Predicting Failure Before It Happens

Our client is a regional water and wastewater utility operating more than 60 pumping and lift stations, with roughly 1,400 critical rotating assets — pumps, motors, valves and blowers — plus the field IT and OT devices that control them. Much of the equipment is remote, unstaffed and expected to run around the clock.

Maintenance was largely reactive. Crews were dispatched when an alarm tripped or a station went offline, often at night and at overtime rates, with spare parts sourced in a hurry. Asset records sat in a mix of spreadsheets, a legacy maintenance database and technicians' notebooks, so nobody could answer simple questions such as how old a pump was, how often it had failed, or what it had cost to keep running.

The cost of that approach is well documented. Siemens' True Cost of Downtime 2024 report estimated that the world's 500 largest companies lose about $1.4 trillion a year to unplanned downtime — around 11% of their revenue. The U.S. Department of Energy's Operations & Maintenance Best Practices Guide (prepared by PNNL for FEMP) notes that preventive maintenance can save 12–18% compared with purely reactive maintenance, and that a working predictive programme can save a further 8–12% on top of preventive alone.

Vyrazu Labs deployed Asseto to connect vibration, temperature and current sensors on critical equipment to a predictive maintenance engine, and to manage every asset — mechanical and IT — through one lifecycle record covering purchase, assignment, maintenance history, audits and retirement. Maintenance now happens when data says it is needed, not after something breaks.