MtsPSO-PID Neural Network Decoupling Control in Power Plant Boiler

In allusion to boiler combustion process nonlinear, multivariable, strong coupling to a large extent reduces the system regulating quality, the application of extreme disturbance simplify Particle Swarm (mtsPSO) algorithm combining with PID neural network methods is proposed. For the weight learning of neural network is easy to fall into local optimum, so using improved particle swarm algorithm to modify PID neural network weights, and by the addition of extreme value disturbance operator to accelerate particles to jump out of local extreme point. Through simulation analysis, mtsPSO algorithm to optimize PID neural network, very good decoupling effect to a smaller number of populations and the evolution of the number of generations.