Solutions / ELMS — AI-powered loss detection
Energy Loss Management System.
Electricity theft costs utilities globally an estimated USD 89.3 billion annually. South African municipalities face significant non-technical losses that erode revenue and infrastructure capacity. Detect technical losses and non-technical losses — theft, tampering and illegal connections — with deep-learning models trained on live metering data.
- Theft and tamper. Deep-learning detection of non-technical losses on live metering data.
- Audit trail. Automated alerts municipalities can act on.
- Cloud-based. Adapts to new theft patterns without CCTV or manual sweeps.
Capabilities
Stop leakage with an audit trail.
Technical loss (TL) detection: line faults and equipment inefficiencies
Non-technical loss (NTL) detection: power theft, meter tampering and illegal connections
Pattern analysis and AI model training on metering data from smart meters and TMDs
Automated alerts with a clear audit trail the utility can use for investigation
How it works
01
AMI hardware streams interval data into ELMS.
02
Deep-learning models — an IP-patented methodology — identify distribution-network irregularities.
03
Unlike CCTV or manual checks, ELMS is data-driven, reflects new theft patterns dynamically, and runs as a cost-effective cloud service.
What an engagement covers
- Loss-data diagnostic or scoped ELMS pilot
- Prioritise anomalous circuits for investigation; enforcement and recovery stay with the utility
Platform: NuriFlex hardware and software, delivered and supported locally by Joule.
How Joule delivers it.
All referencesAI
Deep-learning detection of theft, tampering and technical losses on live metering data.
Cloud
A cost-effective service that adapts to new theft patterns without CCTV or manual sweeps.