基于知识的管道系统泄漏检测方法

The Leakage Detection Methods of the Piping System Based on Knowledge

提出了3种基于知识的管道系统泄漏检测方法,分别是:通过对负压波波形分段符号化处理,采用上下文无关法表示波形结构模式的模式识别诊断系统;以选取合适维数的功率谱特征向量而建立的神经网络泄漏检测系统;与泄漏机理数学模型结合的泄漏诊断专家系统。

This paper puts forward three leakage detection methods of the piping system based on the knowledge. The first is the diagnosis system which symbolizes the negative wave form in segment mode and uses the context-free grammar of symbolic representation of the wave form to identify the mode. The second is the neural network detection system which uses a neural network with the input matrix composed by stress wave characteristics in time domain or frequency domain to classify various situations of the pipe- line. And the last is the leakage diagnosis expert system which combines the leakage mechanism mathematics model.