In view of the time-consuming

low efficiency and low accuracy of manual acquisition and adjustment of operating parameters in the process of power plant units operation management

this paper designs an optimization and control system of power plant units operation based on data mining. Firstly

the data mining algorithm is used to integrate and analyze the parameters of the historical and real-time operation data of the units

and the key parameters affecting the units operation are obtained

which are used as the indicators of the unit health status evaluation. Then

the long-term ans short-term memory(LSTM) neural network model is employed to train the eigenvalues determined in the health state feature acquisition module

and the change trend of parameters with time is predicted

so as to realize the intelligent control of units. Finally

a set of gas-steam combined cycle generator operation control system is developed and applied to a power plant in Zhejiang Province. The operation results show that

this system can guide the optimal operation of the unit and improve the operation reliability and economy of the unit. After optimization

the output power of the unit increased by 0.412 5%

the annual power output increased by 4 806 MW·h

and the annual revenue of the unit increased by 3.26 million yuan.