Linepack storage valuation under price uncertainty

• Published in: Energy (Oxford) 2013-04, Vol.52, p.155-164
• Main Authors: Arvesen, Ø., Medbø, V., Fleten, S.-E., Tomasgard, A., Westgaard, S.
• Format: Article
• Language: English
• Subjects: Applied sciences; Computer simulation; Contracts; Electric power generation; Electric power plants; Electricity prices; Energy; Exact sciences and technology; Flexibility; Gas pipelines; Gas storage valuation; least squares; Linepack; Monte Carlo method; Natural gas; Natural gas prices; pipelines; Power plant; power plants; prices; Purchasing; Regime-switching models; uncertainty
• ISSN: 0360-5442
• DOI: 10.1016/j.energy.2012.12.031

Natural gas flows in pipelines as a consequence of the pressure difference at the inlet and outlet. Adjusting these pressures makes it possible to inject natural gas at one rate and withdraw at a different rate, hence using the pipeline as storage as well as transport. We study the value of using the so called pipeline linepack as a short-term gas storage and how this functionality may offset the discrepancy between the low flexibility in take-or-pay contracts and the high inherent flexibility of a gas-fired power plant. To value the storage option, we consider a cycling power plant facing volatile power prices while purchasing gas on a take-or-pay contract. We estimate a Markov regime-switching model for power prices and a mean reverting jump diffusion model for gas prices. Applying Least Squares Monte Carlo simulation to the operation of the linepack storage, we find that the storage option indeed has significant value for the plant, enabling it to better exploit the sometimes extreme price fluctuations. Finally, we show how power price volatility and jump frequency are the main value drivers, and that the size of the storage increases the value up to a point where no additional flexibility is used. ► Linepack, i.e., storage of natural gas en route in long pipelines, is valued for the first time. ► We find significant storage value for a North Sea case and a German gas-fired power plant. ► Power and natural gas prices are modelled realistically, as related stochastic processes with mean reversion and spikes. ► Storage operation is valued under uncertainty yielding close to exact values, without heuristics.