Change Leader: Moving Asset Management in a Better Direction
Infrastructure is an organism; an ever-changing amalgam of past, present and future. For most, the task of developing, delivering and maintaining the built environment relies on logical inference rather than data-driven determinations drawn from the interconnected whole. Didem Cataloglu, CEO of DIREXYON Technologies, takes aim at that distinction as the company she leads dives into the challenge of asset management for large-scale infrastructure.
“The common denominator among our verticals is each owns and maintains a large volume of assets to be operated for a very long time,” says Cataloglu. DIREXYON’s asset-management software applies advanced financial modeling to streamline processes, optimize spending and actualize a more-sustainable future. “Generally, utility providers, transportation providers and governments already have data about what it is, what it costs, how it performs, how it is maintained and when it is likely to expire. What they lack is a way to leverage it all into a synchronized set of facts that helps them make decisions.”
Today’s infrastructure owners include municipalities, counties, states and federal agencies as well as for-profit utilities in water, power, telecom and transportation. Many such entities have been diligently infusing technological innovations into their assets wherever possible for decades. Today, sensors draw real-time data from assets. Yet synthesizing all these data into actionable intelligence remains challenging for a host of reasons. Budget, political and public pressures are among the many arbitrary influences that shape inference-based decision-making. Influences that can be made irrelevant when the power of artificial intelligence and machine learning are brought to bear on computational equations to derive data-driven determinations.
Modeling Technologies
DIREXYON’s platform is based on stochastic-combinatorial modeling, which allows user-configurable asset investment planning and portfolio-management solutions. To break that down, “stochastic” means to have a random probability distribution pattern that may be analyzed statistically but may not be predicted precisely. And “combinatorial” means relating to the selection of a given number of elements from a larger number without regard for their arrangement. Through combinatorial optimization, an infrastructure owner’s asset management can be modeled to maximize investment value while satisfying all constraints placed on the process.
“The first thing we do is create a digital twin of the client’s assets,” shares Cataloglu. The digital twin is a virtual model of the owner’s infrastructure inventory. Every data point on every piece of equipment from the date of purchase and installation to dates and descriptions of routine maintenance to repairs is used to establish a baseline. The process also appreciates that despite technological advances being infused through time, the operation and maintenance of assets is managed by human hands.
The platform allows users to modify their inventory as changes occur as well as factor in fuel prices, commodities, labor markets and supply chain uncertainties. Through machine learning, the model can adapt to situational circumstances without having to follow explicit instructions. Using artificial intelligence, a system’s many disparate elements are assembled as labeled data that are analyzed for correlations and patterns used to make predictions about future states.
Tuned to Clients
“The City of Montreal was spending millions annually on maintaining water, wastewater, and roads and sidewalks, yet much of their asset portfolio still suffered from chronic underinvestment and maintenance deficits,” adds Cataloglu. “By modeling its infrastructure in our platform and taking a proactive approach to infrastructure interventions, Montreal achieved a 30-percent reduction in annual capital and operating expenses.”
Montreal’s infrastructure consists of 4,000 kilometers of water mains, 4,800 kilometers of sewer mains and 4,000 kilometers of roads that serve approximately 1.6 million people. Due to their advanced age, the city intended to replace most of its underground networks of water and sewer pipes. Likewise, the area’s harsh freeze-thaw cycle combined with advanced subsoil deterioration was contributing to a sharp increase in the number of potholes. Most of the city’s leadership believed that rather than trying to rehabilitate their infrastructure, complete reconstruction was the best option.
The DIREXYON Suite has to be hyper-tuned to the particularities of each owner’s universe of assets and those responsible for maintaining the datasets. In Montreal, that meant opening roadways to replace pipes factored in rehabilitating that segment of road in the same process. The criticality of each network component was modeled for age, materials, dimensions, failure history and other variables to prioritize interventions in the most-crucial segments while managing risks related to the number of assets in poor condition.
“Beyond the dollars and cents of this, there is confidence,” says Cataloglu. “Through machine learning, we can anticipate the likelihood of an asset failing by detecting patterns that have occurred historically. This allows the owner to understand the evolution of an asset into the future under existing conditions. When we apply constraints, like the year-over-year level of investment, the program can simulate the infrastructure’s risk profile.”
About Sean O'Keefe
Sean O'Keefe is an architecture and construction writer who crafts stories and content based on 20 years of experience and a keen interest in the people who make projects happen; email: [email protected].
