改进模糊神经网络PID的瓦斯掺混浓度控制  

Improved fuzzy neural network PID for gas mixing concentration control

作　　者：张昭昭 代强 朱应钦 ZHANG Zhaozhao;DAI Qiang;ZHU Yingqin(College of Computer Science and Technology,Xi'an Uniersity of Saience and Technology,Xi'an 710054,China)

机构地区：[1]西安科技大学计算机科学与技术学院,陕西西安710054

出　　处：《西安科技大学学报》2023年第2期388-397,共10页Journal of Xi’an University of Science and Technology

基　　金：陕西省自然科学基础研究计划陕煤联合基金项目(2019JLZ-08);陕西省自然科学基础研究计划资助项目(2020JM-522,2021JM-396)。

摘　　要：为解决瓦斯发电过程中瓦斯掺混浓度控制系统存在一定的时变与非线性导致控制精度不佳的问题,提出一种基于改进粒子群算法的模糊神经网络PID控制算法(IPSO-FNN-PID)。首先,根据瓦斯掺混配比原理及历史数据建立瓦斯掺混浓度控制系统的传递函数数学模型。其次,为解决模糊神经网络选定随机初始网络参数导致网络输出结果差异较大的问题,采用遗传算法中的交叉、变异操作改进传统粒子群优化算法,以此提高传统粒子群优化算法的搜索性能,利用改进后的粒子群算法优化模糊神经网络的初始网络参数。最后,用优化后的模糊神经网络实现PID控制参数的整定,基于真实瓦斯浓度数据进行试验,并与传统PID,模糊PID和模糊神经网络PID进行了对比。结果表明:基于IPSO-FNN-PID控制算法在超调量及调节时间等方面均优于另外3种控制算法,能够实现瓦斯掺混过程中2种动态实时变化瓦斯浓度的精确控制。In order to solve the problem of poor control accuracy due to time-varying and nonlinear of gas mixing concentration control system in gas power generation process,a fuzzy neural network PID control algorithm(IPSO-FNN-PID)is proposed based on improved particle swarm optimization.Firstly,according to the principle of gas mixing ratio and historical data,the mathematical model of transfer function of gas mixing concentration control system is established.Secondly,in order to solve the problem that the random initial network parameters selected by the fuzzy neural network lead to large differences in the network output results,the crossover and mutation operations in the genetic algorithm are used to improve the traditional particle swarm optimization algorithm,so as to promote the search performance of the traditional particle swarm optimization algorithm.The initial network parameters of the fuzzy neural network are optimized by the improved particle swarm optimization.Finally,the optimized fuzzy neural network is used to achieve the tuning of PID control parameters.Tests are carried out based on real gas concentration data,and compared with those by traditional PID,fuzzy PID and fuzzy neural network PID.The results indicate that the IPSO-FNN-PID control algorithm is superior to the other three control algorithms in overshoot and regulation time,and could realize the accurate control of two kinds of dynamic and real-time changing gas concentrations in the process of gas mixing.

关 键 词：瓦斯掺混 浓度控制 模糊神经网络PID 改进粒子群算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]