Okay, so, with this in hand, how do we write down the shape function, the basic functions, right? So, again, let me now draw this. And again, because I'm going to draw the basic functions here, I'm going to draw this triangle, though it is, of course, the same triangle I had in the previous slide. I'm drawing it now in perspective view. Okay, all right, so, in this, In this view, I'm going to first write out the basis functions, and then sketch them out. Remember this is A = 1, that is A = 2, and this is A = 3, okay? So in that setting, it's really easy. N1 (xi 1 xi 2 xi 3) = xi 1, N2 is xi 2, N3 is xi 3. Okay, and in particular, it is the convenience of writing out these basis functions in this manner, in this, so to speak, completely symmetric manner. That also motivates the definition of the third of these coordinates, xi 3, even though the span of the space is just 2. Right, and so the dimensionality of the space, sorry, is just 2, okay? Right, so let's sketch these, right, so I'll sketch just one of them to begin with, and, So I'm going to sketch N1, okay? As you can see from that, N1 goes to 1, right, at xi 1 equals 1, right? At xi 1 = 0, which is the xi 2 axis, it drops to 0, right, so to 0 all along the 3-2 axis, right? And it goes down linearly, okay, along the line, right, along this line here. All right, along that line, right, the N1. Of course, it's already equal to xi 1, right, along that line, it slopes down linearly from 1 to 0. Okay, that is xi 1, and likewise we can construct xi 2 and xi 3. Xi 2 is 1 there, slopes down linearly to 0 at the other two nodes, right, and that goes down to 0. And likewise, you see xi 3 or N3, sorry, is, 1 at that node, goes down to 0 linearly. Okay? So that is N3. We should write, okay, N2, Is that basis function, and, N1, I've wrote in green, N1, I drew in green. Okay, so those are our basis functions, and the one thing to note, of course, is that, You will have observed that I referred to these basic functions as being linear, rather than bilinear, okay. And it should be completely clear why, right, they are indeed just linear, not bilinear. Okay so that's it, once we have these basis functions, everything proceeds just as before. We just need to pay attention to the fact that the number of nodes in our element is not four. Okay, even for the simplest of this class of elements, it's not four, it's three. And we can go ahead, and construct our entire finite element formulation. One can go into higher-order triangular elements also, there are quadratic triangular elements which involve mid-side nodes, as well. And one can define the base functions for them also, as an extension of what we've done here. I'm not going to get into it, because really, if you do need quadratic triangles, you are probably better off just using bilinear quadrilaterals anyway. Okay, so these elements have certain advantages, the primary advantage of them is of course simplicity, okay, so simplex elements in general, They're simpler than, Quadrilateral and, And hexahedra. Right, of course, I haven't put down the simplex element in 3D, but I'll do that pretty soon. They're simpler, but here's a drawback, right, there's an obvious drawback, and what do you think that could be? Just the basis, right, I mean, you're using a lower-order basis here. You start out with a linear basis instead of a bilinear basis, your representation of functions is already a little poorer. There are issues of incompleteness of polynomials, and so on, which arise. So in general, they don't work as well as the corresponding bilinear, biquadratic elements and so on, okay, but nevertheless, they are simple to use, and they are convenient. They also, of course, for certain problems of generating meshes, they do present an easier task. But that has largely been overcome in the finite element community. All right, so we have this, all right, let me just put down the simplex element in 3D, and then we'll have got this covered as well. Okay, so in 3D, All right, so we have tetrahedra. Okay, so a general element in the physical domain may look like that. All right, it will have four nodes, right, the simplest linear tetrahedra Right, I'm showing you the simplest linear tetrahedra. So this would be our Omega e, and it would be constructed, as you may imagine, from a pairing domain, right, which is, Also in 3D, but it is a tetrahedron in 3D, right, sort of a right tetrahedron in 3D, so it is, Okay, and, I guess in order to make it completely clear, I should make some of these lines dashed. Yep, that's really the only line we needed to make there. Okay, so the coordinate directions would be xi 1, xi 2, now we would probably have a xi 3 coordinate dimension, right? And here in the physical domain, we may have numbered them as whatever, A equals 1, 2, maybe 3 and 4, right. What happens here is that we have 4 nodes, of course. Right, so we will label them here at A = 1, A = 2, A = 3, and here we will get A = 4. Okay, so convenience again, we will define now xi 4 = 1- xi 1- xi 2- xi 3. Okay, all right, using the same sort of idea, in that case we use this idea of area coordinates. Now we can I guess effectively call them volume coordinates, right, but the idea is just the same. Now all we have N1(xi 1, xi 2, xi 3, xi 4) = xi 1, right, until we come down to N4, And that's it, right, we just go ahead, we now know that the simplest of our basis functions is linear, there are four nodes to the elements, four degrees of freedom. We just go ahead, and construct our entire finite-element formulation just as before, okay, calculation of derivatives and everything, it's straightforward. One remark I could make, is that now when we work with linear triangles, or linear simplex elements, What is the order of any gradient, now, of our basis functions? Therefore, the order of our representation of any gradient field, Is constant, right? Right, so if we take any of these linear basis functions, and then compute its derivative with respect to any physical coordinate, right, what we will see is that it leads to constant gradients, all right? And this, in some cases, can be an advantage for simplicity of calculation. But of course, if your best representation of gradients within the element is a constant, you are obviously losing some fidelity of representation of functions, and so forth, okay. You're just not able to represent higher order functions, right, so there are those sorts of drawbacks as well. So when it comes to integrating with these types of simplex elements, of course, there are special numerical quadrature rules, numerical integration rules. Just as we did in the case of triangles, we can also define higher order tetrahedra. Okay, just as we do for triangles, right? And one can construct a triquadratic tetrahedra, sorry, not triquadratic, cubic tetrahedra, and so on. One can also define, there are numerical quadrature, or numerical integrations, rules defined for them. The optimality of Gaussian quadrature, however, does not hold for triangular elements, right? There are other numerical integration rules that are better used, so Gaussian quadrature is no longer optimal. Nevertheless, these can be defined, these have been defined, and they're actually available in much of the finite-element literature. Okay, but we can stop this segment here, and indeed, this unit. Hopefully it allowed us a unified representation of basis functions, definitely of the linear, bilinear, and trilinear. And therefore quadratics, biquadratics, and triquadratics, and so on, based on Lagrange polynomial functions. And we've also seen how simplex elements are defined, the simplest versions of simplex elements, the linears. We also saw, in the course of this unit, the extension of quadrature to three dimensions, two and three dimensions, for basis functions that are constructed from Lagrange polynomials. All right, we'll end this segment here, when we return, we'll start up a new unit.