Virtual Power Plant Contributing to Primary Frequency Containment Using Demand Response - A case study of a commercial building

Abstract: The Swedish power grid is experiencing an ever-growing challenge to efficiently balance the grid. An increasing share of intermittent generation, decreased system inertia, continued electrification and more distributed energy resources result in an increasing need for flexibility. Smart grid technologies have, amongst other applications, shown to be a possible solution to this increased need for flexibility. The virtual power plant, one of many smart grid technologies, uses a heterogenous portfolio of aggregated distributed energy resources, combines their respective characteristics efficiently, in order to create a single operating profile which can be used to offer services to the grid operator, such as system balancing. However, since their potential is relatively unknown, as with any smart grid technology, it is important to determine their prerequisites to offer balancing services to the system operator. In this thesis the potential for a virtual power plant to contribute to the balancing of the Swedish power grid is analyzed. The analysis is based on a case study of a commercial building which is the basis for the implemented virtual power plant. The virtual power plant, and the energy resources provided by the commercial building, are modeled in MATLAB and Simulink. The energy resources included are air handling units and chillers in a large scale heating, ventilation and air-conditioning system. The models are implemented by performing system identification or system estimation methods. Moreover, an energy storage system is modeled in order to compensate for the activation and deactivation sequences of the chillers in the proposed control scheme. A control system of the virtual power plant is proposed in order to efficiently control the assets and to participate on the primary frequency containment market FCR-N. The control system uses a regulation controller, implemented with filter to regulate the load of the assets. Furthermore, a scheduling unit is implemented, which together with a load corridor, schedules and places bids with the corresponding available capacity of the assets, on FCR-N. The scheduling also activates the assets and the regulation controller whenever frequency containment is scheduled. The performance of the virtual power plant is verified by simulating a prequalifying test for the balancing market, which is developed by the Swedish system operator and transmission system operator Svenska Kraftnät. The results of the prequalifying test proved that the air handling units were viable assets in regard to their responsiveness. The chiller also proved to be viable but given the proposed control system and the compensation provided by the energy storage system. Moreover, a simulation of the virtual power plant performing primary frequency containment based on historical data was simulated. This highlighted how problematic it is to quantify the performance of balancing resources. But despite this, the results indicated that the proposed implementation of the virtual power plant, based on the case study, and its control system could perform primary frequency containment and thus, contribute to increased flexibility in the Swedish power system.