Home Decarbonizer: Greening Household Energy Consumption Using Temporal Shifting

Residential electricity consumption contributes significantly to global carbon emissions, primarily due to our reliance on fossil fuel sources for power generation. This paper focuses on mitigating household carbon footprint by scheduling flexible electric loads based on the predicted carbon intensity of electricity. Our hypothesis is that past carbon intensity data often follows daily patterns, and these patterns can be used to forecast future carbon intensity and schedule daily chores during predicted periods of lower intensity. In this paper, we develop a day-ahead carbon intensity predictor and a carbon-aware scheduling algorithm that recommends optimal scheduling times for flexible household loads. Our results show the benefits of our approach for several regions. In the California region, for example, our carbon intensity predictions have an MAPE of 3.09%, and our carbon-aware scheduler can reduce carbon emissions by 32.4%.