Investments in a Combined Energy Network Model: Substitution between Natural Gas and Electricity?

• Author: Abrell, Jan

1. INTRODUCTION

   Energy systems throughout the world are expected to undergo a transition in the coming decades. Industrialized countries aim to switch increasingly from fossil-based to renewable energy-fueled systems, and it is forecast that a significant increase in energy demand will occur in developing regions. Both developments will require a large amount of investment in energy production and transport infrastructure; e.g., IEA (2011a) estimates global investment needs of approximately 38 trillion USD up until 2035. As energy markets are interlinked due to their substitutability for specific utilizations (e.g., heating oil or gas vs. biomass or electricity) and the direct usage of energy fuels in downstream markets (e.g., coal, oil and gas as fuel input for electricity) the relations between markets need to be taken into account in estimates of future developments. In addition energy markets often rely on network structures which add a spatial layer to the problem.

   The interaction of energy markets is particularly relevant for natural gas and electricity systems. The increasing importance of emission reductions raises the need for a shift from coal-based to natural gas-fired units. Similarly, the increased utilization of intermittent renewable generation units increases the need for more flexible generation units as back-up capacities which are mainly assumed to be gas-fired. These developments are likely to increase demand for natural gas in the electricity sector, which can increase the need for investment in gas infrastructure. On the gas supply side the 'fracking boom' in the United States, the general global prospects for unconventional gas, and the further increases in the LNG infrastructure are likely to lead to shifts in the global natural gas market dynamics and consequently gas prices (IEA, 2011b). Furthermore, the conflict between Russia and the Ukraine has awakened security of supply concerns in Europe's energy markets. These changes in the natural gas market feed back into the electricity market and impact the prospects of gas as fuel option.

   Given these medium- and long-term challenges, this article presents a combined natural gas and electricity market model framework which focuses on investment options while accounting for the network characteristics of both markets. The model follows existing single market representations of natural gas and electricity networks and is formulated as equilibrium approach. Using the model, we highlight the interaction of natural gas and electricity investment alternatives, first, using an illustrative network setup and, second, using a stylized European network.

   Focusing on the above-described linkage between gas and electricity markets via the fuel option in power generation, the interaction can be traced to three main elements: gas production costs, pipeline topology, and the topology of the electricity transmission network. The first aspect of gas production costs basically summarizes the competition with other fuels such as coal or nuclear generation. Based on overall market conditions, this impacts relative prices in electricity generation and thereby determines the potential revenues for gas-fired power generation. This occurred in 2012, for example, when the price spread between gas and coal in the U.S. market led to a shift from coal- to gas-fired generation. (1) The second aspect of pipeline topology captures the impact of costly and potentially limited transport alternatives within the gas market. Even though the overall price level may make gas a competitive electricity generation option, this may not hold for all locations within the market, as some locations are more costly to supply due to transport cost depending on the pipeline network structure. Examples for the impact of pipeline constraints are price divergences between the different natural gas trading hubs in Europe which in a competitive market represent the difference in transport cost. This aspect also includes the potential need to construct the necessary pipelines to supply specific locations.

   The third aspect of electricity network topology captures the influence of the structure of the electricity transmission grid on local prices. Electricity transmission fundamentally differs from pipeline transmission: While the system operator can decide about the flows along a pipeline network, in an electricity network flows are determined by physical laws and the sole choice of the system operator is the injection and withdrawal at the different nodes in the system. Put differently, a change in the topology of the transmission system or a change in the supply and demand patterns at a given node leads to a change of the network flows along all lines in the system. This behavior of electricity flows is known as the loop flow problem.

   Seen from a financial perspective, the decision on how to supply local electricity demand is impacted by the fuel price differences, potential network limitations on the supply and electricity side, and the current market conditions in the electricity market. Naturally, to derive consistent evaluations about economically efficient developments in the electricity market, it is necessary to include the supply-market side as well as the network characteristics in the analysis.

   Over the last two decades, research has focused on the analysis of both electricity (e.g., Leuthold et al., 2012; Moest and Keles, 2010) and natural gas markets (e.g., Lise and Hobbs, 2009; Egging et al., 2010), as well as on modeling the overall energy system (see, e.g., Pfenniger et al., 2014, for a review). Within this strand of research, the coupling of electricity and natural gas markets has gained increasing attention in recent years, and has predominantly focused on the incorporation of natural gas constraints in the short-run electricity market dispatch via different modeling approaches (see, e.g., Rubio et al., 2008; Liu et al., 2009; Damavandi et al., 2011; Erdener et al., 2014). Most of these approaches focus on the technical interaction of both markets. Medium- to long-term analyses with endogenous investment representations are rather limited so far. Unsihuay-Vila et al. (2010) develop a coupled optimal investment model for natural gas and electricity networks which excludes loop-flow aspects and apply it to the Brazilian network. Linert and Lochner (2012) combine an electricity model (DIME) and a natural gas model (TIGER) with a stylized transmission approach based on cross-border transmission restrictions and apply it to the European markets, and provide long-term assessments on gas-fired generation capacities. Bakkem et al. (2007) develop a coupled model design for multiple energy infrastructures. However, the investments are externally defined and ranked by the model approach, but no endogenous optimal investment is obtained. Geidel and Andersson (2006) use a hub-based approach for structural optimization regarding the conversion technologies in the hubs (i.e., plant technologies), but not for the network connecting the hubs. Finally, Chaudry et al. (2014) develop a combined network extension model for natural gas and electricity with detailed network flow representations. Although, the physics of power flows are explicitly included in the model formulation, it is not indicated whether their model accounts for feedback effects due to investment in network capacity. The application to the U.K. system relies on a linear setup of the British electricity transmission system and therefore does not account for the externalities stemming from electricity transmission.

   The present article extends this discussion on the interaction of natural gas and electricity markets by including investment decisions in a combined natural gas and electricity model. In Section 2, we first provide the underlying mathematical formulation based on the static model of Abrell and Weigt (2012). The natural gas market is represented including gas producers, traders, consumers, and the pipeline operator. Similarly, the electricity market model includes generators, consumers, and the system operator. We allow for investments in natural gas and electricity transmission infrastructure, as well as in electricity generation capacity. The models are then coupled via the market clearing condition for natural gas: On the one hand, demand for natural gas as a fuel input of power generation becomes endogenous in the market clearing for natural gas. On the other, the natural gas price is included as a endogenous price variable in the generators' profit optimization.

   In Section 3, the model framework is applied to a stylized four node network. The stylized network provides insights on the ability of the developed model to capture the substitution effect between investments and visualizes the impact of loop flows on market results. We first provide a simple linear network setup ranging from production which is distantly located from demand but cheap in terms of production costs to production facilities which are closely located to demand, but which exhibit high production costs. As demand coverage at a network location includes both the pure production costs and the transport costs, a tradeoff between both elements will be reached based on the underlying cost structures. The same principal also holds when the linear structure is omitted, and a meshed structure is used thereby introducing the loop flow problem. Due to the nature of the power flows, the resulting investment pattern exhibits non-trivial deviations from the linear setup. Nevertheless, both cases provide the basic investment pattern, that with higher shares of transport investment costs, additional production options which are closer to demand are realized including natural gas extension alternatives.

   In Section 4, we apply the model in a stylized European market framework. Each European country is represented by one node in the natural gas...