BP神经网络优化参数的螺杆点胶阀无模型自适应控制技术  

Screw Dispensing Valve Model Free Adaptive Control Technology Based on BP Neural Network Optimization Parameters

作　　者：朱燕飞 楚友洋 李传江[1] ZHU Yan-fei;CHU You-yang;LI Chuan-jiang(The College of Information,Mechanical and Electrical Engineering,Shanghai Normal Universuty,Fengxian,Shanghai 201400)

机构地区：[1]上海师范大学信息与机电工程学院,上海奉贤201400

出　　处：《液压与气动》2023年第9期175-181,共7页Chinese Hydraulics & Pneumatics

基　　金：上海市自然科学基金(22ZR1445300)。

摘　　要：螺杆点胶阀因点胶效率高、精度高、胶点均匀性高等优点,广泛应用于微电子封装等领域,而电机转速的控制性能是决定螺杆点胶阀点胶质量的重要因素。针对点胶阀在外部干扰和负载扰动情况下,PID速度控制效果不理想,参数难以整定的问题,采用BP神经网络-无模型自适应控制(BP-MFAC)算法,通过BP神经网络在线整定无模型自适应控制器参数来实现对电机速度的自适应控制,无需人工整定参数。仿真和实验对比结果表明,该算法相较于PID控制算法,在电机速度控制上具有更小的超调量和稳态误差,更短的调节时间;在外部干扰和负载扰动的情况下,具有更好的抗扰动能力。The screw dispensing valve has high dispensing efficiency,high precision and high uniformity,is widely used in microelectronic packaging and other fields.Motor speed control performance importantly influences screw valve dispensing quality.Under external interference and load disturbance,PID controller effect of screw dispensing valve is unsatisfactory.And the PID controller parameters are difficult to be adjusted.To solve these problems,a BP neural network-model free adaptive control(BP-MFAC) algorithm is proposed.It adopts BP neural network to adjust the model free adaptive controller parameters online.The algorithm achieves adaptive control of motor speed,without manual parameter tuning.The simulate and experimental results show that compared with PID control algorithm,this algorithm has smaller overshoot and steady-state error,less setting time in motor speed control.It has better anti-disturbance ability under external disturbance and load disturbance.

关 键 词：螺杆点胶阀 PID参数整定 BP神经网络-无模型自适应控制 外部干扰 负载扰动 

分 类 号：TH137[机械工程—机械制造及自动化]