本論文研究的內容主要有兩個重點，其一為建立一滿足浴缸形失效率曲線之失效機制，進而建立一適用廣泛之動態可靠度模型；另一個則是探討有修理行為時的失效數據處理與分析，並依據前述模型就整體行為來探討修理對失效機制的影響。本文所介紹之動態可靠度模型係以可靠度為變數，並考慮(1)隨機失效、(2)損傷累積失效、(3)人－機介面影響、及(4)適應過程等四項失效機制構成。文中除了討論模型中各參數與不同失效機制間之關係，並以不同型態之數據驗證其可用性與通用性。除此之外並以常見之疲勞損傷過程，深入討論疲勞損傷過程與殘餘強度衰減之關係，在研究過程中藉用蒙地卡羅模擬法來模擬材料強度在隨機應力分佈的使用條件下之退化過程及其動態可靠度，做為進一步動態可靠度模型分析所需之數據；至於含修理行為之不完整失效紀錄數據分析則是以模糊集合與模糊關係理論處理，重建出合理之完整數據。同時提出等效動態可靠度概念來表現含修理之可靠度行為，再進一步以動態可靠度模型分析修理行為對動態可靠度退化的影響。從種種不同的實例中證實本文所發展的動態可靠度模型確實可以通用於各種不同狀況之數據，而且模型的數學形式簡單，並且可以由嵌合的參數說明動態可靠度退化過程中之各種物理現象，相較於傳統的模型而言，其對於工程設計分析較有其優越之處。 There are two issues investigated in this dissertation. The first one is to propose a general form of dynamic reliability model with bathtub shape hazard rate function. Another is devoted to develop an analytical method in treating the failure data with repair activities and the repair effectiveness according to the failure mechanisms. The proposed model is represented in terms of reliability and is composed by (1) random failure, (2) cumulative damage, (3) man-machine interference, and (4) adaptation. The relationships between the parameters of proposed model and various failure mechanisms are discussed. Several examples with different types of failure mechanism are taken in the validation of the proposed model. Meanwhile, the fatigue cumulative damage is selected as an example to study the memory characteristics cumulative damage vs. strength degradation. During this example, Monte Carlo simulation is applied to simulate the strength degradation under random distributed stress condition. The simulated results are further analyzed with the proposed dynamic reliability model. On the other hand, fuzzy set operation in the spirit of cumulative damage is adopted to construct the cumulative failure data set with repairs from incomplete information. A concept of equivalent dynamic reliability with repairs (EDRWR) is suggested to represent the dynamic reliability with repairs. In summary, this model is simpler than used models and is meaningful in capturing the physical phenomena during the reliability degradation, i.e. it is powerful in the engineering design analysis.