“…As indicated, the flow turning angle 8 was underpredicted in all of the simulations, as was the loss coefficient C. The values of C obtained with transition suppressed until c 1<i0r4 are slightly lower, and therefore in worse agreement with the experimental result, than the values obtained without using the transition modification. Reduced skin friction on the suction surface upstream of the shock impingement point is the most likely cause of the lower Cr, It is possible that the high loss measured in the experiment was caused by vortex shedding, which can have a very significant effect [Currie, 1996] and has been observed in compressor rotors [Hathaway et al, 1985] and cascades [Paterson and Weingold, 19841, and/or unsteadiness of the passage shock. It should be mentioned that the flow was not separated, and the loss coefficient was only -0.075, when M, was reduced to 1.02 in the experiments.…”