密相气力输送参数的人工神经网络预测

Artificial Neural Network Forecasting of Dense Phase Pneumatic Conveying Parameters

摘要: 在水平管道中，用压缩空气和氢气对煤粉和小米进行密相气力输送实验，利用改进的BP神经网络对不同气量下的固体质量流率进行预测。结果表明，BP网络能对不同实验条件下的固体质量流率进行较好的预测。并画出不同气量下，固体质量流量的等值图。根据此图，可对密相气力输送参数进行初步优化，控制固体的输送量，减少密相气输送中的盲目操作。

Abstract: Dense phase pneumatic conveying experiment of fine coal and millet are carried out using compress air and hydrogen at horizontal pipeline. Solid mass rate at different gas flow is forecasted by using improved BP neural network in dense phase pneumatic conveying. The result shows improved BP neural network can successfully forecast solid mass rate under different experimental conditions and solid mass rate contour diagram at different gas flow is plotted. According to the diagram, parameters of dense phase pneumatic conveying can be optimized, solids mass rate can be controlled, and blind operation of dense phase pneumatic conveying can be decreased.