By Jon Kornik, CEO of Plentify

One of the unintended consequences of load shedding is that when the power is turned back on in an area, most geysers switch back on at the same time and work hard to reheat water that had cooled, leading to sudden surges in demand.

This is called ‘comeback load’, and a new study shows that this phenomenon pushes up maximum geyser energy consumption across some areas by 90%. Considering that geysers are responsible for approximately half of household electricity use, this creates havoc for local municipalities and the grid operator, and impacts consumers as well.

Implications of Comeback Load

Comeback load directly increases electricity costs for large consumers – including municipalities, residential estates, apartment blocks, and retirement villages – who often pay charges based on their maximum demand levels each month.

Ironically, this means that power cuts ultimately translate into higher household energy bills. As per Eskom's Megaflex tariff, which applies to municipalities and large customers such as estates, this comeback load could drive up electricity costs by about R1,300 per geyser per year.

Comeback load, meanwhile, is also damaging electricity distribution networks because it can trip or overload transformers, substations, and other critical infrastructure. Again, this raises system costs and leaves the grid in a more fragile state.

To mitigate the impact of comeback load, some municipalities are carefully sequencing how they return power to individual substations and the areas they serve. While this tempers maximum demand, it does result in extended load shedding periods and higher operational costs.

Further, many municipalities grapple with the problem of which areas to switch back on first, as there are often complaints about inequity and unfairness. In some towns, the last area can be switched on up to two hours after the first, because technicians have to physically drive between substations to switch the power back on.

A recently concluded pilot project showed that there is a better way to deal with this problem.

Project Smart Geyser: A Solution in Action

Under Project Smart Geyser, a 30-month-long study, Plentify partnered with the City of Cape Town, Hessequa Local Municipality, residential estates and retirement villages to gauge what impact artificial intelligence-enabled geyser-management devices would have on curbing comeback load and trimming household energy bills.

We installed 500 of these devices, called HotBots, in households around the municipalities and in multi-family properties. They were coordinated in such a way that the geysers drew power at slightly different times from one another, only turning on those geysers that actually needed to serve hot water, and avoided using energy at the worst times for the grid. This was done only to the extent that homes still had a reliable supply of hot water.

Backed by German development agency GIZ and clean energy financing facility EEP Africa, and with independent oversight from the University of Cape Town’s Measurement and Verification Inspection Body, the study found that the HotBots eliminated comeback load after bouts of load shedding - and went further than that to flatten maximum demand to a point much lower than even without load shedding.

The devices reduced each geyser’s electricity use during peak periods by up to 80%, and cut maximum demand at any point in time across the entire fleet of water heaters by up to 60%.

This slashed demand charges roughly in half, and yielded substantial energy cost savings for participating households, while ensuring that all participants still had hot water when they wanted it.

The project also demonstrated how these devices, when coordinated, can prevent damage to distribution infrastructure and defer the need for major investments in substations and transformers. In fact, with intelligent control, roughly twice as many geysers could be connected to the grid without placing any more strain on infrastructure.

This is illustrated in the chart below, which shows daily maximum demand before the installation of HotBots in purple, and after installation in teal. Not only did HotBots mitigate the biggest spikes in maximum demand, but they also flattened maximum demand, making it much more stable and predictable.