Combine Heat Balance with Data Analytics to Monitor and Detect Changes in Performance
Often plant operational variables change as a group in a manner (pattern) that is recognisable and repeatable. Advanced Pattern Recognition (APR) is a mathematical algorithm that identifies the patterns in historic (training) data, and then uses these learned patterns to detect changes in on-going plant operating data that arise from a developing problem. The SureSense software suite from Expert Microsystems is integrated with MapEx heat balance analysis to to provide the most accurate and cost-effective condition monitoring and diagnostic system possible.
APR does not need to use mathematical, physics or first principle models as the basis to predict a process. Instead, SureSense uses empirical methods to model the expected performance variable values that include the true process value plus amy measurement uncertainty and operational variance. These empirical methods require limited knowledge about the process because the model will learn and infer them directly from the data. As a result, APR can be applied to many types of process variables and process variations, such as operational changes and ambient condition changes. Typically, existing data from the data historian is used to create statistical relationships between relevant signals that can describe key performance, operational or reliability characteristics for key components. Once these relationships or prediction models are derived, they can be used to accurately estimate the predicted values of the various signals while the component is known to be operating normally.
The Predicted signals at “normal” conditions can then be compared to the actual Observed values in real-time by calculating the “Residual” – the difference between the Predicted normal values of the signals and the Observed values of the signals. This Residual should, over time, be statistically zero. If the residual begins to statistically deviate over time, this is a highly sensitive and accurate means for early detection of an anomaly that would lead to a failure.
Most other APR solutions only use this pure mathematical approach to create the prediction models. SureSense can also integrate performance and thermodynamic models into the APR process to holistically detect anomalies. Combining the APR and thermodynamic performance models, anomalies can be detected earlier and more accurately. Plus, the likely root cause of the anomaly can more easily be diagnosed.
Once an anomaly is detected, the underlying problem must be diagnosed properly before it can be resolved. Typically, the remote monitoring team must depend upon subject matter experts (SMEs) to manually provide insights into the problem. SureSense includes a Virtual SME which automates the diagnostic function. The Virtual SME captures the trouble shooting knowledge from the SMEs and SureSense automatically interprets the pattern of anomalies based on the prediction step to arrive at a possible diagnosis of the underlying root cause for the observed anomalies based on the fault characteristics. Most other diagnostic methods use a rule‐based, expert systems. These methods are not well suited for complex diagnostics that may have multiple “symptoms” that might influence multiple faults. SureSense uses a more sophisticated diagnostic methodcapability that can be used to assist the remote monitoring team in diagnostic method based on a Bayesian Belief Network (BBN) model-based approach which has proven to be very effective in diagnosing complex, multi-symptom faults.
The BBN type of diagnostic model determines the probability that a specific fault has occurred on the basis of the fault detector results. The design of the diagnostic model captures a user’s knowledge of the evidence expected in the data when a specific type of fault occurs. The BBN quantifies the belief that a failure mode is occurring based on the current state of the fault detectors, thereby providing the means to diagnose the most probable cause for the fault. Automating the diagnostic function using the Virtual SME allows a much more efficient and cost effective remote monitoring.
APR learns patterns of performance from historical data. The historical data can be enhanced with detailed information that may be critical to recognizing a pattern by providing heat balance outputs in addition to the measured data.
For example heat balance outputs include gas turbine air flow, turbine inlet temperature, first stage nozzle area, exhaust gas flow and composition, steam turbine discharge flow and enthalpy, feedwater heater steam flows, condenser duty, condenser cooling water flow, and many more.
These heat balance outputs are essentially additional measured data that can be input to the APR to improve accuracy, reduce false alarms and enhance early detection of developing problems.