Abstract:

According to the coal slime pipeline blockage problem of coal gangue thermal power plant, after the analysis of the actual scene, it is sure that thick slurry pump master cylinder pressure prediction is the necessary premise of blockage prediction. The thick slurry pump master cylinder pressure prediction model is proposed, which is GSVM (support vector machine with grid method), grid search method is used to parameters optimization of SVM. The simulation results show that parameter optimization time of this GSVM forecast model is 6.55 seconds, its prediction model is stable, and the relative error is less than 3%, which can satisfy the engineering requirement .And compared with prediction model based on GA-SVM(The genetic algorithm support vector machine), The GSVM prediction model is better than GA-SVM one in the optimization time and stability, and can be used in real time pressure prediction of the coal conveying system.

Key words: coal slurry pipeline, blockage, parameter optimization, support vector machine