dGen Overview

The dGen model provides users with a way to run the model on pre-generated agents for a sub-set of states in the United States. Hawaii and Alaska are excluded from the dataset. For additional information on the data needed to run the model, methods used, and other important aspects of configuring a model run, please refer to the home page readme and or the wiki on github.

agent based model

Every dGen analysis starts with a base agent file that uses statistically-sampled agents meant to be comprehensive and representative of the modeled population. They are comprehensive in the sense they are intended to represent the summation of underlying statistics, e.g. the total retail electricity consumed in the state. They are representative in that agents are sampled to represent heterogeneity of the population, e.g. variance in the cost of electricity. As described in ( Sigrin et al. 2018) “during agent creation, each county in United States is seeded with sets of residential, commercial, and industrial agents, each instantiated at population-weighted random locations within the county’s geographic boundaries. Agents are referenced against geographic data sets to establish a load profile, solar resource availability, a feasible utility rate structure, and other techno-economic attributes specific to the agent’s location. Each agent is assigned a weight that is proportional to the number of customers the agent represents in its county. In this context, agents can be understood as statistically representative population clusters and do not represent individual entities.”

dGen Database

diffusion_config - This schema contains tables governing the possible configurations present in the input sheet.

diffusion_iso_rto_data - This schema contains tables related to county-state-ISO/RTO topology mapping, ISO/RTO load profiles by zone and sector, as well as participation factors.

diffusion_load_profiles - This schema contains tables relating to the load profiles used by agents generated by the NREL Buildings team.

diffusion_mapping - This schema contains additional tables related to county-state-ISO/RTO-NERC topology mapping as well as existing installed capacity.

diffusion_resource_solar - This schema contains a table, solar_resource_hourly, which contains the solar capacity factor for a given geographic-azimuth-tilt combination that matches to the same geographic-azimuth-tilt combination found in the pre-generated agents pickle file.

diffusion_shared - This schema contains tables used for inputs in the input sheet. Please browse these tables as the names of these tables are representative of what these data are.

diffusion_solar - This schema contains tables with additional data pertaining to modeling solar constraints, incentives, and costs.

diffusion_template - This schema contains tables that are copied to make a new schema upon completing a dgen model run. Many of these are populated with data from the input sheet, from various joins/functions done within the database, and of course data from the model run.