energy efficiency to reduce residential electricity and natural gas use under climate change - residential energy storage

Since changes in temperature can change heating and cooling loads, climate change may significantly affect consumer demand for building energy.
Warming can also lead to more adoption and use of cooling technology in buildings.
We assessed residents' electricity and gas demand in Los Angeles, California based on multiple climate change forecasts, and investigated the potential for energy efficiency to offset increased demand.
We calibrate residential energy use based on metering data to explain differences in building materials and appliances.
In the case of rising temperatures, we found that in the absence of policy intervention, the demand for electricity for residents may increase by up to 41-87% between 2020 and 2060.
However, aggressive policies designed to upgrade heating/cooling systems and appliances may result in a 28% increase in electricity consumption, which may avoid installing a new generation of capacity.
Therefore, we propose to actively improve energy efficiency in combination with low energy efficiency
Carbon producers offset the projected growth in residential energy demand.
To quantify the relationship between energy consumption and climate change, we have developed a model to predict residential energy use between 2020 and 2060 in Latin America and the Caribbean.
Our model is a firm bottom in space and time.
We evaluated the use of residential energy based on actual consumption data.
Using survey data and physical information about building inventory, we created 84 prototypes
Building Simulation in building energy consumption optimization (BEopt)
Software representing all residential buildings in LAC.
At BEopt, we use energy Plus, a state-of-the-
The art building simulation software developed by the US Department of Energy serves as the main simulation engine.
According to the year of construction, we classify (
It's a single family, a townhouse, etc)
And climate zones.
For each prototype, we include 21 heating and 13 cooling techniques ().
Next, we use the census block group to expand the scale of electricity and gas consumption to the county level while maintaining space details (CBG).
Then, we use the CBGs subset calibration model that belongs to the Los Angeles Hydropower Department (LADWP)
Service area using 1 year power data.
One assumption of the model is that the usage patterns are related to the type of building, for example, homeowners in the same climate zone similar to older homes will use similar set temperatures at home.
Then, we predict that in the case of climate change and population increase, the electricity and gas consumption of LAC residents.
We have also developed scenarios for different equipment and building efficiencies to examine the possibility of offsetting projected energy growth.
In developing the prototype, we used three major sources of information: the household appliance saturation survey (RASS)
, LAC assessor database and California assessor manual.
RASS is from California-
Specific equipment surveys managed by the California Energy board capture various variables for building thermal performance and equipment use.
The latest RASS survey in 2008 contains 6,500 LAC survey responses, which we use to inform of the electrical appliance distribution and some material properties of the building in each prototype ().
The LAC Assessor Office maintains a database of each building in the LAC, mainly for tax purposes, and we use the dimensions, classification, location and quality information of these buildings when developing prototypes.
In addition, the California assessor Handbook provides "typical" features of buildings of different quality grades and serves as a reference for assessing the value of California real estate.
We use this manual as a supplement to the assessor database and RASS to add more details about thermal performance for each assessor.
We summarize the data from each source.
In Latin America and the Caribbean, the climate varies greatly between coastal and inland regions;
Therefore, we distinguish prototypes according to five climate zones (and ).
The California Energy Commission has developed these areas specifically for buildings in order to comply with California's energy efficiency standards for title 24 buildings ().
We developed custom building prototypes based on previous work and then subdivided according to climate area, building cycle and residential building type.
We considered seven major time periods, five climate zones and four types of buildings with 140 potential prototype categories (7 × 5 × 4=140)
, But we summarize this further to ensure that each prototype has sufficient survey response in RASS.
This resulted in a total of 84 prototypes ().
We then divide all residential buildings in LAC in the Los Angeles assessor database into each of the 84 categories.
We use the features in the assessor database as the specification of the prototype (
Average building size, for example)
, Grouping also allows scaling of final simulation results to county level.
For each of the 84 categories, we have compiled a profile of a typical building shape (
Ratio of circumference to area)
, The main materials in the frame, average size and quality grade in the assessor database ().
California assessors use quality grade names to represent greater quality and family value.
In some cases, this also means an increase in thermal performance.
For prototypes that are located in the same climate zone, with the same primary quality grade and similar floor area, we combine them together to save time in calculation.
We maintain all 84 prototypes for the equipment evaluation, but for the building simulation, we use a condensed 51 simulation model ().
For these 51 categories, we use data from three sources about the thermal performance of buildings to develop models in BEopt ().
For HVAC technology, we use 21 different heating techniques and 13 different cooling techniques in each prototype ().
We did this to get a more representative "weighted average" at the end of HVAC"
Consume energy.
In RASS, for example, the findings show that the main heating technique for most prototypes is an NG furnace with an efficiency of 78%.
However, if we simulate this technology only for all prototypes, we will not be able to capture the true variability of the heating technology (
And related energy use)that exists.
Instead, we run all the techniques in BEopt and weighted the energy consumption by Prototype category based on RASS survey results.
The BEopt output is in units per hour (
The same is true of the BEopt core simulation engine EnergyPlus)
, We summarize it to the annual resolution so that the calibration is consistent with the LADWP data.
In order to obtain the total consumption of residential building inventory, we standardized HVAC terminals
Use the consumption per square foot by prototype and multiply it by the square foot of each prototype category in each CBG.
The RASS survey reports the frequency of use of HVAC equipment during the day, in many prototype categories, non-
The proportion of residents with HVAC equipment is negligible, but it is not used most of the time.
We take advantage of this
Use the percentage to adjust the typical energy consumption of the prototype.
In addition, we use 51 analog prototypes to simulate lighting, standardized per square foot, but we maintain electrical appliances at each prototype level.
In the aggregation, the device type is maintained so that the final
The final model can determine the use and track it in the forecast.
We run the simulation integration for 2011-2012 and develop custom weather files for BEopt for LADWP climate conditions to match the calibration data set.
We calibrate according to the median annual household electricity consumption in the LADWP service area summarized by CBG.
Researchers at the University of California, Los Angeles, obtained the data as part of the California Energy Commission research program.
The University of California, Los Angeles, has removed some CBGs that may violate the secrecy of account holders.
In total, there are 2501 CBGs in the data set that can be used (
More than 90% of the LADWP service area)
There are a total of 6,422 CBGs in LAC.
The data for LADWP is from July 2011 to June 2012.
BEopt uses energy plus weather (EPW)
Document, which includes a series of climate variables such as temperature, humidity, solar radiation, snow, precipitation and rainfall ().
We created a custom EPW for each of the five climate zones of this period, leveraging climate input from local weather stations and publicly available solar radiation databases, once we run this collection with appropriate weather data and extend it to the county level, we are able to extract a subset of CBGs that exist within the LADWP service area.
The goal of calibration is (1)
Make the total simulated electricity consumption equal to the electricity consumption reported by LADWP and (2)
To end-
Use in the model a percentage of consumption similar to reported in RASS.
To identify prototypes that need to adjust the thermal properties, we compare the standardized heating and cooling power consumption from the model to the final prototype
Prototype consumption using RASS reports.
RASS model not measurement terminal
Use consumption, but this is still useful to identify which prototypes are above or below expectations.
To determine the priority of the prototype to be modified, We weighted the average gross floor area of all buildings mapped to that prototype against the deviation of RASS.
Priority is given to high coverage prototypes by building area, as they have the greatest impact on the model.
Once we have identified the modified prototype, we modify the thermal properties of the shell within the uncertainty of the input data source, for example, changing the efficiency of the pipe, changing the floor or increasing the insulation.
For electrical appliances, we use a linear proportional factor to adjust consumption, rather than adjusting the type distribution within the homeuse breakdown.
The entire calibration process is based on electricity consumption, as this is the data we use to verify, but NG includes a large amount of energy use consumed by LAC homes, mainly heating in water and space.
There is no NG data on time for school, but we keep NG results to complement the power modeling.
End of final calibration
The consumption of the base year is located.
We develop customized EPW files for each of LAC's five climate zones to predict changes in building performance under GCM temperature changes.
These files are the weather data input for the building simulation software BEopt.
In tradition (that is, non-forecasting)
Application, EPW file by combining historical weather data (
Usually up to 30 years).
Then, considering that there is no change in the climate of the location, this can be used as a criterion for predicting and comparing the Building Performance of a single location.
In order to use EPWs for forecasting, a unique EPW file must be developed for each forecast and year.
The fifth assessment report on inter-governmental Climate Change uses four different atmospheric carbon concentration predictions known as RCPs.
Each RCP was developed by an independent modeling team and specified by their 2100 level of radiation forcing.
For example, the most optimistic scenario RCP 2.
6 was developed by the image modeling team of the Dutch Environmental Assessment Agency with a radiation forcing peak of 3.
Around 2050, but so far this year, 2100 of emissions have been reduced to enough to force 2. 6u2009Wu2009m by 2100 (ref. ). RCP 4.
5 was developed by the Mingkang team of the Pacific Northwest National Laboratory in the United States, representing the scenario of stabilizing radiation forcing to 4.
5 u2009 W u2009 m before 2100 (ref. ). RCP 6.
The 0 mode radiation is forced to stabilize at 6.
2100, created by the AIM modeling team of the National Institute of Environment in Japan.
The most pessimistic scenario, RCP 8.
5, developed by the Information Team of the Austrian Institute for International Applied Systems Analysis, including an increasing 2100 of greenhouse gas emissions (ref. ).
In this study, we used 10 GCMs for each of the four RCPs to capture a range of future climate scenarios that could have an impact on residential energy consumption.
In order to maintain the spatial differences between the climate zones, we use the statistically reduced CMIP5 (
Corrected by deviation from constructed analogs)
Temperature Prediction.
For California, these data were provided in 1950 at a resolution of 12 kilometers × 12 kilometers per day, 2099 kilometers from the U. S. Bureau of Agriculture and reclamation.
In order to obtain representative temperature predictions for each model operation and climate region, we obtain them from all grid points operating within the climate region and average them at each time point.
We then "deform" the daily trajectory to get the temperature distribution per hour. Belcher .
The first proposed "deformation" is a way to create hourly profiles using daily climate predictions, which is necessary to build simulation software.
We use Sailor's modification of Belcher's original proposed method to change the temperature trajectory of each model's operation: where is the temperature at any time in the future, and DTR are the daily temperature range (
Differences in daily maximum and minimum temperatures)
For the model and base file, the temperature of the base file for that hour, respectively, is the minimum value of the day in the base file and the daily minimum value in the climate model.
For our study, the basic weather file is EPWs developed by CEC for each of the 16 climate zones in California and can be used as the default file for BEopt.
In practice, this deformation transformation matches the highest and lowest daily temperatures of the GCM and scales the intermediate time based on the EPW mode.
For the temperature trajectory of each deformation, we use 4-
H weighted average to smooth the continuity between days.
We created a total of 1,640 EPW files (
10 GCMs × 4 RCPs × 41 years)
Each of the five climate zones.
These files can then be run using our 51 calibrated prototypes.
To incorporate our simulations, each climate model must accurately predict the average number of cooling days (CDDs)
Between 1970 and 2000 within positive and negative 10%.
We provide a summary of the GCMs included.
The California Treasury predicts that the population of Latin America and the Caribbean will grow from nine. 8 (today)to 11.
2060 5 millionref. )
This will require the construction of new residential units to accommodate more residents.
Tracking population growth through the Southern California government housing forecasting Association to 2030 (
Including changing the size of the family)We develop bi-
Ten-year housing growth rate in Latin America and the Caribbean ().
We apply these population-based housing growth rates to all scenarios, starting with building inventory in the assessor database.
In a previous study, we evaluated the historical building turnover trend of LAC and developed a building turnover model based on the initial year of the building.
In addition to housing growth, we include these building turnover rates in the stock model, replacing old years with newer years of the same classification and climate zones.
For scenario 1 and scenario 2, we use the same rate as the previous paper, and in scenario 3 and scenario 4 we increase the turnover to 10 times the natural growth rate, to encourage the upgrading of building turnover and building housing ().
We do not separately model the building housing upgrade (
For example, improved windows, insulation, etc)
However, we use the increased turnover as an agent for the Shell upgrade because the updated prototype has a more efficient hot shell.
In applying population growth and building turnover, we distribute these changes in space based on the location of existing residential units.
In terms of population growth, in fact, new buildings may appear in CBGs with a small population, rather than encrypting existing areas.
We simulated all prototypes under 10 GCMs and 39 heating and cooling devices, and then, we averaged the results to obtain an average prediction of the total energy, differences in GCMs represent the variability of predictions.
We also maintain spatial resolution in the simulation set to study the spatial differences in energy demand changes under climate change.
Using the custom weather files we developed for each GCM run, we used the batch processing of EnergyPlus to simulate our 51 prototype models.
EnergyPlus is the main simulation engine of BEopt;
So once we 've created the model, we can customize the EnergyPlus input files and run them directly in the EnergyPlus.
This saves processing time and also customizes the analog output format.
We conducted a total of 83,640 simulations: 10 GCMs × 4 RCPs × 41x51 prototypes.
Through model output, we post the data in Python and store it in an SQLite database.
In the prototype calibration, we run models using 28 different heating and cooling techniques, and we include an additional 11 techniques for forecasting ().
In order to run the complete integration, this will result in 3,261,960 building simulations, which is limited in computing.
To capture differences in heating and cooling techniques without running so many simulations, we run a full set of techniques for a single GCM, and establish a linear relationship between the heating/cooling load of the building and the energy consumption generated by the technology in this prototype category.
In our tests, this method captures the energy actually used by the technology within 4%.
We use these factors in post-processing to calculate the energy consumption of each case in 3,261,960 cases without having to run each simulation.
Similar to the calibration phase, we calculate the weighted average of HVAC technology based on the number of residential units and the popularity of the technology in each prototype category.
The main difference is that these factors are temporarily dynamic in the prediction, as the adoption of technology and the number of residential units change over time.
In order to obtain the total number of operations, we predict the average power and NG of all GCM operations within the RCP.
The maximum and minimum energy values of all models per year are the uncertain boundaries of this RCP.
In the forecast, we evaluated the adoption of equipment technology and the improvement of efficiency to test the potential to offset demand growth.
We have developed four scenarios to test each climate prediction :(1)
Business as usual ,(2)
As usual, high electrified ,(3)
Moderate efficiency intervention and (4)
Active efficiency intervention.
For each case, we include dynamic building turnover based on LAC inventory modeling, inventory expansion due to population growth, appliances (
Water heater, TV, oven, etc)
And heating/cooling equipment.
These variables vary dynamically in each prototype category each year between 2020 and 2060 in order to have different predictions for power and NG per hour during this period.
For heating/cooling equipment, we use weighted average consumption of different technologies in each prototype, not just using the most popular technologies (
NG furnace for heating, for example)
, Which captures the variability of the technology used throughout the LAC ().
Scenario 1 includes existing and proposed policies as well as normal building and equipment turnaround.
Scenario 2 includes the same assumptions, except for all retired water and space heating equipment, which is replaced by an electric version regardless of the original fuel type.
This situation captures positive electrified, and previous studies have claimed that it is necessary to achieve 80% of greenhouse gas emission reduction by 2050 (refs , ).
Scenario 3 includes heavy-duty electrified (
Such as scene 2)
However, in addition to the existing standards, the modest efficiency improvement of home appliances and the increase in building turnover.
Scenario 4 also starts with a lot of electricity, but includes efficiency and building benefits in addition to current technology.
In all cases, we include an increase in air flow.
Conditioning saturation (
That is, the proportion of residential units with air conditioning)
As previous studies have found, saturation is closely related to temperature.
Our mitigation strategy is based on an enhanced version of existing policies at the state and federal levels for electrical and NG equipment.
As part of our analysis, we estimate the cost of conservative energy.
Complete background can be found in.
In the next section, we outline the assumptions behind each scenario, including current policies and publications that support the development of these assumptions.
We are summing up efficiency measures.
We use a distribution based on the age of the device to turn around the existing heating and cooling equipment in each prototype category (
Ex: technology accounted for 80% by 2020, 90% by 2030, etc).
These turnaround time frames are constant for all scenarios.
We then create a replacement matrix that gives a technical distribution to replace any retired device in each category.
For example, a 76% efficient NG boiler may be replaced by an electric furnace 10% of the time, 30% of the time is replaced by a 80% efficient NG furnace, and 10% of the time is replaced by a 85% efficient NG furnace, 10% of the time by SEER 14 air-
Source heat pump, 10% of the time by SEER 15 air-
Source heat pump and 10% of the time by SEER 19 air-
Source heat pump.
In Scenario 1, we determine these replacement rates based on the purchase trends reported in the US Department of Energy Building Energy Database.
In Scenario 2, we also use buying trends, but instead use NG and propane as an option for new equipment replacement, we distribute only new purchases in electrical technology.
In Scenario 3, we limit all heating replacements to heat pumps only, and in scenario 4, all heating and cooling equipment is replaced only by the most efficient heat pump in the model.
In addition to the scrapping of aging HVAC equipment, we have added additional cooling equipment based on previous studies of air conditioning saturation rate and temperature.
Sailors and Pavlova, based on data from cities across the United States, established an empirical relationship between CDD and the percentage saturation of air conditioners.
In this equation, air
The regulated saturation of a given year is the initial saturation, CDD is the number of cooling days for future years, and CDD is the initial quantity of CDD.
We apply this relationship to each climate zone, as well as to the annual rate of saturation for all four average RCPs.
Because CDD may change year by year in the forecast, we will fill the saturation in advance so that the air-
The regulated saturation rate for the coming year cannot be lower than the previous year.
For example, if the CDD for 2034 projects is less than 2033, we apply the saturation of 2033 to 2034, because the person who purchased the air conditioner in 2033 will not discard the air conditioner next year.
Then we apply the saturation to the prototype of that climate area of that year.
In addition to lighting and plug loads, we assume that the number of appliances used is linearly proportional to the number of residential units.
We will not change the distribution of appliances throughout the scene, only energy consumption, except to capture the "plug load" category used by various appliances in the home (
For example, mobile phones, electronic products, mixers, etc).
This section discusses unit changes in electrical appliances, not the total amount consumed per category.
On 2012, the lighting section of the Energy Independence and Safety Act came into effect, regulating the power consumption of incandescent lamps in the United States.
This not only improves the efficiency of incandescent lamps, but also reduces the cost of alternative bulbs such as compact fluorescent lighting and lighting
LEDs (LEDs).
The US Department of Energy predicts that by 2030, these changes, especially the adoption of led, will reduce the lighting power consumption of residents by 53%, below the level of 2013 (ref. ).
Inferred market penetration of lamps and other future lighting technologies applied to lighting power consumption decreased by 80% in the case of households from 2060 as a space nuclear power source scheme and a monthly, 90% and 95% month scheme.
CEC proposed the computer and display regulations that will come into effect on 2017 and 2018.
The rule will set performance standards for laptops, desktops, and displays, and set standby energy consumption targets.
The Central Election Commission expects the rule to reduce the consumption of desktop computers by 60% and laptops by 10%.
In our prediction, we assume that by 2060, the computer and TV energy of scenario 1 and scenario 2 decreased by 30%, the energy of scenario 3 decreased by 50%, and the energy of scenario 4 decreased by 70%.
We include nine different types of water heaters with four different fuels (
Electricity, gas, propane and solar energy).
In scenario development, you can use (1)
Switch to another water heater or (2)
Change the efficiency of the nine categories.
For type switching in Scenario 1, we use the purchase profile for the domestic hot water market, scenario 2 uses the same distribution, but the restricted purchase only includes the type of power and solar products.
In scheme 3, by 2060, we increased the proportion of heat pumps and solar water heaters to 30% of the total replacement inventory, and in scenario 4, these technologies account for 70% of the inventory, the remaining 30% is an electric tank-free system.
On 2015, the energy efficiency standards department of hot water heaters began to implement.
Small residential water heater (