ABSTRACT
FE software. The model has to take into account that improving some parameters can negatively influence other parameters. For instance significant thickness reduction of HPC plate may cause structural cracking. The numerical model with estimated geometrical proportions is a starting point for its design and optimization procedure. The single objective simulated annealing was developed by Kirkpatrick et al. (1983). A reasonable design solution could be reached by using a combination of more than single objective by using multi-objective design optimization (M-ODO) (Konak et al. 2006). The optimal set procedure is preferred in the M-ODO. In the movement from one solution to another solution, there is always an amount of loses in one objective to achieve some gain in the other objective. The M-ODO gives a set of solutions and the best solution measured regarding to all objective functions. Engineers always like to get only one value from the set of solutions. Therefore, solving M-ODO problems can be conducted by both searching and decision making. Finding the optimum solution involves extended iterative calculations and, thus, need of powerful calculation tools. Optimization of all design parameters at the same time would lead to high number of iteration and to very complicated ways to find the convergence. Therefore, all the design constrains should be carefully set before its optimization process. The proper selection of constrains leads to the faster convergence solution and significant reduction of computation time. Figure 1 shows the proposed design optimization methodology of HPCSPs. Its optimization procedure is consists of the four phases as follows:
1. The structural FE software Abaqus is used to perform structural optimization. It automates the design process for accuracy and speed purposes.
