In this paper we model the economic feasibility of compressed air energy storage (CAES) to improve wind power integration by means of a profit-maximizing algorithm. The Base Case is a wind park with 100 MW of installed capacity and no storage facility. In Variant 1 we add a central CAES system with 90 MW of compressor and 180 MW of generation capacity. The compressed air is stored in a cavern. The CAES system is operated independently of the wind park such that profits from peak power sales on the spot market and the reserve power market are maximized. Variant 2 is an integrated, decentralized CAES system, where each wind turbine is equipped with a compressor but no generator. The compressed air is stored in a cavern and converted into electricity by a turbine, again maximizing profit as a peak power plant. Both variants are modeled for conventional diabatic and the more advanced adiabatic systems. We find that the economics of the systems studied depend on how intensively the spot market and the market for minute reserve are used. Unless a minute reserve market exists, where hourly contracts can be traded, none of the CAES power plants studied is economically feasible. CAES plants can be operated economically if combined trade in the spot market and the minute market is enabled, and provided that some support scheme is in place (e.g., such as the German Renewable Energies Act). A centralized CAES plant is found to be more attractive than a wind power plant with integraed CAES, irrespective of whether a feed-in tariff scheme also exists for the integrated plants. Diabatic CAES runs out to be more than adiabatic CAES. (C) 2011 Elsevier Ltd. All rights reserved.