| The combined natural gas and electricity system with wind energy is one of the most promising models to reduce cost and improve energy efficiency. Wind energy is clean notwithstanding its high uncertainty. Natural gas power generation is also known to be sustainable, cost-effective and fast-response. It plays a significant role in the electricity system and can be used to compensate for wind energy deficiency. Thus, a combined natural gas and electricity system with wind energy is studied to learn how the natural gas fueled power generators can help to incorporate the wind energy in the power system.;The thesis takes the viewpoint of the operator of a combined electricity and gas system, whose goal is to minimize the total operating cost, generating cost and non-served energy cost of the combined system, while satisfying all the operational and equilibrium constraints of each element of the gas system and the electricity system. For simplicity, we assume gas and wind are the only sources of electricity. A mixed integer nonlinear optimization model is formulated to investigate the hourly unit commitment and dispatch solution for the electricity system as well as the natural gas system's hourly working schedule in a single day. Additional linear constraints are formulated with binary variables to approximate the nonlinear constraints of gas flow in natural gas pipelines, and the resulting mixed integer linear optimization model can be solved efficiently to optimality. In addition, a two-stage stochastic optimization model is applied to incorporate the uncertainty of wind power production and learn the operational decisions of pipelines, gas supply, and generating unit commitment in the first stage. In the second stage, decisions on the hourly schedule of gas-fueled power generators, power flows and the gas storage flows are made given various wind energy scenarios.;The computational results of a deterministic model with fixed reserve constraints and the here and now (stochastic optimization) model are compared, along with a wait and see model that gives a lower bound on the expected cost of the stochastic optimization model. In the deterministic optimization model with reserves, the planning problem is solved in the day-ahead market in view of the wind energy forecast and the first-stage variables are fixed to those optimal values. Then, in the real time market, under various scenarios for actual wind energy supply, a dispatch model without reserves is solved to test how well those day-ahead decisions perform. The numerical results for a small test case illustrate that the gas fueled power generators and gas storage are able to counteract the wind energy deficiency to satisfy demand at a lower and more stable cost by incorporating various wind energy scenarios in the stochastic optimization model. |
