To the problem that the existing methods are difficult to accurately predict and analyze the heat consumption of steam turbines with complex non-linear characteristics

this paper presents an improved prediction model of steam turbine heat consumption based on fuzzy support vector machine(FSVM). Firstly

the gap statistic algorithm is used to calculate the optimal number of clusters to avoid the uncertainty of the number of clusters.Then

the kernel fuzzy C mean clustering(KFCM) algorithm is applied to divide the heat consumption data

generate cluster subsamples

and replace it into the FSVM optimized by particle swarm optimization(PSO)

and establish a thermal consumption rate prediction model based on the FSVM. Finally

this model is employed to predict the heat consumption rate of an ultra-supercritical 660 MW unit steam turbine based on the data collected in the field

and the results are compared with that of the conventional support vector machine. The research results show that the improved FSVM method has higher prediction accuracy and stronger generalization ability.