An improved correlation is presented for the deviation in axial turbines at subsonic flow conditions. Development of the correlation began with the assumption that the deviation is primarily determined by the blade loading (pressure difference) towards the trailing edge. The blade row parameters which influence this pressure difference are identified. A functional form for the correlation was then assumed and the unknown coefficients in the function were determined based on a database of known values of the deviation. This database includes in-house experimental results, estimates of deviation obtained from Navier-Stokes simulations, as well as cases from the literature for which sufficient geometric data are provided. The evaluation of the empirical coefficients of the correlation was treated as an optimization problem. The optimization was performed using a genetic algorithm which proved very effective and robust. The new deviation correlation appears to be significantly more successful than existing correlations.

This content is only available via PDF.