在可修復系統可靠度的研究上，一般常使用Power Law Process模型或是Log Linear Process模型，來描述一可修復系統的失效過程。而Pulcini(2001)所提出的Bounded Intensity Process模型，則對於一系統有隨著運轉時間愈久，其失效強度趨於常數值的現象發生時，能較Power Law Process模型與Log Linear Process模型，更有效地描述該系統的失效過程。 而本文將針對一退化的可修復系統，在上述三種不同的模型下，分別使用失效及時間設限二種收集失效時間點的方式；研究由Bain等(1985)所考慮的時間趨勢檢定法，分別為，Likelihood Ratio檢定法、F檢定法、Laplace檢定法與MIL-HDBK-189檢定法的檢定力表現為何。並且於F檢定法中，在資料切割點的選定上，隨著不同的模型，做一些適度的變換。 最後，本文將引入由Kumar與Klefsjö(1992)所提到關於LHD水壓系統發生失效的部分資料；透過由LHD的壽命，老(old)、中(medium old)、新(new)三群分類中，各挑選一組資料，分別為，LHD 1、LHD 9及LHD 17；以上述三個模型分別來分析其失效過程的結果。並且企圖歸納出以何種模型來描述此類水壓系統的失效過程較為適當。 On the reliability study of the repairable system, we often use the Power Law Process, the Log Linear Process to describe the failure process of the system. Pulcini(2001) finds the Bounded Intensity Process model, which is more efficiently used to describe the failure process. Because the operating goes, the failure intensity of system goes to a constant value possibly. This article will consider a deteriorating repairable system and use two kinds of methods of collecting data denoted as failure censored and time censored. Under the previous methods, we want to study the power of the following tests denoted as Likelihood Ratio test, F test, Laplace test, and MIL-HDBK-189 test given by Bain, etc.(1985). And, under the F test, we make proper changes on the selection of the cutting of the data with the different models. Finally, we consider the failure data of the hydraulic system of LHD machines, given by Kumar and Klefsjö(1992). In particular, we use the previous three models to analyze the failure data relative to the machines denoted as LHD 1, LHD 9, and LHD 17. And we try to conclude a proper model to describe the kind of the hydraulic system from the individual analysis.
