摘要
为了能对隧道掘进机(tunnel boring machine,TBM)拆装装置作业时的安全做出有效预警,通过研究两级信息融合建立最优的安全预警模型,为TBM拆装装置吊装作业提供安全保障。一级融合将数据通过层次分析法-熵权法算法融合得出安全状态系数;二级融合建立灰色模型(grey model,GM)、差分自回归移动平均(autoregressive integrated moving average model,ARIMA)、长短期记忆网络(long short-term memory,LSTM)模型,通过3个单项预测模型构建4个简单平均组合模型和4个最优加权组合模型,对拆装装置作业时的安全状态系数进行预测分析,通过相关系数R、平均绝对误差(mean absolute error,MAE)、平均相对误差(mean absolute percentage error,MAPE)、均方根误差(root mean square error,RMSE)4个评价指标以及后期预测数据的相对误差对预测模型精度进行比较,选出最优组合模型。结果表明:最优加权组合模型的评价指标、后期数据相对误差、模型拟合效果明显优于单项与简单平均模型;通过两级信息融合,构建了权重为(0.21,0.10,0.69)的TBM拆装装置作业时的最优加权组合预警模型GM-ARIMA-LSTM。可见创建的二级信息融合安全预警模型在TBM拆装装置作业时能有效判断装置的安全状态,对危险做出及时预警。
In order to make effective early warning for the safety of tunnel boring machine(TBM)disassembly and assembly device operation,an optimal safety early warning model was established by studying two-level information fusion to provide safety guarantee for the hoisting operation of TBM disassembly and assembly device.The first-level fusion fuses the data through the analytic hierarchy process-entropy weight method algorithm to obtain the safety state coefficient.Second level fusion establishes grey model(GM),autoregressive integrated moving average model(ARIMA),and long short term memory(LSTM)prediction models.The precision of the prediction model was compared by the four evaluation indicators,namely mean absolute error(MAE),mean absolute percentage error(MAPE),root mean square error(RMSE)and the relative error of the later prediction data,and the optimal combination model was selected.The research results show that the evaluation index,the relative error of the later data,and the model fitting effect of the optimal weighted combination model are significantly better than the single item and the simple average model.Through two-level information fusion,the optimal weighted combined early warning model GM-ARIMA-LSTM is constructed for the operation of TBM disassembly and assembly devices with weights of(0.21,0.10,0.69).It can be seen that the created two-level information fusion security early warning model can effectively judge the security status of the device when the TBM disassembles and assembles the device,and gives timely early warning to the danger.
作者
安小宇
王德健
李楠
李刚
时安琪
杨洋
陈傲松
AN Xiao-yu;WANG De-jian;LI Nan;LI Gang;SHI An-qi;YANG Yang;CHEN Ao-song(College of Electrical and Information Engineering,Zhengzhou University of Light Industry,Zhengzhou 450002,China;Shield Manufacturing Co.,Ltd.,China Railway Engineering Equipment Group,Zhengzhou 450016,China)
出处
《科学技术与工程》
北大核心
2023年第1期422-428,共7页
Science Technology and Engineering
基金
郑州市重大科技创新专项(2020CXZX0066)
河南省专业学位研究生精品教学案例项目(YJS2021AL024)。
关键词
安全预警模型
两级信息融合
安全状态系数
最优加权组合模型
security early warning model
two-level information fusion
security state coefficient
optimal weighted combination model
