基于Transformer模型的电力隧道开挖工法预测研究

Prediction Research on Excavation Method for Power Tunnel by Transformer Model

在电力资源供应不足与绿色城市建设的双重影响下,电力隧道已经成为城镇化建设的必要选择,而根据已有工程条件确定电力隧道开挖工法更是施工稳定与取得经济效益的关键。以我国多个省份提供的电力隧道开挖工法应用情况为基础构建了数据集,应用Transformer模型对数据集之间的因果序列关系进行了学习与预测研究。研究结果表明,与传统机器学习模型相比,Transformer模型做到了不定项预测,同时在电力隧道开挖工法的最优项预测方面也取得了较好的结果,其准确率、精确率、召回率和F1值分别为98.25%、98.53%、98.45%和98.47%。Transformer模型在北京市某些已有工程实例中也得到了进一步的验证,具有极大的发展潜力。相关结论可为类似研究提供参考。

Under the influence of insufficient supply of power resources and green city construction,a power tunnel has become a necessary choice for urban construction.That determining the excavation method for a power tunnel by the existing engineering conditions is the key to construction stability and economic benefits.Based on the application of power tunnel excavation methods of several provinces in China,a data set is constructed.The causality sequence relationship between data sets are studied and predicted by the Transformer model.The research results show that compared with the traditional machine learning model,the Transformer model can achieve uncertain term prediction.Better results in the optimal term prediction is obtained for power tunnel excavation construction method,with the accuracy,precision,recall rate,and F1 values of 98.25%,98.53%,98.45%,and 98.47%respectively.Being further verified in some existing projects in Beijing,the Transformer model has great development potential.The relevant conclusions can provide a reference for similar research.