改进GWO的小波神经网络温控系统设计  

Design of Temperature Control System Based on Improved GWO Wavelet Neural Network

作　　者：陈珂乐 任天平[1] 郭帅 李保强 CHEN Kele;REN Tianping;GUO Shuai;LI Baoqiang(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou Henan 450001,China;SIPPR Engineering Group Co.,Ltd.,Zhengzhou Henan 450007,China)

机构地区：[1]郑州大学机械与动力工程学院,河南郑州450001 [2]机械工业第六设计研究院有限公司,河南郑州450007

出　　处：《机床与液压》2023年第9期97-102,共6页Machine Tool & Hydraulics

基　　金：河南省高等学校重点科研项目(19B460012)。

摘　　要：针对目前铸件砂芯表干炉温度控制性能差、燃烧效率低,设计一种新型热风循环温控系统。该系统以变限幅双交叉燃烧策略为基础,采用改进灰狼优化(GWO)算法的小波神经网络对PID控制参数进行自适应调整。系统仿真表明:与传统PID控制相比,超调量接近于0,系统调节时间减少了50%,温度切换控制速度提高了47%。最后通过砂芯烘干试验验证,与传统比值串级PID控制相比,变限幅双交叉燃烧策略和改进GWO小波神经网络PID对炉温的控制效果有很大的提升。Aiming at the poor temperature control performance and low combustion efficiency of casting sand core surface drying furnace,a new hot air circulation temperature control system was designed.Based on the variable limiting amplitude double cross combustion strategy,the wavelet neural network with improved gray wolf optimization(GWO)algorithm was used to adaptively adjust the PID control parameters.The system simulation shows that compared with the traditional PID control,the overshoot is close to 0,the system regulation time is reduced by 50%,and the temperature switching control speed is increased by 47%.Finally,through the sand core drying test,compared with the traditional ratio cascade PID control,the variable limiting amplitude double cross combustion strategy and the improved GWO wavelet neural network PID have a great improvement on the control effect of furnace temperature.

关 键 词：砂芯表干炉 变限幅双交叉 改进灰狼算法 小波神经网络 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]