CFD and its various uses
Let’s discuss a cool technology, used by student-run projects like TeamKART, F1 giants like Merc, academics from various higher educational institutes like IIT Kharagpur, and even R&D professionals in industry giants like Boeing. It is Computational Fluid Dynamics, popularly known as CFD simulations! CFD is emerging as one of the most popular research fields across physics and engineering with a lot of scopes and might be a promising field to those readers inclined to research.
A good number of the readers of this blog might have seen various “colour pictures” of F1 cars but might not have understood what those pictures exactly mean. Also, it happens that many people tend to associate CFD with its most famous application, i.e., aerodynamics. But as it turns out, CFD is a broad field not only limited to aerodynamic applications! Another common misconception about CFD is: “Upload the geometry into XYZ software, input some numbers, press some buttons, and DONE!” It might be a nasty surprise to many among you (especially those who have only heard buzzwords like “CFD”, “Meshing”, “Residuals” etc., and haven’t actually tried to do a simulation) that CFD ain’t as simple as that. We hope this blog will serve as a good introduction to the world of CFD.
What is CFD?
CFD is a branch of fluid mechanics that uses numerical analysis and data structures to solve problems that involve fluid flows which are governed by the Navier-Stokes Equations… No, we are not going to be your professors! Let us keep things simple (Though if one plans to learn CFD to greater depths, they inevitably have to face this jargon. Can’t help it) So for the time being, treat CFD as a way to calculate the velocity, pressure, and other parameters of a fluid at various points in a fluid flow by solving a certain system of equations (basically the conservation of mass, momentum, and energy) using computers. Depending upon the situation, the computer could be your homie lappy or big boys like the PARAM SHAKTI supercomputing system here at IIT Kharagpur, capable of doing 1.66 PetaFLOPS!
In simple words, in CFD we divide the system (here, the Computer-Aided Design or simply, a model of the system generated using a computer) into multiple small grids (this process is called meshing) and then numerically solve the governing differential in each of these grids, taking into account all the boundary conditions. Something as generalised and powerful as this can find its utility anywhere and everywhere we deal with fluids.
In fact, the colour pictures of the CFD simulations you see related to F1 aero and similar stuff actually describe the pressure or velocity distribution of the airflow around the car. Usually, the red regions correspond to regions of high pressure or velocity and the blue ones correspond to regions of low pressure or velocity. Also, CFD is not merely a tool for creating a visualisation of the fluid flow around or inside a body. When supplied with the required data, it can actually give good estimates of the forces and moments experienced by bodies interacting with the fluid flow, and this is what makes CFD a favourite tool for scientists and engineers alike!
Some of the popular commercial software used for CFD simulations are ANSYS, COMSOL, Autodesk, Altair CFD, and Simscale (the last two happen to be sponsors of TeamKART). OpenFOAM is a very popular open-source software used for CFD. In fact, there are even more powerful open-source software but unlike OpenFOAM, these software are currently restricted to academia (it takes time for the industry professionals to “copy” stuff from the academia)
CFD and Aerodynamics
Nowadays, CFD plays a major role in the aerodynamic development of vehicles (from cars to airplanes), wind turbines, tall buildings, bridges (you might want to watch this if you haven’t: Tacoma bridge), etc. Some of you might be surprised to know that CFD is used to improve the performance of golf balls too! We, at TeamKART, do a lot of CFD analysis. This is due to the following reasons:
- The cost associated (in terms of money as well as time) with traditional physical testing methods like wind tunnel testing
- It is simply impractical to make models of various parts, test them one by one and throw them away if optimization is not achieved
- All you need for testing models using CFD is a good PC.
- A lot of institutions and organizations might have little to no access to a proper wind tunnel facility.
CFD has its share of disadvantages though. Of course, there are issues like inaccuracies arising from oversimplification of the real-life situations faced by the car on the track. Also, full-body simulation of a car requires removing smaller features from the car and making approximate solid bodies of the nosecone, sidepods, etc. introducing more inaccuracies. But overall, CFD turns out to be a good tool for aerodynamic optimization and validation.
TeamKART and other manufacturers concerned with aerodynamic performance use CFD software to visualise the airflow around the car, find the downforce (negative lift) and drag experienced by various components, as well as the dependence of downforce and drag on the vehicle orientation and optimise the aerodynamic package accordingly.
CFD and Powertrain
Now, we discuss another application of CFD, i.e., the powertrain. Powertrain refers to the engine and the transmission of a vehicle and CFD has important applications in this field. Due to the various chemical reactions and rapid heat transfer happening in the powertrain components, especially the Internal Combustion engine, CFD has to be often coupled with various chemical reactions as well as thermodynamic mechanisms to be useful as a tool to analyse combustion and associated phenomena.
TeamKART and the automobile manufacturers in general, uses CFD to analyse IC engine characteristics, intake system analysis, and model fuel tank sloshing (A good analogy for sloshing would be the movement of water inside a half-filled cup when we carry it) Coming to IC engine characteristics, the CFD results help us in deciding several factors like efficiency, torque-power characteristics, and mass-air flow rate while designing the engine. The CFD analysis of the intake gives the pressure distribution across it and helps us to understand how turbulent the flow inside is. This is important as turbulence promotes better air-fuel mixing, which in turn increases the efficiency of fuel combustion. Additionally, the volumetric efficiency, which depends on how much air flows into the cylinder during the intake stroke, may be affected due to turbulence. So, the trade-off can also be decided based on CFD results. The modelling of fuel tank sloshing is also important as it generates dynamic load as well as a small noise (especially when it approaches resonance).
Other CFD Applications
CFD is used very widely in almost all kinds of industry, from turbomachinery, chemical reactors, power generation, automotive sector, chemical manufacturing, polymer processing, petroleum exploration, medical research, and even in the food industry. Guess what, it can also aid in weather forecasting!! With the development of computers having enhanced computational power and memory, it is now possible to simulate fluids and heat transfers on a large scale. As a result today, fluid simulation is an essential tool to simulate meteorological phenomena, from daily short-term weather forecasts to long-term climate change. These large-scale fluid simulations are also useful in the study of geophysical flows.
The concept of CFD is also now being widely used in astrophysics, especially in Astrophysical fluid dynamics. Astrophysical fluid dynamics is a modern branch of astronomy involving fluid mechanics that deals with the motion of fluids, like the gases which the stars are made up of or any fluid which is found in outer space (mainly plasma, which can be treated as a fluid with magnetic properties).
Even though CFD is still an emerging field in biomedicine, it has a very important role to play in medicine and healthcare. Many simulations and clinical results have been used to study congenital heart disease, heart failure, ventricle function, and related topics. With decreasing hardware costs and rapid computing times, researchers and medical scientists are trying to increasingly use this reliable CFD tool to deliver more accurate results. CFD is turning out to be an important methodology to understand the pathophysiology of the development and progression of diseases and for establishing and creating treatment modalities in the cardiovascular field. Thus, collaborations between mechanical engineers and clinical and medical scientists are essential.
CFD also provides a means by which the fundamental mechanics of fluids can be studied. By using massively parallel supercomputers, CFD is being widely used to study how fluids behave in complex scenarios. Complex phenomena such as boundary layer transition, turbulence, sound generation, multiphase flows, etc, which have tremendous relevance and applications in our day-to-day lives (and which we often don’t tend to notice) are being investigated through this method.
In short, CFD is very useful whenever we are dealing with any kind of fluid. From large-scale industrial processes to tiny microscopic flows within the plasma of a single cell, CFD can be used to validate, investigate, predict and understand the hidden mysteries of nature.