Prediction method for bad geological type of shield tunneling based on Xgboost

Abstract

The invention provides a prediction method for bad geological type of shield tunneling based on Xgboost, which is used for solving the technical problems of low prediction accuracy and low timelinessexisting in the prior art. The realization steps are as follows: preprocessing the PDV historical data of a shield tunneling machine; acquiring the key features of multiple pieces of tunneling parameter data after preprocessing; constructing a bad geological prediction data package; establishing a bad geological prediction model of an Xgboost algorithm; evaluating the bad geological prediction model of the Xgboost algorithm; and predicting the bad geological types in shield tunneling. The invention extracts the key data feature set of tunneling parameters which can characterize the stratum change through a random forest algorithm feature extraction model, and predicts The bad the bad geological types through the bad geological prediction model of the Xgboost algorithm, thus improving the accuracy and timeliness of bad geological prediction. The method can be used to monitor and analyze the geological conditions of surrounding rock in the shield tunneling process in real time.

Description

technical field [0001] The invention belongs to the technical field of industrial big data, and relates to a method for predicting unfavorable geological types of shield tunnel construction, in particular to a method for predicting unfavorable geological types of shield tunnel construction based on Xgboost, which can be used for real-time monitoring and monitoring during shield tunnel construction Analyze the geological conditions of the surrounding rock of the excavation face. Background technique [0002] With the rapid development of China's economy, a large number of people have poured into large and medium-sized cities, the urban planning area is limited, and the existing transportation conditions can no longer meet people's transportation needs. The acceleration of urbanization, the surge in the population of large cities and the congestion of urban road traffic have made urban rail transit increasingly highly valued by governments of all countries. Facing the increas...

AI Technical Summary

Problems solved by technology

Although the BP neural network algorithm used in this method has the ability of nonlinear mapping and good generalization ability, the BP neural network is dependent on the training samples and is easy to form local minima during the training process, and here The method uses mechanism analysis to extract key influencing factors, which may ignore other key influencing factors that can characterize formation identification, resulting in the model not being able to achieve the optimal accuracy when the sample sampling is not typical, and the convergence speed of the BP neural network algorithm is very slow. Slow, which makes the model not time-sensitive, and cannot satisfy the real-time prediction of the stratum

Images