Using high frequency consumption data to identify demand response potential for solar energy integration

Abstract:

California’s renewable portfolio standard (2012) requires the state to get 33% of its electricity from renewable sources by 2020. Increased share of variable renewable sources such as solar and wind in the California electricity system may require more grid flexibility to insure reliable power services. Such grid flexibility can be potentially provided by changes in end use electricity consumptions in response to grid conditions (demand-response). In the solar case, residential consumption in the late afternoon can be used as reserve capacity to balance the drop in solar generation.

This study presents our initial attempt to identify, from a behavior perspective, residential demand response potentials in relation to solar ramp events using a data-driven approach. Based on hourly residential energy consumption data, we derive representative daily load shapes focusing on discretionary consumption with an innovative clustering analysis technique. We aggregate the representative load shapes into behavior groups in terms of the timing and rhythm of energy use in the context of solar ramp events. Households of different behavior groups that are active during hours with high solar ramp rates are identified for capturing demand response potential. Insights into the nature and predictability of response to demand-response programs are provided.