摘要
针对动物缺氧实验中气体浓度控制这一时变非线性的过程,将BP神经网络与传统PID控制相结合虽然可以取得较好的控制效果,但是也存在着网络收敛速度慢、稳定性较差等问题.基于此,提出了一种基于改进的遗传算法优化的BP神经网络PID控制器.首先,该控制器对遗传算法的收敛速度和稳定性进行改进,利用改进后的遗传算法优化BP神经网络的权重初始值;然后,用优化后的BP神经网络实现PID控制参数的在线调整;最后,在MATLAB中对两种控制器进行仿真实验,结果显示,与传统的BP神经网络PID控制器相比,改进后的BP神经网络PID控制器具有更好的控制性能.
The process of gas concentration control in animal hypoxia experiment is time-varying and nonlinear,Combining BP neural network with traditional PID control can achieve better control results,but there are still some problems such as slow convergence speed and poor stability.To solve these problems,a new BP neural network PID controller,optimized by improved genetic algorithm,is proposed.The convergence speed and stability of genetic algorithm are improved in order to optimize the initial weights of BP neural network,then the optimized BP neural network was used to realize on-line adjustment of PID parameters in this controller.Finally,the conventional and improved controllers are simulated in MATLAB,the results show that the improved BP neural network PID controller has better control performance,compared with the conventional BP neural network PID controller.
作者
李航
杜璠
胡晓兵
周韶武
LI Hang;DU Fan;HU Xiao-Bing;ZHOU Shao-Wu(School of Mechanical Engineering, Sichuan University, Chengdu 610065, China;Sichuan Aerospace Changzheng Equipment Manufacturing Co., Ltd, Chengdu 610065, China)
出处
《四川大学学报(自然科学版)》
CAS
CSCD
北大核心
2020年第6期1103-1109,共7页
Journal of Sichuan University(Natural Science Edition)
基金
中国制造2025四川行动计划(2017ZZ018,2018ZZ011)。
关键词
PID控制
BP神经网络
遗传算法
气体浓度控制
智能控制
PID control
BP neural network
Genetic algorithm
Gas concentration control
Intelligent control