Blog
Detecting energy theft or loss at the source: Trilliant’s Non-Technical Loss Analytics
By Greg Myers
Global Vice President, Product Management
Non-Technical Loss (NTL) is a critical and growing issue facing electricity distribution utilities around the world, and one we hear about from customers on an ongoing basis. It primarily results from energy fraud, theft, or misconfigured/malfunctioning meters.
As the cost of energy rises, along with the cost of living overall, individuals as well as businesses may tamper with their meter in an effort to reduce the amount they owe. In fact, a high percentage of NTL occurs at the point of service, where electricity users interfere with their meter to either partially or completely bypass it, preventing the meter from accurately measuring their energy consumption.
The impact of NTL on utilities is significant. A 2021 report indicated that globally it’s costing utilities more than US$100 billion in lost revenue, and that number continues to grow. However, NTL has even broader impacts, because it can also affect the reliability, security, and performance of the entire grid.
In order to resolve this issue, the end users causing the loss have to be identified with high accuracy. Traditionally, detecting this illegal activity has been either abstract or costly. You could assess your overall energy balances, as these indicate the asset as well as the date and time associated with high losses, but these don’t identify the specific meter or customer associated with the assets causing the loss. This issue is amplified as utilities need up to date distribution network mapping that is not always generally available. The distribution network is regularly altered in efforts to restore power to customers during outages as quickly as possible.
Workers could be sent out to manually inspect random meters, hoping to find the ones that are at fault or have been tampered with, but this is a very time-consuming (and costly) endeavor.
At Trilliant, we have a different approach. What if you could instead automatically catch the energy theft at the meter?
A unique solution
Non-Technical Loss Analytics from Trilliant is an Analytics-as-a-Service solution that identifies premises with a high probability of energy theft or loss at the source, throughout a utility’s entire service territory. It operates on the principle that tampered or misconfigured meters record anomalies in voltages and energy (power) readings. Data from smart meters provide the necessary information for the Non-Technical Loss Analytics engine to identify the location of each meter that’s exhibiting anomalous behavior, enabling utilities to visit the address(es) to verify and remediate the theft.
Here’s an overview of how it works:
- Service points (meters) are grouped into clusters
- An unsupervised AI learning algorithm and fuzzy logic determine expected behavior based on load profile patterns for each hour of the day and each day of the week
- Service points that display anomalies in voltages and energy use are detected and identified based on readings from the day before
- A theft- or loss-probability score is assigned to each anomalous service point
- The locations and other information about suspect service points are displayed on an easy-to-use dashboard
- Results are reported in Trilliant’s Smart Grid Manager, Microsoft Power BI, and other third-party applications
Non-Technical Loss Analytics from Trilliant helps to address critical issues:
- Utilities are able to recover the revenue they were losing to energy thieves or misconfigurations by detecting non-technical losses — at the source.
- Labor costs are reduced because technicians only need to be sent out to addresses that have shown a high probability of having compromised meters.
- An added benefit: Energy forecasting is improved with a more accurate picture of how much energy is being consumed, and how much needs to be produced.
If you’re ready to learn more, reach out to us at info@trilliant.com. We’d be happy to share a demo and show you first-hand how our powerful analytics solution can work for you.