中文摘要：

针对普通压电泵实现精密输送难问题,提出了基于神经网络预测控制的压电泵主动控制方法。在Matlab/simulnk环境下利用BP神经网络对压电泵实验样机的输入输出特性进行了辨识,建立了压电泵的模型。通过仿真分析研究了神经网络预测控制在压电泵精密输送中的应用,仿真结果表明,神经网络预测控制系统具有很好的动态特性及稳态精度,可较好的实现压电泵输出流量的主动控制。为实现压电泵的精密输送提供了一种新的可能的思路,也为压电泵向智能化方向发展做了准备,具有重要的工程应用价值。

英文摘要：

We advanced an active control method of the piezoelectric pump based on the neural network predictive control in order to solve the problem that the common piezoelectric pump is difficult to realize the precise transmission.The BP neural network was used to identify the input-output characteristics of a piezoelectric pump prototype by the Mstlab/simulnk software.The model of the piezoelectric pump has been established.The application of the neural network predictive control to the precise transmission of the piezoelectric pump has been studied through the simulation analysis.The simulation results showed that the neural network prediction control system has good dynamic performance and steady-state accuracy which can be used to control the output flow of a piezoelectric pump actively.The proposed active control method based on the neural network predictive control provided a new possibility of the precise transmission of the piezoelectric pump and opened up a new way for the intellectualization of the piezoelectric pump.The research results have important values in the engineering applications.