Flexible and resilient operation of power systems with renewable integration

Abstract

With the increasing requirements of environment, economy and society on power grid, it is particularly important to build a flexible and resilient power grid. By analyzing resilience construct, the process for enhancing power grid resilience could cover three stages: normal operation stage before events, abnormal operation stage under unfolding events and system restoration stage. The thesis focuses on the former two stages and proposes methods from three different aspects to improve flexible and resilient operation performance of power systems. The first two proposed techniques are for the first stage, while the third technique for the second stage. The demand response has been regarded as an effective method to reduce the peak energy demand in main power-grid. The flexible loads can be dispatched as a kind of operating resources to make up for the shortage of controllable resources in operation of power system. To solve the problems of manually controlled loads, this thesis proposes an automatic demand response (ADR) method implemented by a back to back PWM converter. The loads are connected to the power grid through such PWM converters. The proposed ADR is realized by the flexibly controlled output voltages of the power converters. Due to concerns of environmental pollution and energy crisis, renewable energy sources, especially wind power, are integrated increasingly into power systems worldwide. Unit commitment and generation dispatch problems are extensively studied to evaluate and realize its environmental benefits, such as carbon emission reduction. However, production costing is rarely visited. It also deserves careful study since decisions including coal type selections and order quantities, which have great impacts on future system operation and resulting emissions, are made. In this paper, the production costing problem is formulated as a mixed-integer linear programming (MILP) problem with thorough consideration of fuel issues. Wind power is incorporated to realize its environmental benefits in terms of emissions of CO2, SOx and particulate matter (PM). In the second stage, fast fault detection is an effective approach to enhance power system resilience. Ground fault through transition resistance is common in power systems highly impacting the correct action of protection. To improve the reliability of power systems, a random matrix based method is used in this thesis as a backup protection to detect the high-resistance grounding fault, on the basis of successive PMU voltage measurements. A vector model is proposed consisting of matrix decomposition, Gaussian noise and the fluctuation components of PMU voltage measurements. Under a failure scenario, the sample covariance matrix combined by such successive models is influenced exclusively by the perturbation matrix related to the change of admittance matrix. This remarkable characteristic can be used to detect high-resistance grounding fault.