Grey Box Building Models for Demand Response Energy Trading
Recent work between UMD and Earth Networks has given us the capability to accurately predict the response of a home to changing external weather conditions by using an internet enabled thermostat to observe the home temperature and HVAC system response (calls for cooling or heating) over a period of time. Once the response of the home to future weather can be reliably predicted, this capability can be used to precool or preheat the home to shift electrical and natural gas loads away from peak demand periods to levelize energy use for utilities and/or minimize the cost of energy to the consumer. Timing of the HVAC system can be controlled to ensure the home is comfortable whenever the occupants are scheduled to be at home. Additionally, home energy audits can be performed since the response of the home to specified weather conditions (windy, sunny, cold, hot) can be modeled, allowing us to determine the quality of the insulation within the home, the tightness to infiltration, solar load, etc. The above capabilities have been packaged into a suite of programs (Modeler, Whatiffer, Scheduler) collectively referred to as e5.
The capability to model the thermal response of millions of homes to short-term weather predictions allows utilities to flatten their load curves through precooling (preheating in winter) and can allow consumers to avoid peak costs when variable pricing is in effect. The usefulness of this technique relies on highly accurate, localized weather forecasting. The capability to model the thermal response of millions of homes to short-term weather predictions allows utilities to flatten their load curves through precooling (preheating in winter) and can allow consumers to avoid peak costs when variable pricing is in effect. The usefulness of this technique relies on highly accurate, localized weather forecasting. Earth Networks has the world’s largest weather network (weather.weatherbug.com) with over 8000 tracking stations that continuously monitor and stream live weather data. Earth Networks is planning to enter the wholesale electricity markets by trading electrical HVAC capacity in response to electric price signals or in response to grid stability signals (demand response). Their success in these markets, and e5’s competitive advantage, are dependent on the accuracy of the thermodynamic models to predict, given a near-term weather forecast, a structure’s internal temperature, and the HVAC power profile required to maintain the internal temperature within a comfortable range while shifting HVAC load for economic gain. We are planning to start trials this June in Texas and California.