Identification of nozzle parameters in 3S device using artificial intelligence and meta-heuristic optimization scheme

The supersonic separator is a device used to separate the unwanted components in a natural gas. The separated components are drained through the condensate drain which is placed before the normal shockwave in the convergent divergent nozzle. In this work, optimal area ratio and operating pressure ratio of the nozzle are obtained to develop the shock at a specific location with an acceptable pressure recovery coefficient. Here, the non-conventional optimization scheme genetic algorithm is employed to identify the optimal parameters. Further, the mathematical modelling is replaced by a neural network in the optimization scheme to reduce the computational cost. The complete work is carried out in the Matlab platform. The optimal parameters obtained using neural network are well in agreement with the mathematical model and the computational cost is drastically reduced.