中文摘要：

以一种橡胶阀双腔体并联泵为例，在输入110V正弦交流信号时，测得该压电薄膜泵的输出流量及出口压力的值。在实验数据的基础上，应用BP神经网络建立了压电泵输入输出系统的映射关系，并利用Matlab／Simulink提供的模块搭建了压电泵的模型。结果表明，通过BP神经网络得到的输出特性值具有较高的精度，能客观地反映压电泵输出特性随输入频率的变化关系。

英文摘要：

Taking a rubber valve dual-chamber parallel pump as an example, the output flow and pressure of this piezoelectric film pump were measured after imputing 110 V AC sine signal to it. Based on the experimental data, mapping relations of the piezoelectric pump input-output system were established by using the model of BP neural networks and the model of this piezoelectric pump was established using modules provided by Matlab/Simulink. Research shows that the values of output characteristics are very accurate by applying BP neural networks, which can objectively show the relations of output characteristics and input frequencies of a piezoelectric pump.