Optimal Control of Operation Parameters During EPB Shield Tunnelling Based on Artificial Neural Network Model

Abstract

Shield machine is a widely utilized excavation equipment in the tunnel construction process, especially in the complex geological environment. Optimal control on the operation parameters is vital to maintain the quality of the tunnel construction and to avoid the serious sinkhole hazards. This study therefore established an artificial neural network (ANN) model to predict the total thrust force and the advance speed of the earth pressure balance (EPB) shield machine. A set of data from the seventh section of Wuhan metro line six in China was utilized to train the ANN model. Then, the parameter filtering procedure was performed to identify the key factors controlling the tunnelling performance and hence to improve the prediction accuracy. Finally, the established ANN model was adopted to predict the operation parameters of shield machine (i.e., total thrust force and advance speed). The good agreement between the predicted and measured data demonstrates the precision and advantage of this ANN model, which poses a guiding significance in dynamically identifying the environmental conditions and adjusting the operating parameters of an EPB shield machine during tunnelling.