Stochastic model for natural gas portfolio optimization of a power producer

This paper develops a risk constrained expected fuel cost minimization model for a gas turbine power producer in India. The decision strategy of the generator is to optimize the natural gas portfolio comprising long term fixed contracts and short term purchases to minimize the yearly fuel cost. The model considers stochastic behavior of load demand and gas price in short term gas market. Constraints in the gas contracts like maximum off-take are considered here. Two Stage Stochastic programming framework is applied to estimate the optimal portfolio of natural gas, ensuring that load demand is met throughout the year. The risk trade off is achieved by constraining the objective function with Conditional Value at Risk (CVaR). A numerical example of a generating station is illustrated to show the fuel cost minimization strategy by this optimization.