Western Kentucky University
On the Development of a Faster Algorithm for Evaluating the Cost Effectiveness of Electrical Utility Demand Response Programs
Institution
Western Kentucky University
Faculty Advisor/ Mentor
Jeffrey Ward
Abstract
Electrical utility companies define a demand response program as one in which a consumer may opt-in to shed load at peak hours in order to curtail high demand. This program allows electrical utilities to avoid constructing new power plants, allows transmission lines to assure system reliability, and compensates the consumer for their efforts. As demand response begins to see more attention from electrical utilities, the need for developing faster algorithms to calculate the value of participation and the resulting reductions in loss of load probability (LOLP) arises. In this study, a few well-known algorithmic paradigms are used, each containing their own set of enhancements. In the end, a unique algorithm, which links together dynamic programming and a greedy search, rises above the rest in terms of speed (0.095s) and near optimality (3.9 - 0.0%). The increase in speed and minimal optimality gap allows energy executives to explore demand response opportunities in an efficient manner. Further explorations of pre-processed searches have the potential to provide an even greater optimality.
On the Development of a Faster Algorithm for Evaluating the Cost Effectiveness of Electrical Utility Demand Response Programs
Electrical utility companies define a demand response program as one in which a consumer may opt-in to shed load at peak hours in order to curtail high demand. This program allows electrical utilities to avoid constructing new power plants, allows transmission lines to assure system reliability, and compensates the consumer for their efforts. As demand response begins to see more attention from electrical utilities, the need for developing faster algorithms to calculate the value of participation and the resulting reductions in loss of load probability (LOLP) arises. In this study, a few well-known algorithmic paradigms are used, each containing their own set of enhancements. In the end, a unique algorithm, which links together dynamic programming and a greedy search, rises above the rest in terms of speed (0.095s) and near optimality (3.9 - 0.0%). The increase in speed and minimal optimality gap allows energy executives to explore demand response opportunities in an efficient manner. Further explorations of pre-processed searches have the potential to provide an even greater optimality.