The U.S. Energy Information Administration (EIA) has long been a key source for electrical market data. In the past, much of the EIA’s data have been useful for long-term planning, but have suffered from long lag times and cumbersome manual downloads. Some data have not been published until months or even years after the time period they describe. For example, a generator which began operating in May of 2012 might not have appeared in the EIA’s primary resource list (the EIA-860) until October or November of 2013. Historically, these issues have limited the usefulness of EIA data for many modeling purposes.
However, over the last 2 years, the EIA has made several improvements to the management and delivery of their datasets which some longtime modelers may not be aware of. These enhancements include the EIA-860M, the new Excel Add-in, and the U.S. Electric System Operating Data application. Together, these enhancements greatly expand the list of tasks for which EIA data may be useful.
The EIA-860 is a comprehensive list of grid-connected generators in the U.S. with capacity greater than 1 MW. No data set is perfect, but the EIA-860 has characteristics which are attractive to anyone concerned with data quality. EIA-860 data are collected directly from plant owners who are legally required to respond, it is expressed in consistent terms nationwide, and it is vetted by EIA staff prior to release. While thorough and generally accurate, this process is slow and has only been conducted once each year, leading to lag times of 10-22 months.
In July of 2015, the EIA quietly started publishing data from a new monthly survey, the EIA-860M. This survey is sent to plant owners which reported capacity coming online or retiring in the near future as reported in the most recent EIA-860. The EIA-860M keeps track of these expected changes, and gives plant owners a chance to update the EIA on their progress mid-year. Much of this information has previously been available through the Electric Power Monthly reports, but the EIA-860M combines these data with similar information from the full EIA-860 to create a comprehensive list of active generators. Here are a few things to keep in mind when working with the EIA-860M:
- It includes a smaller set of unit characteristics than the full EIA-860
- It has a lag of 2-3 month, so responses for May are posted late-July
- Like the EIA-860, the Retired list for the EIA-860M is not comprehensive. Only entities with operating plants are required to file with the EIA. So, if a company shuts down its last plant, it no longer responds to the EIA-860 or EIA-860M surveys, and its retired plants will not show up in the Retired list
- Unlike the EIA-860, the EIA-860M is not vetted prior to release. In order to maintain a timely publishing schedule, the EIA-860M is posted “as-is” and is subject to update without notification
Despite these limitations, the EIA-860M is a relatively thorough and current census of existing and planned generating capacity in the US. It is a welcome addition to the EIA’s current offerings.
Electric System Operating Data
The EIA has taken their first step into the world of intra-day reporting with the new U.S. Electric System Operating Data viewer. While the tool is still in Open Beta, and comes with a fair number of known issues, it promises to be an excellent source for very near-term information about the bulk electrical grid of the U.S.
Figure 1: EIA Operating Data – Status Map
Since July of 2015, the EIA has been collecting hourly data from all 66 Balancing Authorities operating in the U.S., including:
- Day-ahead demand forecasts
- Actual demand
- Net generation
- Interchange with surrounding Balancing Authorities
When everything is working smoothly, the EIA posts these data with a lag of only 80 minutes! These same data are available for download in table form and include API codes for pulling them directly into an Excel workbook using the add-in described below. The EIA also includes a series of pre-made charts and reports on daily supply-demand balance, discrepancies between forecast and actual demand, and much more.
Even for long-term planners, the new datasets collected by the EIA will likely be useful. Never before has the EIA published such granular demand and interchange data. The interchange data in particular has historically been very difficult to find from a publicly available source. Also, Balancing Authorities are much more useful footprints for modeling purposes than states, which is how the EIA partitions much of their information currently. Although it is still in its infancy, the Electric System Operating Data tool promises to open many avenues of analysis which were previously infeasible.
Released in February of 2015, the EIA Excel Add-in is useful for importing frequently updated data series into an existing process. While the EIA Interactive Table Viewer is handy for browsing and pulling individual data series, the data almost always need some sort of manipulation or conversion before being input into production cost models such as AURORAxmp. Whether you are converting between nominal and real dollars, changing units, extrapolating growth rates, or combining EIA data with other sources, a series of computations are usually required between raw data and useful inputs. The new Excel add-in allows a user to construct an Excel workbook with all the necessary conversions which can be updated to the latest EIA data with a single click.
Figure 2: EIA Excel Add-in Ribbon
Economic data series from the St. Louis Federal Reserve are also available through the same add-in, allowing the user to pull in indicators such as inflation or exchange rates alongside energy-specific data from the EIA. Not only does this save time, it ensures that the correct data series is queried each time the data are updated.
The EIA has always been a key data source for energy analysts, and they are rapidly evolving to become even better. Staying up to date with their latest offerings can reveal relatively easy solutions for some of the toughest data management and upkeep issues encountered by power system modelers.