Distributed computing system for a distillation column

Abstract

The use of a distributed memory message-passing system for dynamic chemical process simulation was developed. The computational performance of parallel system depends on the size of distillation column and the data communication. The distillation model was based on MESH equations. The case study example is performed using MPI message-passing system on three AMD DURON 600 MHz. Multicomputers. In this study, the dynamic behaviors of the binary and multicomponent distillation column were investigated. The multicomponent distillation was solved by the Bubble point method. The distillation column was partitioned into subsystems for each processor. For each subsystem, the serial computing portion were carried out without waiting computation the values from other processors for each time horizon integration. The comparative results between serial and parallel computations suggest that there is an optimum size of the distillation column and that of the transferred data to obtain the better result with parallel processing.