Modern engineering components generally work under aging and dynamic cumulative damage processes. To prevent failures of such components, the proportional hazards model (PHM) was proposed to integrate both processes for health prognostics. However, the existing PHMs use constant damage rate within monitoring intervals for machine health estimation and still lack consideration of dynamic operational conditions, which fails to model the practical degradation situations. This article presents a prognostic model using a new PHM to consider aging and environment-varying cumulative damage for engineering machines. A dynamic multistate process with practical transition mechanisms under varying operational conditions is presented to model the cumulative damage progress. To address the difficulties in prognostics with PHMs, a matrix-based approximation method with low computational load is developed to compute important health measures such as conditional reliability, mean residual life (MRL) and residual life distribution. A prognostic scheme featuring online prediction and dynamic updating is presented. The particularity of the proposed model is that it considers dynamic environments and can be applied to a large number of deteriorating states. The proposed approach is illustrated using a case of pump under different operating environments, and comparison with other advanced PHM is given to validate the applicability and effectiveness of the proposed approach.

中文翻译：

---

老化和动态累积损坏机器的健康措施预测


现代工程部件普遍在老化和动态累积损伤过程下工作。为了防止此类组件发生故障，提出了比例风险模型（PHM）来集成两个过程以进行健康预测。然而，现有的 PHM 在监测间隔内使用恒定的损坏率来进行机器健康评估，并且仍然缺乏对动态运行条件的考虑，无法对实际的退化情况进行建模。本文提出了一种使用新 PHM 的预测模型，以考虑工程机械的老化和环境变化累积损坏。提出了在不同操作条件下具有实际转换机制的动态多状态过程，以对累积损伤过程进行建模。为了解决 PHM 预测的困难，开发了一种低计算量的基于矩阵的近似方法来计算重要的健康指标，例如条件可靠性、平均剩余寿命 (MRL) 和剩余寿命分布。提出了一种具有在线预测和动态更新的预测方案。该模型的特殊性在于它考虑了动态环境并且可以应用于大量恶化状态。该方法以不同运行环境下的泵为例进行说明，并与其他先进的 PHM 进行比较，以验证该方法的适用性和有效性。