为了提高俯采工况薄煤层采煤机的装煤性能,针对煤层倾角和装煤率,运动参数之间的关系难以量化以及普通BP神经网络预测精度较低等问题,提出一种基于改进DBO-BPNN的装煤率预测模型。其以BP神经网络为基础,采用改进DBO优化算法优化其初始权重和阈值,提高其预测精度和稳定性。为了验证该预测模型的预测效果,使用EDEM离散元仿真软件建立俯采工况薄煤层采煤机装煤仿真模型收集数据进行模型训练,并和其他算法优化的BP神经网络模型进行对比,结果表明使用改进的DBO算法优化的BP神经网络在俯采工况薄煤层采煤机滚筒装煤率预测精度方面有较大的优势,平均预测误差仅为2.0525%,是一种有效的装煤率预测办法,可以为俯采工况薄煤层采煤机的参数优化提供一些帮助。

In order to improve the coal loading performance of thin coal seam coal winning machines underdownward mining conditions, a coal loading rate prediction model based on improved DBO-BPNN was proposed toaddress the difficulties in quantifying the relationship between coal seam inclination angle and loading rate, motionparameters, and low prediction accuracy of ordinary BP neural networks. It is based on BP neural network and usesan improved DBO optimization algorithm to optimize its initial weights and thresholds, improving its predictionaccuracy and stability. In order to verify the predictive effect of the prediction model, the EDEM discrete elementsimulation software was used to establish a simulation model for the loading of thin coal seam coal winning machinesunder downward mining conditions, collect data for model training, and compare it with BP neural network modelsoptimized by other algorithms. The results showed that the BP neural network optimized by the improved DBOalgorithm had a significant advantage in predicting the accuracy of the loading rate of thin coal seam coal winningmachines under downward mining conditions. The average prediction error is only 2. 0525%, which is an effectivemethod for predicting the loading rate and can provide some assistance for parameter optimization of thin coal seamcoal winning machines under downward mining conditions.