Applications of Non-Intrusive Load Monitoring and Solar Energy Disaggregation

A method for remotely setting, controlling, or modifying settings on a programmable communicating thermostat (PCT) in order to customize settings to a specific house and user, comprising:

• receiving at a remote processor information entered into the PCT by the user, the information comprising temperature set points and start and end times, the information received via a network connection between the PCT and the processor;
• receiving at the remote processor:
• non-electrical information associated with the specific house or user, non-electrical information received from a plurality of information sources, including publicly available database and weather data;
• and
• energy usage data of the specific house, received from a utility, Smart Meter, or measuring device;
• performing by the remote processor energy disaggregation on the energy usage data, the energy disaggregation comprising determining any contribution from solar panels and adjusting for such contribution;
• determining by the remote processor a plurality of custom schedules for the PCT based upon the information entered by the user, the non-electrical information associated with the specific house or user, and disaggregated energy usage data;
• revising by the remote processor, the custom schedule for the PCT based upon additional user input or seasonal changes;
• providing the custom schedule for the PCT to the PCT.

A method of energy disaggregation, comprising:

• receiving at a processor a whole house profile of energy usage;
• selectively communicating with a database comprising:
• non-electrical information comprising information received from a programmable communicating thermostat (PCT), including:
• one or more temperature set points;
• one or more start times and end times of the one or more set points;
• and
• training data;
• and
• determining by the processor based on the whole house profile, the non-electrical information and the training data, individual appliance load profiles for one or more appliances, by at least receiving the whole house profile and determining, using a trained classifier based upon the training data, features present in the whole house profile indicative of incidences of occurrence and/or usage of the individual appliance.