Optimal planning of power systems with renewable energy integration

Abstract

In the last decade, the growth of renewable energy capacity has been increased rapidly in power systems. Large-scale renewable energy integration of the power systems may bring many economic interests and environmental benefits. However, due to the stochastic characteristics of renewable generation output, widespread installation of renewable generators will pose some great challenges to the normal power system operation. For example, power flow patterns of transmission lines will significantly change and become inevitably fluctuating due to high renewable penetration. This may result in some potential negative effects, including but are not limited to line congestion, increased active power loss and large voltage deviation. To deal with these issues, the traditional power system expansion can be taken into consideration, such as power system line expansion and reconstruction, new electrical plants installation and existing facilities upgrade. However, these system expansion options are usually investment-intensive and time-consuming and may cause environmental problems. In this regard, this thesis focuses on dealing with negative concerns caused by renewable energy integration via optimal advanced flexible AC transmission systems (FACTS) devices planning in transmission networks, and enhancing renewable energy hosting capacity via optimal advanced electrical devices in distribution networks, respectively. The thesis firstly focuses on planning in transmission networks, which is to cope with the negative effects introduced by high wind power penetration. To tackle the negative impacts caused by wind energy integration, a stochastic optimal TCSC location-allocation model is proposed. This planning model is formulated as a two-stage optimization program, where the planning decisions including sites and sizes for TCSC devices installation are determined in the first stage and the second stage is to minimize the expected operation cost of transmission systems under various wind-load uncertainty scenarios. The proposed planning model is firstly formulated as a mixed integer nonlinear programming (MINLP), and then both the linearization technique as well as the approximation approach are used to transform this MINLP to a mixed integer linear programming (MILP), which can be directly solved by commercial solvers such as CPLEX and GUROBI. The TCSC planning model considers uncertainties of wind energy output and load demand, which are represented by wind-load scenarios. These scenarios are originally generated by using classical copula theory and then reduced by a well-established backward-reduction algorithm. Finally, a modified IEEE 57-bus transmission system is utilized to verify the effectiveness of the proposed planning model. The thesis secondly focuses on planning in distribution networks, which is to improve the ability of distribution networks to accommodate more photovoltaic (PV) generations. To improve PV hosting capacity, a two-stage optimal var compensator (SVC) planning model is proposed. In detail, the first stage is to determine the PV hosing capacity of the given sites and SVC location-allocation decisions and the second stage is to minimize the operation cost of SVC devices in all considered uncertainty scenarios. Besides, the concept of PV accommodation capability (PVAC) is proposed to describe the amount of PV generation that can be reliably accommodated at a certain node of a distribution network within a certain time period. To enhance the daily PVAC, this thesis proposes two-stage MILP based voltage regulator (VR) placement model, where the hourly PVAC and VR allocation decisions are determined in the first stage and a stochastic programming based feasibility checking model is developed to ensure the network constraints security in the second stage. These two planning problems are both intractable due to numerous operation scenarios involved as well as the time-coupling constraints. In this regard, to reduce the computational complexity, a Benders decomposition algorithm based solution method is developed to solve the proposed two-stage stochastic problems. IEEE distribution systems are utilized to verify the effectiveness of the proposed planning model and solution method.