This study optimises and compares the operation of a conventional gas-fired power generation company with its operation in combination with wind power and compressed air energy storage (CAES). A mixed integer non-linear programming (MINLP) formulation is developed for the optimisation problem. Limits in ramp rate, capacity, and minimum on/off time, as well as start-up cost constraints, are considered for the modelling of conventional units. Injected and produced power constraints, storage, air balance and CAES-operation limits are considered in the CAES modelling. Two objective functions (profit maximisation and cost minimisation) are modelled. Without considering capital costs, it is found that the use of CAES results in 43% higher operational profits and 6.7% lower costs in a market environment. (C) 2012 Elsevier Ltd. All rights reserved.