基于动态贝叶斯网络的设备故障预测方法研究

设备维修方式由事后维修和定期维修逐步向基于状态维修进行转变，而故障预测（剩余寿命预测）则是基于状态维修的关键。本文介绍了动态贝叶斯网络的基本原理，在此基础上提出了利用动态贝叶斯网络进行故障预测的方法和步骤，最后将本方法应用于流体控制器的故障预测，结果验证了本方法的有效性。

Study on the Equipment Fault Prognosis Based on Dynamic Bayesian Networks

Morden maintenance techniques have gradually changed from correction maintenance, schedule maintenance to condition based maintenance （CBM）. Failure prognosis （remaining useful life）is the key component of CBM. The contents of research introduce the keystone of Dynamic Bayesian Networks（DBN）, At the basis of these work, we put forward means and steps for failure prediction which make use of DBN. Finally, the proposed frameworks are validated through the failure prediction of fluid distributor. The result indicates that the frameworks is avilability.