Reliability based design optimization of casting process parameters using Markov chain model

Abstract

Foundries in India are facing lower productivity issues due to low quality of castings and wrong practices followed, therein. Minimizing rejection as well as production costs is the key area of concern in these foundries. Casting process parameters are important in the optimization of casting process. In casting process optimization, simulation has become a significant tool used by many researchers/practitioners. Design optimization using reliability has also emerged as a promising tool for optimization, using the probabilistic approach. This approach works on probability of failures through various parameters. Use of reliability-based design optimization (RBDO) in case of casting process optimization can provide some novel results. Markov process is a predictive technique for generation of a stochastic model with sequence of possible events. This research work focused on the development of a Markov chain model for casting failure. Various casting defects, such as cold shut, inclusion, etc. were considered in the study. Process parameters in case of casting process, such as mould hardness and, pouring temperature etc. are also considered. A Markov chain model provides the mapping of various casting failures with casting process parameters. With the probabilities of various process parameters and casting failures, a predictive model for casting process failure can be generated.