基于BP神经网络的冲击波压力传感器组件动态特性分段建模方法研究

A segment modeling method for the dynamic characteristics of shock wave pressure sensor assembly based on BP neural network

受爆炸场中寄生效应的影响,需要采取相应抑制措施对冲击波压力传感器进行改造。为研究冲击波压力传感器组件的动态特性,基于双膜激波管对冲击波压力传感器组件进行动态校准,获得了阶跃响应信号;采用微分法求取了传感器组件动态特性非参数模型;根据冲击波压力传感器组件动态特性非参数模型变化规律,在频域内对其进行合理分段,并基于BP神经网络分段建模方法得到传感器组件动态特性模型;通过实例分析与比较,证明了基于BP神经网络的分段建模方法能够有效提高模型精度和建模效率。

Affected by the parasitic effect in explosion fields, appropriate measures need to be taken to reform theshock wave pressure sensors. To research the dynamic characteristics of shock wave pressure sensor assembly, based on double - diaphragm shock tube, the pressure sensor assembly of the shock wave can be calibrated and give a step response signal. Then the dynamic characteristics of the sensor assembly,s nonparametric model can be obtained by differential methods. According to the non-parametric model ’ s variation of the shock wave pressure sensor assembly,s dynamic characteristics,the dynamic characteristics should be reasonably segmented in frequency domain, and each part of the dynamic characteristics can be modeled based on BP neural network. Through analysis and comparison of an example, it can be proved that the segment modeling method based on BP neural network can effectively improve the accuracy and efficiency of modeling.