Abstract
The advance rate (AR) is a significant parameter in shield tunneling construction, which has a major impact on construction efficiency. From a practical perspective, it’s helpful to establish a predictive model of the AR, which takes into account the instantaneous parameters as well as the past operations. However, for shield tunneling in mixed ground conditions, most researches focused on the average values of AR per ring and neglect the influence of past operations. This article presents a long short-term memory (LSTM) recurrent neural network model, which was developed for the slurry shield tunneling in a mixed ground of round gravel and mudstone in Nanning metro. A temporal aggregated random forest is employed to rank the importance of the explanatory features. The model performances in different ground conditions are investigated. The results show that the LSTM model can be effectively implemented for the AR prediction. A high correlation is observed between predicted and measured AR with a correlation coefficient ( $R^{2}$ ) of 0.93. The LSTM based AR predictive model is compared with the random forest (RF) model, the deep feedforward network (DFN) model, and the support vector regression (SVR) model. The comparison shows that the LSTM model has the best performances compared to other models. With one-fourth features, we can achieve a 95% prediction accuracy measured by the $R^{2}$ in the proposed model.
Highlights
Owing to the growing demand for efficient inter/intra-city transportation in China, shield tunneling in mixed ground conditions has been becoming increasingly common [1], [2]
Motivated by recent advances in deep learning methods and their performance in time series prediction, we presented an long short-term memory (LSTM) based deep learning model for the advance rate (AR) prediction on large data set of a tunnel section in the mixed ground condition
Inspired by the feature selection approach employed in the spatio-temporal deep learning model for the passenger demand prediction [55], we use the temporal aggregated random forest (RF) based feature evaluation approach to indirectly verify the importance of the input programmable logic controller (PLC) parameters on AR prediction
Summary
Owing to the growing demand for efficient inter/intra-city transportation in China, shield tunneling in mixed ground conditions has been becoming increasingly common [1], [2]. Elbaz et al employed ANFIS with genetic algorithm (GA) [22] and improved PSO [23] for the AR prediction of the shield in the tunnel section of Guangzhou Metro Line 9 in China, which performed well with R2 = 0.88 in small data set (200 samples in totals). LSTM based deep learning methods can obtain good results in time series prediction, the implementation of these methods requires a sufficient set of past data. Few kinds of research conducted numerical experiments on large scale data set when they employed the deep learning methods in shield tunneling parameters prediction. Motivated by recent advances in deep learning methods and their performance in time series prediction, we presented an LSTM based deep learning model for the AR prediction on large data set of a tunnel section in the mixed ground condition. The maximum rotation speed of the cutterhead is 3 rpm, and the maximum AR of the shield machine is 50 mm/min
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call