论文部分内容阅读
提出了由动态贝叶斯网络表达的隐半马尔可夫模型用于齿轮磨损状态识别的新方法,通过计算待识别磨损特征向量的概率值来确定齿轮磨损状态。根据磨损特征之间的非线性关系这一特性,应用曲线距离分析方法对特征进行降维。最后,利用5种不同工况下的齿轮磨损实验数据对模型进行验证。结果表明,该模型可以有效地识别齿轮磨损状态,识别正确率可以达到94.5%,为齿轮箱的健康管理提供了科学依据。
A new method of Hidden Markov Markov model based on dynamic Bayesian network for gear wear state identification is proposed. The gear wear state is determined by calculating the probability value of the wear feature vector to be identified. According to the non-linear relationship between the wear characteristics, the curve distance analysis method is used to reduce the dimension of the feature. Finally, the model was verified by using gear wear test data under five different conditions. The results show that the model can effectively identify the gear wear status, and the recognition accuracy can reach 94.5%, which provides a scientific basis for the gear box health management.