TBM Digital Twins in Tunneling: A New Era for Real-Time Performance Control

Mechanized tunneling has long relied on the collection of vast amounts of operational and geotechnical data, but until recently, this information remained fragmented—distributed across TBM logs, monitoring reports, and design models. The emergence of Digital Twins marks a shift from simple data acquisition to data intelligence, creating a real-time virtual counterpart of the tunnel that evolves in synchrony with the excavation process.

In this context, the digital twin integrates the TBM’s mechanical parameters, geological and geotechnical data, and instrumentation measurements into a continuously updated digital environment. This connection allows for a dynamic representation of the machine–ground interaction, where variations in penetration rate, torque, thrust, or face pressure are instantly reflected in the model. By coupling operational behaviour with analytical models, the digital twin enables engineers to quantify ground response, evaluate performance trends, and anticipate deviations before they affect production or stability.

For tunneling projects facing variable or complex geology, this real-time capability represents a paradigm shift. Instead of operating reactively—adjusting parameters after issues occur—teams can now base decisions on continuous, data-driven feedback, improving both the safety and efficiency of excavation operations.

Real-Time Performance Monitoring and Adaptive Modelling with DAARWIN

This new approach to digitalization in tunneling becomes operational through DAARWIN, a data-integration and analysis platform designed to create the digital twin of the TBM in real time. DAARWIN connects directly to the TBM’s data acquisition systems, capturing key operational variables such as thrust, torque, penetration rate, total advance, face pressure, and cutterhead rotation. All parameters are centralised in a single environment where they are continuously updated and time-synchronised with geotechnical and monitoring data.

Through AI-powered predictive models, DAARWIN estimates the expected penetration rate and torque under given ground conditions. These models are trained on historical datasets and continuously recalibrated during excavation, providing a predictive layer that helps detect anomalies, anticipate performance changes, and optimise operational parameters. Complementing this, the platform performs automated geotechnical backanalysis, comparing measured field behaviour against the results of numerical or analytical models. Ground parameters are adjusted dynamically at the end of each ring or shift, ensuring that the geotechnical model evolves consistently with the TBM’s actual performance.

This process closes the loop between design, monitoring, and operation, turning the digital twin into an intelligent system that not only visualises but also interprets excavation behaviour. It provides engineering teams with an immediate understanding of how the ground is responding, allowing early mitigation of risks such as excessive settlements, face instability, or unexpected pressure changes.

Towards Predictive and Self-Learning TBM Operations

The integration of real-time monitoring, predictive modelling, and continuous backanalysis sets the foundation for predictive and adaptive control in tunneling. With DAARWIN, engineers can analyse long-term performance trends through custom plots and multidimensional data visualisations, correlating TBM behaviour with geological transitions, tool wear, and advance rate variations. The platform’s incident logging system enables systematic classification of operational and geological events, creating a structured database that supports machine learning-based performance optimisation for future drives.

In practice, this means that every excavation becomes a source of cumulative knowledge: data from previous rings, sections, and even past projects can be reintroduced into predictive algorithms to improve the performance of subsequent drives. The digital twin thus evolves into a self-learning system, capable of reproducing and anticipating patterns that would otherwise remain hidden within isolated datasets.

As the industry moves toward fully digitalized project delivery, the implementation of tools like DAARWIN demonstrates that real-time digital twins are not only feasible but essential for high-performance tunneling. They enable a continuous feedback loop between the field and the model, transform monitoring into predictive insight, and open the path to autonomous and optimised TBM operation—a genuine step forward in the engineering of underground infrastructure.

Try DAARWIN TBM now https://www.saalg.com/tbm-data-management