Research on diagnosis of abnormal causes in quality of weld based on the Bayesian net model

In the process of major equipment manufacturing, the quality of weld is an important influencing factor of the product. The diagnosis of abnormal causes in the quality of weld is the premise and important procedure in the quality management. In this paper, a new diagnosis method of abnormal causes for the weld quality based on Bayesian net model is proposed. Firstly, the quality characters and abnormal influencing factors are defined by analyzing the measuring data of weld. Meanwhile, to build the Bayesian net model, the quality characters and abnormal influencing factors is used as the symptom node and the abnormal cause node respectively. Then the priori probability of Bayesian net model can be obtained from the historical fault cases and expert knowledge. In the end, the diagnosis of abnormal causes is realized by the priori probability reasoning. A case study is provided to illustrate the validity of the proposed method.