基于概率神经网络的燃气调压器故障诊断模型

Gas pressure regulator fault diagnosis model based on probabilistic neural network

燃气调压器故障诊断对城镇燃气管网的安全运行有着十分重要的意义,通过对燃气调压器进行故障诊断为其正常运行提供可靠的依据显得极为重要。鉴于此,针对主流燃气调压器提出一种基于概率神经网络(PNN)的故障诊断模型,主要通过搭建的概率神经网络对同一燃气调压器的4种不同类型状态(关闭压力高、用气高峰压力低、喘振故障和正常状态)进行诊断。仿真结果表明,与目前已经应用于工程实践的燃气调压器故障诊断算法如经验模态分解(EMD)算法相比,该方法在故障诊断率和诊断用时两个指标上有明显优势,应用于工程实际问题具有良好的有效性和实用性。

The fault diagnosis of gas regulator is very important to the safe operation of urban gas pipeline network. It is very necessary to provide reliable basis for its normal operation by fault diagnosis of gas regulator. In view of this,the fault diagnosis model based on probabilistic neural network(PNN) is introduced for the main gas regulator fault. The PNN is used to diagnose four different states(high shutoff pressure,low peak pressure,surge fault and normal state) of the same gas regulator. The simulation results show that the method has more obvious advantages in the fault diagnosis rate and the time required for fault diagnosis in comparison with the existing gas regulator fault diagnosis algorithms such as the empirical mode algorithm (E MD). It is effective and practical for the practical engineering problems.