Pulse suppression control method of bellows pump based on flow prediction
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.
bellows pump(BsP)flowrate pulsationfluid-solid coupling simulationneural network flowrate prediction modelCOMSOLSimulinkfeedforward control
万方数据
维普
