Planning models for electric transmission network expansion
Abstract
The electric transmission expansion problem involves the timely addition of transmission system components in order to ensure the transmission system constraints are satisfied over a defined planning horizon. Existing transmission planning models adopt several assumptions to help solve this problem. The main assumption is the use of the simplified linear DC power flow equations, instead of the more accurate but nonlinear AC power flow equations. While the DC power flow equations simplify the planning model, they consider only the real power requirements and ignore the voltage and reactive power requirements of the AC electric power system. These voltage and reactive power requirements are tackled in a separate second-stage planning process. Furthermore, most of the existing planning models have been single period models which are used to help plan the transmission system for worst case scenarios. The assumptions adopted by these existing models are being challenged in the emerging competitive electricity environment where greater efficiencies in system operation and planning are being encouraged. The goal of this research is to develop and investigate the potential benefits of minimum cost electric transmission network planning models that utilize fewer assumptions than the current models. Two such multi-period models are developed. The first is a mixed-integer linear programming (MILP) model that utilizes the DC power flow equations while the second model is a mixed-integer nonlinear programming (MINLP) model that utilizes the AC power flow equations. A candidate generation heuristic procedure is developed for the DC-based expansion model to reduce its solution times on larger problem instances. This MILP model with the heuristic procedure is shown to provide lower cost expansion solutions than the traditional nonlinear DC-based expansion model which is extended into a multi-period framework. A Generalized Benders decomposition approach is used to tackle the MINLP AC-based expansion model. Even though the optimality of the solution is not guaranteed, this model is shown to provide lower cost solutions than the traditional two-stage methods used for AC system expansion. Furthermore, these cost benefits are shown to increase when the system is deficient in reactive power supply capability and reactive power expansion is required.
Degree
Ph.D.
Advisors
Rardin, Purdue University.
Subject Area
Electrical engineering|Operations research|Industrial engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.