基于长短时记忆网络的盾构机刀盘扭矩实时预测

盾构机是土质隧道开挖的优质工程机械,被广泛应用在地铁建设。刀盘扭矩是保证盾构正常推进的关键参数,能被精确实时预测对预防灾难事故、确保施工正常推进具有极高的指导意义。针对现有扭矩预测多为计算平均值的问题,提出一种基于长短时记忆(Long-Short Term Memory,LSTM)网络的扭矩实时预测模型。首先通过分析盾构机状态参数与扭矩的相关性,选择一组关键状态参数,降低输入维度;然后建立LSTM扭矩预测模型;最后利用该模型在归一化后的实际数据集上进行验证并与BP网络模型对比。试验结果表明,该模型在测试集上均方差为0.002 81,平均绝对误差为0.036 7,均优于BP网络;该模型具有良好的预测能力与泛化性能,能够很好地拟合关键状态参数与刀盘扭矩之间的非线性关系。

Real-time Prediction of Torque of Cutterhead of Shield Machine Based on LSTM

Shield tunneling machine is a great construction machinery for full-face excavation of soil tunnels (tunnels), which is widely used in domestic urban subway construction. Torque of cutterhead is one of the key parameters to ensure normal shield driving. Precise prediction and monitoring of torque of cutterhead has a very high significance for preventing disaster accidents and ensuring the normal progress of construction. In order to solve the problem that the current calculation of cutter head torque of shield machine is mostly static prediction, a real-time prediction model of cutter head torque of shield machine based on Long-Short Term Memory (LSTM) is proposed. Firstly, a set of key state parameters are chosen to reduce the input dimension by analyzing the correlation between the state parameters and the torque of the shield machine. Then a real-time prediction model of the cutter head torque of LSTM shield machine is established. Finally, the model is validated on the normalized data set of shield machine operation and compared with BP network model. The experimental results show that the mean square error of the model is 0.002 81, and the mean absolute error is 0.036 7, which is better than BP network. The model has good prediction ability and generalization performance, and it can well fit the non-linear relationship between the key state parameters and the cutter head torque.