Section A: Introductionto CFDExistenceof modern technologies had eased the application of science for humans. Withprogression of computational performance, Computational Fluid Dynamics (CFD)software was introduced to the modern society to study on flow of fluid withoutany physical subject to be tested on. This software was implemented to a widevariety of industries which includes automotive, aerospace, astrophysics,chemical manufacturing and power generators. Before the introduction of CFD,researchers had to rely solely on their knowledge of mathematics and fluidmechanics to estimate an outcome of fluid flows.
While the software is of aidto modern society, CFD would need a computer to process its simulation and dataretrieval of the desired simulation. Therefore, a substantial amount ofcomputational performance was needed to ensure the data was processed quicklyand accurately to meet current industries standard of efficiency.CFDsoftware is a tool where knowledge of fluid mechanics, mathematics and computerscience were applied in simulating a fluid flow motion. These motions arederived by mathematical equations and represented in computer programminglanguage only readable by computers.
The results were then converted to quantifiabledata that is displayed in readable human language. During the simulation inCFD, all progress was predicted using numerical data as from derivation fromall available mathematical equations in fluid dynamic flow.Whilefluid is solved in mathematical solutions, there are still fundamental stepsrequired to be obeyed. (Fawehinmi et al., 2005). There are threemethods in this problem solving through Experimental, Analytical andComputational solving where data are validated. Back when there is nocomputational aid such as CFD, Experimental solving uses a more tedioustechnique which is to build a scaled down replica model representing the realobject of study and the flow properties are to be measured and recorded.
Thistechnique is frequently used, where experimental data were all recorded, andthe results were compared to theoretical values derived from mathematicalequations from fluid dynamics. It is different from Analytical solving asanalytical does not require any physical model to obtain theoretical data. Themethod involved for analytical is by comparing measurements of the desired objectof study in mathematical modelling to related equations from equations of fluiddynamics. But Analytical solving applies to a limited number of simplified flowgeometries.
There were recorded errors from the above methods of solving whichleads to search for improvement in error reduction while accurately obtainresults subjected to fluid dynamics. Throughyears of testing and perfecting, CFD software is now able to display resultsaccurately while having little to no error. It applies to any fluid flow, be itsimple or complex fluid flow.
With such accuracy, computational solving methodhad to be ideal choice of studying fluid flow problems as the results were morereliable albeit the other methods, Experimental and Analytical solving approachwere still being used. Fluidflows are controlled and influenced by partial differential equations that isrepresented by laws of Conservation of Mass, Momentum and Energy. CFD uses theabove governing equations and below represented by following Navier-Stokes equationsbased from conservation laws:a) Conservationof Mass (from continuity equation) b) Conservationof Momentum (from Momentum equation) Where; i: Local change of timeii: Momentum convectioniii: Surface forceiv: Molecular-dependent momentum exchange(diffusion)v: Mass force c) Conservationof Energy (from Energy equation) Where;i: Local energy change with timeii: Convection termiii: Pressure workiv: Heat flux (diffusion)v: Irreversible transfer of mechanical energy intoheatWhen all theconservation equations were applied to Navier-Stokes equation, the followingsimplified general form is formed: SinceCFD was used due to its high efficiency and low risks, it also improves safetywhile saving on production costs. An example would be the use of wind tunnel todetermine efficiency of a car model cutting through the air (fluid flow) willincrease costs in Research and Development sector of a company. By using CFD asthe base to study fluid flow on the desired model, less electricity would beused on simulation compared to a physical wind tunnel.
Costs of material willalso be saved since no models were needed and changes to the car model toimprove efficiency and performance can be done without wastage of anymaterials. Besides that, CFD was also to be used in ensure minimal losses inpiping by simulating the pipe designs with different angles to reduce losses atbends of pipe, attachment of valves, taper and other pipe attachments. (Gabryjonczyk, 2013).Therefore, this CFD software can solve many fluid problems, preventing designfailure while ensuring safety of its design for prototype or production models.
Hereis the order of the process of CFD software based from the 5th slideof (Choudhary, 2015): Figure 1: Flow chart ofComputational Fluid Dynamics Thecycle of CFD is not that complex to understand to begin with as it all beginswith the problem related to fluid where humans tend to solve from theirknowledge of Fluid Mechanics. This data is then brought into Navier-StokesEquations to determine the nature of the flow. This data is then translatedinto Discretized Form as a computer only understood programming language. Inthis form, the computer can analyse and compute the entered data, Grid by Gridfor each mesh on the model. After the meshing process, simulation is executedto show every result obtained from fluid flow. The results are then convertedfrom programming language to human language for quantifying and displayed foruser to compare and analyse the data.
CFD Analysis ProcessThereare few main components of CFD design cycle. Below are the following importantsteps: 1. ProblemstatementFirst, states the problem to be solved and identifythe flow problems.
Notice the physical phenomenon need to be considered in thisanalysis. Besides, create the geometry of the object (domain) and thosenecessary operating conditions. After that, notice the internal obstacles,internal-surface and free surface of the object. Identify the type of flowwhether it’s laminar, turbulent or it’s a steady or unsteady flow. 2. MathematicalmodelFirst, a proper flow model, good viewpoint andreference frame are chosen.
Identify the forces that influence the fluidmotion. Second, problem is set with the computational domain. Writes out allthe related formula for the conservation law of mass, momentum and energy.
Furthermore, equation is simplified for reducing the computational effort.Lastly, specify those necessary boundary conditions. 3. DiscretizationmethodFirst, Partial differential equations system (PDE) istransformed into a set of algebraic equations.
Through this, it is createdbecome approximately, become discretized versions and because of those smallpart element. It became more easily to be solved. In fact, there are many waysto discretize the partial equations but in the end. The goal of all the methodsare the same, that is turn a calculus problem that human cannot solve into analgebra problem which human can solve. Below is some example of discretizing method:· Discretization of spatialvolumes through finite difference, finite element and finite volume.· Discretization of gridtopology through structured, unstructured cartesian and generalizedunstructured. 4. CFDSimulationFirst, once the object in 2D geometry such asrepresentative cross section is created.
It is necessary tofamiliar those parameters in the simulation.Once simulate it, there are something that can obtained from this simulation.For example, investigation of effect of fluid properties on the overall object,what proper meshing and what appropriate boundary layer meshing to be used andestimate the accuracy that expected from a 3D model. Furthermore, the qualityof simulation results based on some few factors such as assumptions made beforethe simulation, mathematical model and quality, size of mesh and so on.
5. PostprocessingAfter CFD Simulation, the next step is to extract theimportant information from the computed flow field. The important informationsuch as the calculation of derived quantities and integral parameters,verification of CFD model and visualization of 1D, 2D and 3D value. 6. Errorand uncertaintyResults of CFD simulation is depends on theaccumulation of error and the level of variability. The source of causinguncertainty is might be not enough of knowledge about CFD.
Error might be causeby other reasons. If comparing both of it. Error is more easily to find out compareto uncertainty.
This is because, error have many ways to determine, estimatingand solving them. But for uncertainty, there is no proper way to find it outbecause it is because of lack of knowledge about CFD. So, it might remain theproblem that undiscovered ever and in the end. It might also cause otherdifferent big problems. (Patel, 2013) Advantages of CFDAgain,Computational Fluid Dynamics (CFD) is a software or can be said as system topredict the fluid flow, heat transfer, mass transfer, chemical reaction and soon. Furthermore, there are a lot of benefits about it. First, development cost.It is very expensive that using real or physical experiments and tests to getsome important engineering data for design purpose.
Thus, this software is usedto get those important data and it is relatively inexpensive in cost. Data getfrom CFD sometime is more accurate as it is using computer to calculate.Second, execution time. CFD simulation can run or so call test in a shortperiod of time, this make engineering data introduced early before the designprocess. Third, overall data.
Experiments or tests can only get those data inlimited location and it is not so detailed, but CFD allows to get more precisedata and it is very comprehensive. Lastly, simulation under many conditions. Inreal experiment, some fluid flow or heat transfer process cannot be easilycontrol or tested under different kind of conditions. But in CFD, it providesthis ability. Disadvantages of CFDEventhough there are a lot of advantages about CFD. However, there are also havedisadvantages of it. First, for an CFD investor, they have no power to vote orsay in an engineering company since CFD is not their underlying asset. Second,for CFD system, it is very hard to get a perfect grid for a complex geometry,sometimes it can but very time consuming.
Besides, it is also very difficult toget the accurate flow motion result over the whole thing in simulation. This isbecause, those data required a good quality in grid session, perfect grid isvery hard to get for a complex geometry. Lastly, this system is not so friendlyto all people and it is quite complicated to be used as well. Thus, it has totake a long time to keep one practice in how to use this software and how toanalyse the results.
Application of CFD ComputationalFluid dynamic can be used in many ways for predicting the important data todesign something. Below are some examples of application of CFD:· Smoke control system· Classroom· Swimming pool ventilation· Aerospace· Automotive· Electronic Infact, there are still a lot of application about CFD but not necessary to showall of it. Smoke control system is one of the most common usage of CFDapplication. CFD has the capability to simulate the smoke flow in a space withcomplex geometry.
This idea is keeping the smoke away from the objects or to besaid as occupants for saving their life when accident happens. From here showsthat, why CFD is so important in simulation before designing the building forsafety issues. (Saeidi, 1998) Figure 2 Figure3 Figure2 shows the smoke away from the building when accident happens.
Figure 3 showsit’s result of simulation in CFD.