Estimating Missing Interval Data – How Best to Fill Gaps
Almost everyone doing load research does some type of load data editing, usually filling gaps in the data or correcting clearly erroneous intervals. To assess several different methods of estimating filled values, we artificially created missing data of various gap lengths using actual 15-minute residential load data. The data methods examined include Mean and Linear Interpolation techniques, estimating missing values using day type averages, and three methods using the EXPAND procedure built into SAS. Because we have the original data for the missing intervals, we were able to assess the accuracy and bias associated with each filling technique across gaps of different lengths. We summarize how the various methods performed for different gap lengths, and make recommendations for practitioners dealing with interval data problems.
Smart Grid Pricing Pilots – What Did We Get for Our Money?
The Federal government made a large investment in Smart Grid Investment Grant (SGIG) Consumer Behavior Studies funded through the DOE. In this report, we attempt to answer the question: “What did we get for our money?” We review SGIG Consumer Behavior studies that are currently being conducted or have recently been completed. These studies examine the response of residential and small commercial customers to time-based rate programs that are implemented in conjunction with the deployment of AMI and customer systems such as IHDs and PCTs. The report includes brief descriptions of the pilots being conducted, highlights key findings from pilot experiences to date, and provides a set of recommendations based on the lessons learned from these pilots.
Domains Analysis – Slicing, Dicing, and Combining Load Research Samples
In this paper we describe the technique to estimate load shapes for groups of customers, or domains, that differ from the groups for which samples were originally designed. These domains can include future rate classes as well as subsets of populations, such as those with and without an end use (i.e. central AC or electric water heat) or demographic groups, to name just a few. We explain the method, and then walk through an example, using actual rate class load data, that creates hybrid rate class load shapes based on the combination of the rates from two different service territories.
Long-Run Savings from Energy Efficiency Measures
This report focuses on codes and standards, both current and future, and on widely accepted overall trends in technology and EE savings to provide a picture of how EE might look in the long-run. It first discusses how codes and standards are developed and describes the current efficiency standards for some common residential and commercial equipment. Then, it moves to the future to investigate how efficiency levels might change and how industry professionals can use this information to adapt their own forecasts.
Fast DR: Demand Response and Ancillary Services
This report explores the relatively new concept of using of Fast-DR resources as an ancillary service which can be integrated into ISO and RTO wholesale markets on par with generation. We investigate the different types of ancillary services, and how Fast-DR can provide those services to the market. We also summarize the recent Participating Load Pilot in the CA ISO taking a close look at results, barriers, and lessons learned.
Price Elasticity of Demand in Econometric Models
The own-price elasticity of demand for electricity is notoriously difficult to estimate in econometric models outside of carefully structured studies. However, elasticity is often of interest, especially to management. In this report, we will explore why price elasticity is so difficult to estimate and suggest viable solutions for forecasters and modelers.
The Brave New World – Developing and Evaluating Integrated DSM Programs
Faced with options ranging from energy efficiency (EE) programs to demand response (DR) programs to new rate structures, how are consumers to decide what is best for their homes and businesses? From the utility perspective, the emerging world of smart grid technologies and devices all must work in an integrated fashion to incent appropriate customer behaviors and deliver reliable demand-side resources. In this emerging landscape, it no longer makes sense to think of energy efficiency separately from demand response, and utilities and their regulators are now moving toward integrating these programs to take advantage of synergies and better serve customers. We suggest ways the industry can address these obstacles, including modifying regulatory goals and structures, integrating programs and incentives, educating consumers, and implementing codes and standard that advance not only EE but DR as well.
Demand Response – It’s a Resource, So Treat It Like One!
This report focuses on demand response (DR) as a resource and how impacts should be counted. The primary focus of the report is how DR events are treated in cost-of-service allocation, but it also includes program design and load forecasting/resource planning for context and background. We talked with ten utilities across the country to explore their experiences and thoughts on the counting of DR impacts. We also used actual data to examine, through in-depth examples, the consequences to cost-of-service of counting DR impacts in different ways.
Business Strategy: Impact of PHEV’s on Utility Load Analysis
This report explores some of the many ways that the increasing number of plug-in hybrid electric vehicles (PHEVs) will affect the utility industry. As they become more prevalent, PHEVs will impact customer energy use, load shapes, and utility system operations. If increased electricity usage is not managed through pricing and grid control, PHEV charging has the potential to negatively impact system peaks. On the other hand, if managed appropriately, it can increase off-peak sales and load factors, which are both beneficial to utilities.
Best Practices: OG&E Weather Normalization Case Study
This report explores the hourly weather normalization model developed by OG&E and Itron to address a PUC mandate to use the system peak hour for cost allocation. Most of the time, daily or even monthly weather normalization is adequate; however; due to a mandate from the public utility commission (PUC) for a one coincident peak (1CP) allocator, OG&E chose to develop an hourly weather normalization model in record time.
Permanent Load Shifting with Ice Bear
This document looks at the Ice Bear technology and how it enables commercial customers to shift load to off-peak hours. It also takes a look at two utility programs that use the Ice Bear technology.
Load Forecasting Workshop Highlights
This Update presents highlights from our second annual Load Forecasting Workshop held October 6–7, 2008 in Boulder, Colorado. The theme of this year’s workshop was integrating demand side management program impacts into energy and demand forecasts.
Load Forecasting Salary Survey
This Update and the accompanying PowerPoint document present the results of the Load Forecasting Salary Survey. Subscribers can use these results to benchmark salaries and see how they compare with other utilities.
DR Policy, Planning, and Programs – Highlights from 2008 Spring/Summer Conferences
This Update presents conference highlights related to demand response (DR) policy and program design and implementation. The industry conferences held during the spring and summer of 2008 included a wide variety of presentations. Instead of summarizing the presentations from each conference separately, we have chosen to group the conference highlights by topic, focusing on one area for this Update.
Can’t We All Just Get Along? IT and Load Research Departments Play Nicely in the Post-AMI World
AMR and AMI implementations challenge utilities by changing the way meter data is collected and analyzed and by forcing changes in the data storage, retrieval, and archival processes. This report pays special attention to the way load research and IT departments interact to address the challenges presented by collecting, accessing, and validating the much larger quantities of interval data gathered by AMR and AMI systems.
Who Has What? Predictive Modeling Using Customer Billing Data
Predictive modeling uses information that is known for a sample, combined with information that is known for the population overall, to create a statistical model for predicting something about all members of the population. This report first discusses three approaches to predictive modeling. Then it provides two case studies involving utility residential customers and showing how this approach predicted the presence of central air-conditioning in one case and electric heat in another.
What’s New with Demand Response
In this Update, we provide highlights and note trends emerging from the U.S. Demand Response Coordinating Committee’s (DRCC’s) fourth National Town Meeting on Demand Response, held in April 2007. We also discuss the growing use of backup generators to provide demand response (DR), including programs run by utilities and third-party aggregators. Finally, we summarize a couple of recent noteworthy reports on DR.
Energy Efficiency and Demand Response: Separate Efforts or Two Ends of a Continuum?
The typical approach to energy efficiency and demand response is to identify technologies and programs to pursue on a piecemeal basis. Utilities and regulators rarely assess how these programs interact with one another, which can limit opportunities and, hamstring implementers, and confuse customers. In this report, we advocate looking at these programs in an integrated fashion that considers energy company objectives, the customers’ needs and wants, and how to achieve better outcomes.
Managing AMI Data
With the installation of automated meter reading (AMR) and advanced metering infrastructure (AMI) systems comes the promise of more and better customer information, particularly interval data. But are these systems delivering on those promises? We share insights gathered from utilities that have installed AMI systems and provide recommendations on how load researchers should get involved in the AMI decision-making and implementation process.
Gas Load Forecasting
How can gas load forecasters overcome challenges due to volatile prices, “demand destruction,” and insufficient historical data for modeling price elasticity? This report offers insights garnered from discussions with forecasters, as well as suggestions for improving business processes.
How Hot Was It?
“How hot was it?” – or more appropriately perhaps, “How hot will it be?” – is more than a rhetorical question when it comes to forecasting energy use or developing load shapes. This report reviews the various methods analysts use to develop typical and extreme weather year data and provides insights on their pros, cons, and applications.
Load Impacts of Distributed Energy
New applications for distributed energy have been emerging in both the residential and C/I sectors. This report reviews these technologies and analyzes how their increased use could affect individual and system loads.
Planning and Design Issues for Residential Direct Load Control Programs
Ensuring a new demand response program will cost-effectively deliver sustainable savings requires methodical planning. This report details steps for successful program design, using four load control programs as case studies.
The Impact of Tankless Water Heaters
Tankless water heaters are growing in popularity, thanks to their greater efficiency, compact size, and promise of “never running out of hot water.” However, the high demand of some electrical units raises potential issues for utility distribution systems.
Distribution Equipment Sizing
Applying load research data – beyond just the peak load – enables properly sizing distribution system equipment to reduce costs and insure reliability.
Comparing Load Shapes
Two load shapes may appear different, but from a statistical point of view, are they really different? How much of a difference matters?
Residential Air Conditioning: Impacts and Responses
U.S. air conditioning use continues to rise, with repercussions upon load shapes, peak demands, load forecasting, and distribution systems. What options do utilities have for addressing this trend?
Interval Data Over AMR
Automated meter reading (AMR) has the potential to allow cost-effectively collecting interval data, but can also create problems for load researchers. This report outlines how to take advantage of AMR’s potential benefits while avoiding its pitfalls.
Building Simulation Tools
Today’s building simulation tools, with user-friendly interfaces and dynamic defaulting, can be applied to a wide range of load research and forecasting functions.
Elasticity and Time-of-Use Pricing
We consider the economic concept of elasticity, which describes how changing prices affect product demand, and discuss its application to energy use, particularly time-of-use and critical peak pricing programs.
Education Sector Energy Use and Characteristics
This reference provides data on energy use and market characteristics for primary / secondary schools and colleges/universities. It includes information on facility characteristics and age, load shapes, decision making, and interest in energy services.
Characteristics and Energy Use of Restaurants
This report is a reference guide to the restaurant segment, with data on market characteristics, energy use, load shapes, and end-use equipment.
How can you identify your company’s most valuable customers? We describe a model for prioritizing customers and consider a case study.
Measurement and Verification for Demand Response
This report introduces the types of demand response programs in the industry, contrasts demand response with traditional load management and demand-side management programs, and provides an in-depth study of measurement and verification for demand response.
Puget Sound Energy and Residential Time-of-Use Rates – What Happened?
In this report, we explore what went wrong, what went right, and what we can learn from Puget Sound Energy’s residential time-of-use rate, which was offered to customers from May 2001 to November 2002.
This report highlights selected presentations from the AEIC Load Research conference held in July 2002 in Las Vegas, NV, and the WLRA Fall meeting, held in September 2002 in Portland, OR.
Benchmarking the Load Profiling Process
Across the U.S., many utilities are engaged in load profiling for settlement and reconciliation in open energy markets. This Technical Brief summarizes what processes and systems they’re using.
The Impact of Consumer Electronics on Household Electricity Use
Residential electricity use per household continues to increase despite stringent efficiency standards for many major household appliances. How much energy do consumer electronics use and how much will they use in the future? We believe this is an important question so we set out to answer it in this report.
Using U.S. National Energy Forecasts to Supplement Utility Forecasts
In this report, we summarize the results of the AEO 2002 forecast and several key issues facing energy markets. Also included is a guide to information resources.
Weather Response and Profitability
This report shows that the relationship between weather and profitability is based on how weather influences energy demand and how load affects supply price. Our analysis is based on an assumed relationship between weather, load, and supply price; and on historical load data, weather, and hourly supply price in California.
What’s So Good About Better Load Profiling?
Better load profiling is discussed from the perspective of the different stakeholders in the market.
Simple Mantras for Maximizing Savings
In this paper we present six recommendations from our evaluators’ unique perspectives that can help your utility increase the impact of energy efficiency programs. These recommendations incorporate simple strategies to bridge the savings gap from projected savings during program design to adjusted savings during program evaluation. Throughout the paper, we share several examples to illustrate successful as well as unsuccessful practices in the business of energy efficiency.
Residential EMP Update
We have updated our Residential Energy Market Profiles (EMP). The EMP provide summary information about U.S. energy use for major customer segments by end use and fuel, and are a key resource for forecasting, marketing, and program planning. This report summarizes the information in annual spreadsheets for 2009-2013 delivered with the report.
The Evolution of Peak Time Rebate – Past, Present, and Future
This report explores the evolution of PTR by reviewing the many published studies and evaluations that have tested PTR over the last eight years, and by talking with those in the trenches implementing and piloting PTR. We provide a literature review and summary of the existing pilots that tested PTR rates. Next, we dig into some of the issues that have been called out recently, including baselines and customer engagement, by bringing together primary and secondary research in the field. Finally, we offer some observations and recommendations related to PTR and its implementation as a residential pricing program.
Innovative Uses of Smart Meter Data – Supercharging Data Analytics
In this report we discuss smart meter deployments, the challenges utilities have faced during the grid upgrade, and explore the benefits that have been and could be realized. We split our discussion of the applications of smart meter data into two groups, established practices and innovative applications. We describe several of each type, and include some ideas that have yet not been tried, but have the potential to add new value.
Integrating DSM into Energy Forecasts – Issues and Potential Solutions
This report addresses the topic of integrating Demand Side Management (DSM) impacts into load forecasting models. In general, we consider the entire continuum of demand side activities part of DSM, including demand response (DR), energy efficiency (EE), and permanent load shifting. However, specific details presented here primarily focus on the more challenging task of integrating energy efficiency into forecasts. First, we explore the issues surrounding why it is important to integrate DSM impacts into load forecasting models. Second, we address the more traditional methods currently used in the industry to integrate DSM impacts and the pros and cons of each method. Finally, we look to those in the industry using new or innovative methods to integrate DSM impacts.
Residential End-Use Metering – Can AMI Get In the Door?
This report focuses on end-use metering in the residential sector and how the advent of AMI might enable more end-use studies in the future. We also address some of the technical challenges utilities face in end-use metering and how AMI may or may not address those challenges. We gathered information and opinions on end-use metering from both utilities and vendors through in-depth interviews. We also asked both vendors and utilities about the technical challenges that we will face integrating AMI and smart meters with end-use metering. Finally, we include two case studies of end-use metering projects currently being conducted in the industry.
AMI Implementation Update: Load Research Using AMI Data – Are We There Yet?
For this report we interviewed individuals at utilities that are in the process of – or have just completed – installing AMI systems throughout their service territories. We provide insight into their experiences installing and implementing their AMI systems, as well as their experiences collecting, validating, storing, and using interval data for load research.
Putting the “Control” in Direct Load Control
This report examines the results of five industry studies testing various curtailment strategies. Many residential direct load control (DLC) programs are operated via switches installed on customers’ air conditioners, but with increasing acceptance of programmable communicating thermostats (PCTs), many utilities are offering PCTs as alternatives to switches. The utility industry sometimes looks at the load impacts of different curtailment technologies on DLC programs since both the magnitude and the shape of the impact will differ based on technology. We go one step further. We examine the impact on energy savings, customer comfort, and customer acceptance different cycling strategies, for switches, and reset strategies, for thermostats.
Peak Time Rebate’s Dirty Little Secret
This report reveals Peak Time Rebate’s (PTR’s) potential “dirty little secret” – many customers will get paid for doing nothing, based on normal random variations in their electricity use. PTR programs are known as carrot-only or win-win programs and are more easily approved by commissions; but while the program is good for customers it can pose problems for utilities.
Energy Efficiency Provisions in the American Recovery and Reinvestment Act of 2009
This Perspective discusses the American Recovery and Reinvestment Act (ARRA) of 2009 signed into law by President Obama, which has created quite a buzz in the energy industry. Many are calling the bill the green new deal, which it may be, considering it contains more than a $40 billion investment in the energy industry geared toward clean, efficient, and independent energy for America. The spending is focused on renewables, energy efficiency, smart grid, and alternate fuels.
Beyond the Bill: In-Home Displays Deliver Energy Savings
The fourth in a series of reports on in-home displays, this report examines the impact on residential electricity consumption of in-home display units. We summarize the results of three industry pilots and look at the results achieved by a utility program that incorporates an IHD. We also address the implications of these findings to both the residential electricity market and to utility companies considering IHD programs.
Residential EMP Update
We have updated our Residential Energy Market Profiles (EMP). The EMP provide summary information about U.S. energy use for major customer segments by end use and fuel, and are a key resource for forecasting, marketing, and program planning. This report summarizes the information in annual spreadsheets for 2001-2007 delivered with the report.
To Call or Not to Call: When to Call Demand Response Events
Residential demand response programs come in many types and sizes, but they all have one thing in common: targeting demand reduction on particular event days. It is vital to carefully consider how event days will be called to maximize the probability of calling on the highest load days and achieve the largest load impact. We examined seven fully deployed demand response programs and two pilot programs across the United States and Canada to uncover a variety of methods for calling and allocating event days or hours.
Load Research Salary Survey
This Update and the PowerPoint document present the results of the Load Research Salary Survey. Load research requires a unique set of skills and experience that are not typically required elsewhere within the utility or in industry in general. As a result, human resources departments frequently have a difficult time correctly assigning load research positions to a predetermined set of utility job descriptions or pay grades, or even to compare with similar positions in other industries.
What’s Up in Load Research
This Update provides highlights from the Western Load Research Association (WLRA) biannual conference held in late September in San Francisco. This Update also provides commentary on the recent acquisition of LODESTAR Corp. by Oracle and how it will affect LODESTAR users and future customers.
Load Profiling During Extreme Conditions
Industry experts give their perspectives on load profiling for extreme conditions.
Time-of-Use and Critical Peak Pricing: Considerations for Program Design
This report identifies key considerations for residential time-of-use (TOU) and critical peak pricing (CPP) rate design. It discusses how enabling technologies such as programmable communicating thermostats can support these rates. It summarizes results of existing program evaluations and provides insights from interviews with 11 utilities that have implemented pilots or programs.
FERC’s Assessment of Demand Response and Advanced Metering
This Perspective examines and summarizes the Federal Energy Regulatory Commission (FERC) report “Assessment of Demand Response and Advanced Metering,” which the FERC prepared as mandated by The Energy Policy Act of 2005.
National Town Meetings on Demand Response: Highlights and Analysis
This Update provides highlights from a series of national “town meetings” on demand response (DR) to foster information sharing among the leading organizations involved in implementing DR programs. This report focuses on the third town meeting but also includes related insights from the earlier sessions.
Energy Market Profiles 2006
The Energy Market Profiles (EMP) provide a snapshot of annual energy use by end use and fuel and region for major customer segments in the commercial, residential, and manufacturing segments. These spreadsheets are both a valuable reference for quickly defining energy-use patterns for individual customer segments and provide a solid foundation for a variety of market analysis tasks, including program planning and evaluation.
WLRA and AEIC Conference Summaries
This Update shares our insights from the Western Load Research Association Spring 2006 conference and the Load Research Committee of the Association of Edison Illuminating Companies Annual Workshop, where presentations highlighted load research practices and policies, demand response, pricing, and weather analysis.
Energy Use Update: Smart Metering Mandates in the Energy Policy Act of 2005
This Update investigates the implications of Section 1252 of the Energy Policy Act of 2005, including the requirement that, within 18 months of the law’s enactment, utilities must offer customers the option of being on a time-based rate.
Load Research Sample Design
By making informed decisions during the sample design process, load researchers can help ensure that a study achieves its objectives and cost-effectively provides valid results. This report discusses the issues affecting sample design and includes illustrated, step-by-step procedures, based on an actual load research sample design.
Energy Use Update: Real Time Energy Feedback, Enabling Technology for Demand Response
This Update provides highlights from a recent conference on the effects of providing energy use feedback to customers, available technologies for doing so, and pilot programs that are putting feedback to work. We also look at the role of enabling technology in demand response programs for small businesses.
Demand-Side Options for T&D Relief
Faced with high costs for expanding T&D assets in certain locations, utilities are experimenting with using demand-side resources to provide targeted relief to constrained T&D systems.
Trends in Residential Energy Use
Spurred by ongoing population growth and new electricity-consuming appliances, U.S. residential energy use continues to climb upward. This report drills down through the data to understand the trends over the past 20 years, and how energy use is likely to change between now and 2010.
New Options for Interval Load Data Collection
New choices are emerging for interval data collections tools. This report reports on the evolution of these tools and shares insights from interviews with the major companies offering products for the North American market.
Best Practices in Load Research
We summarize load research practices, used to support regulated cost-of-service based ratemaking, and highlight activities at utilities that represent a cross-section of the industry.
My Meters Can Do That?
With recent technological advances in metering and communications, load research samples are an untapped goldmine of valuable data and corporate innovation. In this report we investigate how – with no additional cost or with a relatively small additional investment in new meter technology – the load research unit can provide information to support system management, customer relations, and regulatory obligations.
Gas Prices – What to Do?
After a wild ride since February, natural gas prices have dropped somewhat, but are still well above the level of the last few years. How will this affect energy end users? And what does this mean for energy utilities? This perspective synthesizes information from a wide range of sources to examine impacts and discuss what utilities can do.
Load Shapes for Small Commercial Customers
This report draws on our EnergyShape™ database to provide typical-day load shapes and related information by segment and region for small commercial customers.
The Load Profiling Forum – What Have We Learned?
This report is a retrospective of the Load Profiling Forum, which covers what the industry has learned about load profiling topics ranging from segmentation to benchmarking – and how we can apply this information to the broader practice of load research.
Energy Market Profiles – A Reference Guide
This report provides an overview of the Energy Market Profiles for each sector and describes the sources and method for data development. It provides an introduction to the Energy Market Profile Viewer, an online application that provides convenient access to all the EMP data.
You’ve Got a Great Product! Now Who’s Going to Buy It?
Market research and market assessment are necessary for new product development. Learn how targeted market research provides both qualitative insights into customer response and quantitative information for market assessment.
A Visual Overview of U.S. Commercial-Sector Energy Use
This report graphically summarizes U.S. commercial-sector energy use by region and customer segment.
Sample Design for Load Profiling
This Technical Brief describes the sample design process for data collection studies to support load profiling.
Energy Use Outlook 2001
At the Energy Use Outlook Conference 2001, major energy players offered diverse views on the future of energy-use analysis and information. This report summarizes the conference presentations and the lively roundtable discussion on load profiling.
The Role of Load Profiling in Load Curtailment for Mass Markets
Learn about load profiling and load curtailment for residential customers in regulated and deregulated environments. This Technical Brief details the situation.
Class Averages and Individual Customers
In a new approach to recognizing volatility, this report looks at methods load researchers use for tasks that involve class averages vs. those that need an understanding of the loads of individual customers or small groups of customers.
|Back to Top
|Spring 2011 WLRA-AEIC Conference Summaries, April 2011
(GEP-LAMS-M-001) Google PowerMeter Memo, July 2011
|Summary of IEE Whitepaper: Assessment of Electricity Savings in the US Achievable through new Appliance/Equipment Efficiency Standards and Building Efficiency Codes – 2010-2025,
November 2011 (GEP-LAMS-M -003)
|Back to Top
Reports in Progress
Calculating Energy and Demand Impacts from Behavioral Programs
Many utilities are now offering some type of behavioral or informational programs to their residential and small commercial customers. These programs might include periodic reports, such as an Opower program, or may include access to Smart Meter usage data via a web portal. In most cases, utilities are making an effort to quantify the energy and sometimes demand savings related to these programs. In this report, we will provide an overview of some of the methods that can be used to estimate savings, and address some of the savings related issues that often arise. For example, behavioral programs are often associated with or encourage participation in other utility EE or DR programs, but separating out the effects of the behavioral program from other utility programs can be tricky.
Designing Pilots to Get Results
From dynamic pricing programs to behavioral programs, utilities across the country are offering customers more choices than ever, but deciding what combination of technology, rates, and information will optimize results for your customers can be challenging. This is why, before implementing a full-scale program of any kind, utilities often design pilots to test program design aspects, marketing strategy, pricing schemes, and estimate customers’ response. It is critical that all aspects of the pilot program are carefully considered as the pilot is designed so that the results are valid and can be applied to the population at large. In this report, we will discuss the aspects of designing pilots that help ensure statistically valid results. We will also share our experiences designing and evaluating utility pilot programs.
Deciphering Changes in Customer Energy Use
From Plug Loads to Personal Income – Both load researchers and load forecasters see that the way residential customers use energy is changing. Inside the home lighting and big ticket appliances are becoming highly efficient, while a larger percentage of total usage than ever is falling under the umbrella of “digital” or “miscellaneous” loads. Outside the home, customers are responding to the state of the economy, higher energy awareness, and a new flood of utility programs aimed at conservation. In this paper, we will explore changing residential usage patterns from a load research point of view, focusing on end-use related drivers, and from a forecasting view, looking at how customers are interacting with the external economic and utility signals to change their consumption habits.
|Back to Top
Fixed Effects Can’t Fix Everything: Working with Panel Data
In the utility evaluation world we often work with panel data, or datasets that consist of multiple observations from many customers over a period of time. In general, the accepted approach to building models with this type of data is to use a fixed effect model to capture and isolate customer specific effects without having to include (or obtain) detailed demographic and appliance related information for each customer. While using a fixed effect model can be very useful, there are many situations where a fixed effect simply cannot capture all the differences between customers, which can cause problems. In this report, we explore some common issues that can arise when working with panel data, and discuss some potential solutions through our experience using illustrative case studies.
EV Charging Load Shapes: What Do they Look Like and What Are their Implications?
Electric vehicles have been a hot topic for some time now, but what do their charging load shapes actually look like, and how do EV owners respond to price based incentives to charge off peak? In this report, we will take a closer look at electric vehicle charging load shapes under different pricing and charging scenarios. We will also explore the correlation between EVs and rooftop photovoltaics, and how charging behavior might be different for those with PVs and those without.
The goal of a utility forecast is to estimate, as accurately as possible, future sales and load. Often, this involves a single “base case” prediction, potentially with “high-growth” and “low-growth” alternates. This type of forecast is easy to use and understand, but doesn’t provide much insight into the possible variation or uncertainty in future outcomes. Instead of saying “next year’s peak demand will be X MW,” we can provide a probability distribution that recognizes the uncertainty of all the inputs to the forecast (weather, economics, etc.) and the uncertainty of the forecast model itself. This distribution can be used to consider the risk associated with various future resource options. We’ll take a look at what those who integrate probability distributions into their forecasts are doing, both at utilities and in other industries, and discuss methods to create probabilistic forecasts.
Load Forecasting Analysis Software
There are now several options for software tools to perform forecasts and other econometric analyses. We’ll take a look at what is available in the market, check in with the vendors of these products, and talk to users to get a sense of the real capabilities of the tools and their ease of use. Based on this feedback and our experience, we’ll provide unbiased reporting on the strengths and weaknesses of each of the tools.
Understanding Power Quality
Beyond simply understanding how much apparent power (kWh) is consumed by a customer, utilities are also interested in understanding more about other aspects of power quality, including power factor, real power, harmonics, and the like. We’ll look at what some utilities are doing to measure interval data beyond kW and kWh, why they are measuring it, and what they are doing with the results.
Forecasting for Transmission and Distribution Planning
Load forecasts are important inputs when sizing future transmission and distribution systems and can be used to evaluate the need for improvements to current infrastructure. In this report we will take a closer look at the role of load forecasting in transmission and distribution planning. We will look at different models being used and possible new applications of forecasting techniques specific to T&D.
Targeting Residential Customers for Dynamic Pricing
As utilities move from pilot projects into the realm of full-deployment, designing recruiting strategies for dynamic rates is a real concern. Ideally the utility wants to target not only customers willing to participate in the program, but customers who will actually shift load and provide real demand response. Many evaluators in the industry use propensity models to design matched control groups for evaluation, but what about using a propensity model to target potential participants that have the ability to shift load and save money on a dynamic rate based on interval data? In this paper, we will use data from a utility in their second year of full-deployment to attempt to identify savers in the population using a propensity model that can then be applied to the entire residential population.
Low-Income and Low Usage
Many utilities offer baseline or lifeline rates, which offer lower prices for the first block of electricity and/or natural gas for some or all customers. Other utilities discount prices offered to qualifying low-income customers. But how do energy use and load shapes for low-income customers compare with the rest of the residential population? Policymakers don’t generally consider these factors when implementing income-based rates, but perhaps they should.
How Well do Enabling Technologies Enable?
Many utilities are now using various technologies to enable price based DR programs like dynamic pricing, and demand/capacity bidding. We will investigate the logistics of these technologies and review current pilot programs to uncover both effects on load impacts and best practices in this evolving area.
Analysis and Expansion Primer
As a sequel to our 2005 Load Research Sample Design Primer, we’ll provide a detailed description of analysis and expansion methods to use once data collection begins. We’ll again include a step-by-step example, walking you through the process of estimating class load shapes from a stratified sample.
Sample Rotation – How Important Is It?
While industry best practices call for replacing load research samples frequently, about every three years or so, most load researchers are unable to do so because of budgetary or time restrictions. We know that the accuracy of samples degrades over time, and the potential for bias exists because new homes don’t have the same load characteristics as older homes. But what is the impact of leaving samples in the field longer? What are the benefits and limitations of techniques like post-stratification and reweighting? And how do developments such as AMI/AMR impact this question?
Can DR Fill in the Gaps? Addressing the Variability of Renewable Generation
Renewable energy is becoming a larger part of the utility generation portfolio, driven by regulatory mandates and global climate change concerns. A primary challenge to the adoption of two of the leading renewable sources, wind and solar, is their variability. One potential solution to fluctuating output is demand response. Can DR act as a reliability tool and potentially supplant the need for fossil fuel generation to back up wind and solar? How large a role can demand response programs play in the integration of variable renewables into the grid?
|Back to Top
For More Information Please Contact: