基于人工神经网络的岩石截割参数预测

Prediction of Cutting Characteristics of Rocks Based on Artificial Neural Network

鉴于前人推导的镐形截齿破岩截割阻力和截割比能耗的理论公式计算值与实际值相差较大以及最优截槽宽没有定量表示,文中选取岩石密度、单轴抗压强度、抗拉强度、静态弹性模量等为影响因子,建立了BP预测网络模型,并利用此模型对我国常见的4种岩石镐形齿截割参数进行了预测。检验及预测的结果表明建立的预测网络运行稳定,预测结果良好,对截割力的预测优于理论计算结果,对截槽宽和截割厚度最优比值、截割比能耗的预测结果良好,相对现有理论的计算和经验公式计算精度有了很大提高,能更好的满足工程要求。

Because there is large error between theoretical value and experiment value of cutter forces and the specific energy,and the best line spacing hasn＇t quantitative expression,but they are very important design and selection parameters of the mining machinery,which influence the design work to a larger extent,so selecting the density,compressive strength,tensile strength,static young＇s modulus of the rock as the impact factors to establish the BP network model for predicting cutting characteristics and predicting the four common rock of china by the net in the paper.The test and prediction indicate that the network run basically stable and gives good results,the prediction of cutting forces is better than theoretical calculations.Prediction of the ratio between the line spacing and depth of cutting and specific energy are able to better meet engineering requirements,which have a great improvement than the existing method to a large extent.