Data Mining-Based Operation Optimization of Large Coal-Fired Power Plants

Abstract

Large coal-fired power generation is a complex process characterized as nonlinear and coupling correlation between the levels of equipment, subsystems and function modules. It is therefore difficult to describe the energy-consumption behaviour and optimize the operation parameters under different operation conditions and boundary conditions with conventional methods. With data mining methods such as Support Vector Regression (SVR) and Genetic Algorithm (GA), a huge amount of practical operation data stored in the plant-level Supervisory Information System (SIS) were used to model the energy consumption and optimize the operation parameters for less coal consumption. The results show that the power coal rate reduced significantly under the combination of SVR and GA. The optimal operation program has a practical feasibility, and the whole optimizing process can supply model basis for large coal-fired power units.