In order to accurately describe the relationship between the controllable parameters and the heat consumption rate

the prediction model for heat consumption rate was established. Five major variables of the model were determined and applied as the input parameters by using the steam turbine heat consumption rate equation and the partial correlation analysis method. Through comparative analysis

the RBF_kernel was selected as the kernel function of the model. The punishment factor μ and nuclear radial range σ of the least squares support vector machine(LSSVM) were optimized by the fruit fly optimization algorithm(FOA)

and the FOA-LSSVM black box model was built. The mathematical model for predicting the heat consumption rate of steam turbine was established based on the FOA-LSSVM. It was compared with the prediction model of heat consumption rate based on LSSVM and RBF neural network

and its robustness was analyzed. The results show that the prediction model of heat consumption rate of steam turbine based on FOA-LSSVM has the advantages of high precision

strong generalization ability and strong robustness. The model provides a theoretical basis for thermal economic diagnosis and parameter optimization of the unit

and has a certain value of engineering application.