基于PSO-SVR的土压平衡盾构施工进度优化

针对隧道掘进机(TBM)利用率预测研究匮乏的问题，建立数据驱动的利用率预测模型并进一步对施工进度展开优化. 结合新加坡某地铁隧道项目数据，研究地质类型、司机操作与载荷对TBM利用率的影响，提出基于支持向量回归(SVR)的利用率预测方法，并以施工进度最大为目标展开操作参数优化. 利用SVR建立掘进环利用率与地质类型、载荷、操作参数的映射模型；建立以施工进度最大为目标，以地质类型、载荷、操作参数为约束边界的优化方程；利用粒子群优化(PSO)寻找特定地质类型下最优的操作参数. 结果表明：SVR模型在验证集和测试集上的R²分别为0.729和0.625，均优于多元线性回归、决策树、k最近邻、随机森林、AdaBoost和XGBoost模型；PSO能准确地找出最优的操作参数.

Advance rate optimization of earth pressure balance shield based on PSO-SVR

A data-driven utilization factor prediction model was established and the advance rate was further optimized, in view of the lack of research on the utilization factor prediction of tunnel boring machine (TBM). The effects of geological type, driver operation and loads on TBM utilization factor were studied on the basis of on-site data collected from a metro tunnel project in Singapore. A utilization factor prediction method based on support vector regression (SVR) was proposed, and the operational parameters optimization was further carried out with the goal of maximizing the advance rate. SVR was used to build the mapping model between the utilization factor of tunneling cycle and the geological type, loads and operational parameters. Then, optimization equations were established with the maximum advance rate as the objective and the geological type, loads and operational parameters as the constraints. Finally, particle swarm optimization (PSO) was applied to optimize the equations and find the optimal operational parameters under specific geological type. Results showed that R² of SVR model on the validation set and the test set were 0.729 and 0.625 respectively, which were better than that of multiple linear regression, decision tree, k-nearest neighbors, random forest, AdaBoost and XGBoost models. Moreover, PSO can accurately find out the optimal operational parameters.