Multi-objective optimization of shield operational parameters to enhance shield performance for surface settlement control

Ground surface settlement is the main problem in tunnel construction process, which is influenced by tunnel geometries, uncertainty geological condition and shield operational control parameters. However, adjusting the shield operational control parameters to reduce the ground surface settlement may
resulting in a reduction of penetration rate, which led to an increasing of time and budget for construction. Therefore, this study aims to identify optimal shield operational parameters within the framework of multi-objective optimization through the application of non-dominated sorting genetic algorithms (NSGAII), simultaneously minimizing ground surface settlement and maximizing penetration rate. The long-short term memory (LSTM) is employed for time series prediction of penetration rate and ground surface settlement for using as fitness function in NSGAII. And, the LSTM models have RMSE equal to 3.585 mm/min and 6.696 mm for penetration rate and ground surface settlement prediction respectively.