Many utilities and “virtual generation” companies offer Demand Response programs to customers. Demand Response (DR) programs help customers move their electric use out of high-cost system peak times and into the evening or weekend when more electric supply is available at a lower cost. This virtual capacity at system peak times reduces the need to build peaking generation.
Organizations that offer Demand Response programs want to know how much energy they reduced during peak times. The benefits of these peak reductions to the system are compared to the costs of the program to determine if there are net benefits from the effort. In other words, is the program cost-effective?
Determining program net benefits is less straight-forward than one may think when first considering the question. A major complication is that the key benefit, peak reduction, is not easily measurable. While we can use meters and loggers to measure energy use at any given time, to measure peak reduction we have to estimate what the level of use would have been without the program and then compare it to the actual level of use and figure out the difference. Mathematical modeling methods are needed to estimate what energy use would have been. Also, measuring the energy use of a specific appliance can be very expensive. For these reasons and others, sampling is often a big part of measuring peak reductions.
The need for sampling speaks to the need for skills in statistics, survey methods and data collection as well as an understanding of how energy is used or changed. All of these components are usually part of any Demand Response program evaluation effort. Individual evaluation projects that Daniel or Mary have worked on are listed on this page. Click on the project name to get more details on what they did.
Direct Load Control Assessment. Prepared program design recommendations and completed net benefits tests for a statewide Residential Direct Load Control program in New Jersey. In addition to developing a net benefits calculator in Excel, DSMore software was used to make a probabilistic assessment of future market prices for electricity based on the known variability of weather in the region. Daniel reviewed new control and communication technologies that could potentially be used to lower the cost of Direct Load Control programs. (2007)
Direct Load Control Program Evaluations. Mary led many impact and process evaluations for Direct Load Control programs. This work included programs in Wisconsin, North Carolina, Texas, Missouri, Kansas, Pennsylvania, Maryland, New Jersey and Delaware. One of Mary's DLC studies was cited in a California regulatory proceeding as being one of a set of “good examples of the application of various evaluation methods” for demand response programs. (2004 through 2011)
Good examples of the application of various evaluation methods: http://www.calmac.org/events/FinalDecision_AttachementA.pdf
Daniel prepared the logger data for these studies. This included automated downloading and compiling of the individual logger files which saved a great deal of time, as well as validation of the readings to make sure the recorded kW usage was reasonable given the size of the air-conditioner being monitored. For studies with 5-minute data, Daniel developed a moving template method to identify individual responses and start times of randomized-start control events.
KCPL MPower Curtailment. Mary evaluated annual impacts from the Large Customer MPower Curtailment Program for Kansas City Power and Light for three years. Daniel prepared the data. (2009 to 2011)
California Self-Generation Incentive Program. Mary performed data-mining and statistical analysis for a performance investigation into California's Self-Generation Incentive Program for combined heat and power facilities. (2010)
PECO Conservation Voltage Reduction. Mary contributed to the development and application of an innovative method for evaluating impacts from a Voltage Reduction Program for Philadelphia Electric Company (PECO). Daniel prepared, merged and cleaned five-minute SCADA data for the analysis. This required understanding substation schematics so a database could be built that correctly accounted for all power flows and relationships between the components and transducers at each substation in the study. (2010)
Puget Sound C&I Demand Response Pilot. Mary managed an evaluation of Puget Sound Energy's Commercial and Industrial Demand Response Pilot which included both impact and process evaluation of two seasons of load curtailment – summer and winter – since PSE is a winter-peaking utility. (2009-2010)
CPUC DR Load Impact Protocols. Mary provided analytical support for the California Public Utilities Commission proceeding R.07-01-041 to develop protocols for demand response load impact evaluation. This included thorough review of all documents related to the development of the protocols that were filed by the utilities and other parties, preparing summaries of issues and differences between parties, and helping prepare staff reports. (2007)
California Flex Your Power Program. Daniel used regression analysis to study impacts of the Flex Your Power program, a program that sent peak power alerts to radio and other broadcast media, asking customers to voluntarily reduce electric use temporarily. He combined and analyzed summer usage data for all Residential customers of the three largest California utilities as part of this project. (2007-2008)