基于BP神经网络的煤矿干式钻孔产尘率预测模型

On Prediction of Dry Drilling's Dust Production Rate Based on BP Neural Network

分析总结了干式钻孔孔口产尘率的4个主要影响因素:煤体硬度、粉碎性能指标、固气比和含水率,建立了一个预测干式钻孔孔口产尘率的BP神经网络模型.根据在不同硬度和粉碎性能指标、不同含水率的煤体以及不同固气比条件下采集的数据,对建立的BP神经网络模型进行训练、验证和误差分析.结果表明,网络训练所得预测值与孔口产尘量实测值之间的最大绝对误差为4.200"10-3,相对误差的最大值为3.06%.由本文所建立的BP神经网络模型所得到的预测值与实测值吻合度很高,故该模型为防尘设备参数的选择和防治职业危害危害提供了理论依据.

This paper analyzes and summarizes the four significant factors i.e. the coefficient of coal hardness, the indicator of crushing performance, solids loading pneumatic, the moisture content and the drilling time that affect the dust capacity of dry hole orifice. In addition, a BP neural network model is built based on the properties of the coal dust capacity of dry hole orifice. The field measurement data under different conditions are used for building the training samples, and the accuracy of the model is tested. Differences between prediction and actual results are studied. Results show the maximum absolute error is 4.200 ~ 10-3, and the maximum relative error is 3.06%. The results obtained by the BP neural network model have well matched with the measured data, and the BP neural network model of production rate of dust has provided a theoretical basis for the parameters of equipment and controlling occupational hazards.