As zdas04 said, you need to estimate the power on a stage by stage basis (I normally use polytropic equations, but as it is air, then adiabatic is OK). This gives you the "gas power". You need to add on an allowance for mechanical losses (bearings, seals). This will give you an estimate for...