In Pune, Larsen & Toubro (L&T) serves as the water operator tasked with enabling continuous, pressurised supply to approximately 300,000 consumers connected to approximately 1800 km under the Pune Municipal Corporation.
SmartTerra was deployed in approximately 150 km of distribution network serving 15,000 consumer connections across 12 DMAs. Each of these DMAs had inlet/outlet EMFs and network pressure sensors were deployed in a lift and shift manner. Consumer meters are AMR meters.
THE CHALLENGE
- Extremely high levels of non-revenue water, ranging from 50% to 75% in certain zones, with a mix of leaks, DMA breaches, non-working meters, unauthorised connections, etc.
- Intermittent supply conditions, with water available for only one to four hours per day. Manual leak pinpointing had to be done within these few hours during noisy day time conditions.
- Asset data quality was uneven and baseline information on losses was limited.
- Hydraulic modeling did not reflect real operating conditions.
- Database entry/integration problems meant that the SCADA and meter data from the MDMS was not balancing.
THE SOLUTION
SmartTerra deployed NetCity Localizer, a combined ML and Hydraulic modeling approach, to flag pipe segments with high probability of losses. NetCity combines consumer meter data with available DMA flow and pressure data to perform dynamic ML+hydraulic modeling tailored to intermittent supply patterns.
NetCity is coupled with a mobile-app based field workflows to guide targeted on-ground investigations using tools such as ground microphones, acoustic correlators and reference consumer meters. The mobile app has GIS‑enabled dashboard and workflows for real‑time data-access, feedback-capture and field decision‑making.
RESULTS
- Localised NRW issues to less than 30% of network length on average.
- NRW reduction by 43.2% points, volume amounting to 10.8 MLD, across 12 DMAs.
- Reduced NRW to less than 20% in 3 DMAs in the project timeline of ~1.5 years, at a per-DMA timeline of 4 months.
- Accuracy of localisation analysis at 60% – 85%.
- By narrowing the search area, field teams were able to deploy resources more efficiently, reducing time spent on broad, low-yield surveys.
