基于PSO和BP网络的粉状炸药配方的预测控制

BP神经网络具有良好的非线性处理能力,粒子群优化算法(PSO)提高神经网络的学习效率、保证神经网络全局收敛。针对粉状炸药配方控制系统中存在的强耦合、非线性及参数不确定等问题,建立一种基于粒子群和BP神经网络的炸药配方预测控制模型。仿真和现场运行结果表明,该配方控制系统具有良好的自学习和自适应能力,取得了良好的控制效果,满足实际生产要求。

Prediction Control of Powdery Dynamite Reciping Based on PSO Combined with BP Neural Network

BP neural network has good performance of nonlinearity processing.Particle swarm optimization can raise the learning efficiency of neural network,and ensure global convergence.To the problem that powdery dynamite reciping control system is nonlinear,coupling,and parameter uncertainty,a dynamite reciping predictive control model based on PSO and BP neural network is presented.The results of simulation and engineering application show that the control system has good ability of self-adaptive and self-learning.It has good control results,and the the solution obtained can meet the production practice expectation.