Water and wastewater utilities generate large volumes of operational data, including both real time and historical information. This data has the potential to significantly improve asset reliability, operational efficiency and decision-making. However, many utilities face the challenge of converting this data collected into meaningful insights. Analytics software, such as GE Vernova’s Proficy CSense, now enables engineers and operators to use industrial data directly, without requiring advanced coding or data science expertise. With this approach, utilities can identify operational issues, predict failures, and implement preventive actions, resulting in measurable improvements in reliability and performance.

The Challenge

A mid-sized water utility was facing recurring pump failures, which led to costly unplanned downtime and prolonged service interruptions. Although the utility had accumulated significant operational and maintenance data, engineers struggled to analyse this information effectively to detect early warning signs. The recurring issue was eventually traced to a small but critical bolt within the pump assembly, which was susceptible to corrosion. Due to its hidden position, the bolt was difficult to inspect visually, and its deterioration was typically discovered only after a major breakdown occurred. When the bolt corroded, it caused the impeller to become unstable, generating excess vibration that stressed the motor and coupling. Eventually, the bolt head separated, resulting in the impeller dropping out of its housing and leading to complete pump failure. This problem caused weeks of lost production, high maintenance costs, and inefficient resource utilisation due to emergency repairs.

The Solution and Result

To address these challenges, the utility implemented GE Vernova’s Proficy CSense analytics software. This platform provided a self-service environment that enabled engineers to analyse existing operational data without the need to write any code. By utilising data from the plant’s historian and maintenance records, the engineers trained an analytical model capable of detecting subtle changes in vibration patterns that indicated early-stage failures. They applied statistical methods, such as principal component analysis, to isolate the key variables affecting system performance while filtering out irrelevant data points.

The refined model continuously monitored vibration signals and identified patterns associated with early corrosion of critical bo

lts. This predictive capability provided up to sixteen days’ advance notice of potential pump failures, allowing maintenance to be scheduled proactively and avoiding costly downtime.

The implementation of Proficy CSense enabled the utility to shift from reactive to predictive maintenance, reducing downtime from several weeks to just one day. Maintenance teams were able to plan work more effectively, minimise disruption to the water supply, and extend the service life of critical equipment. The project demonstrated that predictive analytics can deliver substantial value using existing data without the need for complex programming or additional infrastructure. It also enhanced engineers’ confidence in applying analytics to asset management, enabling data-driven decisions that improved overall operational reliability.

For more information on how analytics can enhance operational efficiency and reliability in water and wastewater utilities, please contact IPD .