The Strategy Of Predictive Generation And Flexible Load Control In Smart Grids

Nowdays, smart grid has faced numerous challenges, such as intermittent renewable resources, increased power demand, distributed generators and so on, which raise complexity and difficulty in maintaining a reliable power system operation. In addition, a population of flexible loads exist in the modern power system, which can be used to participate in the frequency and power regulation through direct or indirect control strategies. By controlling loads, the traditional pattern that the generation follows demand is changed, and the system resources can be reallocated economically. This paper mainly studies some novel generation and flexible load control strategies to participate the primary and secondary frequency control. The main results of this thesis are listed as follows.Firstly, in the supply side, the model predictive control strategy is applied into the real power regulation for the AGC units. Based on the models of one single area and multi-area power system, this paper presents two model predictive frequency control strategies to allocate active power set points for different generators. One is implemented by considering units’economic costs and very short-term load forecasting (VSTLF), and another by considering the CPS1/CPS2. The effectiveness of the proposed strategies is tested in different simulation scenarios. Simulation results show that the first strategy can significantly reduce system frequency oscillation and stabilize system frequency, and the second strategy is beneficial for the frequency improvement of the interconnected power system.Secondly, in the demand side, an estimation of the power and distribution of the aggregated heterogeneous thermostatically controlled loads (heTCLs) in the steady state based on the statistical theory is presented. It is derived that the aggregate power of the heTCLs is approximately propor-tional to the difference between the ambient temperature and the average setpoint temperature of the energy buildings, which provides the underpinning theory for the widely used notion of heating or cooling degree day (H/CDD). This knowledge is also essential for making energy managemen-t decisions and load forecasting. Furthermore, the mathematical expressions of the ON and OFF probability densities of the aggregated heTCLs are given, and their summation is only determined by the distributions of the setpoint temperature and the temperature deadband. Numeric simulations show the validity of the results.Thirdly, a decentralized frequency-based control strategy for a population of heTCLs is pro-posed for primary frequency control service. The frequency-based control strategy is implemented by local controllers of heTCLs in decentralized way without causing new consumption oscillations; the relationship between the ambient temperature and the steady power consumption of numerous heTCLs is used to estimated the aggregate power of heTCLs. In the strategy, each TCL responds randomly according to the system frequency and its operation state guaranteeing the fairness among TCLs. Moreover, the proposed strategy is integrated into the primary control of the power system to provide short-term frequency regulation service. Simulation results show that it can reduce the frequency deviation when a sudden load change occurs, and guarantees the satisfaction and fairness for the end-users.Finally, based on hierarchical MPC and bid-based dynamic economic dispatch, the price-responsive loads are used to participate in the secondary frequency control. At the upper level, the Independent System Operator (ISO) solves a multi-periods economic dispatch problem based on the bids of gener-ation and load companies, then gives the economic setpoints for generators and forecasts the Market Clearing Price (MCP) for the next period. At the lower level, the model predictive generation control is applied for generators to follow the reference power given by the upper level with considering the tuning cost of each unit; in the demand side, an electricity price incentive model is proposed for the first time by considering the control cost and utility of the load agent, which gives an optimal power consumption for flexible load agent based on MCP and the contract. The simulation results show that the proposed strategy can flat the demand profile, reduce the tuning costs of generators and stabilize the system frequency.