Diagnostics, Prognostics, and Condition Monitoring of Turbomachines

Condition monitoring is the process of collecting, interpreting, and monitoring the different operating parameters of turbomachinery, such as vibration, temperature, acoustic emissions, performance data, etc. The ultimate goal is identifying a change that might indicate a developing fault or problem.

A condition-based maintenance plan or predictive-maintenance plan develops a smart strategy that shuts down turbomachinery and repairs or replaces parts as required rather than relying on historical data and associated crude-life estimation. Comprehensive condition monitoring allows a turbomachine to continue operating uninterrupted until a fault or issue is detected.

DIAGNOSTICS AND PROGNOSTICS

Diagnostics identify a turbomachine’s current health and condition using applied logic, analytics, and experience to determine “cause and effect.” They are used to determine the causes of symptoms, mitigations, and solutions for the turbomachinery’s current situation; in other words, diagnostics include anomaly detection, fault isolation, fault classification, and uncertainty.

Prognostics is an engineering discipline focused on predicting the time at which a system, turbomachine, or component will no longer perform its intended function. The lack of performance is usually a failure or a stop in the operation. This method is different from rough-life estimation based on some historical data (often used in preventive maintenance). Prognostics science is a prediction of the remaining lifetime based on condition monitoring and details of possible failure mechanisms. Prior knowledge of the machine’s current health status (diagnostics) is required to develop prognostics capabilities. The predicted life (remaining useful life) is necessary data in the decision-making for maintenance planning and contingency mitigation. In addition to the estimation of remaining useful life, the prognostic results should include the uncertainty of the prediction and incipient fault detection.

Prognostics science forecasts future performance by assessing the extent of deviation or degradation from its expected normal operating conditions. It is based on analyzing failure modes and detecting early signs of degradation, wear, aging, and fault conditions. An effective prognostic solution requires knowledge of the failure mechanisms likely to cause degradation, leading to eventual failures. Therefore, it is necessary to have accurate and reliable data and information on the possible failures, including the failure modes, causes, and mechanisms, to identify the parameters to be monitored.

CONDITION-BASED MAINTENANCE PLAN

A condition-based maintenance plan can be optimized depending on the actual condition monitoring, diagnostics, and prognosis data. The main benefits of a condition-based maintenance strategy are:

• time and cost reductions due to better efficiencies and effectiveness
• optimized component life compared to time-based strategies and other traditional maintenance methods
• reduced mean maintenance time due to fault localization
• reduced operational disruption
• increased safety/reliability due to the reduction of unexpected failures

EVALUATING HEALTH AND CONDITION: PHYSICS-BASED MODELS

Data-driven models are based on statistical and machine-learning techniques and do not rely on the knowledge of the physics that govern the system or its degradation mechanisms. The main advantage of these techniques is their potential to be used in several turbomachines, as their physics is not required. This is sometimes used for cases where physical models are too complicated or unknown; however, there are some problems with data-driven models, such as the risk of overfitting and the necessity of large training data sets.

Physics-based models are preferred because they synergize with models used during the turbomachinery design phase. The physical concept can also be derived and understood, which is helpful when evaluating the condition monitoring, diagnostics, and prognostics processes and their results.

Hybrid models integrate different systems using various approaches depending on the task; for example, they combine data-driven and physics-based models. Normally, for diagnostics, a fault is detected by comparing the outputs of the physics-based model with the measurements from the real condition monitoring system.

For prognostics derived from physics-based models, degradation models represent the degradation mechanisms of the turbomachinery or system, and the remaining useful life can be estimated as an output of these models. For example, degradation models that represent crack growth can be used to estimate the remaining useful life of a component subject to fatigue and crack(s).