1 1.2. Main concepts 3 PHM 3. Diagnostics 4. Prognostics 5. Maintenance Optimisation & logistics 6. Decision-making and policy deployment 2. Data & Information collection Reliability Modelling Markov process-based Prognostics 1. System/ Component/ Process (Section 1.2.3) (Section 1.2.2) Maintenance Optimisation (Section 1.2.4) Figure 1.2: The main stages of the PHM paradigm. The main stages in PHM are depicted in Figure 1.2. Stage 1 identifies the system, component, or process of interest. Stage 2 involves the design and implementation of infrastructure that enables the collection of data and relevant information, such as monitoring systems. Stage 3, known as diagnostics, addresses what is wrong? and focuses on the current system condition. Here, models detect, isolate, locate, quantify, and classify anomalies and failure modes. Stage 4, known as prognostics, answers how long until an event or state is reached? and focuses on the future condition of the system. Here, models attempt to characterise the system’s future performance. Stage 5, referred to as maintenance optimisation & logistics, focuses on the set of algorithms that seek the optimal set of actions to control andmeet functionality requirements. Lastly, Stage 6 focuses on the decision-making and deployment of the controlling policy. Once Stage 6 is reached, the cycle restarts, even with new goals given former stages, e.g., evaluating the e!ectiveness of a maintenance action. The arrows in Figure 1.2 refer to the logical progression between stages rather than a strict sequence of steps. For instance, a non-nominal behaviour identified in Stage 3 may prompt a maintenance action (Stage 5), such as performing an inspection, thus skipping Stage 4—Prognostics. Additionally, PHM can serve as a design tool. For instance, the requirements for employing a specific type of prognostic model in Stage 4 help determine part of the data to be collected in Stage 2.
RkJQdWJsaXNoZXIy MjY0ODMw