========================== Input Database Methodology ========================== At its core, the input database can be broken down into four fundamental parts: geographies, drivers, subsectors, and temporal shapes. Geographies ----------- Geographies: EnergyPATHWAYS allows for data input and running on a variety of geographical levels (country-wide, state-wide, city-wide, regional, etc). Map keys are used to map data between these different levels. Examples include population, number of households, land area, and steel production. For instance, if iron and steel sector energy demand is taken in on the national level, but the model is run on the state level, the demand can be broken down by state in proportion to the map key of steel production. All regions and map keys are outlined in the spatial join table (GeographiesSpatialJoin.csv). See Figure below, repurposed from the Evolved Energy Research EnergyPATHWAYS website, for another example. .. image:: ../images/image5.png :alt: Geography Rescaling Example :width: 5.73224in :height: 3.01555in *Geography Rescaling Example* *Here, VMTs known by census region are mapped to electricity interconnection region according to fraction of households in the census region that are in each electricity interconnection region.* *Source: Evolved Energy Research EnergyPATHWAYS Website (Retrieved June 27, 2024)* Demand Drivers -------------- Demand drivers like population and gross state product are often correlated with energy demand. The file DemandDrivers.csv allows for historical and projected inputs of such drivers into the model. *Single Driver Example*: Imagine we are projecting iron and steel energy demand to 2060 and specify a driver of population. Historical annual energy demand input for iron and steel will be divided by the population input for each year. These per capita values will then be projected into the future using the extrapolation method specified by the user (e.g., linear regression, exponential fit). Finally, the projected per capita values to 2060 will be multiplied by the respective population projection for each year to obtain a final-energy demand value. See Figure below. .. image:: ../images/image6.png :alt: Single Demand Driver Visual Example :width: 6.08429in :height: 1.16161in *Single Demand Driver Visual Example* *Multiple Drivers Example*: Multiple drivers can be employed together in a parallel manner. For each new driver, there is just an additional division and multiplication step. The model divides by each driver, projects forward, and then multiplies by each driver to get a total energy demand. For instance, if the drivers of population *and* gsp (gross state product) are specified, the model divides annual energy demand values by both population and gsp, obtaining values in units of energy per person per dollar. These values are then projected into the future and multiplied by the future population and gsp projections to obtain final-energy demand projections. See Figure below. .. image:: ../images/image7.png :alt: Multiple Demand Driver Visual Example :width: 5.53514in :height: 2.32819in *Multiple Demand Driver Visual Example* Subsector Data Input -------------------- The EnergyPATHWAYS model groups data on two main levels: sectors and subsectors. These can be defined by the user. Sectors are the primary groupings of energy demand and are often analogous to the different sectors of society. In the Net Zero Australia study, four sectors were used: transportation, residential, productive (industry), and commercial. Subsectors are the secondary groupings of energy demand and are located within each sector. For instance, in the Net Zero Australia study, the iron and steel industry was a subsector of the productive sector. After loading in geographic and demand driver data, as outlined in Sections :ref:`Geographies` and :ref:`Demand Drivers`, the user can define the sectors and subsectors and begin to input individual subsector data. The model can be run independently for each subsector, so they can be tested as they are added. Subsector modeling is done using one of four core methods: (the first two being the most common) 1. **Energy Demand** (see `Net Zero America Annex A.2 `__\ *, 2020, p. 56*)\ **:** This method is the simplest employed by the model. It is commonly used for the productive sector, where subsector stock and service demand data are limited. Instead, users can input direct historical energy demand by fuel type in DemandEnergyDemands.csv. Historical energy demand data are used with user specified drivers and growth methods to project energy demand into the future. a. Users can further influence projection with efficiency measures (e.g., 1% annual growth in energy efficiency) or fuel switching measures (e.g., push a transition from coal to hydrogen in iron and steel sector). See Sections :ref:`DemandEnergyEfficiencyMeasures.csv` to :ref:`DemandFuelSwitchingMeasuresImpact.csv` for details on the configuration of such measures. 2. **Stock and Service Demand** (see `Net Zero America Annex A.2 `__\ *, 2020, p. 48-54*)\ **:** This method is the most explicit representation of energy demand in the model and requires the most amount of input data. a. The main data requirements include: 1. Historical service demand data (e.g., passenger-km per year) input into the model and tied to relevant drivers in DemandServiceDemands.csv. 2. Relevant technologies and efficiencies specified in files such as DemandTechs.csv. Each subsector using the Stock and Service Method has a set of user-inputted demand technologies that are used to meet the projected annual service demand. A demand technology can be defined as a technology that uses an energy carrier (e.g., electricity, coal) to meet a service demand (e.g., passenger-km). For example, in the passenger vehicles subsector, a light duty gasoline automobile and light duty electric vehicle are two examples of demand technologies. 3. Initial technology stock specified in DemandStock.csv, with relevant sales measures specified in DemandSalesShareMeasures.csv. For an example of how sales share inputs are utilized, see `Net Zero America Annex A.2 `__\ *, 2020, p. 50-52.* b. Method: The model uses initial stock and sales measures to perform rollover calculations and obtain the technology stock for each future year. Service demand is allocated to stock technologies. Using technology efficiency (e.g., mpg) and service demand, final energy demand is calculated. 3. **Stock and Energy Demand** (see `Net Zero America Annex A.2 `__\ *, 2020, p. 54-55*)\ **:** This method accounts for subsectors where service demand data are not readily available, but technology stock and energy consumption data are. Instead, users can input direct energy demand and technology stock data, and the model will back solve for service demand and run in a similar method to Stock and Service. a. The main data requirements include: 1. Historical annual energy demand data by demand technology input into DemandEnergyDemands.csv. 2. Relevant technologies and efficiencies specified in files such as DemandTechs.csv and DemandTechsMainEfficiency.csv. 3. Initial technology stock specified in DemandStock.csv, with relevant sales measures specified in DemandSalesShareMeasures.csv. For an example of how sales share inputs are utilized, see `Net Zero America Annex A.2 `__\ *, 2020, p. 50-52.* b. Model Calculation Method: 1. For each year that energy demand data is *explicitly* inputted into DemandEnergyDemands.csv, divide (or multiply, depending on units) by that year's technology efficiency (from DemandTechsMainEfficiency.csv) to obtain the service demand value for that year. 2. Project service demand values into missing years using the interpolation and extrapolation methods specified in DemandEnergyDemands.csv. 3. Run the "stock and service" method as usual to obtain energy demand projections. 4. **Service and Service Efficiency** (see `Net Zero America Annex A.2 `__\ *, 2020, p. 55*)\ **:** This method accounts for subsectors that have service demand data but lack the technology or stock data necessary to perform a full rollover calculation. Instead, service efficiency terms are input into DemandServiceEfficiency.csv and, combined with service demands, allow for an energy demand calculation. Temporal Shape Data Input ------------------------- EnergyPATHWAYS has the capability to craft temporal (hourly) demand curves for each energy carrier on the subsector level. The model can be instructed to do this from the config.ini file discussed in Section :ref:`Key Run Files`, where the user can specify which energy carriers to produce subsector-level temporal energy demand curves for. Demand shapes are based on input temporal shapes in the ShapeData folder of the database. All shapes are listed in Shapes.csv (Section :ref:`Shapes.csv`). For example, the file residential_electricity.csv contains residential electricity demand for every hour of the year for different geographical regions. Subsectors are assigned shapes in one of two ways: 1. **Subsector Level**: The user can specify a demand shape for an individual subsector in the "shape" column of DemandSubsectors.csv. The shape name is from the list in Shapes.csv (Section :ref:`Shapes.csv`). This shape will be used for ALL final-energy carriers' hourly profiles if generated. For example, if a shape is assigned to the iron and steel subsector, coal demand, electricity demand, and all other energy carriers’ temporal demand will follow this shape. When a shape is forced on the subsector level, there is an additional data field (in the shape csv file) known as the dispatch_feeder. The user must use DispatchFeedersAllocation.csv (Section :ref:`DispatchFeedersAllocation.csv`) to specify dispatch_feeder rates. This controls what proportion of the energy demand is met by each type of dispatch_feeder shape values. For instance, imagine the dispatch_feeder types are “productive” and “residential” in the passenger vehicle subsector’s shape. This means that separate shape data is available with the flag “productive” and “residential.” The user must use DispathFeedersAllocation.csv to control what proportion of the subsector’s shape is determined by the “productive” flagged data versus the “residential” flagged data (“1” means fully by one feeder; “0.5” and “0.5” means a 50-50 split, etc.). 2. **Sector Level**: If a subsector-specific shape is not specified, then it defaults to the sector level shape determined from DispatchFeedersAllocation.csv (Section :ref:`DispatchFeedersAllocation.csv`). Each sector is assigned a shape in the DemandSectors.csv file (e.g., Residential sector is assigned the shape residential_electricity from the list in Shapes.csv). In DispatchFeedersAllocation.csv, the user breaks down what proportion of energy in each subsector comes from each sector's shape. For example, the user can specify that, for the passenger vehicle subsector, 50% of the demand is met by the residential sector shape and 50% is met by the productive sector shape. The model will combine the two shape curves from those two sectors (which were specified in DemandSectors.csv) to produce one for the subsector. As in the first method, this shape will be used for all energy carriers in the subsector.