基于融合策略粒子群优化BP神经网络的磨煤机出粉量优化  

Optimization of pulverized coal mill output based on particle swarm optimization BP neural network with fusion strategy

作　　者：李浩然 李力 陆金桂[1] LI Haoran;LI Li;LU Jingui(School of Mechanical and Power Engineering, Nanjing Tech University, Nanjing 211800, China)

机构地区：[1]南京工业大学机械与动力工程学院,江苏南京211800

出　　处：《南京工业大学学报（自然科学版）》2022年第2期180-186,共7页Journal of Nanjing Tech University(Natural Science Edition)

基　　金：国家科技支撑计划(2013BAF02B11)。

摘　　要：为了优化磨煤机系统参数,提高磨煤机出粉量,结合磨煤机系统参数和出粉量建立误差反向传播算法(BP)神经网络磨煤机出粉量模型,对参数进行软测量。为提高软测量准确度和对参数优化的效果,提出一种新的基于融合策略改进的粒子群优化BP算法,用于磨煤机系统参数和出粉量数据之间的非线性映射,建立估算模型。将估算模型应用于磨煤机出粉量的软测量中,为验证基于融合策略改进的粒子群优化BP算法的可靠性,将该算法与传统粒子群优化BP算法和传统BP算法对磨煤机出粉量的计算进行对比。结果表明:基于融合策略改进的粒子群优化BP算法模型对磨煤机出粉量有较好的软测量能力,预测值与实际值平均相对误差仅为3.8139%。In order to optimize the system parameters of coal mill and increase the pulverized coal output of pulverized coal mill,back propagation(BP)neural network pulverized coal mill output model was established combining with pulverized coal mill system parameters and pulverized coal output,and soft measurement was conducted on the parameters.To improve the accuracy of soft measurement and the effect of parameter optimization,a new improved particle swarm optimization BP algorithm based on fusion strategy for the non-linear mapping between pulverized coal mill system parameters and pulverized coal output data was proposed,and the estimation model was also established.Estimation model was applied in the soft measurement of pulverized coal output.To verify the reliability of the improved optimization BP algorithm,the improved optimization BP algorithm was compared with the traditional particle swarm optimization BP algorithm and the traditional BP algorithm for pulverized coal output of coal mill.Results showed that the improved particle swarm optimization BP model based on fusion strategy had better soft measurement ability of pulverized coal output,the average relative error between predicted and actual values was only 3.8139%.

关 键 词：磨煤机 出粉量 软测量 BP神经网络 参数优化 粒子群算法 融合策略 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TH122[自动化与计算机技术—控制科学与工程]