On-Line Performance Monitoring With Signal Validation and Uncertainty Estimation

Methods are developed for performance monitoring of power plant components. On-line uncertainty estimation algorithms are developed. Since this can give new information about the validity of the measurements, proper use of performance monitoring can be achieved. A sequential fault detection method is introduced for the detection of small faults. This signal validation program gives an additional check for good inputs to the performance monitoring. Low-order models are solved using optimal estimation theory to get analytic measurements. The above algorithms are applied to heat transfer loops with real measurement data.