Frequency control via demand response in smart grid

In order to have a reliable microgrid (MG) system, we need to keep the frequency within an acceptable range. However, due to disturbances in a MG system (such as a sudden load change), it can experience major or minor deviations in frequency, which needs to be reduced within seconds to provide the system stability. In order to maintain the balance between energy supply and demand, traditionally, generation side controllers are utilized to stabilize the power system frequency. These systems add high operational cost, which is not desired for power system operators. With the introduction of smart grid, more and more renewable energy sources are to be used in the power system. The intermittent behavior of these energy resources, as well as high operation cost of conventional controllers, has led to research for new alternatives. In a smart grid environment, demand response (DR) programs can be considered as a promising alternative to the conventional controllers, to e ciently contribute to the frequency regulation by switching responsive loads on or o . DR programs can reduce the amount of energy reserve required and, hence, are more cost efficient. Moreover, they can act very fast and can provide a wide range of operation time from a few seconds to several minutes. Thermostatically controlled loads (TCLs) are proper candidates to participate in frequency regulation programs. However, individual TCLs do not have a noticeable impact on frequency due to small size. They should be aggregated in order to have a considerable effect on frequency. Nevertheless, there are still many challenges which should be addressed in order to make use of TCLs for frequency control in smart grid. In this regard, proper aggregated load models and control algorithms for TCLs contributing to this service need to be investigated. In this thesis, we present an aggregation model for TCLs and a control strategy to coordinate power provided from DR participants with that of generation side of the MG to keep system frequency within its desired range. For the aggregation model considered in this study, a state space model is used to take into account the interdependency of TCLs' temperature participating in DR programs. The model groups TCLs into clusters, each controlled by an aggregator. A minimum off/on period is considered for individual TCLs to avoid frequent switching of these devices. A control strategy is presented to control frequency by coordinating the generation and demand side