By Greer Gosnell
Published: December 9, 2024
From 2018-2020 and 2022-2024, members of our team partnered with UK universities (Imperial College London and LSE), a government DISCOM in Haryana, and Tata Power to implement field experimental research studies. The study with UK universities provided a proof of concept for the technology and algorithm, while the study with the government DISCOM and the study with Tata Power shifted the research from the lab to the field – the former in a pilot trial in Panchkula and the latter in a trial comprising over 1000 households in Mumbai and New Delhi. The studies investigate the potential impact of automation and incentives on demand management facilitated by inexpensive smart energy technologies in the residential sector.
To access the student housing paper click here!
To access the Tata Power paper click here!
Our team’s early research aimed to establish a proof-of-concept for future research into and implementation of an inexpensive demand response solution in homes and businesses. To this end, we piloted a program – which we called POWBAL, short for ‘power balancing’ – that distributed smart plugs to 144 occupants of university residences and randomized the frequency (once a day or up to eight times a day) and duration (15 or 30 minutes) of automated events in which the plugs discontinued power to their connected devices. Rewards for allowing these events to take place were in the form of a fortnightly lottery, where participants’ chances of winning the lottery were directly proportional to their contributions to the virtual power plant.
Results suggested that users responded more positively to longer events that were less frequent, and that contributions to the virtual power plant increased with reward size. In the research output, we explore an early algorithm that indicates that such a program could reduce carbon and wholesale energy costs by about 10% on average in the UK.
However, our ambition for impact had set our sights beyond the bounds of the Western world, where electricity grid emissions have generally become less dependent on fossil fuels and where energy reliability is high. To maximize impact, we followed our instinct to target a geography with a triple bottom line: climate benefits from greenhouse gas emissions reductions, economic benefits from improved energy reliability, and health benefits from a decline in harmful air pollutants. A fateful conversation between two of our founders redirected our research to India.
POWBAL web platform used during the trials
In 2019, a team of researchers (who would later become members – including all three founders – of the pow-dr team) successfully applied for funding from JPAL’S King Climate Action Initiative to run pilot trials of the technology in India. The team did not have existing research relationships with DISCOMs in India, so we built from the ground up, starting with relationship building efforts with a DISCOM in Haryana, who was eager to undertake the research with us after a few months of conversations.
The research design required our field research team to administer surveys and distribute smart plugs via a door-to-door approach. Unfortunately, we were struck with some bad luck: the COVID-19 pandemic halted our operations, and restrictions outside of our control meant that we were largely unable to recruit at the scale we would have liked. Nevertheless, the research team identified a few opportune windows in which to approach households, and we were ultimately able to register a few dozen households from whom we collected data. This pilot providing promising insights, in that we observed household compliance with >90% of automated events (what we now call ‘pow-downs’). In other words, households did not override the randomly timed 30-minute events, but instead demonstrated that there was latent energy use flexibility that could contribute to load balancing.
A few months after our team started working with JPAL, Dr. Shefali Khanna joined Ralf’s team at Imperial College as a postdoctoral researcher. Shefali had worked with Tata Power previously, and the ongoing load balancing research aligned with Tata’s objectives. Within a few conversations, yet another POWBAL research partnership was born.
In this version of the research, Tata engineers installed Wi-Fi-controlled smart switches in participants’ homes. Switches could connect to one appliance, and in this context, 80% of households opted to connect their switches to a room air conditioning unit. As in the UK trial, participants experienced automated thirty-minute events in which a participant’s smart energy device paused supply to their connected appliance, with customers earning rewards based on each kWh of electricity they contributed to the virtual power plant during these events.
The findings ‘in the wild’ also supported the solution: during the automated events, household energy use declined by an average of 8.5%, with no ‘rebound effects’ to compensate for these reductions. This reduction translates into a 5.4MWh virtual power plant over 17 months with only ~1000 households participating. Household cost and carbon emissions savings declined by 2-3%, with a negative mitigation cost (or ‘marginal abatement cost,’ for the climate economics enthusiasts) – in other words, instead of society paying to reduce emissions, it actually saved $23 per ton of CO2, on average. In short, it’s a win-win.
While the trials continue to support the notion that automated demand response has a promising role to play in facilitating clean and just energy transitions globally, much work remains to understand the cost-effectiveness of the pow-dr concept relative to other crucial measures – such as energy efficiency and battery storage – and to bring the technology to scale in critical markets.
What is the role of automated demand response in clean energy transitions?
To this end, we are:
Watch this space – more to come as the journey continues!
.png){kind=small}
