Load Frequency Control Design for Improving Frequency Quality of Power Systems

Abstract

Load frequency control (LFC) has been widely recognized as a feasible tool in power system frequency regulation. With the rapidly increasing scale of power grids and proportion of renewables in power systems, the frequency control has been potentially impaired because this occurrence may raise more power fluctuations and diminish power system inertia. In light of this, there is a necessity to accommodate the current frequency control methods to the emerging integrated power systems. Against this background, this thesis addresses the frequency control problems from two aspects: one of ameliorating the performance of load frequency control immediately after frequency excursions, and the other of mitigating the power fluctuations in wind energy penetrated power systems. On the one hand, the traditional frequency control can be enhanced by tuning appropriate parameters through optimization tools. Meanwhile, by scheduling proper power dispatch for each generator unit, which refers to optimal power dispatch, the system also sees improvement. Specifically, with the commercialization of power industries, the optimal power dispatch is related to operating economy in electricity markets, where frequency ancillary services are provided. Therefore, the frequency ancillary service, as well as the economic factor is considered to comprehensively compose the reformed load frequency control design. In consideration of this, to start with, an optimal power dispatch scheme (OPDS) method is presented. This method incorporates power dispatch within load frequency control through the proposed optimization framework. Sequentially, an optimal frequency ancillary service strategy (OFASS) considering the operational costs is developed in frequency control ancillary services (FCAS) markets. In these markets, power industries tender for carrying out the ancillary services of load frequency control, and part of them are elected based on the cost and performance. As a part of the OFASS, the OPDS is used to propound the optimal scheme based on the determinate generating units to pursue a balance between minimizing the cost and facilitate the frequency control through reconstructing a comprehensive objective function. Genetic algorithm (GA) is utilized as the optimization tool. Simulations based on the singlearea control system and two-area system are carried out to verify the efficacy of this work in MATLAB Simulink. On the other hand, as an indispensable renewable resource, wind energy calls for further attention, of which the noticeable integration in power grids aggravates the power fluctuations in wind plants’ side due to its intractable characteristic, provoking the issue of power system frequency stability compared with the conventional power resources. Therefore, this thesis presents a novel control method to help manage the scarcely predictable and intensively fluctuating wind power to maintain the frequency quality. In this thesis, a discrete control is initially developed to curtail the undesirable wind power, which is caused by the wind speed rising in a short time, under the context of growing wind energy penetration. The potential overproduction of wind power is evaluated and categorized into three types by a classification mechanism, which monitors the instant change of wind speeds and assesses the offloading scheme. Specifically, the wind turbines are coordinated through using the pitch control in accordance with the given scheme. Meanwhile, a battery energy storage system (BESS) is equipped to compensate the fluctuating active power through storing and releasing the energy, contributing to frequency regulation. Consequentially, the discrete control and the BESS are incorporated into the BESS cooperating with the discrete control (BCDC) method to ameliorate the frequency quality in integrating wind power. Finally, Monte Carlo (MC) method is used to simulate the variable wind speed samples while validating the feasibility and performance of the work in MATLAB Simulink and DIgSILENT PowerFactory.