Maintenance is important for the service of products and it is different from repair because repair focuses on the time node when the products break down, which is a qualitative problem, but maintenance pays more attention to the continuity of machine’s work, and thus it is not a qualitative problem. Nevertheless, almost all the study methods are of qualitative methods because they only qualitatively divided the health state of machines into several levels, which are not fit to comprehensively explore the degradation principle of machines and the relationship between degradation and maintenance. To discover the degradation principle of machines influenced by maintenance, a quantitative study method is proposed by calculating the Health Index (HI) based on fuzzy analytic hierarchy process (FAHP) and convolutional neural network (CNN). Finally, a case study is used to demonstrate the implementation and potential applications of the proposed method, in which two major maintenance methods in prognostic and health management (PHM), i.e. time-based maintenance (TBM) and condition-based maintenance (CBM) are studied. The results show that the application of the proposed method leads to a significant increase in the life of machines. This study puts forward a new method to study the degradation principle of machines and will lead to the development of PHM using the HI.

维护对产品的服务很重要，它与维修不同，因为维修着眼于产品出现故障的时间点，这是一个定性问题，但是维护则更注重机器工作的连续性，因此并不是一个定性的问题。然而，几乎所有的研究方法都是定性方法，因为它们只是定性地将机器的健康状态划分为几个级别，因此不适合全面探讨机器的退化原理以及退化与维护之间的关系。为了发现受维护影响的机器的退化原理，提出了一种基于模糊层次分析法和卷积神经网络的健康指数（HI）的定量研究方法。最后，通过案例研究来说明该方法的实施和潜在应用，其中研究了两种主要的预后和健康管理（PHM）维护方法，即基于时间的维护（TBM）和基于条件的维护（CBM） 。结果表明，所提出的方法的应用导致机器寿命的显着增加。这项研究提出了一种研究机器退化原理的新方法，并将导致使用HI的PHM的发展。结果表明，所提出的方法的应用导致机器寿命的显着增加。这项研究提出了一种研究机器退化原理的新方法，并将导致使用HI的PHM的发展。结果表明，所提出的方法的应用导致机器寿命的显着增加。这项研究提出了一种研究机器退化原理的新方法，并将导致使用HI的PHM的发展。