基于流量预测的波纹管泵流量脉动抑制控制方法  

Pulse suppression control method of bellows pump based on flow prediction

作　　者：顾晓伟 周明昊 申慧敏[1] GU Xiaowei;ZHOU Minghao;SHEN Huimin(School of Mechatronics Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学机械工程学院,上海200093

出　　处：《机电工程》2024年第9期1684-1691,共8页Journal of Mechanical & Electrical Engineering

基　　金：国家自然科学基金资助项目(52175055)。

摘　　要：随着晶圆内部线宽减小,在晶圆制造的湿法清洗工艺中对超纯水杂质颗粒的数量和直径提出了更严格的要求,以提高晶圆生产的产量,因此,对作为流控部件的波纹管泵(BsP)有更高的要求。针对波纹管泵运行过程中流量脉动导致的超洁净水颗粒杂质增加和晶圆良品率下降的问题,提出了一种基于神经网络流量预测模型的流量脉动抑制控制方法。首先,阐述了波纹管泵的工作原理和流量脉动产生的原因;然后,介绍了波纹管泵流固耦合模型的搭建方法,使用神经网络流量预测模型建立了流量脉动前馈控制的理论方法;最后,使用COMSOL仿真进行了流量预测模型验证研究,基于Simulink仿真平台开展了前馈控制应用于流量脉动抑制的验证研究。研究结果表明:在流量为[3.39,5.01]L/min时,预测的绝对流量误差范围为[8.0×10^(-1),1.6×10^(-1)]L/min;当流量为5.0 L/min时,最大流量脉动可从1.6 L/min降至4.0×10^(-1) L/min。在仿真验证中,该方法对波纹管泵脉动抑制有显著的效果,还可以有效提高波纹管的性能。With decreasing internal line width of wafers,stricter requirements were placed on ultraclean water impurity particles in the wet cleaning process of wafer fabrication,increasing the demands on the bellows pump(BsP) as a flow control component.Aiming at the problem of increased ultra-clean water particle impurities and reduced wafer yield caused by flow pulsation during the operation of the BsP,a flow pulsation suppression control method based on a neural network flow prediction model was proposed.In this study,a description of the working principle of the BsP and the reasons behind flow pulsation were provided.The construction method of the flow-solid coupling model of the BsP and the theoretical approach to establishing flow pulsation feed-forward control through the neural network flow prediction model were introduced.Finally,the flow prediction model was validated using the COMSOL simulation platform,and the application of feed-forward control for flow pulsation suppression was validated using the Simulink simulation platform.The experimental results show that the predicted absolute flowrate error falls within [8.0×10~(-1),1.6×10~(-1)]L/min for flowrate ranging from [3.39,5.01]L/min.When applying this model to BsP control,a simulation control model was established in Simulink.The results indicate that the maximum flow rate fluctuation can be reduced from 1.6 L/min to 4.0×10~(-1) L/min when the flow rate is 5.0 L/min.The use of the method of this study in simulation verification has significant effect on the bellows pump pulsation suppression.The application of this method can effectively improve the performance of the BsP.

关 键 词：波纹管泵(BsP) 流量脉动 流固耦合仿真 神经网络流量预测模型 COMSOL SIMULINK 前馈控制 

分 类 号：TH38[机械工程—机械制造及自动化] TH703[电子电信—微电子学与固体电子学] TN405