Online real-time prediction of propulsion speed for EPB shield machine by SSA-GRU

Given the extremely complex working environment of the shield machine, precise control of the digging parameters is the guarantee for the shield operation's safety. Therefore, the paper presents a sparrow search algorithm-gate recurrent unit (SSA-GRU) based online prediction approach for shield machine propulsion speed. Firstly, the construction data are correlated based on the Pearson correlation coefficient, to obtain the boring parameters that are highly correlated with the propulsion speed and are considered as input variables for prediction model. Secondly, SSA is utilised to find the optimal hyperparameters of model. Finally, a prediction model is established based on optimal hyperparameters found by SSA, which more precisely exploits the nonlinear relationship from input features with propulsion speed, and accurately predicts propulsion speed. Simulation findings demonstrate that SSA-GRU model can precisely predict propulsion speed, and the prediction performance is superior to that of other models, effectively maintaining the stability of the excavation surface.