A method to model the system-wide impacts of residential heating electrification under various future load scenarios in Texas
The push to decarbonize the world's economies has led to discussions about fully electrifying the residential sector. This transition involves replacing any appliances that use on-site fuels (e.g. propane natural gas, fuel oil) to versions that use electricity. For space heating, this transition includes replacing a fossil fuel powered furnace with an electric heater such as an air-source heat pump (ASHP) or a mini-split heat pump (MSHP). Widespread installation of heat pumps could have large impacts on the electrical infrastructure in regions that typically rely heavily on fossil fuel furnaces for heat. The projected change in load on the electrical grid are valuable to evaluate now, in anticipation of increased adoption of electrified heating units. This dissertation develops novel methods that (1) quantify energy consumption and peak demand from residential electrification, (2) validate an aggregate building energy model utilized for residential electrification modeling, and (3) assess the system-wide effects of residential electrification on a synchronous electric grid. First, a method was developed to quantify the energy consumption and peak demand of a residential sector with fully electrified space heating. This method was applied to the synchronous Texas electric grid operated by the Electric Reliability Council of Texas (ERCOT). The method utilizes the National Renewable Energy Laboratory’s (NREL) ResStock tool to develop geographically representative housing stock models and the physics-based EnergyPlus modeling software to create an aggregate building stock energy model that represents the residential sector in the ERCOT operating region. In this aggregate building energy model, all natural gas and other fossil-fuel furnaces are replaced with reversible electric heat pumps of varying efficiencies that can provide heating in the winter and cooling in the summer. Spatially-resolved actual meteorological weather data are integrated with the building stock energy model to simulate a specific year (2016) of hourly-resolved energy usage in the ERCOT region. Annual electricity consumption, peak hourly power demand for each day, and load duration curves for each of 17 regions within ERCOT are estimated for a variety of electrification scenarios and an as-is base scenario. From the base scenario, the absolute winter peak electrical power demand in the residential sector could increase by as much as 36%, or 12 GW. These results indicate that grid capacity would need to increase by 10 GW (a 25% increase for the residential sector) to accommodate a winter peaking residential sector. Though winter electricity consumption would increase for home heating, the annual amount of electricity consumption would stay roughly the same or decrease because the higher efficiency heat pumps provide more efficient cooling than the conventional air conditioners they also replace. Using average 2018 emissions rates, the analysis shows a change to standard efficiency heat pumps would reduce CO₂ emissions 4.1% and NO [subscript x] emissions 5.8% from the residential sector. There is no significant change in SO [subscript x] emissions in the standard efficiency scenario, but in the high and ultra-high aggregate efficiency scenarios, SO [subscript x] emissions are reduced by 8.3% and 15.0% respectively. Second, a method to validate the aggregate building stock energy model was developed. Publicly available measured electricity and natural gas consumption data from the Texas and ERCOT residential sectors were compared to energy consumption data produced by the aggregate building stock energy model. The validation analysis includes comparisons on annual energy consumption, monthly electricity consumption, consumption trends via geographic location, and consumption trends via heating fuel type. This validation analysis revealed relatively accurate results from the aggregate building stock energy model during months with historically mild weather in 2018: March, April, November, and December. During the winter peak month of January, the model overestimated consumption by 10% and during the summer peak month of July, the model underestimated summer consumption by 20%. This deviation was seen in most of the other hot months of 2018: June, August, and September. Geographically, the model performed accurately in low population regions. Higher population regions like those that contain Houston or Dallas/Fort Worth saw larger errors between the measured data and the modeled data. Homes with heating powered by a fuel other than electricity saw relatively large deviations in consumption from measured data---approximately 0.5 kWh less energy per household. During the summer, the aggregate building stock energy model's average occupancy behavior was shifted 2 hours earlier than the measured data's average occupancy behavior. These deviations likely stem from differences in actual occupant behavior and how occupant behavior is represented in the model. Lastly, a method to assess the impacts of residential electrification on the electric grid was formulated. This method used a multi-nodal UC&D grid model of the ERCOT wholesale market with load data of electrification scenarios modeled by an aggregate building stock energy model. The results from the grid model show that residential electrification would increase the grid's dependency on peaker power plants, like natural gas boiler generators, during the system's winter peak. For some electrification scenarios, the natural gas boiler plants' capacity factors increased by 350% to 400% compared to the base case. In the summer, dependency on peakers is reduced. This reduction was especially large in scenarios with an added 10 GW of solar capacity where the capacity factor of natural gas boiler plants' was reduced over 75% compared to the base case. Additional findings include a reduction in CO₂ emissions ranging between 2.9% and 22.1% for all electrification scenarios. There is an increase in CO₂ emissions for the winter peak day in the majority of electrification scenarios because of the added amount of load from electric heating. Despite this increase in load, the electrification scenarios cause relatively low amounts of congestion during the winter peak timeframe, adding as much as 7 hours of line congestion over a 15 day span.