Using dynamical uncertainty models estimating uncertainty bounds on power plant performance prediction

Predicting the performance of large scale plants can be difficult due to model uncertainties etc, meaning that one can be almost certain that the prediction will diverge from the plant performance with time. In this paper output multiplicative uncertainty models are used as dynamical models of the prediction error. These proposed dynamical uncertainty models result in an upper and lower bound on the predicted performance of the plant. The dynamical uncertainty models are used to estimate the uncertainty of the predicted performance of a coal-fired power plant. The proposed scheme, which uses dynamical models, is applied to two different sets of measured plant data. The computed uncertainty bounds cover the measured plant output, while the nominal prediction is outside these uncertainty bounds for some samples in these examples.