Predictive maintenance
September 3, 2021
2 months

EMBRACING DATA SCIENCES TO BETTER PRESERVE AND REPAIR STRUCTURES

Bridges are essential links in road networks. Any undetected structural weakness can quickly become a threat to their operation and the safety of road users. The ageing of a large stock of concrete structures around the world, subject to aggressive environments and ever-increasing traffic loads, requires adopting a different approach to better maintain and preserve them.

The construction industry has adopted later than other industries the latest technological trends, but it is now fully committed to succeed in its digital transformation and collecting the benefits of new technologies for the business. Thus, the use of the Internet of Things (IoT) emerges as an innovative method for the maintenance of transport infrastructures through the installation of multiple sensors collecting in real-time a large amount of data regarding their structural health.

As this technology is less applicable to existing bridges where the starting condition is often not known, VSL has decided to explore another path using big data concepts alongside machine learning and artificial intelligence (AI). The critical and informed insight they can provide will allow to enhance the quality of condition assessments, increase warning times in case of major deterioration and ultimately help our clients to minimize maintenance costs while maximizing availability of the traffic network.

Under Bouygues Construction’s innovation programme, VSL has teamed up with the Chair in Construction 4.0 at Centrale Lille in France to research the use of these disruptive technologies with a particular focus on prestressed concrete structures, making use of existing consolidated data sources.

The objective is to develop a tool that can forecast the future evolution of the structure’s condition and hence help to guide decision making on predictive maintenance. A team of 4 researchers is working for a period of 2 years on the project working in close cooperation with VSL’s specialists.

Predictive maintenance: the value of data

The potential of big data is undeniable: it provides companies with a method to collect, analyze, and apply vast amounts of information to help solve problems and create added value for their clients. In the construction sector, it poses at the same time a significant challenge when it comes to collecting, sharing, structuring and using the data generated across the long life of bridges.

“In the coming years, one of the biggest challenges of the construction business will be how to create true value from data.” Andreas Schwarz, VSL Repair & Preservation Manager.

As a specialist of structural systems and components, VSL has collected a lot of data over the years and the research team can leverage both internal and external data sources to build predictive models and improve the power and accuracy of their predictions.

The first step of the work consists of using artificial intelligence to detect deterioration patterns in consolidated bridge data collected from inspection reports, photos, non-destructive testing (NDT) results and general information.

The second part will lead to investigate and work on algorithms that can predict the evolution of these patterns over time to serve as a key differentiator in the repairs and preservation market.

“The application of IA to predict the sustainability of bridges is really important when the experience of the human knowledge in theory and in practice is taken into account.” Professor Zoubeir Lafhaj, Chairholder of the Chair in Construction 4.0 at Centrale Lille.