To save energy and reduce consumption

according to the historical operating data characteristics of coalfired generating units

an economic load dispatch method based on data mining technology is proposed. Firstly

the sliding window technology is used to extract the steady-state data in the historical operating data and the steadystate data is divided into different working conditions based on the fuzzy C-mean clustering algorithm. Secondly

based on the method of Gaussian process regression

the prediction models of the coal consumption rate of power supply under different working conditions are developed. Thus

the plant-level economic dispatch model with the minimum average coal consumption rate is constructed

and the dispatching output of each unit is solved by genetic algorithm. Finally

the effectiveness of the proposed method is verified by the simulation example. The results show that

the proposed method can establish the prediction model of coal consumption rate of power supply accurately. Moreover

compared with the normal AGC mode to single unit

the proposed method is more reasonable for the load distribution among the units

which reduces the average coal consumption rate of the whole plant and improves the coal saving capacity of the whole plant. The maximum coal saving is 1.77 g/(kW·h).