Leveraging data for demand response programs
- February 10, 2016
- 839 views
By Tom Kerber
In Q2 2015, 16 percent of all U.S. broadband households owned one or more smart home devices. Many of these smart devices offer energy management capabilities including demand response. Today, utilities are evaluating the benefits of integrating smart devices into utility programs.
There are several ways that utilities can leverage data to improve the success of DR events: leveraging data from smart thermostats, leveraging data from multiple connected devices, and leveraging data from smart meters.
Energy modeling and load shifting
One application that uses data from smart devices is energy modeling. The cost of heating and cooling a home is a function of its equipment efficiency, building efficiency, and resident operation. Companies use weather information, basic information about the home size and layout, and thermostat operating history to build mathematical models of equipment performance and building performance. They then use those models to determine the best unoccupied temperature for each day and automatically make small changes to the system.
Energy modeling can enhance the effectiveness of thermostat-based demand response programs. For instance, when a DR event is taking place, some customers manually override the thermostat set point changes. In many cases, the manual adjustments result in longer equipment runtimes during the demand response events. When occupants become uncomfortable and discover that their thermostat set points have been changed, a common reaction is to push back by choosing a set point that is lower than the normal set point. The AC system must run longer to reach this new set point and consumes more energy than it would have without the DR event. Traditional thermostat-based demand response programs have override rates as high as 10 percent.
Load-shifting solutions use a combination of physical and behavioral models to define the pre-cooling strategy and event strategy. Pre-cooling uses the thermal mass within the home as a thermal battery storing energy that will then be used to cool the home during the DR event. The combination of the physical and behavioral models radically improves all metrics. For utilities that are considering installing and managing smart thermostats, energy modeling can maximize the return on that investment.
Leveraging data from multiple smart devices
Smart devices such as smart thermostats, lighting, water heaters, electric vehicles and EV chargers, energy storage systems, solar power inverters, and other Internet of Things assets can all be connected to DR programs.
Predictive data analytics across all devices can be performed on an individual house level and then aggregated for DR events. Companies such as Autogrid provide the utility with the ability to view customer profiles on a map of its territory, including areas with high demand. It can then use that information to target DR calls down to the substation level and device type. The platform can create groups by substation, by residential load profile, or on the individual device level. For instance, one neighborhood has a high concentration of pool pumps and electric vehicle charging stations. Another neighborhood has a high concentration of air conditioners and electric hot water heaters. One DR event can cycle off the largest energy consumer for each neighborhood.
Leveraging smart meter data
Disaggregation is another application that uses data from smart devices. Disaggregation of residential meter data from a single whole-home data point into a listing of devices in the home and their associated consumption can be accomplished using different technical approaches. The first approach uses existing smart meter data stored in the meter data management system, which generally stores 15-minute interval data. This is a huge improvement over one reading per month. And from a disaggregation perspective, 15-minute data also makes it possible to determine the top three loads in a home. Collecting 15-second interval data allows the disaggregation of 10 appliances. Sampling the data at higher data rates, say at a kHz sample rate, allows disaggregation analytics engines to see harmonics and use those to break out even more devices. Collecting the real and reactive power components also provides more variables to more precisely identify individual loads.
Utilities can use disaggregation to enhance the effectiveness of DR programs by pushing notifications with specific behavioral changes that will provide the greatest amount of energy reduction for the specific house.
Connectivity and data analytics provide many new opportunities for enhanced energy management capabilities. As more smart devices with energy management features enter the home, utilities will have to decide where they see themselves fitting into the smart home environment. Will they partner with device manufacturers or service providers? Will they compete in the market and sell or distribute their own smart devices or platforms? Ultimately, using smart device data to achieve personalization will be key to the success of utilities.