Is low or zero electricity demand growth the new reality?
This author no longer works for DNV GL.
Can end-use information shed some light?
The U.S. Department of Energy’s Energy Information Administration (EIA) held its annual conference earlier this month in Washington, D.C., to discuss issues facing policymakers and energy markets. The conference covered a broad set of topic areas that on the surface may seem disparate, but are actually deeply connected. Resource extraction, gas prices at the pump, renewables integration, demand side management, and revenues all play in the balance between the forces of energy supply and end-use customer demand.
This graphic presented by Jim Diefenderfer of EIA illustrates that declining load growth is not a new trend: year-on-year growth in kWh of electricity sales, while mostly positive, has been in decline for several decades and EIA’s current projection is that U.S. growth will remain steady at around 1 percent over the next few decades, tracking below that of GDP.
Load is now declining without GDP growth of around 1 percent. How has it come to this?
There is no consensus on why this is happening, but Paul Sotkiewicz of PJM Interconnection highlighted several factors. (PJM is a regional transmission organization (RTO) that coordinates the movement of wholesale electricity in all or parts of 13 states and the District of Columbia.)
- Income effects – median income has been in decline in real terms
- Increased energy efficiency from both programs and naturally occurring trends in the market
- Accounts are in decline in some areas
- We have been approaching a plateau in electrification of end uses.
Interestingly, higher energy prices are not among the factors, at least in PJM, as the wholesale power cost has not rebounded with the economic recovery. This is due in large part to the low cost of natural gas, which has steadily increased its share in the overall electricity supply picture in recent years. As the following figure illustrates, gas generation resources now account for the highest share of cleared capacity in PJM auctions, with coal in relative decline.
How has this affected the industry?
Reduced Congestion – Implications of this scenario include lower congestion levels. In fact, the lower load growth was a key reason for the cancellation of a 500 kV transmission line across the PJM system.
Reduced Emissions – 26,000 MW of generation has been retired in the PJM system since 2009 (14 percent of the overall fleet) with the reasons due as much to lower demand as to new environmental regulations.
While this has certainly been a challenging operating scenario for the electric industry, it is important to consider where we could be right now. Without the significant new entry on the supply side from low-cost, low-emission natural gas generators, and without strong gains on the demand side in energy efficiency and demand response resources, we would likely be facing an energy crisis coinciding with the economic rebound, as Daniel Yergin of IHS noted. The consequences to the industry from economic effects on U.S. consumers of an energy crisis could have been more severe.
As we move forward in this complex new reality, it will be critical that EIA and the industry continue activities designed to better understand how much, when, and why customers use electricity.
EIA is currently releasing results of the 2012 Commercial Buildings Energy Consumption Survey (CBECS) in stages. Characteristic summaries of the U.S. commercial building stock are now available, providing insights such as changes in the average age and growth rate in the number of buildings and overall floor space. In the coming months, more details will be released including energy consumption and expenditures attributed to various commercial end uses, estimated through an improved end-use modeling methodology.
Individual states, such as Massachusetts, California, and New York, are also closely studying customer end-uses of energy. DNV GL is currently working with each of these states in their data collection and analysis efforts. For example, we are helping Massachusetts Electric and Gas Energy Efficiency Program administrators to characterize energy use in commercial and industrial buildings through comprehensive on-site investigations of 800 buildings across the state.
On the residential side of conference presentations, Sidhant Gupta of Microsoft Research presented recent developments on a not-so-new concept for capturing the end-use consumption within households. The technology aims to disaggregate the total metered load for a household into load profiles of individual end uses in the home. This is achieved by linking frequency signatures of the various devices consuming power within the household with changes in total load in the home. This Non-Intrusive Appliance Load Monitoring (NIALM) concept—non-intrusive compared to using meters to log power consumption for all devices in the home—emerged from research out of MIT in the 1980s and is now being supported by the Electric Power Research Institute (EPRI).
Unfortunately, NIALM has not taken off due to several technical challenges. We shall see if, in the coming years, companies working on this technology are able to develop a NIALM application for consumers that overcomes challenges such as:
- Making the product easy to install safely by consumers. Most consumers do not have the training to safely install current transformers in the electric panel of their home.
- Training the system. Will consumers understand how to provide information needed to train the machine learning algorithms, and will they take the time to do this for all end uses?
- Ensuring that the end-use load estimates are accurate. It is not straightforward to translate certain frequency profiles into power consumption.
- Privacy issues. Who will own the load data and what are the risks associated with real-time load monitoring?
- Translating the data collected into actionable information. Data alone is not information. Value messaging to customers will need to be developed and the industry will need to develop or refine approaches for accounting for detailed customer usage in their planning activities.
Why is this important? There are potentially several important applications for this technology, but related to the theme of understanding the customer, consider the fact that for the 2009 Residential Energy Consumption Survey (RECS), EIA was able to publish electricity consumption disaggregated into just five end-use categories, with the “Other” category comprising 55 percent of annual electricity consumption, using a purely statistical modeling approach, as shown in the snapshot below of EIA’s table of end-use electricity consumption estimates for the U.S.
Having much deeper insight into the end-use load usage behind the meter will be a big help to utilities, regulators, and economists seeking to better understand the underlying reasons why electricity sales growth has been in decline and whether or not this trend will continue, has reached bottom or is going to begin to reverse.
To this end, DNV GL has collaborated with Oak Ridge National Laboratory (ORNL) and EPRI to develop a whitepaper, End-Use Data Development – It’s time to Better Understand our Customers. In the whitepaper, we outline a construct for developing regional collaborative studies to improve our understanding of the end-use make-up of our customers. We are driven by our vision of an industry rich in meaningful data and insight on how and when consumers are using electricity.