设备状态智能诊断模型的自更新机制及其UML建模

A mechanism for self-updating of the machine condition monitoring and intelligent diagnostic model for its modeling based on UML

为解决在设备状态监测应用中建立智能诊断模型经常面临历史样本数据空间有限的问题，研究了智能诊断模型的自更新机制，并采用统一建模语言对其进行了分析建模。在此基础上，给出了该机制的实施架构。该机制的基本思想是用实际设备状态监测过程中的监测数据来更新智能诊断模型。在此机制作用下，通过在设备状态监测过程中跟踪设备状态的变化，一个基于有限的设备状态样本空间训练的智能诊断模型能够在模型失效的情况下，通过学习新设备状态下的监测数据不断提升其诊断能力。该机制的可行性和有效性通过实例应用得到了验证。

A mechanism for self-updating of an intelligent diagnostic model for machine condition monitoring is proposed and modeled with the unified modeling language （UML） to overcome the common problem that the historical sampie space is always limited when building the intelligent diagnostic model. A framework for implementing such a mechanism is also presented. The mechanism is based on the common procedure in building an intelligent diagnos- tic model. The underlying idea is to update the intelligent diagnostic model with the monitored data under a new machine condition. An initial diagnostic model built with the limited historical sample space is about to be updated with the monitored data under a new machine condition when the current intelligent diagnostic model is detected not competent for the diagnosing tasks of the new machine condition. Finally, the proposed mechanism is verified by applying it to a bearing condition monitoring example.