A Quantitative Interpretation Method and Equipment for Tunnel Advance Drilling Based on rs-xgboost

Abstract

The invention relates to the field of tunnel engineering, in particular to an RS-XGBoost-based quantitative interpretation method and equipment for advanced drilling of tunnels. The invention randomly samples the tunnel to be excavated, obtains the drilling data of the tunnel to be excavated and performs preliminary processing, then inputs it into a pre-built RS-XGBoost model for quantitative interpretation, and outputs the quantitative interpretation result. By combining the powerful nonlinear data analysis performance of the XGBoost machine learning model and the efficient hyperparameter optimization capability of RS random search, the difficulty of model establishment is greatly reduced while ensuring the accuracy of identification and classification of unfavorable geological bodies in tunnels. It avoids the adverse effects of manual parameter adjustment; at the same time, it proposes to use the types of unfavorable geological bodies as the interpretation results of the machine learning model, and to use several types of unfavorable geological bodies that are more common and more harmful to tunnels as the quantitative intelligent interpretation results, and according to Interpret the results and adjust the excavation method and support measures in time to guide the tunnel construction on site.

Description

technical field [0001] The invention relates to the field of tunnel engineering, in particular to an RS-XGBoost-based quantitative interpretation method and equipment for tunnel advanced drilling. Background technique [0002] Since entering the 21st century, with the rapid development of my country's transportation industry, the construction scale of highway tunnels has also become increasingly large. According to statistics, by the end of 2020, there were 21,316 highway tunnels across the country with a length of 21,999,300 meters, including 1,394 extra-long tunnels with a length of 6,235,500 meters and 5,541 long tunnels with a length of 9,633,200 meters. fastest growing country. In the process of the overall construction of the tunnel gradually changing to the direction of large buried depth and long tunnel line, the characteristics of remote site selection, high stress, strong karst, high water pressure, and complex structure have gradually become prominent. Various un...

AI Technical Summary

Problems solved by technology

[0009] However, there are still two problems in the existing research: First, most of the prediction results are based on the grade of the surrounding rock or the property of the stratum, because the grade of the surrounding rock commonly used at present is a relatively large concept, and different unfavorable geological bodies and different stratum lithologies They may all be of the same grade of surrounding rock, and the grade of surrounding rock is basically determined at the tunnel design stage, which often results in inconsistencies with the actual excavation conditions on site. Although the prediction accuracy is high, it has limited guiding significance for tunnel construction; The traditional machine learning model has few hyperparameters, and manual parameter tuning can basically meet the needs, but the accuracy rate is low; the existing XGBoost model has excellent theoretical performance and high accuracy rate, but there are many hyperparameters that need to be adjusted, and manual parameter tuning cannot give full play model performance

Images