压力容器声发射信号人工神经网络模式识别方法的研究

采用人工神经网络模式识别技术对现场压力容器各种声发射源信号特征参数进行了模式识别分析，提出了采用人工神经网络分类方法对压力容器声发射源信号进行定量分析的概念，从而找到了评价声发射源严重程度的方法，设计和培训的人工神经网络可以给出一个多种因素产生的复合声发射源中裂纹扩展、氧化夹渣断裂、残余应力释放和机械摩擦信号所占的百分比，这一结果使声发射技术对压力容器安全状态的无损评价成为可能。

INVESTIGATION OF ARTIFICIAL NEURAL NETWORK PATTERN RECOGNITION OF ACOUSTIC EMISSION SIGNALS FOR PRESSURE VESSELS

The artificial neural network pattern recognition technique was employed to analyze AE source signals of pressure vessels in site. A concept for quantitative analysis of AE sources of pressure vessels by artificial neural network classification was given and the method for evaluating the severity of an AE source was thus found. The artificial neural network designed and trained gave the percentage of crack growing, slag inclusion cracking, residual stress releasing and structure rubbing signals for a complex AE source. The result made it possible to evaluate the safety condition of pressure vessels by acoustic emission testing.