Managing load aggregations and distributed generation resources for peak shaving
University of New Brunswick
Power consumption variation during the day may lead to imbalances in energy supply and demand. To maintain the balance, utilities need to generate more anergy, typically at a higher cost. Demand side management is an effective solution that has been used by utilities to avoid utilizing more expensive, or fossil generation when higher demand is expected. The use of direct load control (DLC) with aggregation of high number of small loads has been successfully used to reduce electricity costs for energy companies. Virtual power plants (VPPs) are effective methods to strategically coordinate the use of multiple aggregators and small renewable resources. This study aims to find a method to reduce the electricity peak demand by VPPs optimally managing a fleet of aggregators, diesel generators, solar energy, and battery systems. In this thesis, a simple but effective threshold mechanism is used for peak load shaving. Diversity factors and genetic algorithms are used for identifying optimal scheduling and management of the load aggregations and resources. As a real example, the available capacity for peak shaving at Saint John Energy Company (SJE) is used. This capacity includes several load aggregators: electric water heaters (EWHs), heat pumps (HPs), baseboard heaters (BBHs), and electric thermal storage systems (ETSs); as well as other energy resources such as a utility battery (UB), diesel generators (DGs), and residential batteries (RBs). Two optimization scenarios are simulated with MatLab® taking into consideration common utility’s constraints and disregarding them to identify possible improvements in demand reduction.