Precision versus intelligence: Autonomous supporting pressure balance control for slurry shield tunnel boring machines

• Published in: Automation in Construction Vol. 114; p. 103173
• Main Authors: Zhang, Yakun, Gong, Guofang, Yang, Huayong, Li, Wenjing, Liu, Jian
• Format: Journal Article
• Language: English; Japanese
• Published: Amsterdam Elsevier B.V 01.06.2020; Elsevier BV
• Subjects: Artificial intelligence; Artificial neural networks; Autonomous control; Boring machines; Control systems; Controllability; Controllers; Cyber-physical system; Deep learning; Dynamic models; Finite state machines; Intelligent TBMs; Model predictive control; Neural networks; Operators; Predictive control; Singular value decomposition; Slurries; Slurry shield machine; Stability; Tunnel boring machine; Tunnel construction; Tunneling shields; Autonomous control; Deep learning; Model predictive control; Tunnel boring machine; Cyber-physical system; Slurry shield machine; Intelligent TBMs; Artificial intelligence
• ISSN: 0926-5805; 1872-7891
• DOI: 10.1016/j.autcon.2020.103173

This paper presents a method for the autonomous control of supporting pressure balance for a slurry shield tunnel boring machine. The mechanism of multi-system coupling interactions of the slurry supporting process is revealed by establishing the dynamic model of the process. Furthermore, the degree of controllability of the manipulated inputs is analyzed and verified theoretically using singular value decomposition. Based on the analysis of the supporting process dynamics, a cyber-physical system (CPS)-based hierarchical autonomous control scheme is proposed. The execution level digital optimal controllers are designed and auto-tuned. The discrete event-driven control logic is also included in the execution level and modeled as a finite state machine. For comparison purpose, the coordination level controller is implemented using a hybrid switched model predictive controller and a deep neural network, respectively. Various artificial neural networks with different hyper-parameters are trained and compared using big data. The performance of the proposed autonomous control methodology is tested and compared with human operators by using randomly extracted construction field data. The test results show that the autonomous control system with switched model predictive controller outperforms that with the deep neural network and human operators. The results validate the feasibility and effectiveness of the proposed autonomous control methodology. •The mechanism of multi-system coupling interactions of the slurry supporting process is revealed.•A cyber-physical system-based hierarchical autonomous supporting pressure balance control scheme is proposed and validated.•Advanced control-based and artificial intelligence-based solutions for the coordination level are developed and compared.