Reactor Temperature Prediction Method Based on CPSO-RBF-BP Neural Network

A neural network model based on a chaotic particle swarm optimization (CPSO) radial basis function-back propagation (RBF-BP) was suggested to improve the accuracy of reactor temperature prediction. The training efficiency RBF-BP is influenced some degree by large randomness initial weight and threshold. To address impact threshold uncertainty combined network, this paper proposes using algorithm correct network’s threshold, as well optimize speed up prediction accuracy. measured acquired on-site enterprises confirmed compared predicted results BP, RBF-BP, PSO-RBF-BP models. Finally, Matlab simulation tests were performed, experimental data revealed that CPSO-RBF-BP in had root-mean-square error 17.3%, an average absolute 11.4%, fitting value 99.791%. Prediction superior those model’s validity feasibility established. study findings may provide reference values for reactor’s