Computational models of natural gas markets for gas-fired generators

Abstract

Climate change is a major factor reforming the world's energy landscape today, and as electricity consumes 40% of total energy, huge efforts are being undertaken to reduce the carbon footprint within the electricity sector. The electric sector has been taking steps to reform the grid, retiring carbon-intensive coal plants, increasing renewable penetration, and introducing cyber elements end-to-end for monitoring, estimating, and controlling devices, systems, and markets. Due to retirements of coal plants, discovery of shale gas leading to low natural gas prices, and geopolitical motives to reduce dependence on foreign oil, natural gas is becoming a major fuel source for electricity around the United States. In addition, with increasingly intermittent renewable sources in the grid, there is a need for a readily available, clean, and flexible back-up fuel; natural gas is sought after in New England to serve this purpose as a reliable and guaranteed fuel in times when wind turbines and solar panels cannot produce. While research has been conducted advocating natural gas pipeline expansion projects to ensure this reliability, not enough attention has been paid to the overall market structure in the natural gas and electricity infrastructures which can also impact reliable delivery of gas and therefore efficient interdependency between the two infrastructures. This thesis explores the market structures in natural gas and electricity, the interdependence of natural gas and electricity prices with increasing reliance on natural gas as the penetration of renewable energy resources (RER) increases in order to complement their intermittencies, possible volatilities in these prices with varying penetration rates in RER, and alternatives to existing market structures that improve reliability and reduce volatility in electricity and gas prices. In particular, the thesis will attempt to answer the following two questions: What will the generation mix look like in 2030 and how will this impact gas and electricity prices? How do Gas-Fired Generator (GFG) bids for gas change between 2015 and 2030? In order to answer these questions, a computational model is determined using regression analysis tools and an auction model. Data from the New England region in terms of prices, generation, and demand is used to determine these models.