The demand-side revolution has begun. Aired by moderator Aaron Leopold, this was the key message at a fast-paced talk entitled “Building resilient incomes to ensure sustainable business models” on the second day of Digital Africa Utility Week and POWERGEN Africa on Wednesday.
Leopold is the CEO of the Africa Mini-grid Developers Association (AMDA), which represents the decentralised utility sector serving remote populations in Africa.
Speaker John Kidenda, software and analytics director at micro-grid developer PowerGen Renewable Energy, based in Nairobi, Kenya, said that trends are shifting and that funders need to engage with communities to better gauge demand.
“The volume of funding is increasing day by day,” said Kidenda. “Where you find a challenge though, is that oftentimes the funding is not as demand-driven as it needs to be. Businesses come with ideas about what they should be funding. They tell you: ‘Here’s $100,000 or $500,000 to build water pumps in these rural villages.’ And oftentimes, the conversation around whether that’s actually what is needed in these villages, and whether setting up a water business in this location is a viable business model to pursue, has not been fleshed out.”
Demand-side economics holds that economic growth and employment are best achieved through a high demand for products and services. Demand-side management entails actively influencing electricity demand on mini-grids so it matches electricity generation.
Kidenda continued: “So I think what I would say is, as the volume of financing increases, what needs to happen is a transformation in the nature of financing, so that it actually enables developers like ourselves to ask the questions of the communities that we’re serving. These questions will inform the kinds of appliances that will be financed, in a way that makes them accessible to these communities.”
New e-Guide tool
During the discussion, much interest was sparked around a new freely available tool that can help predict electrical consumption in areas presently without electricity, by looking at satellite imagery.
Developed by researchers at the University of Massachusetts Amherst, in the United States, the Electricity Growth and Use in Developing Economies (e-GUIDE) is an API (application programming interface) used to measure estimates on residential consumption in Kenya, with Uganda and Rwanda soon to follow.
Lead researcher Jay Taneja explained: “What we’ve built is a tool that can essentially give some insight into what a new customer might consume in a place that does not currently have electricity supply. So how we built this, we actually worked very closely with a utility partner – Kenya Power – to look at previous consumption data. Then we trained a machine learning model, neural network, which is able to look at satellite imagery in those locations, where we have good data, and then transfer that knowledge to locations where we don’t necessarily have consumption data – but we do have satellite imagery.
“And so that satellite imagery is the link here that really allows us to understand in that new place, what is the likelihood of the number of high-consuming customers, lower-consuming customers? How might a system be designed in that new place?”
Essentially the e-GUIDE poses a solution for regulators and businesses, helping them to better understand how to roll out their systems.
“For more efficient investment of resources so they are not overbuilding in certain communities, and underbuilding in others,” said Taneja.
He added that as more data is fed into the machine, it becomes more accurate: “One of the great benefits of a machine learning model is that it gets increasingly better in a diverse array of situations. And so, as you train this model with more and more examples from more and more countries, you can apply the model with higher confidence.”
The researchers hope that by late 2021, the API will provide consumption estimates around Africa.
Speaking from Abuja, Nigeria, Ifeoma Malo, CEO of the Clean Technology Hub, was enthusiastic about the idea: “If there’s a way that we can collaborate with Jay and his group to sort of feed into that data, we would be very happy to do so,” she told the panelists.