Turning AC exhaust
into power
Design and Development of a Device to Harvest AC Compressor Exhaust Heat for Sustainable Electrical Energy Production.
About the paper
Written with Dr. Om Prakash Sahu, Associate Professor at the School of Computer Science, University of Petroleum and Energy Studies, Dehradun, together with Khyati Verma and Lekha Sahu of the Department of Computer Science and Engineering, SSIPMT Raipur.
How I wrote it
One and a half percent. That is the number I brought to Dr. Sahu after four months of work. I expected him to end the project.
At eight, I filled a balloon with hydrogen using aluminium foil and drain cleaner, and watched it burn. That was my first real experiment with materials and reactions, a thread I picked up again years later, testing metals for a thermoelectric device.
The idea was simple to say out loud: air conditioners and refrigerators throw away heat, 40 °C to 70 °C of it, through the compressor exhaust, and nobody uses it. Under Dr. Om Sahu, I built a device to catch that heat and turn it into electricity, using Peltier modules and the Seebeck effect. Simple to say. Not simple to build.
Nearly four hundred hours went into it, spread over four months. I tested copper against stainless steel for the catchment plate; stainless steel won on performance by a whisker, and by a mile on cost. This is a win-win situation, I remember writing, because a cheaper metal winning was not what I expected. I wired six Peltier modules in series, then in parallel, even though the theory said the two should perform the same. Series won, and I still do not fully understand why! I only trust what the graphs showed me.
Then came the number that worried me. Carnot's theorem sets a ceiling on how efficient a system like this can ever be, based on the temperature difference available. Mine was small: a compressor at 40 °C to 70 °C, a room at 25 °C to 35 °C. That ceiling sat around 5%. My device reached 1 to 2%. To charge one AA battery would take it 24 to 30 hours.
When I brought this to Dr. Sahu, expecting a change or worse, termination of the project, he changed my perspective. There are millions of air conditioners running right now, in every city, wasting heat the same way mine was. Even 1% efficiency, multiplied across all of them, is not nothing: it is scalable impact. That reframed how I judge engineering. Not by how clean a number looks on its own, but by what it could do at scale.
The paper itself went through as many versions as the device did.
In October 2026, the paper will be presented at the IEEE TEMSET symposium. I still think about that balloon, at eight, and how it is really the same question I am still asking: what can I make from heat nobody wants.