Predictive Analytics

Driven by AI technology, modern software provides an advance warning of imminent failures

- JONATHAN WOOD

Cities are struggling to keep up with the increasing number of infrastructure repairs needed today. In many parts of the world, aging structures, patched transit systems and outdated utility networks are among the assets that require attention. “State of Good Repair” is the goal for all but municipalities are often forced to prioritise and focus on the most critical assets first.

But how can engineers and officials predict which physical assets are most likely to fail — or when? Where will be the greatest safety risks or the most disruptive occurrences that impact commerce and constituents?

Fortunately, modern software solutions can help. Artificial intelligence (AI) and machine learning (ML) software can predict likely occurrences, based on data patterns and driving influences. Reporting tools help identify early warning signs and anticipate probable outcomes, while there is still time to take proactive measures and intervene.

Scheduling preventive maintenance or replacing aging parts before they fail can prevent high-impact disruption.

Predictive science, not science fiction

Predictive science is a capability that today’s advanced enterprise asset management (EAM) solutions provide. Data from inspections and collected through embedded sensors and IoT can be analysed to find anomalies or data points that fall outside of predefined acceptable limits.

These early warning signs can point to future breaches or failures. The outlying data can trigger automatic responses such as shutting down equipment that may be unsafe. Early intervention, like dispatching a technician, ordering back-up parts or rerouting activities, can change the likely outcome.