RBF Neural Networks and Genetic Algorithms Based Optimization Control of Aluminum Powder Nitrogen Atomization Process

Abstract

Aluminum powder nitrogen atomizing process is with nonlinearities, large time delay, strong coupling and severe uncertainty, thus it is difficult to obtain the deterministic model and implement process optimization control by conventional methods. In this paper, the optimization control of aluminum powder nitrogen atomization process is presented to improve the fine powder rate. The process model of nitrogen atomization is established using RBF neural networks and the set values of control variables are optimized dynamically by means of implement of the optimization strategy based on enhanced genetic algorithms. Comparisons of the aluminum powder particle size distribution before and after optimization illustrate that the implement of process optimization control can improve the effect of nitrogen atomization and promote the percentages of ultra-fine aluminum powder greatly.