The Carbon Neutrality: Smart Controls sub-team aims to reduce errors in Cornell's record of its campus energy consumption by detecting meter faults and anomalies. In the fall, SC focused on quality control, including detecting missing data, repetitions, sudden spikes in heating or cooling, and discrepancies from previous years. Going forward, SC plans to 1) improve baseline algorithms, 2) use ML packages to make the model predictive (e.g., based on weather patterns), 3) collaborate with other faculty and staff working in smart buildings, 4) establish methods of communicating when meters need to be fixed, and 5) develop more advanced model selection techniques based on biophysical formulas.



Also available in: Atom