 Yeah, so our next session will be by Professor Dheeraj Gaur of Shivnathar University. So just a second. Yes, I can see that your presentation is already there. So Professor Dheeraj is a professor in the Department of Chemical Engineering, School of Engineering, Shivnathar University. So, sir, over to you. You can start your session. Okay. Thanks a lot. Good afternoon. Namaskar, everyone. So welcome, everyone, to this workshop, and we are taking few courses, a few sessions on mashing. So Professor Hari Krishna has already taken one session on introduction about mashing the purpose, what are the different types and so on. And you also had some hands-on on mashing in the previous sessions. I would like just to add more information, more understanding of what exactly mashing is and how it is going inside the open form. So in this, I will discuss about the mass specification and the energy constraints, how the numbering is done with the vertex space and edge, how the mesh is generated, and in the end, why do we require so many different types of mesh? So each of the things is important since you are doing only open form, where you have to write certain code. So you should have a visualization of what is happening inside and why it is required. So the major source of content here, I have taken it from open form itself. So for details, you can actually go to this link and actually go through it for more details. So let's start with the classic. So the first thing, how to specify mesh and what are the constraints to when you do it. So the basic thing of classic start will be points between zero dimensions. So we need to have time and they are located in 3D space. And what we do in open form is that we create a list and the numbering of each point starts from zero. Unlike in other cases, if we say one, two, three, it starts from zero. So if we have three points, it will be numbered as zero, one, and two. Okay, now there are things that you cannot have two different points. The base of the mesh is the points which are zero dimension. They are located in space and they are numbered from zero and they are stored in the, with the numbering. Now, the thing is that you cannot have two different points with two different numbers, but exactly the same identical location, same identical position. You can only one point with the same identical position. Besides, another constraint is that there cannot be any point that is not part of at least one phase. Okay, so that means each point must be connected and it must be part of the phase. There cannot be isolated points which are there. So sometime what happens we may create many points, but we forget to connect them and that will may cause any issue, give you warning that there is a issue. Then from these points we want to generate spaces, which is basically two dimensions. So we cannot jump from 1D, 0D to 2D. So what do you need is how to generate spaces. It's basically then using the ordered list of points. So ordered list point means the point may have its own numbering, but you need to tell in which order you need to put so as to create phase. Now what is the importance of order is that when you connect two neighboring points to in the list, it creates edge, edge is one dimensional structure just like a line segment. Okay, so once you are connecting edges, when the edges are connecting back to origin, it completes the thing and phase is created. And once phase is created, it's like a plane and every plane is having a direction which is represented by a normal vector. And as it's a direction normal vector, we use the right hand rule. And you can remember that like A4 sheet is there. So you can write in both sides. So just to visualization, you can say one side is blue and another is red. So blue is in the positive direction or which is the preferred direction and the opposite direction or negative whichever you want to as a reference is the red one. So blue we are talking in the positive direction. So we can see this here like this is the phase created from points. So point 0 going from 1 to 2, 3, 4. This is how you write an open form and it creates a phase. Now remember this thing instead of writing like this, let's say I change the order. Instead I write the order, let's say 0, 3, 2, 1 and then 4. So do you think despite mentioning all the five points, will it create any real phase? The answer is no because it will change the formation of edges. From 0 it will go to 3, 3 to 2, 2 to 1 and 1 to 4 and 4 to 0. So these two additional edges and the sequence will not create a realistic phase or a surface. So order is important. Another thing is that in the same five points depending on the you are mentioning like 0, 1 and 2. Let's say I stop at this point itself. Then what will happen? It will create a phase from 0, 1, 2 and 2, 0. So this can be one different phase. I can create another phase 0, 2, 3 and 0. So like this I can create a phase by minimum three points. I can use more points but minimum three points are required to create any phase. And remember from the geometry, three points are always coplanar. Only when four clicks are there, there may or may not be enough same plane, which will add additional complexity. Now moving to further things of the complexity of faces, then there are two different criteria of faces we see in the machine. What is called internal faces. What is meant by that is that an internal face is between the cells. So when you are creating several cells, the common face between both of them is the internal face. Now it has to become part of both the face. Okay. And since this is common, so the face normal point into the cell with a larger label. So larger label means the cell which is the direction of normal on the surface would be from the cell which is formed first towards the cell which is formed later. Okay. So this is how we are doing. And from the previous case, when we are talking about faces and the direction, we always follow right hand rules. So all the faces that should be made should have the normal towards the positive coordinate direction we are chosen. So it has to have the consistency. Okay. So now these are the internal faces. Now another type is the boundary face. So internal face will always see two faces, two cells on both sides. It will be a common surface. Whereas the boundary faces are the one, which will have only one cell on one side, but other side there will be no cell of the same block. There may be another block, but it's not part of the same block. And hence it has to act as a boundary. Okay. So it will coincide with the boundary of the domain and attest by one cell only as a boundary patch. And the face points, so that they are normal. If you look at the normal, they will point outside the computational domain. So when it is pointing outside the domain, it will not encounter any cell within the computational domain. Now, from this, we want to generate. So we have come from 0D point. We have come to the edge 1D, and from edge we generate 2D faces. Now together we create this cell, which is 3D structure. So here they can be mixed of faces in arbitrary order, no particular order is necessary, but there needs to be certain important things, properties, so as to ensure proper cells. And what are those? So first is the contiguous. That means they must, when you are creating cell, it must completely cover the computational domain. First thing, that means no portion of the domain should left without any cell. And second, they must not overlap one another. That means they must share the faces, but not the face should not be present in some of the cell, inside the cell. It must be at the surface. Otherwise this will not be realistic for normal machine. Second thing is the convex. Now what is the meaning? Well, we simply say cell and center inside the cell, but the convex meaning is coming from geometry. A basic definition says, if you take any two random points in the given structure, so let's say there is a plane given of any given boundary, you take any two random points including boundary, and you create a straight line between the two points. At no point, this line will go outside the plane, outside the structure. So such structures are called convex. If it is going outside at even a single point, then that geometry is not convex. Okay, so this is to ensure that all points are in a way that every point connecting line between them is within the system cell itself. Then another point property should be closed. Then the cell must be closed. And what does it mean? There are two parts. One is called geometrical close. So all face vectors, area vectors are oriented to point outward of the cell. So their sum should equal to zero vector. So if you have created the faces consistently in the positive direction, so if you look at the cube, let's say you create a cube in x direction. So looking from outside, one direction in the beginning, one side would be red, and another side would be blue. So when you add both of them, one is positive, another is negative, the area sum will be zero. If both are blue, then it will be an unrealistic thing. That's why you need to create in a uniform way. Second is the topological. This was the geometrical. That means the direction of faces should be in the same proper way. Second is the topological. Topological dictates that all the edges in the cell are used exactly by two faces of the cells. What does it mean in fact, as you see the boxes, like there are sweet boxes, some sweet boxes you create by simply folding the sides. So each side is one face. So like you call it close when all the sides properly connect to each other. So there you will see each edge is a part of two attaching faces. So that ensures that it is not open. The cell is closed. If a edge is shared by more than two faces on one side and one face on another, then what we do, that we create an additional point there at the meeting point and create instead of one single edge, we create more edge. So as to avoid this issue. Another is the orthogonality. Orthogonality is basically that two cells are, if there are two consecutive cells and you join them from center to center and you see the face, normal of the face. So angle between the normal of the face and the two lines connecting the center, their angle should not be more than 90. Because in that case the flux from one cell to another would be 0. So that is the orthogonality. It's also part of the skewness. So we want to restrict the skewness to certain level. So now this leads to the another thing called as boundary. Once you have created the inner things you have matched the inner faces now you have to define at the outer, which defines the boundary. So there are list of patches in open form. That is the name of the boundary, like inlet, outlet, wall. So like this there are conditions which you do. So clearly one must contain only boundary faces and no internal faces. So boundary must be at the outer surface, not in the inside. And to ensure that all surfaces, external surfaces have been matched with the boundary. The sum of all boundary face area vector should equal to 0. So this condition ensures that you have not left any face which has not been given any boundary condition. So this is a mathematical way you ensure that all faces are properly given the boundary condition. Now, when we have done this, the numberings are there in the open form using the cell shapes. So you can see, we can take a head roll. So you can see in the left most first we create a vertex 0. So this will be the origin. So 0, then in the positive x1 direction or x direction you create 1. So that is 1. y direction or x2, you create 0.2. Then back to 0.3, which is in the negative x direction, but in the same and then from 3, it will automatically come to 0 to create and you see if you see in this positive y direction, this is counterclockwise because the theta is positive counterclockwise. And hence it will be able to create this face whose normal is in the positive direction, positive z direction. Now, to create this size, we define 4th which is in the positive z direction. And from here, again you follow the same sequence 5, 6, and 7. Once you are able to do, then it completes the hexahedron. Now, if you see it creates 6 faces. So faces, you can see this is face 0, then 1. The front face is then 2, at the back face is 3, the bottom face is 4, and the top face is 5. And the edges, if you see from this, this is the 0 edge, 1 edge, 2. So in this direction, 0, 1, 2, 3. Then 4, 5, 6, 7 and 8, 9, 10, 11. So for x direction, keeping in the x direction, varying in the y and z, you are doing in the beginning, then keeping common in variation in x, z. And then keeping z constant you are doing in the other direction. So this is how you are having. So now you can actually visualize how the numbering is happening, why the sequence is important. Now, you can create different type of geometries with this thing. So for example, you merge 3 with 0. So you can create a wedge time. What will be the thing basically here? 0, 1, 2 and 0. So because 3 is now merged as 0. And then this is the lower face and the top face should be then 3, 4, 5, 6. Because now there are only 6, 7 points only. So you can easily understand based on this that the third point has now shifted to 0 location and accordingly faces will be created. So this is how these are the faces. 5 faces. So 0, 1, 2, 3, 4 and 5. So these are the faces it will be creating. And accordingly these are there. So if you want to create a print, you can see here even the 6 have merged with 3. So what will you write? 0, 1, 2, 0 3, 4, 5, 3. So this will create this one. So this is for the structure grid. If you want to go let's say for pyramid, you know what to do. Okay. So tetrahedron and tetrahedron. So these are the different types it is creating and the idea behind that. So we will block mesh generation. As I said the thing that is taken is that the point 0 is taken as the origin. Then the first point is in the positive x direction. The second point is the positive y direction x2, then x3 and back and then similarly for 4, 5, 6, 7. So you can see here that this edge line is curved whereas the edge stand is straight line. So normally it should have been a straight line but you can actually make it arc type also. You can give what should be the curvature of this line and the thing is as I said 3 points are always coplanar, coplanar. But the fourth point may not be coplanar and here that is the case. It may happen that because of this these 4 points may not remain coplanar and hence the center of the cell of this face may not be physically on this area but its projection in this direction would be on the cell phase. Okay. So this is how the flash is generated. Now there is another thing that you might have seen in the previous sessions or will be coming in. What is the expansion, cell expansion ratio? Now as we find the cell expansion ratio is the delta e upon delta s. That means the last cell and the first cell. The ratio of that is the expansion. So if it is expanding this will be more than one. If it is contracting it will be less than one. So that is how it is defined. Now if you look closely basically what happens is the cells that are being created is following a geometric ratio series. So you must all of you must be knowing geometric series. So like if this first cell is the length a then having a constant factor r the next cell would be a r and then next is a r square and so on. So each consecutive cell would be having a ratio of r. So here you can see technically the expansion factor or ratio is last cell which is a r to the power n minus 1 upon first cell which is a. So this is the total expansion factor. So you can actually create on your own. You can write equation accordingly. And you can see that if r is more than one the cell will expand or consecutive cells if r is less than one the consecutive cells will shorten. Now why do we require this at all that we will see in later part of the presentation. Now next come once you have created mesh then you have actually create several blocks. So you may require to create several blocks not just one block. Now what to do when you create multiple blocks then there are issues in this case you can see these are inner faces whereas these outer what you can see is the boundary faces. Now if you are bringing them together then you need to match the face. So when you are matching the face they must they are formed from the same set of vertices. And the faces do not form an external boundary and combine each collocated pair into a single internal face that connects cells from the two blocks. So you can merge two blocks you can create them separately and then merge. What happens is that sometimes when you are generating geometry so you generate different blocks separately but then you bring them together and merge them into a single block. So when you want to merge them you make a single block you need to match faces exactly because if they are not matching the information passed through them will not be similar. So that is the requirement. Now another thing that may require is the face merge. Like in this case what happens here is a group of faces from one patch from the one block are connected to another to form the patch. And they create a new set of internal faces connecting the two blocks. So now here when you merge this what happens this external boundary is the new face rather than this and this together separately. So you can merge the faces when you have several faces which have the same, let's say same boundary conditions. So let's say you have a cylinder and outer surface is the four different surfaces are there each of them is let's say wall boundary conditions. So instead of working on them independently you can merge them together and create a single boundary. So like this you can merge the faces. So as I said you can create a wedge or another type of mesh depending on the requirement from 8. So 8 vertices is the common and what you can do here is 0, 1, 2, 3 in the first case and to make this mesh type 4, 5 then 6 is written as 5 and 7 is written as 4. So 4, 5, 5, 4 so that will create this type. So whatever the type you need or it requires to be generated it can be done accordingly. So actually there are several softwares are there now commercial software and others are there which actually are doing all these jobs on their own so you can make it more user friendly but at the end of the day at the back end procedure remains same okay they are charging you because they are doing on your part making your life easy and in those cases where this is not the case you can do on your own and they are free of cost. So it brings naturally brings all these things together to us that okay we have seen so many different types of meshing so different types of the thing together and how will you decide which mesh to choose in what case and how what should be the number so first of all is to have the adequate number and type of meshing that is the very basic reason that why we have so many different types because both the number and the type okay as professor had a question I will discuss structured and structure grid in structure grid there are different types we have seen in an structure grid, tetrahedron, trisman other types are great so we need to have an adequate number and why do we have that thing first to cover the geometry because we want to cover our geometry properly second is to get the current and conservative numerical solution okay because mesh is nothing but the spatial representation of the basic equation we are using the differential equation so if you remember differential equation is nothing but when we say delta by delta x tends to 0 okay in reality delta x never tends to 0 there is a limitation so that limitation that delta x is finite is the visual representation in the form of mesh so as mesh tends to be smaller technically we are saying delta x tends to be smaller and approaching the differential equation okay so as we are approaching that we will get the result similar to the differential equation and third is to simulate different types of physical phenomena which may require the different type of machine we will see and later chapter later session will be covering that so let's see this case so now you see this is the surface we want to generate the machine now if the machine previous session we have discussed if the geometry is simple and when I use the word simple it means generally straight lines which are orthogonal so structure mashing in straightforward case no issue with that so we are not interested in going for another type but unfortunately all geometries are not simple like this is the you can see straight lines are there but angles are some angles are there are curvatures so would we be able to cover it with the structure machine so let's start with this in this case so snapping smash is the thing which is going to generate so let's say start with this you can see now when we have a structure group then to cover this we start the grid the cells small cell larger cell and all that we start trying to cover every part of the geometry so for doing the procedure we create more and more to cover the boundary so we need to do so we created small cell somewhere small cell somewhere larger cell so very basic idea is that wherever the curvature or the gradient is large you need to require smaller machine okay the mathematical part is the Taylor series when the because we truncate it to the two terms making it the first order so the idea is that the delta x is small but when the gradient is very large then you need to make delta x even smaller further to make this approximation similarly here so you can see when you remove the remaining mesh which is not required to be computed it leaves the structure like this you can see clearly that using this mesh this type of structured mesh we are not able to cover the geometry in proper way the curvature is not taken up properly in this direction somewhere it is outer side somewhere it is inside so how would you cover it further then you further restrict the area which you require machine and then we further refine and change the thing truncate this part so you can see now those portions to cover or map the geometry boundaries properly from this it has been changed so some portions are being removed in the sense those vertices are moved to the point so different machine is generated ok this is why you require different type of machine then to further capture certain places where let's say here the drag is more you want to see variations are high so you create additional surface additional meshing here ok so another example can be taken which is I have taken from this reference you can see that wherever curvature is there uniform meshing structure meshing is not a very good idea because it does not cover curvature properly so here you can see here we have used unstructured grid ok which is much easier to cover geometry the curvature is the least now one of the important thing that you can visualize here is there seems to be solid sphere and cylinder but what are the things you can simulate through them so depending on what is the properties of individual blocks so they are different blocks ok we have made them different blocks ok from this we have removed this part from this and this is there so their interaction is that the surfaces where they need to match so depending on the properties inside in each block you can simulate different things in the same geometry for example let's say I define air inside this block and there is a liquid inside this block so it's a physical case that bubble is in a liquid and if I simulate gravity then because of buoyancy it should rise up ok if there is no gravity I am basically simulating the zero gravity condition accordingly you will see these facts ok in the same if I make the reverse for example in the same sphere if I simulate if I declare the property of liquid and the this channel or the cylinder I declare it as air so this will simulate the condition of drop falling in the air in a channel ok so that is how I can simulate different things I can simulate well let's say this is oil this is water so it's like an oil drop in the water so same geometry can do several things it can be considered static also if this is a metal and this is another metal I can see if in a heat transfer how will they move how will they expand or think so just because certain geometry there does not mean that by default the machine or the system will know that what to do with it we have to define and accordingly it will be work out ok so if we have to create a cavity here we don't have to give anything we don't have to specify here so that will be recognized there is now material inside it so we can simulate a cavity also in the same place so as we see one of the reason that different type of geometry another numerical solution so in numerical solution one portion that has been covered earlier is the grade independence test now why do you require grade independence test already covered by professor Harikishi now the important outcome of that grade independence test is the number of machine because you need to have a certain number of machine at certain places certain place means those areas where you expect the variation gradients to be higher so you need to capture them so when you do that the overall number of machine may increase now what it do is that it increases your computational time and resources ok now the thing is you cannot avoid grade independence test to get a correct answer if somebody thinks that I want to reduce the computational time and resources so let me do it at a lower grade it won't help there are more better there are better ways more intelligent ways to achieve in a lower so if you are using a direct solver just giving you idea how much it is increasing so like say in case of the structured grade they are direct solvers we are using let's say if we are using direct solver then the theory says that in one direction it depends on the n cube that means if you increase the number by twice the computational time will increase by 2 to the power 3 that is 8 times so if you are increasing the number in each direction by twice that means you are increasing number of mesh by 8 times the overall time will increase by 24 times 8 into yeah so 8 into 8 times which is 64 times so you can see that one thing which you are able to do let's say in one hour earlier now it will take 64 hours to complete so that's a very substantial increase so we cannot choose the mesh blindly we have to be careful while we cannot bypass this thing so we can use structured grade we can use unstructured grade or a better thing can be hybrid mesh if the structure allows which is basically a combination of structure and unstructured grade so when we use hybrid grade we can be more judiciously placed more bashing around the place where gradients are supposed to be high if we know beforehand or we expect from our understanding that in certain flow locations the gradients high gradients means high variation of let's say velocity pressure or any variable we are looking at this will be high and we can avoid we can provide coarse meshing at place where the variation is smaller we can also use the symmetry for a given condition to reduce the total number of mesh we will see how so let's say a simple case from this reference so previously before that when it was generated automatically it says the nodes are 1400 and elements are 1309 so nodes are the points vertex elements are the cells so we see these are the cells now when we make it properly arranged let's say in a structured form here whereas on these points we make it increased so although this will give you a better result than this and with the lesser of nodes which is 1073 and elements are reduced to 980 so we can see the substantial reduction in the computational resources and time despite getting and besides getting a better result by judiciously putting the meshing we can also see that the same geometry can be matched with the different types so you can easily see in one and one or the other condition meshing numbers can be created much greater like in this case much less whereas here it is much high in these cases so while you will get the same result but the computational time may be too high okay so that is the thing another thing is like this geometry this is a circular or the curvature come that is are always tricky one because using structure grid we definitely like to do but then you will not be able to capture the curvature easily in the structure grid here what you are saying is the visualization but from the grid perspective it is very coarse so if you want to capture the curvature well the number of structure grid in the axial direction would be too high on the other hand the same thing can be easily achieved using the unstructured grid in a less number of meshing so one of the tricky part is that which one to use depends on the post processing part of the software you are using sometimes certain capabilities are applicable only to one type of meshing so let's say there is a one so one software lets say you want to trace you want to track the particle from the theoretical particle entry to exit it may be available only for structure grid so if you want to do this thing that okay how the particle is moving after the simulation if the software does not allow for structure grid you have to do it for structure grid only so one has to see what gridding is that what are the what post processing is required and based on that you can make a judicious choice another example from the console you can see is here this is the rim of the tire now you can see why using unstructured grid the number of mesh would be too high whereas using hybrid some places you are where the curvatures are high you use the unstructured grid where the curvature is left and the variation expected to be left we use coarse grid so the overall number of meshes much reduced we can further reduce the meshing by exploiting the symmetry so you can see it is symmetric in this direction so if the force or the thing we are looking at is a uniform throughout does not affect the direction then we can take one-fifth of this and we can apply symmetry boundary condition on both sides okay in case of turbine and such rotary dynamic we can apply cyclic boundary condition knowing that one part an outlet of one side becomes the inlet for the another in this case the process is not symmetric let's say there is a force on this direction only then we cannot apply or use the symmetry boundary symmetry of this thing so we have to be we have to be careful while exploiting the symmetry another thing is like in this case so here you can see there are two case two things so it's a microfluidic channel inside the slab how do you know which is inlet and outlet so you have to define boundary condition at this or this point second by itself the system does not know that which one is the where the fluid should flow and where it should not so you can see it that this is a glass tube in a pool of liquid or you can see it another way that this is solid and the fluid is flowing in this one so depending on the type of boundary condition and block properties you can simulate different conditions so there are two different colors because these two they represent two different blocks and you can see if you more that these two blocks are connecting to each other touching each other their face must match okay then only you will be able to simulate say transfer of heat from one block to another or any other thing which we are planning to simulate easily so that brings the last slide which is so there are different types like in the triangular or tetrahedron or even a polygon type this may be quadrilateral hexagonal type and we can use tetra hybrid in one direction and unstructured in another so here you can see an example where same type of mashing is used in so one structure is structured here tetrahedral here okay here you are making even finer mashing using tetrahedral whereas on the other side in this one same geometry is matched with the polyhedral and structure grid so you can see and think of that in both the cases number of mashing may be large in one smaller in another you may be able to cover the geometry properly okay the results as you make any mesh smaller and smaller you will get the same result but at the end of the day what important is that you are able to get the results in a reasonable time with a reasonable effort available to you you cannot plan that okay I have a infinite time infinite resource I will be able to deliver these are the practical that leads to a limited choice so that is also my side thank you very much so if anybody wants to ask some questions you can go ahead for an airfoil case which type of mesh would you recommend well I am not expert with the airfoil type but anyway the thing is as I said depending on what is the level what are the things that you are looking for the simply flow what level the you need to see it may be different at different location where you expect more variation so if the flow is less the Reynolds number is less the mashing may be more like this one as you can see in this one at the back as the flow number increases at the back also you may require more and more finer mashing it depends what capability of the software your software is having it also depends the computational power you have it also depends the time you have to solve the problem so in last session previous session professor Ayakrishnan said in structural grid it may be slower compared to unstructured grid so let's say time is not the constraint so you will see the computational power in the case of unstructured grid generally they are indirect software which are iterative in nature so it has a much lesser computational requirement they are in iterative so they take longer time but they can do in a lesser way so it all depends it will ultimately bring down to that what are the constraints you are working in so that is how it is to be selected I am not expert in that but I try to answer to some extent ok sir thank you hello sir how can I inslide different kind of message in open form today I have inslide block mesh or hex type mesh in open form but in you are visualizing hybrid mesh and poly how can I inslide I think there may be a separate session for that file if there would be a session for that yes sir would there be a separate session to teach them the unstructured grid and the hybrid machine in this course no sir it would be one for the unstructured grid so I would suggest please be in contact with the team and they may be able to help you guide in this how to do it actually my purpose of my talk was to make you aware more about that these things can be done and not as to how to do it exactly so for greater detail I would suggest anyone who is thinking of using hybrid grid I think they better be in contact with the foresee team and they would be happy to help you out so if you want to use unstructured grid you can use salome or make a mesh in ANSI then you can import in open form that you can do okay sir hello sir this is rosen from rosen so sir I am asking sir if I don't want to use any of the commercial software like ANSI then what is the possibility of generating this kind of polygon kind of mesh so if you are using any commercial software sir I don't want to use any commercial software and I want to create this polygon kind of mesh then do we have some kind of some open software or using the open form itself can be how to create such kind of mesh polygon mesh okay Anshuman can you answer for that actually you can use blender to create mesh yeah you can use blender but we prefer salome salome is a free open source software that we use time and again to create unstructured mesh and then import it in open form okay what is it called salome S-A-L-O-M-E salome okay thanks so can you explain the mesh for heat transfer from the surface which you have explained let me see second last slide is this one I think this is what you are talking so in this if you are heat transfer you are talking in this portion I think see you can simulate several things depending what you are planning to simulate so let's say this is a channel and the liquid is flowing here so in this block the blue one you are defining the properties of the liquid okay now in this liquid this is in contact with so the other the green one is to be defined as the solid so you will define it the property of solid now what solid let's say copper, glass whatever that you want to say so basically we are now saying or we are trying to say that in this liquid one side the liquid is answering the liquid or let's say gas or any fluid it is flowing through it and exiting at the other end and while flowing through it is interacting with the solid structure the green one now what are the things you can simulate so you can simulate simply flow you can simulate the heat transfer you can simulate both at the same time so like this whatever you want to do so the boundary condition will be accordingly so for example you want to simulate flow only so the boundary condition for flow condition would be so at the wall the wall boundary condition now if you want to simulate heat transfer also then the boundary condition has to be there for the temperature now whether it is the heat flux whether it is a constant temperature or you want to calculate based on convection that you need to define so if you define that boundary condition it will take it as the heat transfer condition and then you have to define then what is the initial temperature what is the temperature at the surface or the flux accordingly things will evolve now you want to couple it then based on heat transfer let's say the temperature inside the fluid is increasing so if you do not simulate the changing in density and viscosity of the fluid along with the temperature or decrease whatever the case may be then you are basically decoupling the heat transfer with the momentum transfer which may not give you proper result or under the condition it may be acceptable that the rise or change in the temperature may not have a significant impact there in such cases you can use it and you may still get the reasonable correct answer but in the cases where viscosity density is changing are sensitive to temperature there you need to simulate model them as a variable of temperature accordingly it will evolve with temperature and accordingly flow and heat transfer will there and that's where you can so it will become a coupled problem heat transfer and momentum transfer and it will consume more time despite having the same number of things but it will consume more time to evolve into rightful solution so I hope that is clear but if there is anything I can help sir can you if you wanted to know how much time the liquid would spend inside this would you do that so you want to know the time basically it's the residence time so basically what is the total volume of this structure divide by the average velocity so that will give you the residence time average residence time one thing is this if you already know it then it's fine if you know want to calculate then you can do the experiment also what is the RTD of this system that would be then transient study of this one then it won't be a steady state so that would be a different solver then solver would be different depending on the process you want to simulate the solvers would be accordingly but the mashing mashing I am just here I am concentrating on mashing so mashing will remain similar only the thing what you want to simulate would be thing if you want evolution of solution with time then you naturally have to use a transient solver anything else I may help if no more question is there I will say thank you everyone we thank you a lot it was a very good informative session thank you very much ma'am thank you everybody