Unlock all 54 lessons of Advanced CFD using OpenFOAMFull course ₹40,000 · paid
Simplified Navier-Stokes Models, Convection and Diffusion
A free lesson from Skill-Lync's Advanced CFD using OpenFOAM course, on Grow by Skill-Lync · Last updated 16 Nov 2026 · Session: “Week 1 Session 2”
What will you learn in this free Advanced CFD using OpenFOAM lesson?
The lesson covers simplified forms of the Navier-Stokes equations, including incompressible flow below Mach 0.3, thin shear layer, parabolized, Euler and potential flow. It then separates convection from diffusion with OpenFOAM examples and introduces what CFD is and its three main elements.
- Incompressible, thin shear layer, parabolized and Euler models
- Potential flow for setting up initial conditions
- Convection versus diffusion, with pipe and methane examples
- Preprocessor, solver and postprocessor in CFD codes
What comes after lesson 2?
The next free lesson is lesson 3, “CFD Workflow, Solution Quality Checks and OpenFOAM Setup” (26 min). Later in Skill-Lync's Advanced CFD using OpenFOAM course (paid): lesson 4, Week 2 Session 1 Part 1; lesson 5, Week 2 Session 1 Part 2; lesson 6, Week 2 Session 2 Part 1. The full course has 54 lessons (18 h 30 min of video); see the course overview or its price and syllabus.
Lesson transcript
Machine transcript of the lesson audio (lightly cleaned). Technical terms may contain errors; the video is the reference.
So, as we saw, the individual components and sort of a very basic derivation of the conservation laws and fitting them into Navier-Stokes equations, simplification of Navier-Stokes equations. So, solving for the entire set of equations in your Navier-Stokes equations is a bit troublesome. By that, we mean that by solving for each and individual unknown parameter, we are increasing the complexity of our solution and hence increasing our computational time. So, to avoid that, we have some of the simplified Navier-Stokes equation models. These are based on a few approximations and assumptions that we make which eliminate some of the parameters.
Very basic examples are if we consider our flow to be of, let's say, low speed. At low speed, or at least as a thumb-off rule, lower than 0.3 times Mach number, the flow can be considered incompressible. And when flow is incompressible, the rate of change of density with respect to time and with respect to the spatial components can be ignored. Similarly, there are a few other examples such as the thin-share layer approximation. In this, we assume that the gradients of flow quantities in the normal direction to surface are the sole contributors to viscous stress.
And every other gradient of the flow quantity in other directions is essentially negligent and 0. The other example is parabolized Navier-Stokes equations. In parabolized Navier-Stokes equations, we assume that the flow is essentially steady, hence neglecting the time terms. And we also assume that flow moves in essentially just one direction and cross-flow is non-existent. So, gradients of that one direction will stay and every other direction will just be assumed as 0. So, then we have Euler equations. In Euler equations, we neglect our viscous effects completely. Instance for a high Reynolds number of flows, when boundary layer is comparably thin to the dimensions of the body, we can neglect the viscous effects and hence employ just the simple Euler equations.
There are other few examples, such as the potential flow. Potential flow is one of the most simplest flow models in the CFD realm. In potential flow, the fluid is assumed to be viscid. And also, we assume that the velocity field is irrotational. So, right here, our gradient of velocity potential becomes 0. Divergence of gradient of the velocity potential tends to be 0. And hence, we have a very simplified model for our Navier-Stokes equations. Potential flow is a very important sort of stepping stone towards generating your initial conditions for your analysis.
It is advised that before running any simulation, you first run a sort of crude simulation of a potential flow inside your domain, which then helps set up your initial conditions properly, and then just run your solution regularly as you do. So, in Navier-Stokes equations, we have some of the very basic phenomena that are happening, such as convection and diffusion. There is a thin line of difference between the two. In convection, we say that convection is essentially movement within a fluid by the tendency of difference in the state variables.
That is to say, difference between two points in terms of pressure, density, temperature, etc., causes movement within that fluid. And the state variables sort of transport from one point to another. Convection is much more prevalent and closely studied in heat transfer problems, where transport of temperature from one point to another becomes of paramount importance. Then we have diffusion. Diffusion is essentially the same as convection, and it differs in one very important aspect. Diffusion is a result of random molecular motions. It is very random in nature and just transports from high concentration to low concentration so as to sort of come to an equilibrium state of uniform concentration.
So, in Navier-Stokes equations, as we saw, these are the convective terms where you are transporting your, in here, density, velocity, and so on and so forth. And in here, we have the viscous stresses, which are essentially part of the diffusive nature in here. Here, we have a very simple example of convection of a pressure from one point to another. This is a quarter section of a pipe simulated using OpenFOAM. So, the flow comes through here and exits here. So, as we can see, in the center of the pipe, pressure reduces and it is transported from the inlet to the exit.
This is a very crude sort of post-processing of the results of pressure and velocity. At the inlet, the flow enters at 1 meters per second. When it gradually reduces in the middle, then just increases at the outlet to, again, 1 meters per second. Here, we have a very simple example of diffusion. So, in here, again, we have combustion of methane. This is also done using OpenFOAM. And we are measuring the production of carbon monoxide right after its combustion. So, let's take a look. As we can see, carbon monoxide is very random in nature.
And it is traveling from low concentration to high concentration in a random fashion. This is a very simple example of just diffusion. Basics of CFD. So, let's tackle what exactly is computational fluid dynamics. In computational fluid dynamics, we analyze a set of systems which involve fluid flow. And we employ the help of computer-based simulations. CFD has a very wide range of industrial and non-industrial applications. For instance, automobile industry, aerospace industry, even in chemical and many of the industrial applications where we need to analyze fluid flow or interaction of fluids with a structure.
For any CFD code, such as ANSYS Fluent, COMSOL, or the RCCM Plus, OpenFOAM itself, these all CFD codes are structured around a set of numerical algorithms that take these governing equations and then tackle fluid flow problems based on user's input. All CFD codes contain three very main elements called preprocessor, solver, and postprocessor. Before we conduct any CFD study, we must understand a bunch of general considerations and sort of ask ourselves these questions just so we can understand what sort of study we are trying to make and what we are expecting out of it.
So, some of the questions being, what do we think is the conclusion that we get from the CFD analysis? How much accuracy is needed for these results? This impacts our CFD study a lot because if we need a study of higher accuracy, that essentially means that you need to employ more computational time for your study. Next, we have the validation methods. So, validating a CFD study in itself is a difficult task. The paramount evidence with which we can validate our study is by experimental data or a published material.
Then, we also need to understand the time that is taken for the CFD study and how much we can afford to put towards this project. Then, we must also understand the... The amount of computational resources we have at our hand which will dictate how much computational time we have at our hands. Then we must also take into account if we have a sort of distributed computational resources available. This helps in reducing our computational time by just employing much more resources towards our study if required. And in the end, we also need to understand if just one single simulation will be enough to validate our study or if we need much more.