Optimal pressure control for segment assembly in synchronous tunnel boring machines considering time delay

• An S-TBM experimental platform is designed to analyze the status of the hydraulic cylinders. • A BO-KAN deep learning model is developed to improve pressure prediction accuracy. • An optimization framework is proposed to redistribute hydraulic cylinder pressure. • A GDE3 optimization model is employed to balance response time and control precision. • This approach can improve the response of segment assembly and reduce stress concentration. To address the problems of stress concentration and response delay during the segment assembly process in synchronous tunnel boring machines (S-TBMs), this paper proposes an optimization framework for redistributing the assembly hydraulic cylinder pressure. Bayesian optimization Kolmogorov-Arnold networks (BO-KAN) is employed to predict the cylinders’ pressure, while generalized differential evolution 3 (GDE3) is applied for pressure redistribution of the cylinders. The effectiveness of the proposed method is verified through an S-TBM experimental platform. The results indicate that: 1) BO-KAN accurately predicts the pressures of 16 hydraulic cylinders, achieving a high coefficient of determination (R 2 ) values with an average of 0.998. 2) BO-KAN performs superior predictive capability against other models, achieving an R 2 score increase of 0.100 and a composite improvement rate of 1.012 %. 3) GDE3 exhibits superior efficiency during the 1 s delay effect scenario, attaining a 46.955 bar max difference of pressure reduction and a 28.117 % optimization enhancement. 4) GDE3 demonstrates a significant effect in pressure redistribution across different engineering conditions, with the average optimization degrees for the max pressure of the optimal pressure being 11.465 %. The innovation of this study lies in developing a pressure redistribution framework through the integration of BO-KAN and GDE3 taking into account time delay, which significantly mitigates pressure concentration problems and response hysteresis phenomena in S-TBM construction processes.