RTF analysts developed per unit technical potential estimates for demand response impacts from connected thermostats. When demand response events are called, connected thermostats reduce the usage of heating and cooling systems by adjusting thermostat set points: upward during summer to reduce cooling load, and downward during winter to reduce heating load. Importantly, the demand impacts are not coming from the thermostat itself, but from the HVAC technology being controlled. Therefore, impacts are determined by individual household characteristics, such as heating/cooling system technologies, efficiencies, size, and shell, as well as event-specific conditions, such as weather, time of day, and the size of the thermostat setback.
RTF Presentation Connected Thermostat DR Workbook DR Subcommittee Presentation DR Subcommittee Notes
RTF Decision
The analysis and methodology were presented to the RTF at the March 2019 meeting. Due to a robust discussion and time constraints, the RTF was unable to vote upon a proposal at the meeting. However, the following text and caveats was approved subsequently by a majority of voting members with a few expressed reservations:
Limitations
The RTF contract analysts recent work on connected thermostats demand response provides a good starting point for analyzing technical potential of these units, given the tools currently available to the RTF and the resource limitations. The RTF recognizes that while providing a good starting point, there are several limitations in this analysis and opportunities to enhance the work in the future. These include:
- SEEM, as it currently exists today, is not the best tool for estimating demand response savings. Key limitations of the model include:
- The model currently only allows two different temperature set points to be scheduled, which limits the ability to model nuances of a DR event, such as pre-conditioning
- SEEM is an hourly model that does not have the granularity to model inter-hour dynamics.
- For zonal systems, SEEM only has the ability to model the whole house, whereas DR programs may only target the main living area
- Should the RTF do additional demand response in the future, it should explore alternative models and/or enhancements to SEEM
- Additional analysis should update the equipment sizing to more accurately reflect the equipment anticipated for the house/climate need, rather than the regional average sizing used in the current analysis.
- There are limitations to using a single extreme day from TMY weather files, without consideration of the prior few days weather.
- While the current analysis breaks out different insulation quality levels, this may not be practical from a DR program perspective.
- Current modeling does not sufficiently capture systems with backup propane or gas heat. This could be improved in future analysis.
Approach
Based on input from Council staff and the DR Subcommittee, analysts ultimately utilized RTF's version of SEEM to model savings. Thermostat setpoints were adjusted within SEEM in order to simulate 4 hour events utilizing a 3 degree thermostat setback during summer and winter afternoons, and winter mornings. The hourly energy usage for heating/cooling during these simulated DR events were then compared to hourly heating/cooling energy in baseline runs where thermostat temperature settings were not adjusted. These baseline and DR scenarios were run on SF homes for a variety of HVAC system types, including central air conditioning and heat pump for summer cooling, and electric furnace, heat pump and electric zonal resistance for winter heating. The capacity and efficiency of each system type was set up in SEEM to reflect regional averages according to the latest data RBSA data. Also included were scenarios for a range of insulation and infiltration levels, generically summarized as Good, Fair, and Poor, consistent with prior RTF analyses. Additionally, scenarios were also run for each of the 9 heating zone-cooling zone combinations in the region utilizing the most extreme temperature days and cities from each region based on TMY data.
Have questions? Please get in touch.
