Determination of balance operating point of boiler-turbine load-pressure model based on data digging

Boiler and turbine control object in thermal power plant is a strongly nonlinear control system. Multi-model method is effective to solve the problem of nonlinear. While most multi-model controller were designed according to the nonlinear strength of coordinated control system without considering economy performance, operation economic was considered in this paper. The operation condition was classified using thermal power unit history data; then the improved K-means method was used to get K-clusters of the data for the specific operation condition; finally, fuzzy association rule was used to get the linearization balance operating point respectively from each cluster, which provides basis for the design of energy-saving controller.