CAE with CAD to connect simulation results with product design effectively. With this system, an isolated region in the casting during solidification is displayed on the CAD and the riser can be built on that place automatically.
To realise a practical intelligent expert system in the near future, high technologies that allow the computer to recognise the 3D shape of casting and mould automatically, and to analyse, judge and decide each step by itself, will be required. But this will be tough and it will take a long time.
The conference showed that basic and application studies are very active in the Pacific Rim area and holds out the promise of a bright future for more advanced computer simulation of casting. It is planned that future conferences will be held in turn in countries around the rim in alternate years.