With the advent of big data and the tools to interpret it, leaders have been given new ways to optimise their cities’ performance into the future
As recently as ten years ago, asset owners relied on a mix of historical and anecdotal evidence to inform the infrastructure planning process. Thankfully, with the rise of predictive analytics and other more advanced tools of the trade, asset management has come a long way since then.
Organisations have successfully connected common data streams to create an active feedback loop which continually informs them on asset performance and paths to optimise at the lowest possible cost. The embedding of Artificial Intelligence (AI) into that decision-making engine is going to advance these capabilities exponentially over the next few years — once we’ve filtered out some noise.
On what evidence did asset planners base their decisions before predictive analytics? Absent regulatory requirements, they would typically pull in condition assessments, roughly forecast risk and apply common principles such as balance and distribution in order to prioritize projects. The historical and socio-political data was the extent of the information planners could access, greatly limiting their ability to impact cities and systems in a sustainable way.
When failures happened, at a point in the life cycle when assets are most expensive to refurbish or replace, organisations had to scramble to locate resources for emergency repair. Sometimes this even required a diversion of funds. With only historical data to inform decisions, organisations were out of step with the dynamic nature of assets — each degrading at their unique rate according to predetermined degradation profiles and variable use.
With the advent of big data and the tools to interpret it, leaders have been given new ways to optimise their cities’ performance into the future. Funding shortfalls and in-house asset management expertise may not have changed much, but modeling multiple 20-year funding scenarios for a small city — something that used to take at least 48 hours — now takes less than two minutes.
For example, city planners can now say, “Show me what service levels will look like across the network in 15 years if we spend $10 million less here — are we still in business?” They can also ask questions such as, “If we increase taxes by .5 per cent, what is the balance of intervention, location, and timing that will yield the greatest ROI?”
With real-time, object data analysis in place, organisations have essentially begun to apply future versioning to our cities and systems in order to make better decisions in the present.Now with active feedback loops in place, asset-intensive organizations are growing collectively smarter and are empowered to continually optimize their infrastructure and investments. They have also laid the groundwork for the next logical layer of optimization: AI and Machine Learning Real-time data from networks across the asset register can quickly be read and interpreted by AI in order to ensure the most efficient use of a city’s resources over time.
Within the next few years, we will be able to read conditions faster and spin up thousands of models where we now spin up single to double digits at a time. We will be able to compile and interpret consumption patterns across massive data sets — even across multiple countries. Where organisations manually make decisions to optimise their programmes, AI will soon learn what criteria lead to optimal outcomes and apply them automatically. We are already starting to see organisations become more intuitive and adaptive to meet the rapidly changing needs of communities.
In Australia, for example, we’re applying AI in order to more accurately score our road conditions. In China, where complex road conditions present a significant safety issue, they are looking to AI in the form of training robots. The decision may be driven by an appetite for driverless cars, but it is also useful from an infrastructure planning perspective, making roads safer for vehicles with and without drivers. In Redwood, CA they are piloting multiple sensor-based technologies — all of these points of structured data, combined with other non-structured data (think text messages from train conductors) will simply contribute to the feedback loop, ripe for extremely valuable AI interpretation.
There are many other applications of AI and Machine Learning currently in use or being piloted, but AI is not yet a trustworthy last line.
We are able to embed AI along with existing data analysis tools in order to become much more informed about our resources and capabilities, but humans still provide a key intervention point in order to filter out good data from bad, and in order to spot game-changing linkages that defy logic. AI has not eclipsed the human brain in judgement just yet, but we are close. Together, humans and AI will build and maintain more agile, sustainable cities and systems.
Ashay Prabhu an Assetic co-founder with over 20 years’ experience in strategic asset management. He has led the development of condition algorithms, valuation profiles and prediction analytics, and is passionate about applying this science to close the global infrastructure renewal gap. Ashay has a Directorship at the Asia Pacific Institute of Asset Management, is an adjunct professor of Strategic Asset Management at Bond University, a Bachelor of Engineering (Hons), and is a chartered professional member of the Institution of Engineers Australia.
If you enjoyed this, you may wish to view the following:
Back to the future
How Assetic is helping to close the global infrastructure gap
AI provides insight at the edge on the roads
Echelon’s patent-pending technology uses artificial intelligence to enhance safety and reduce outdoor lighting costs