12:54:31 From GV : Hi, thanks for the very interesting presentation. At what point along the characteristic line are the simulations carried out? 12:55:41 From Fang, Jian (STFC,DL,SC) : Hi, Zhongnan, very good simulations. Just one simple question. The LES and RANS zones are fixed or dynamically change with simulation? 12:58:08 From Ashkan : Hi, Zhongnan, way to go! I'm wondering if such methods blending LES and URANS are incorporated into any commercial package. 12:58:48 From Zhong-Nan Wang : @GV thanks. The simulation is running at the approach condition, which is at the half of design speed. That leads to the mismatch of flow angle with blade, hence separation occurs in the fan tip. 12:59:27 From GV : Thank you 12:59:30 From Zhong-Nan Wang : @Jian The zones are fixed in our simulation, following mean flow streamline. 13:01:20 From Zhong-Nan Wang : @ Ashkan No exact method is available in the commercial package for the moment, but you could use UDF functions in Fluent to do a similar job. 13:01:52 From Ashkan : Thanks. 13:09:37 From y-yao : Hi Zhong-Nan, nice talk! at the LES/RANS interface, do you do a sudden switch or apply a narrow region to blend over to avoid oscillations? Yufeng 13:11:28 From Zhong-Nan Wang : @Yunfeng Thanks. The interface is a smooth blending region as you said to avoid oscillations. 13:12:58 From Sylvain Laizet : Hi Chris, quick question: did you investigate different type of turbulence (different intensities, different length scale) for your inflow? 13:13:28 From Sylvain Laizet : would be interesting to know how it affects cooling. 13:13:50 From Sylvain Laizet : Lastly, what type of AI/ML techniques you have in mind for your future work? 13:14:15 From Wei : Hi Chris, nice talk. just wonder whether you compare your three simulations with any experimental data to see the difference? Thanks. 13:14:44 From Zhong-Nan Wang : Hi Chris, very good talk! Just wondering if the coolant plenum size has any effects on the results? 13:15:29 From y-yao : Hi Chris, nice talk! is your simulation still within RANS/URANS framework, if so, for this JICF problem, HTC will be over-predicted in the wake region, right? yufeng 13:16:35 From Hao Xia : @y-yao No, it's WALE LES. 13:16:55 From Christopher Ellis : Hi Sylvain, thank you for the question. We investigated some other developing turbulent boundary layers, but restricted the elevated high FST to this singular case for the time being but we are considering extending the analysis to cover further intensities and length scales. For the ML techniques we are inspired by some of the work in Tensor Basis Neural Networks, but we have pursued simple analysis in previous work we presented at ASME on data-driven turbulent diffusivity coefficients in the HOGGDH closure. 13:18:13 From Sylvain Laizet : thanks! 13:18:56 From Christopher Ellis : Hi Wei, we compare our datasets to the experimental results of Sinha et al. (1989) which is best described by our Boundary Layer Turbulence case. The other cases are divergent from the experimental data. We can use other experimental analysis as a comparative baseline to show that our trends are correct. 13:20:05 From y-yao : I see. Thanks Hao. So I would assume you have run two precursor simu one for cross flow, one for inflow at the bottom? yufeng 13:21:04 From Christopher Ellis : @Zhong-Nan Wang. Hi, I don't think it will have a considerable effect at this size, but there will be a point where the size dominates. The size of this plenum matched other numerical studies of coolant holes. This is an ideal solution where as in a combustor liner flow the coolant flow approaches the coolant hole in a streamwise direction. 13:22:12 From Christopher Ellis : @y-yao Thanks for the questions, no precursor simulations were run. The cross flow statistics are based on experimental data and the inflow is plenum fed to match the experimental setup. 13:22:33 From Hao Xia : Yufeng, you spotted it. Actually, we didn't force any turbulence in the coolant pipe. Partly, it is already quite messing due to the separation in the pipe. 13:23:06 From Zhong-Nan Wang : Thank you, Chris. 13:24:08 From y-yao : I see. Thank you Hao, thank you Chris. 13:24:29 From Christopher Ellis : You are welcome, thank you for the questions. 13:31:19 From Arash Hamzehloo : Hi Jian, thanks for the nice presentation. Have you investigated heat transfer from the wall? Can the DNS framework reproduce the peak heat transfer and its location correctly? Also, have you tried to run your cases in 2D? 13:34:36 From y-yao : nice talk Jian! can I leave this question from Arash for you to reply? yufeng 13:34:37 From Fang, Jian (STFC,DL,SC) : Hi Arash. Thanks for the question. we haven't checked the wall heat flux yet, but will have a look. We have tried simulate 2D cases only for laminar flows. 13:35:09 From Arash Hamzehloo : Thanks! 13:36:25 From y-yao : also to add, we did have studied heat transfer in other high-speed flows. Jian can provide you published papers. yufeng 13:38:34 From Arash Hamzehloo : Thank you! I think capturing the peak heat transfer and its location is the most challenging task in modelling such flows. 13:44:17 From Fang, Jian (STFC,DL,SC) : http://dx.doi.org/10.1063/1.4936576 We have studied a bit on wall heat flux in this paper. unfortunately, no comparison with experimental data. 13:44:43 From Zhong-Nan Wang : Hi Josh, nice work! Just wondering what size of tip gap is compared to your incoming endwall boundary thickness? and How is that compared to the value for a real compressor stage? 13:45:17 From y-yao : Hi Josh, 13:45:41 From y-yao : can you explainhow many grids inside the gaps? yufeng 13:45:48 From Zhong-Nan Wang : Thank you 13:47:27 From Andrei Cimpoeru : Hi! Are the videos going to be available? 13:48:26 From Hao Xia : Thank you. 13:48:28 From Andrew Wheeler : Thanks you