数据驱动的液位传感器故障串联补偿方法研究

Research on Series Compensation Method of Data-Driven Liquid Level Sensor Fault

针对超声波液位传感器故障引起的非线性误差,基于数据驱动建模技术设计了一种逆补偿环节,以期能有效地补偿故障误差对系统的影响。首先借助故障传感器与正常传感器分别获得带故障误差与实际液位的测量值,然后利用LM-BP神经网络的非线性逼近特性设计超声波液位传感器故障误差的逆补偿环节。为验证该补偿方法的有效性与实际可用性,基于OPC、Matlab、Step7及Simatic等技术搭建PCS半实体实验平台,并将设计的逆补偿环节置于搭建的液位控制半实体平台的补偿环节中进行闭环实验。结果表明,该补偿方法对超声波液位传感器故障误差补偿是有效的,并且可将其应用于实际工程中。

Aiming at the nonlinear error caused by the fault of the ultrasonic liquid level sensor,an tnverse compensation link was designed based on the data-driven modeling technology,which could effectively compensate the nfluence of fault error on system.In this paper,the measured values of the fault error and the actual liquid level were obtained by means of the fault sensor and the normal sensor,the above offline data,using nonlinear approximation properties of LM-BP neural link of fault error of ultrasonic liquid level sensor was designed.In order to verify the validity and the practical availability of the compensation method,PCS semi solid experimental platform was built based on OPC,Matlab,Step7and Simatic and other technologies,and the designed inverse compensation link was put in the compensation link of the built liquid level control semi solid platform to carry out closed loop experiment.The results showed that the compensation method was effective for the error compensation of ultrasonic liquid level sensor fault,and could be used in practical engineering.