 Good afternoon. I am talking on the heat transfer in a forward facing step. Actually, this importance of this is the simple geometry. So, what we have is the outline. So, what the governing equations are important. So, what the thing which comes to home and of a curement is that even if it is a simple geometry, it has got a certain contraction and expansion and when that happens the use of normal turbulence models is the eddy viscosity linear and non-linear. How do they behave and what do we do really get out of the thing and which one has to be used for a specific application. Does it meet the requirement or no and combining different turbulence modes with heat transfer what is the results coming that is the idea of this particular work. So, what we are trying basically a K epsilon, K omega, 60 and V2F and as an alias models on this thing and then try to find out with different parameters what do happen. So, this K epsilon model these are just for who are not very familiar with the thing, but probably most of you are familiar then we can just go ahead. So, this is the one which is very commonly used in the industry where kinetic energy and dissipation rate and it is actually has been sold very good for the flow around a bluff body and then these are the governing equations normally we have a question on K and this is question on epsilon and of course, we have the closure coefficients like this and this is the one which has been used and then these are the how we calculate the initial conditions for K and epsilon based on the input parameters what is the boundary conditions and initial conditions we do it. Then we go to the K omega model which is again based on the omega which is the specific dissipation rate rather than the epsilon itself, but it has been evolved because the K epsilon was not able to predict the things properly sometime. So, officially for a flow through a pipe band this gives a better results and has come back to the K epsilon. These are the basic governing equations used for turbulent kinetic and specific dissipation rate which omega use and of course, we also got various this is the menors K low Reynolds number and then we have the SST proposed by mentor which is the combination of K epsilon and K omega model. So, we use K epsilon in the free stream and K omega in the linear volts and we have what happens with that case K omega. But because this has got the good thing of both K epsilon and K omega models and that is why we will be using that one. So, these are the governing equations just for the sake of completeness have included in this course. Now, this is a V2F model which is not commonly used in many applications but is a non-linear viscosity model the previous ones are linear viscosity model this is a non-linear viscosity model and this one can give the catch the turbulence in a much better way. We are trying to incorporate that also in this code and then see what happens with that one. And these are the set of equations which that we mean model in this. So, these are all well documented. So, we are also going for LES this is the LES model. So, we are using the different when LES also with what happens this is basically used for the combustion acoustics and simulations. So, these places this gives you what is a computationally more expensive definitely than the previous models but can give capture the things better. So, with these things these are different models. So, what we are trying to do is we basically use the open source safe record open form and this is a base solver available with that is the symbol form we modify that code to incorporate this various turbulence models and also this does not have a energy equation built into it. So, we added energy equation. So, this was the original structure of the code available and this is what we are you have modified by adding temperature equations and then pressure equations etcetera one to this. And this is actually gives you some tips of what you have actually done into the code which may not be of interest if any question is there of course, we can answer those things. This how the code has got modified and also you added the energy equation into it. So, which was not there earlier. So, this is how we have added the energy equation to that. Now, then also I mean to solve we need a case file. So, we had a prepared a case file for concluding all these models we can try anyone at a time in from that sort of things and then the this is a grid manager etcetera. So, these are standard open form schemes which has been used there. Now, these are three type of parametric studies have been carried out one is by changing the step height. So, step height means this is the channel. So, in this one this is the inlet and this is this portion and all the three cases this portion is heated and this is the no sleep ward. This is the inlet, this is the outlet, this is the geometry which you are having. The first one what you already we are studying the step height. This is the 15 percent, this 30 percent, this 45 percent of this channel woods. So, what happens with that and then we try with different turbulence models and see how the things are behaving that is objective. So, going to the one. So, this is the input parameter which has been used for this case the inlet temperature to this one and the wall temperature or the bottom has been kept free and the velocity has been kept corresponding to Reynolds number of 5000. So, this is the results for k epsilon is just to give you a glimpse of what happens if the step height is changed. Of course, these are the very obvious results for temperatures and then velocity contours and the streamlines. So, this what happened in this one when you have a 30 percent we are having two counter rotating vertices over here and then other cases say, but this is a one k result, but our more interested one is in this. So, what happens suppose you use different turbulence models. So, this is k epsilon, this has been done for 45 percent case of k epsilon, k m, k m, k LST and LAS and V2F. What happens for this case? This is the more important thing. So, if you see it here, these three things may be giving almost the same, but when you come to this one, there is a lot of difference. So, why that is coming and which one is to be here that is still we are investigating. So, this is not that completed. So, that is still investigating on these results. Then this is the results for the velocity contours. So, similar also here we see it may be in V2F and LAS it is the same, but k omega is always behaving in a totally I mean symbol k omega is behaving in a different method. And this is the temperature contours coming for this one. So, here also if you see it LAS even the temperature scales are the same, but LAS is giving a bit different behaviour and in now there is the work is still going on to I mean basically capture the reasons and what is really happening, which one is to be believed, which one is not correct that is still going on and this is the streamlines. So, here also the streamlines the LAS one is capturing in a much different way. All are steady state results, steady state results no transients. It is not average, it is a steady state results whatever I am showing it is not a transients. It has been taken as a 2D of course it is a small strip of 3D, but it is a two and the other two conditions is symmetry boundary condition. So, being symmetry boundary condition it is a 2D. So, this is the streamline plots happening and the velocity and this is also the velocity contours for the three cases. LAS in 2D mode. LAS in 2D mode. So, it is not LAS it is LAS is still there it is it is to couple. It is not LAS it is actually it is not LAS, but what we have done is we have taken a we will try to do it as a 3D with a civic thickness that is what is the plan now it is because LAS say if you in 2D has no meaning. So, we have to 2D has no meaning. So, we have to go into a 3D space and then give it, but since we have considered as a thickness, but the boundary conditions were not given as walls. So, it has been given only as been symmetry. So, it is a 3D analysis, but with symmetry boundary conditions. Yes, it is not. So, then we have studied the effect of Reynolds number. So, with only one particular case three Reynolds numbers have been studied and this is the effect for a different Reynolds number of forward pressure, then velocity and of course, streamlines. So, this is actually with one set of results, but when it comes to the thing, this is the with different with 20 percent. So, Reynolds number 500. So, the pressure condors are being given here. So, they have also we are just what you have said we are just going to the this is the difference is coming in this course when the velocity condors and there is a streamline course here also we definitely get a different thing in V2F is almost totally a different set of results have been observed and then the turbulent this with the vorticity condors in this also. Now subsequently I have also studied the effect of Prandtl number that means that we change the fluid and then with that we define what is what is happening. Sorry, sorry. So, this is the pressure control. So, we have taken two fluids one is air and other is water with that what is the effect coming actually this is from the heat transfer point of view it makes sense to change the Prandtl number which has not been reported earlier. So, then we have the streamline course also for the same case with the same Reynolds number but a different Prandtl number. And then also for the different turbulence models what is the pressure condors that has been specified here. So, the K omega in almost all cases is found to be behaving a different manner than the other two. This is the velocity condors and then streamline plots and velocity plots. So, this is the basic summary of the results. And now what I say is that this and we have carried out this one we have modified the code which is 3D code. In fact, this is changed and you have studied the effect of step height to Reynolds number and Prandtl number. And this indicate that KFC don't sustain V2F models are almost showing similar results but alias is you have to be incorporated in 3D and then we need more investigation to a certain actual type and which has to be brought out. So, we need to find out the friction factor charts, friction factor values and Nusselt number values to really quantify these results. So, that work is still going on. So, it is taken quite some time. Yeah, right, right. So, that results will. The resolution issues we have addressed for one case in 3D the beginning. The grid independence study has been carried out for one case. But then subsequently when you go into this one it has to be put each and every particular case we have to take. So, when you go for a different step otherwise it is fine for a different step heights etc. We need a better resolution that is still. The main difference you were seeing between what you call the AS there and the thermos model there would be solution of features. Which if you have that ready to run transition to thermos maybe they will be closer to the thermos models which we are getting into the so called AS. Okay. Solver is a statistic. So, basically the solver is I mean formulation is unsteady. So, that is my question. Are you reporting us as a time average result or you run a solver? No, solver is not steady state. Solver is unsteady state and we are taking after an invariant solution. So, you do it in time average. Yeah, in time average. So, it is not a solver is not steady state. We are taking after a time average of after you say that things are steady. That time the work is. Heat transfer. It is actually in this one this is not conjugate. We are studying the in this one the heating is right at this place. So, it is not a conjugate in this particular study. So, you are not taking account of the conductivity of this only. No, what right now it is not there is only if you say fluid domain this particular thing is an entirely fluid domain. Boundary condition for the temperature is that we have a specified value at the inlet and this is specified wall temperature specific at this along this length. This is the x direction. It is a fixed temperature along the lower wall and insulated top wall the temperature boundary and there is a inlet temperature and homogeneous Neumann on the outlet on the temperature. Yeah, we have a resource of an experiment which we just need to be useful for while dating. We have carried out a study on the grid for the one one case it has been done. Yes, yes, yes, yes, we have done that. No, the whatever the grid we have. The what you said the you mean the y plus values. The y plus values have been in the order of around 10 for the models which are used. Yes, because usually it means. So, but it's not so easy. You are probably you capture the big oscillation. And in many cases, it's DNS because if you ask how much is your the sub grid is going to be one point the molecular is going to be not so easy. Okay, that's a good resolution. And good and good that he has. Yes, I'll ask a question that you and I have already talked about. Yeah, so maybe other people have some input. Last fall, graduate student interviewed over 100 people individually to see how they used CFD. So, one of the questions that the student asked is that software do you use or what what were the influence was? Well, we use them all. We just want to make sure we get all of them consistent. And then he asked the question, what problems can you do? And they said, we transferred. Then the question was, do you mean all of it? So it was just a kind of interview. And so we had a discussion to resolve. And it's not something that you need a direct American simulation for. How are you going to answer? I mean, maybe you can respond to the question. Actually, what is there that when the students, the community, when they learn the CFD, they just use it by clicking somewhere here. But they really don't get into the necessity, say with variable density or whatever is there. So that is what. And he also has a project in which undergraduates can do CFD problems on a practical sense.