摘要
现有的热泵干燥多侧重于温度控制,不能克服温湿度的耦合作用实现温度与湿度的同时控制,因而提出一种基于预测PI的热泵干燥温湿度解耦控制方法。首先,基于热泵干燥实验数据进行干燥温湿度控制对象的辨识,得到控制对象模型;然后,采用预测比例-积分(proportional integral,PI)算法和动态解耦实现热泵干燥温湿度控制;最后,利用Simulink仿真工具对比分析了PI、比例-积分-微分(proportional integral derivative,PID)、史密斯预估器(Smith predictor)和预测PI(predictive proportional integral)的控制性能、抗干扰能力和解耦效果。结果表明,预测PI控制具有超调量小、振荡小、鲁棒性良好和抗干扰能力强的特点,适合于大惯性、大滞后热泵干燥过程的温湿度多参数控制。
Existing heat pump drying mostly focuses on temperature control,and cannot overcome the coupling effect of temperature and humidity to achieve simultaneous control of temperature and humidity.Therefore,a temperature and humidity decoupling control method for heat pump drying based on predictive PI is proposed.Firstly,based on the experimental data of heat pump drying,identify the drying temperature and humidity control object and obtain the control plant model;Then,the predictive PI algorithm and dynamic decoupling are used to achieve temperature and humidity control for heat pump drying;Finally,the control performance,anti-interference ability,and decoupling effect of PI,PID,Smith predictor,and predictive PI are compared and analyzed using Simulink simulation tools.The results indicate that predictive PI control has the characteristics of small overshoot,small oscillation,good robustness,and strong anti-interference ability,and is suitable for multi parameter control of temperature and humidity in large inertia and large lag heat pump drying processes.
作者
赵海波
张静峰
吴坤
乔玲敏
ZHAO Haibo;ZHANG Jingfeng;WU Kun;QIAO Lingmin(School of Ocean,Yantai University,Yantai 264005,China;School of Civil Engineering,Yantai University,Yantai 264005,China;Department of Automotive and Marine Engineering,Yantai Vocational College,Yantai 264670,China;Yantai Aiclunt New Energy Technology Co.,Ltd.,Yantai 264005,China)
出处
《控制工程》
CSCD
北大核心
2024年第12期2149-2158,共10页
Control Engineering of China
基金
山东省自然科学基金资助项目(ZR2014EL029)
山东省高等学校科技计划项目(J12LB55)
烟台市科技计划项目(2020XDRH097)
烟台市“十三五”海洋经济创新发展示范项目(YHCX-ZB-L-202003)。
关键词
预测PI
热泵干燥
解耦控制
Predicting PI
heat pump drying
decoupling control
