基于多检测模型半导体气敏传感器故障诊断

Fault Diagnosis for Semiconductor Gas Sensor Based on Multiple Measurement Model

半导体气敏传感器的老化与意外“伤害”可能致使传感器失效,为提高检测的可靠性,提出了一种新的的传感器故障诊断方法。该方法对金属氧化物半导体甲烷传感器进行热调制,从传感器的动态输出信号中提取两组特征变量,构建两个不同的甲烷检测模型,提出根据两个模型的检测结果的一致性判断传感器的是否发生故障。应用实例表明,该方法能准确判断金属氧化物半导体甲烷传感器是否发生故障。由于不需要其他冗余硬件设备,因此故障诊断不受其他冗余硬件设备的影响,节省资源,价格便宜,可靠性高。此外,这种方法计算量小,故障诊断对系统的配置要求低。

Aging and sudden ＂damage＂ maybe disable a semiconductor gas sensor. In order to improve the measurement reliability of a measurement system, a new method for fault diagnosis of sensor is presented. Heat modulation is performed to metal oxide semiconductor methane sensor. Two sets of feature variables are abstracted from the dynamic outputs of sensor and are used to build two different measurement models for methane. The consistency between the measurement resuhs of two measurement models is used to judge if fauh happen. Application example shows that it can be judged exactly that if fault happen to metal oxide semiconductor methane sensor using this method. Because other redundant hardware unit is not needed, no effect resuhed from other redundant hardware unit affects fault diagnosis, resource is saved, the cost of fauh diagnosis is low, and reliability is high. Additionally, scheme requirement to fauh diagnosis system is low because count for this method is small in number.