水力致损煤岩比能耗特征及预测模型研究

采用相似材料与模型试验方法,以水泥、河砂、添加剂(丁腈橡胶粉、聚苯乙烯颗粒)等为原料,构造了典型煤岩相似材料,研究了泵压、切割次数、靶距及横移速度对水力致损煤岩比能耗影响特征,并基于粒子群算法(particle swarm optimization,PSO)优化反向传播神经网络(back propagation neural network,BPNN),构建了水力致损煤岩比能耗预测模型。在试验条件下得到:水力致损煤岩比能耗与泵压呈反比关系,且随泵压增加渐近于固定值26.0 J/cm~3;比能耗随切割次数增加表现为类"抛物线"状变化,最低比能耗对应切割次数为4;靶距小于4.0 mm时比能耗变化率较小,靶距大于7.5 mm时比能耗呈指数递增;比能耗随横移速度增加呈现先减后增的变化趋势,当横移速度为5.5 mm/s时比能耗处于最小值。所构建POS-BPNN水力侵彻煤岩比能耗模型的预测值与试验值具有显著的正相关性,可决系数为0.949 7;通过包括平均绝对误差、平均绝对百分比误差及均方根误差的综合指标体系,从多维度检测了所建模型的预测精度,表明该模型可以描述比能耗与水力冲孔参数之间的内在联系与规律。

Specific Energy Consumption Characteristics and Prediction Model of Hydraulic Damaging Coal Rock

According to the similar material and model method, using the cement, the river sand and the additives (the nitrile rubber powder and the polystyrene particle) as raw materials, it constructed the typical coal rock similar material, and analyzed the influences of pump pressures, cutting times, target distances and transverse velocities on the specific energy consumption of hydraulic damaging coal rock. Based on the optimized back propagation neural network (BPNN) by particle swarm optimization (PSO), a specific energy consumption prediction model of hydraulic damaging coal rock was established. In this experiment, the results show that there is an inverse relationship between the specific energy consumption and the pump pressure, the specific energy consumption decreases gradually to a constant of 26.0 J/cm3, as the pump pressure increasing. The specific energy consumption develops as a parabola type trend with the increase in the cutting times, and the minimum specific energy consumption occurs when the cutting time is n=4. As the target distance is in the range of 1 mm to 4 mm, the specific energy consumption just has a slight variation. Whereas, as the target distance is more than 7.5 mm, the specific energy consumption shows an exponential increase. With the increase in the traverse velocity, the specific energy consumption decreases first and then increases, and when the transverse velocity is v=5.5 mm/s, the specific energy consumption reaches the minimum value. The predicted values of the PSO-BPNN prediction model and the experimental values of hydraulic damaging coal rock experiments have a significant positive correlation, for the regression coefficient is R2 =0.949 7. The prediction accuracy was tested by the multiple-indicators, including the mean absolute error, the mean absolute percentage error and the root mean square error, and it is shown that the model can describe the internal relations and rules between the specific energy consumption and the hydraulic parameters well.