基于多头自注意力-Bi-LSTM模型的盾构掘进引发的土体沉降预测

杨明辉,宋牧原,付大喜,郭炎伟,卢贤锥,张文聪,郑伟龙

Prediction of shield tunneling-induced soil settlement based on multi-head self-attention-Bi-LSTM model

Minghui YANG,Muyuan SONG,Daxi FU,Yanwei GUO,Xianzhui LU,Wencong ZHANG,Weilong ZHENG

表 4 不同深度学习模型在不同监测点的土体沉降预测泰勒图参数值（案例1）

Tab.4 Taylor diagram metrics of soil settlement predictions by different deep learning models at various monitoring points

监测点 模型 SD/mm R CRMSD/mm 预测 真实 Y3 1 0.1858 0.1817 0.9481 0.0593 2 0.1915 0.1817 0.8935 0.0866 3 0.1976 0.1817 0.7639 0.1311 Y4 1 0.1751 0.1700 0.9590 0.0497 2 0.1031 0.1700 0.9129 0.0868 3 0.0576 0.1700 0.8100 0.1279 Y5 1 0.0571 0.0584 0.9582 0.0168 2 0.0406 0.0584 0.7815 0.0368 3 0.1747 0.0584 0.7985 0.1327