Data-driven Modeling and Application in Operation Optimization of Coal-fired Power Generation

Abstract

It is significant for the operation optimization to build an accurate and reliable coal rate model of coal-fired power units corresponding to specific operation conditions and working boundaries. A hybrid modeling based on thermodynamic theory and fuzzy rough set (FRS) method was introduced to select the input variables and preprocess the large amount of practical operation data; the coal rate model was built to illustrate the energy-consumption behavior of power units. The realizable optimal benchmark was determined with revolutionary algorithm. The case study and validation result show that the resultant optimum state is effective for operation optimization and coal rate reduction of large coal-fired power units.