Investigation of the relationship between the 3D flow structure and surface heat transfer within a realistic gas turbine blade trailing edge internal serpentine cooling channel

•The heat transfer and flow structures in realistic internal serpentine cooling passage for trailing edge are investigated.•Intense Dean vortex induced by inlet S-duct provides high heat transfer in the subsequent ribbed channel.•Vortex structures in U-bend leave footprints of friction and heat transfer coefficients.•Flow non-uniformity and secondary flow combined parameter (D+E)/2 is highly correlated to the locally averaged heat transfer. Proper internal cooling of the thin trailing edge of gas turbine blades is crucial for blade lifespan extension. In this study, numerical and experimental analyses are conducted to elucidate the relationship between the 3D flow structure and surface heat transfer in a realistic three-pass serpentine channel within the trailing edge of a blade. Magnetic resonance velocimetry is conducted to first experimentally validate the mean velocity field from the Reynolds-averaged Navier Stokes simulation. Heat transfer and friction coefficients are then obtained through numerical analysis. The serpentine channel is investigated at each section. The inlet S-duct contributes to heat-transfer promotion upstream of the first ribbed passage due to an intense vortex pair. This vortex pair generates high streamwise velocity along the pressure and suction side, as well as additional friction due to secondary flow. In the second passage, film cooling holes make the separation bubble attach only to the suction side and create a high streamwise velocity region around the holes by bringing the surrounding flow to the periphery of the holes. This effect also occurs near the ejection holes in the third passage. These flow features within the passages are analyzed using the momentum distortion parameter (D) and secondary flow energy parameter (E). We found that the distortion parameter is not sufficient in predicting the heat transfer trend due to in-plane wall shear. Thus, the estimator is revised as (D+E)/2 to additionally account for secondary flow. This composite parameter shows good performance in estimating heat transfer for all the passages, compared to the friction coefficient which is hard to directly measure.