Digitizing Tunnelling and Geotechnical Engineering with Rajat Gangrade

We recently had a profound conversation with Rajat Gangrade, PhD, PE, Technical Advisor at HNTB and Co-founder of GroundUp on the Digital Construction Podcast. Rajat, with extensive experience from Arup and AECOM, shared his vision for a digital-first approach in tunnelling and geotechnics, highlighting critical advancements, challenges, and opportunities across the project lifecycle, from planning to operations.

The Digital-First Project Lifecycle

Rajat outlined how digitisation transforms every project phase. The detailed design phase focuses on integrating GIS and 3D geological models to create a "single source of truth" or common data environment. The most exciting transformation occurs during construction, where sensors capture real-time data from Tunnel Boring Machines (TBMs) and monitoring instruments, all integrated into a digital twin via cloud-based solutions. Finally, for operation and maintenance, digital archives support predictive maintenance, incident logging, and automated inspections.

Navigating Digital Maturity and Data Challenges

While digital adoption varies globally, Rajat noted that North America's digital maturity in tunnelling is surprisingly low, despite growing awareness. Key barriers include a conservative industry culture, inertia, and uncertainty about the cost-benefit relationship of new technologies, yet this also presents a huge opportunity for innovation. A significant challenge in geotechnics is the underutilisation of site investigation data. Rajat stressed the need for smarter geotechnical database management systems to overcome fragmented data and version control issues, advocating for integrated digital delivery that moves beyond static PDFs to interactive visualisations.

Integrating Uncertainty into Design

Geotechnical engineering inherently deals with uncertainty, especially given the inability to fully excavate ground before design. To address this, Rajat's PhD focused on probabilistic methods. Beyond simply changing single parameters, modern approaches leverage simulations and sophisticated stochastic algorithms to quantify variability and integrate it into design procedures. Critically, communicating this uncertainty to project stakeholders is best achieved through compelling visualisations, such as heat maps of risk along a tunnel alignment, confidence intervals, and probability curves, making complex data intuitively understandable.

Automation: Low-Hanging Fruit with High Hurdles

Automation is identified as a "low-hanging fruit" offering immense opportunities for efficiency in tunnelling. Examples include automating routine design calculations through robust design libraries and enabling software interoperability. Rajat highlighted how scripting with Python can drive numerical modelling tools to run numerous cases and quantify uncertainty. However, adopting automation faces significant limitations: high initial capital investment, tight project deadlines, and the industry's ingrained inertia and conservatism. Prerequisites for successful automation include comprehensive digitisation of data, standardised processes, and a skilled, multidisciplinary workforce capable of integrating diverse digital aspects.

AI: Not Yet "Ready" but Agentic AI is the Future

Despite the current "AI era," Rajat believes the construction industry, particularly in North America, is "not AI ready" for complex tasks. This is due to a lack of standardised guidelines, unproven cost-benefit relationships, and significant inertia. He suggests that AI should initially focus on automating routine tasks to build confidence. Given the unique nature of geotechnical and tunnelling projects, which differ greatly from one another, "Agentic AI" is seen as the future. This form of AI is capable of autonomously learning and adapting to changing project environments, machinery, and design approaches. A critical enabler for this will be cross-industry data sharing, as "if there is no data, there is no AI".

The Vision for the Next Decade

Rajat is highly optimistic about the next 5-10 years. He is particularly excited about the rise of Virtual Design and Construction (VDC), a project management methodology utilising BIM and digital representations to plan, design, and manage projects more effectively. This VDC approach will create "digital archives" of projects, fostering better visualisation, analysis, and optimisation, moving away from siloed traditional methods. The industry will see increased adoption of digital twins and real-time integration of data for monitoring and decision-making. Crucially, cloud-based collaboration will become a standard, streamlining data integration from disparate sources, supported by a significantly more digitally literate incoming workforce.

Key Takeaways:

• Digitalisation offers opportunities across all project phases – from design to operations and maintenance;
• North America's digital maturity in tunnelling is low, presenting a canvas for innovation;
• Underutilised geotechnical site investigation data is a major cause of project failures; smarter database management is crucial;
• Uncertainty in geotechnics can be quantified and communicated effectively using visualisation tools;
• Automation is a "low-hanging fruit" for efficiency, but adoption is hindered by cost, inertia;
• The industry is "not AI ready" for complex tasks; focus should first be on automating routine tasks, with Agentic AI as the long-term goal.