Predicting Film Cooling Performance of Trailing-Edge Cutback Turbine Blades by Detached-eddy Simulation

Abstract

This paper describes numerical studies of the film cooling performance of cutback trailing-edge gas turbine blades using the detached eddy simulation (DES) approach, in order to improve results obtained by industry standard steady RANS and unsteady RANS (URANS). An experimental configuration of blade cooling slot with two rows of long ribs at three different blowing ratios (M = 0.5, 0.8, 1.1) are investigated and three aforementioned numerical approaches based on the same SST k-ω turbulence model are adopted. Simulation flow conditions follow experimental and numerical studies previously done by other researchers. At a low blowing ratio M = 0.5, current numerical predictions have shown that DES results are in good agreement with available experimental measurement of film cooling effectiveness along the trailing-edge cutback surface, whereas both RANS and URANS largely over-predict in the near and far wake regions. At a moderate blowing ratio M = 0.8, DES slightly over-predicts the film cooling effectiveness after a nondimensional streamwise location x/H > 8 (with H denotes the slot height), whereas at a high blowing ratio M = 1.1 it under-predicts after x/H > 5. The effect of lip thickness has also been studied and it was found that a thinner lip thickness t/H = 0.5 (with t denotes the lip thickness) produces an adiabatic film cooling effectiveness of near unity along the trailing-edge cutback surface, indicating that there is need to reduce the lip thickness to achieve a higher level of cooling effectiveness, without compromising other constraints such as blade structure and loadings. NomenclatureA slot = area of slot exit [m 2 ] b r = thickness of internal coolant passage ribs [mm] des C = calibration constant used in DES model (C des = 0.61) D ω = cross-diffusion term f = frequency [kHz] F DES = blending function in DES model 2 G k = generation of turbulence kinetic energy due to the mean velocity gradients G ω = generation of turbulence dissipation rate ω H = height of internal coolant passage slot [mm] k = turbulent kinetic energy k = laterally averaged turbulent kinetic energy L 0,1,2.3 = streamwise lengths of main hot gas domain [mm] L r1,r2 = lengths of row 1 and row 2 ribs [mm] t L = turbulence length scale M ∞ = Mach number M = blowing ratio c m  = mass flow rate of coolant gas [kg/s] q conv = convective heat flux [W/m 2 ] q rad = radiative heat flux [W/m 2 ] q w = wall heat flux [W/m 2 ] R f = fillet radius [mm] s = pitch-wise distance of ribs array [mm] S t = Strouhal number S 1 = the first monitoring point location S 2 = the second monitoring point location S k = source terms in turbulence kinetic equation S ω = source terms in turbulence dissipation equation t = thickness of blade trailing-edge [mm] T aw = adiabatic wall temperature [K] aw T = laterally averaged adiabatic wall temperature [K] T w = isothermal wall temperature [K] T c = coolant gas temperature [K] T hg = hot gas temperature [K] u c = coolant gas velocity [m/s] u hg = hot gas velocity [m/s] X r1,r2 = x-coordinates of row 1 and row 2 ...

Pin-fin shape and orientation effects on wall heat transfer predictions of gas turbine blade

“…A common practice to enhance heat transfer performance of a blade internal coolant passage is to insert small obstacles, i.e., pin-fins, ribs or other objectives in order to increase surface areas as well as to promote the near-wall turbulence intensity level. In the past, blade coolant passage performance has been studied experimentally and numerically for various configurations of staggered and/or in-line arrangements with cylindrical pin-fins [1] [3], elliptical pin-fins [4] [5], streamwise elliptical pin-fins [6] [7], spanwise elliptical pin-fin [8] [9], double in-line ribs array [10] [11]. Han and Rallabandi [12] reviewed the latest developments on the turbine blade cooling techniques.…”

DES study of blade trailing edge cutback cooling performance with various lip thicknesses

“…Most recently, a numerical study using a blade trailing-edge cutback cooling configuration with two rows of long ribs inside the cooling passage was carried out by Effendy et al [27], who applied the SST k-ω turbulence model in steady RANS, unsteady RANS and detached-eddy simulations (DES). It was recognised that DES with the SST k-ω model is more capable of capturing the turbulent flow structures in the mixing region that result in significant improvements on the prediction of the adiabatic film-cooling effectiveness, particularly from a lower blowing ratio of M = 0.5, whereas both RANS and URANS models largely over-predict the film-cooling effectiveness in the near and the far wake regions.…”