Energy Disaggregation

I worked with Prof. Denis Trystram and collaborated with Qarnot, a green cloud computing company, to design an energy disaggregation algorithm for their smart buildings. The Qarnot Computing platform uses computing-heaters and utilizes the waste heat from IT computations in their buidings and offices. Through energy disaggregation, our aim was to better understand the energy requirements at different locations of the building and hence, forecast energy needs. Using this information, we can achieve a more efficient job scheduling, for example, scheduling more jobs where there is greater energy requirement. We developed an integer linear programming approach using the information provided by Qarnot Computing about different devices used in their offices and buildings.