Distributed Generation Capabilities of the National Energy Modeling System
This report describes Berkeley Lab's exploration of how the National Energy Modeling System (NEMS) models distributed generation (DG) and presents possible approaches for improving how DG is modeled. The on-site electric generation capability has been available since the AEO2000 version of NEMS. Berkeley Lab has previously completed research on distributed energy resources (DER) adoption at individual sites and has developed a DER Customer Adoption Model called DER-CAM. Given interest in this area, Berkeley Lab set out to understand how NEMS models small-scale on-site generation to assess how adequately DG is treated in NEMS, and to propose improvements or alternatives. The goal is to determine how well NEMS models the factors influencing DG adoption and to consider alternatives to the current approach. Most small-scale DG adoption takes place in the residential and commercial modules of NEMS. Investment in DG ultimately offsets purchases of electricity, which also eliminates the losses associated with transmission and distribution (T&D). If the DG technology that is chosen is photovoltaics (PV), NEMS assumes renewable energy consumption replaces the energy input to electric generators. If the DG technology is fuel consuming, consumption of fuel in the electric utility sector is replaced by residential or commercial fuel consumption. The waste heat generated from thermal technologies can be used to offset the water heating and space heating energy uses, but there is no thermally activated cooling capability. This study consists of a review of model documentation and a paper by EIA staff, a series of sensitivity runs performed by Berkeley Lab that exercise selected DG parameters in the AEO2002 version of NEMS, and a scoping effort of possible enhancements and alternatives to NEMS current DG capabilities. In general, the treatment of DG in NEMS is rudimentary. The penetration of DG is determined by an economic cash-flow analysis that determines adoption based on the number of years to a positive cash flow. Some important technologies, e.g. thermally activated cooling, are absent, and ceilings on DG adoption are determined by somewhat arbitrary caps on the number of buildings that can adopt DG. These caps are particularly severe for existing buildings, where the maximum penetration for any one technology is 0.25%. On the other hand, competition among technologies is not fully considered, and this may result in double-counting for certain applications. A series of sensitivity runs show greater penetration with net metering enhancements and aggressive tax credits and a more limited response to lowered DG technology costs. Discussion of alternatives to the current code is presented in Section 4. Alternatives or improvements to how DG is modeled in NEMS cover three basic areas: expanding on the existing total market for DG both by changing existing parameters in NEMS and by adding new capabilities, such as for missing technologies; enhancing the cash flow analysis but incorporating aspects of DG economics that are not currently represented, e.g. complex tariffs; and using an external geographic information system (GIS) driven analysis that can better and more intuitively identify niche markets.