Analytics is key to the reduction of losses and revenue assurance for utilities.
In the changing energy market it is becoming ever more important for utilities to ensure that as much as possible of the commodity they produce reaches end users and that as much as possible of that supply is also billed and paid for by them.
There are plenty of possibilities for losses – in the transmission and distribution systems, through theft or non-payment, all adding up ultimately to an impact on the bottom line.
Traditionally it is assumed that losses are a phenomenon primarily of concern to developing countries where infrastructure is old and levels of theft are high. It is true that overall losses, both technical and non-technical, are often the highest in these countries. For example, in sub-Saharan Africa a recent World Bank report highlights losses that range from as high as 48% (in the Central African Republic) with the weighted average for the region (excluding South Africa) coming in at 23% (or 15% including South Africa).
Similar high levels are also seen in parts of India for example.
But ultimately the issue is a global one. For example, in the US losses are estimated by the EIA to be between 5-6%. In the UK, the estimated losses by Ofgem are slightly higher at almost 8%, close to the global average.
Even at these levels, the costs can be significant, adding up to billions of dollars in lost revenue and ultimately higher prices charged to the paying customers. For example, in the US an estimated US$6 billion is stolen annually. In Africa, the T&D losses represent more than 1% of GDP in five countries (Cape Verde, Comoros, The Gambia, São Tomé and Príncipe, Togo), while in three of these bill collection inefficiencies also represent more than 1% of GDP.
Worldwide the loss through theft to electric utilities is estimated at US$85 billion annually. And almost two-thirds of this theft is committed by people who can afford to pay for the service while little more than a third is by the poor or indigent.
The main avenue for reducing losses is through reduction of the non-technical component, i.e. theft and non-collection. Traditionally detection of theft has relied on visual inspection in homes and businesses by utility personnel. But this has now started to change with the enhanced sensing available in the smart grid and especially the application of analytics to this data – besides that the number of utility visits to customers has dropped substantially with the advent of smart metering.
At the basic level, fraud can be detected through billing anomalies or energy balance calculations on feeders, which coupled with other data can narrow down the potential location. But as the techniques for fraud have become more sophisticated so too have had to become the techniques for detecting it.
“We see ourselves as the James Bond of revenue assurance,” says Denis Maia, CEO of the Rio de Janeiro-based Choice Holdings, which has developed machine learning based algorithms, or predictive analytics, for theft detection.
“We have to detect the fraud but the bigger challenge is that the techniques being used are continually becoming more sophisticated. It’s no longer a case of simply bypassing a meter and there are tens if not hundreds of ways of meter tampering. And we even know of companies who hire engineers to come up with new frauds.”
In many ways, the issue of fraud detection is akin to that of cybersecurity, with the difference that in the case of fraud the instance is more likely to be restricted to a single customer account whereas a cyber breach is more likely to be more widespread.
A growing new breed of cybersecurity solutions aim to detect subtle anomalies in grid data. Likewise Choice has evolved a similar approach for its revenue assurance solutions. These are based on utilizing ‘intelligence’ – hence the James Bond association – from a wide variety of sources both within and outside the utility. They are applied on an account by account basis to give a probability of fraud for each customer as well as the potential financial return for them.
“Information is the most important asset a utility has and the aim is to use this to maximize the amount of revenue the utility can recover,” Maia says, noting that the procedure is identical for all utilities, whether electric, gas or water.
For the 27 electric, gas and water utilities representing more than 100 million customers using Choice's Revenue Intelligence solution, he points to an average recovery of US$40 million per year in lost revenues.
Like cyber threats, the risk of fraud is one that isn’t going to go away and utilities need constantly to be on the alert. And clearly the returns will be proportionate to the investment, both financially and in terms of effort, in the solution.
“We know all utilities are striving to limit fraud, but it’s a question of the maturity of the solution,” says Maia. “It’s an endless game we’re playing – just like the never ending need for intelligence agents!”
For more insight, view Engerati’s interview with Denis Maia at European Utility Week 2016.