基于改进PSO和BP混合优化算法的炉温控制系统研究

Study of Temperature Control System Based on Improved PSO and BP Hybrid Optimization Algorithm

针对工业生产中电阻炉温度控制系统所存在的大惯性、大滞后等问题,提出一种应用于电阻妒炉温的基于改进粒子群(PSO,particle swarm optimization)算法的模糊RBF-PID控制策略。在该控制系统中,采用引入惯性权重因子和遗传变异算子的改进粒子群算法对模糊RBF(径向基函数)隶属度函数的初始值进行优化,再用BP(误差反向传播)算法进行细调,并结合模糊推理和RBF;学习能力在线调整PID控制参数,从而达到最优的PID控制效果。仿真结果表明,该算法跟踪快、超调小、不易陷入局部极小值,同时鲁棒性和抗干扰性优于传统PID控制。

In order to solve the problems of large inertia,large time delay in the resistance furnace temperature control system,a fuzzy RBF-PID control strategy based on improved particle swarm optimization（PSO） algorithm applied to the resistance furnace temperature is proposed.The initial value of fuzzy RBF（radial basis function） networks membership function is optimized by the improved PSO algorithm based on inertia weight factor and genetic mutation operator,adjusted imperceptibly by BP（error back propagation） algorithm.Combined fuzzy inference and RBF learning ability to adjust PID control parameters online so as to achieve the optimal PID control effect.The simulation results show that the algorithm is fast tracking,small overshoot,and is not easy to fall into local minimum value,while the robustness and anti-disturbance performance are better than the traditional PID control.