Data underlying the publication: A Bayesian inference-based framework for modelling imperfect post-repair behavior of remaining useful life under uncertainty

Maintenance decisions in condition-based maintenance often involve choosing between replacement and repair. The high costs associated with replacement and the introduced uncertainty from the structure's assembly and disassembly have led to increased exploration of repair methodologies. However, repairs are often imperfect, leading to additional uncertainties in predicting the component's future health condition. In this work, a model that utilizes Bayesian inference to estimate the distribution of the structure's recovery after an imperfect repair based on Remaining Useful Life estimations is proposed. The novelty of the methodology comes from the development of an imperfect repair model by looking at it from an integration point of view under the Prognostics and Health Management umbrella that could work independently of the prognostic and decision-making phases, offering flexibility. Evaluation of the proposed model is conducted through tension-tension fatigue experiments on aerospace-grade aluminium open-hole coupons. Repair is performed via a rectangular carbon fiber reinforced polymer patch placed on each coupon.<br>This is a dataset related to the experimental setup of our work concerning imperfect repairs (paper under review). Additional details can be found in the README file provided.

基于状态的维护（Condition-Based Maintenance, CBM）中的维护决策通常需在更换与维修之间进行抉择。更换操作的高额成本，以及设备结构拆装过程中引入的不确定性，推动了维修方法论的研究探索。然而，维修往往存在不完美性，进而在预测部件未来健康状态时引入额外不确定性。本研究提出一种模型，该模型基于剩余使用寿命（Remaining Useful Life, RUL）预测结果，借助贝叶斯推理（Bayesian Inference）估算不完全维修（Imperfect Repair）后设备结构的恢复状态分布。该方法的创新点在于，从故障预测与健康管理（Prognostics and Health Management, PHM）体系下的集成视角出发构建不完全维修模型，该模型可独立于预测与决策阶段运行，具备出色的灵活性。本研究通过针对航空级铝合金开孔试样的拉-拉疲劳试验对所提模型开展验证，维修操作通过在每个试样表面粘贴矩形碳纤维增强聚合物（Carbon Fiber Reinforced Polymer, CFRP）补片完成。
本数据集与本研究中有关不完全维修的实验设置相关（论文目前处于审稿阶段）。更多详细信息可参阅所附的README文件。