In dynamic complex environments, the degradation of structure systems is generally caused not by a single factor but by multiple ones, and the process is subject to a high level of uncertainty. This article contributes a hybrid physics-model-based and data-driven remaining useful life (RUL) estimation methodology of structure systems considering the influence of multiple causes by using dynamic Bayesian networks (DBNs). The structure model and parameter model of DBNs for the degradation process caused by a single factor are established on the basis of theoretical or empirical physical models, thereby solving the problem of insufficient data. An RUL estimation model is subsequently established by integrating these degradation process models. The RUL value is obtained from the time difference between the detection point and predicted failure point, which is determined using the failure threshold of performance. The sensor data and expert knowledge can be input into the estimation model to update the RUL value whenever necessary. The subsea pipelines in offshore oil and gas subsea production systems are adopted to demonstrate the proposed methodology. The degradation processes with fatigue, corrosion, sand erosion, and internal waves are modeled using DBNs, and the RUL is estimated using a DBN-based RUL methodology.

中文翻译：

---

多重原因影响下的结构系统剩余使用寿命估算：以海底管道为例


在动态复杂环境中，结构系统的退化通常不是由单一因素而是由多种因素引起的，且过程具有高度的不确定性。本文提出了一种基于混合物理模型和数据驱动的结构系统剩余使用寿命 (RUL) 估计方法，通过使用动态贝叶斯网络 (DBN) 考虑多种原因的影响。在理论或经验物理模型的基础上，建立了DBNs单因素退化过程的结构模型和参数模型，解决了数据不足的问题。随后通过整合这些退化过程模型建立RUL估计模型。 RUL值是根据检测点和预测故障点之间的时间差获得的，该时间差是使用性能故障阈值确定的。传感器数据和专家知识可以输入到估计模型中，以便在需要时更新RUL值。采用海上石油和天然气海底生产系统中的海底管道来验证所提出的方法。使用 DBN 对疲劳、腐蚀、沙蚀和内波等退化过程进行建模，并使用基于 DBN 的 RUL 方法来估计 RUL。