Ageing infrastructure is becoming one of the sector's most critical issues. AI and Machine Learning can be powerful tools in the armoury of civil engineers working to address this growing crisis, and the technology's potential in the sector spans the entire asset life cycle. The challenge for civil engineers and asset owners is to identify the right use cases for AI and Machine Learning that enable it to deliver the greatest possible impact on large, complex infrastructure projects. In this blog, we will outline the different stages of a structure’s lifespan and how AI can add real value at each one.
The different phases of the asset life cycle can easily blur into one another, but we’ve presented them here as separate stages to contextualise them and present the possible use cases where AI can be leveraged. Nevertheless, this doesn’t necessarily mean that a use case should be limited to that phase, as there are numerous AI capabilities that can be translated across multiple applications and stages.
Planning the construction of any structure requires bringing in a wide variety of information. This step is fundamentally necessary to understand what can be built, where it can be built, and what constraints need to be considered, such as surveys, ground investigations, and impact and economic feasibility assessments. The challenge of bringing together and making sense of all this data from multiple sources with varying degrees of uncertainty can create significant pain points, even at this early stage. Consequently, most AI use cases at this stage focus on optimising information to enable the most informed planning decisions possible.
Due to this phase's iterative nature and its crossover with the design feasibility stage, this scenario is also relevant to the next stage of the asset life cycle.
Scenario:
A council is planning to build a new bridge. Before construction work can begin, a thorough environmental impact assessment needs to be carried out. This can be time-consuming, but Generative AI can help speed up the process by using text analysis and generation to retrieve relevant information from multiple sources and more efficiently produce the required documentation.
This stage brings together the work done by architects alongside structural, civil, mechanical, electrical, and geotechnical engineers. Multiple use cases can be explored for each of these professions, but in particular, exploring AI applications to enhance Building Information Modelling (BIM) can be significantly productive.
Scenario:
A structure has received planning permission, and technical designs are being finalised. However, different stakeholders need to contribute, data from numerous sources needs to be integrated, and multiple designs need to be considered and tested. Once construction begins, these designs may also need to be adjusted to accommodate unforeseen complications. AI-enabled BIM can streamline this process by checking designs and documentation for inconsistencies, highlighting potential risks, and therefore preventing delays before construction begins.
On-site accidents are one of the largest safety risks in the construction process. Consequently, applying AI to monitor construction site safety can be particularly valuable in reducing the likelihood of accidents. Unfortunately, construction sites can’t be as controlled an environment as a factory, so more variables need to be controlled to mitigate unforeseen events. AI can enable us to understand these variables and the relationships between them in real time, even as they become more complex. In fact, a joint study by Autodesk and FMI found that bad data alone may have cost the global construction industry up to $1.8 trillion in 2020, highlighting just how impactful smarter, real-time decision-making technologies like AI could be.
Scenario:
A large bridge is under construction, and the site is busy with workers and machinery. To monitor activity and maintain safety, CCTV cameras have been installed using AI that finds the optimal sensor placement for maximum coverage. Computer Vision can analyse these images in real time and detect risks that might lead to accidents, such as a construction worker entering an exclusion zone. This can help prevent accidents before they happen.
The operations stage accounts for up to 80% of total asset costs. The potential to leverage AI to lower these costs is, therefore, of significant interest to asset managers. For example, using AI in tandem with other technologies, like IoT sensors and robotics, can make operations processes more efficient and cost-effective. However, one of the main barriers to effective asset management is a lack of quality condition intelligence data. AI can help fill in the gaps and give managers a more comprehensive understanding of their assets to facilitate better, faster decision-making.
Scenario:
A council has identified several road sections under their jurisdiction that require repair. However, the council lacks the resources to repair all the roads at once, and doing so would cause significant disruption, so instead, they need to stagger the repairs. AI can be used to optimise the maintenance scheduling of this work to ensure that the council can repair all the roads in a way that minimises disruption to the public whilst operating within their budgetary and staffing constraints.
Deploying AI to ensure optimised asset maintenance is the principal way to gain value from the technology during this stage, as it enables managers to save time and resources by eliminating unnecessary inspections. As such, AI use cases for the maintenance stage predominantly revolve around understanding and quantifying the asset’s condition and deterioration, often alongside other technologies, to facilitate proactive rather than reactive maintenance. This approach will lower capital expenditure costs and increase the asset’s lifespan.
Scenario:
A bridge is being inspected. During its previous inspection two years ago, some spalling was identified, which has worsened in the intervening period. The original inspection report is stored in a database back at the office, so the inspector would usually have to take photos/notes and compare them to those from the previous inspection. Now they can use a mobile application that takes a photo of the spalling and, with Computer Vision, detects the spalling and quantifies the extent that the defect has changed over time to determine the rate of deterioration and prioritise the urgency of intervention.
In terms of AI use cases, the decommissioning stage is very similar to the construction stage due to similarities in the problems that need to be tackled, specifically questions around how a decommissioning project can be optimally managed while balancing resource, time, and financial constraints, and how the decommissioning of the asset can be made as safe as possible. Additional tasks specific to the decommissioning phase can be enhanced with AI as well, like managing waste disposal, creating an environmental impact assessment, and understanding what can be retrofitted versus what must be demolished.
Scenario:
A factory building has been decommissioned and scheduled for demolition. The company responsible for decommissioning the building wants to know how best to dispose of the materials and whether any of the construction materials can be reused to minimise unnecessary waste. AI can automate early-stage material analysis by using Computer Vision to identify components, linking them to historical building documentation, and applying multi-objective optimisation to balance competing factors such as cost, time, and hazard level.
There are an incredibly large number of AI use cases that can be applied throughout the asset life cycle. Nevertheless, figuring out the best use cases for your organisation will vary based on the number of assets, their type, and their condition, as well as financial constraints, what stage of the life cycle you focus on, data availability, digital maturity, and many other factors.
The first step to successful AI adoption is to reflect on what use cases would bring the most value. Start by implementing and testing smaller use cases as minimal viable products and developing from there. This will ensure value is delivered with AI transformation without requiring a huge budget. AI cannot replace the expertise of asset managers, but it can recreate it digitally to facilitate better, faster, more impactful decisions that will ensure our civil infrastructure lasts as long as we need it to.
Download our whitepaper to learn what civil engineers need to know about AI to build the future.