Estimation of TBM Penetration rate using Gradient Boosting-based Algorithms

Ebrahim Ghorbani, Amoussou Coffi Adoko, Saffet Yağız

2023

Full text Open access: Green Metadata (CSL-JSON)
DOI 10.21203/rs.3.rs-3228718/v1
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

Abstract TBM performance prediction from the rate of penetration (ROP) point of view has yet to draw a lot of attention since it is one of the main challenges for mechanized excavation with tunnel boring machines (TBMs). In this study, five algorithms, Gradient Boosting (GB), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), AdaBoost (AB), and CatBoost (CB) have been conducted to predict the ROP based on the Gradient Boosting theory. Six tunnel cases from different projects were examined to obtain the aim of the research. Dataset developed using those tunnel datasets includes Uniaxial compressive strength (UCS), Rock Type, Distance between Plane of Weakness (DPW), and TBM-related parameter of thrust force (TF). Mentioned Gradient Boosting algorithms were performed to obtain the most accurate results for the study. The developed models showed that XGBoost outperformed the other models, followed by the CatBoost model according to seven different evaluation metrics used to rank the models. After parameter tuning, the GB model outperformed others while those were not improved very much. By using the overall ranking according to the metrics and considering the parameter tuning time, XGBoost and CatBoost presented the first two best performances. Through SHAP values and dependency plots, the features and importance of the inputs showed the TF has the highest impact on the ROP, followed by UCS, Rock Type, and DPW. It is concluded that the XGBoost and CatBoost algorithms could be used for modeling to obtain the TBM penetration for similar rock types.