基于神经网络预测器的静压式油罐计量系统压力传感器的故障诊断

The Fault Diagnose Based on Neural Network Predictor for Pressure Sensor in HTG System

为使静压式油罐计量系统 (HTG)对压力传感器具有容错能力 ,提高系统可靠性 ,本文提出了一种基于神经网络预测器的HTG系统压力传感器故障诊断方法。该方法不但能判断出传感器是否发生故障 ,而且能判定故障类型 ,从而重构数学模型 ,保证监测过程的连续进行。仿真结果表明了该方法的有效性。

In order to make Hydrostatic Tank Gauge system process the fault tolerance and raise its reliability, a method based on neural network predictor is presented in this paper that is used for diagnosing pressure sensors in HTG system. It can determine whether sensors are in fault status and determine also the kind of faults. So the mathematics model can be rebuilt to continue the monitoring process. The simulation indicates it is valid.