Mission Impossible? Why & How to check CFD Academic Project Feasibility
John is a chemical engineer. He is currently enrolled in master’s program at a reputed university. During his master’s program he got introduced to the subject of CFD and it quickly grasped his interest. He read few CFD books and started conducting internet research about CFD and its various applications. During second year of his master’s program he came across some literature over the internet about application of CFD towards microfluidic flows. Coincidently that was the same time when he had to select and define his project topics for master’s thesis. John was very excited to see wonderful applications of CFD towards microfluidic flows.
Excited with this though he informed his Professor by email that he has selected his Masters Project topic and it is application of CFD towards microfluidics. He quickly jumped to the conclusion that he wanted to do this master’s project in CFD and he will apply CFD to study microfluid flows. Next two days he spent enormous time over the internet to search publications and come up with a project idea for his CFD project. By browsing thought many paper abstract he defined his project problem statement and gave it an elaborated title "CFD analysis of microfluidic lubricating layer for abrasive polishing process". Relieved that his project topic was defined so early in his second year he was relaxed now. For next few weeks he focused on his course works. Exam came and went by. Final exams and assignments took up most of this time for the next few months. In the heap of exams and assignment deadline for next many months he forgot about his project topic but was not worried as he thought his project was well ahead of schedule when many of his friends were not even at the point of selecting their project topic. In such busy schedule the entire semester went by. Then came mid semester break and holidays during which he thoroughly enjoyed. Classes started again and this was the last semester of his Master’s program. After few initial months of coursework he began his work on project.
He quickly got access to the latest version of CFD software in his lab and went thought some initial tutorials. He got to know that there are various aspects of the software which are not straightforward and he needed some sought of software training. He joined academic training near his university training centre and got some initial experience of handling the CFD software. Then he began to create his CFD model in the software.
This is the point where he started to have many questions bombarded into this mind. He did not know where to start. What should be his starting geometry? How should he give his mesh? What are the different boundary conditions and how to define the CFD model in general? Although with many such problems he still did not panic. He took an appointment with his professor to discuss his issues. During the discussion one important aspect he learned was that he should think about the real world phenomenon first before starting to build his CFD model. Well he sat relaxed one evening in his lab and began to list down important aspects of his project problem and how it occurs in real world phenomenon. One after another endless questions came to his mind. How does the flow occur in real world? What is the fluid? How small is his entire polishing equipment? How should he get the CAD model for the polishing equipment? Is there also heat transfer taking place? How much will it take for meshing? Is it possible to generate mesh at all for required geometry? How is the flow coming in and how is it going out? What happens to the abrasive particles trapped in flow? How long does it take for the flow to happen? Then will he have to wait that long in the simulation model? The questions never stopped!! The one instant came when it was time to say Press the Panic Button!! John asked himself in panic an important question “Whether at my entire CFD project is feasible???”
How many of us have heard or experienced similar moments in our academic or professional life where we have ended up asking ourselves in panic that whether my CFD project is feasible? Let us rewind our story at start, at the point where it all begins. When you select your CFD project!
What is CFD project feasibility?
CFD project feasibility is the first activity which one should do before selecting a CFD project either in academics or in industry. CFD project feasibility is the process by which you can estimate whether the actual phenomenon in study can be modelled or simulated using CFD process and if not what sensible approximation can be done in order to create a simulation model. Throughout the process it should be kept in mind the CFD model generated can be used to take design decision or end project objective is satisfied.
Why is feasibility so important in the field of CFD?
Feasibility study is the only tool by which you can be assured whether your CFD project is possible or not. CFD is a combination of physics and mathematics to simulate a real world fluid phenomenon. Now during the CFD process there are many uncertainties and it is seldom possible to build a model exactly as the real world phenomenon. There would be modelling approximations, software limitations and most importantly limitations of CFD as a science. Considering this you cannot always say that any phenomenon happening can be exactly capture using CFD simulation unless you go through a feasibility check.
Not doing feasibility check before a CFD project can affect many things.
- It can misguide you to selecting a wrong project
- It can misguide you in deciding timelines for your project
- It can cost you lots of wasteful time
- It can affect the quality of your project
- It can cost your time and resource wasted in learning unnecessary software
- It can cost you a lost opportunity for publication
- It can result in waste of time and resource
- It can result in wrong design
- It can result in wasteful high costs
- It can result in low efficiency in the design process.
Critical elements of project feasibility
Understand the physics
Before taking decision on using any form of CFD simulation, it is critical to understand the real phenomenon in detail. Clearly understand the entire physics of the actual phenomenon. What is the flow process or heat transfer involved and how it is happening? The best way to do this is to be in a position to explain the phenomenon to someone else with expertise in that physics with all details of physics.
See if CFD can really help
Once you have understood the actual phenomenon in detail next step is to see how CFD can help in the design and study of that phenomenon. Before you decide to use CFD see if CFD can really help and make things simpler for you. The criteria for that can be any of the below:
- CFD provides you a way to study or design the phenomenon within less time and within less cost compared to other methods of study
- CFD should provide you more insights into the phenomenon compared to other methods of study
If the answer to any of these questions is yes then CFD might be the right way forward. If not you need to think deeper!
In-house code or commercial software?
Once you have taken the decision to use CFD as a study tool you need to decide whether you have to use an in-house CFD code (which means write a CFD code yourself) or use commercially available CFD software. The criteria for these are fairly simple. If the physics involved is simple and you have the capability to write a program by yourself to model that physics within project duration then this approach might be suitable to you. In short, you need to be a good programmer or programming savvy person! If your project involves complex physics which is out of reach for programming by you, then commercial software should be the choice. Note that commercial software is general purpose software, already validated for wide range of physics and problem. Hence for most of the complex physics it is usually preferred to use commercial CFD software.
Understand the capabilities of CFD software
Before you decide to use any particular software for project it is critical to judge whether the CFD software has enough capabilities to capture the physics involved in your project. For this, you need to read the related case studies, software capabilities brochure and talk with experts of the software. If you already know how to use the CFD software then you can judge the capabilities yourself as well.
Get introductory experience of using CFD software
Many times students decide their CFD project topic and submit the abstracts to their professor in the process to meet semester submissions timelines. This decision of finalizing CFD topics is often based on what students have seen through their literature survey and attended seminars or workshops. But if you don’t know the software, then taking a decision on your project by just observed literature involves a risk. It is always better to get training in the basic usage of software so that you at least have some experience in the software and you can also have some view of your own on the software capabilities and process feasibility.
Can we replicate the physics in CFD?
First question that you should ask yourself when you take up a CFD modelling problem is whether the entire physics can be replicated in CFD simulation. For complex problems this is most often not possible. Hence you have to take some modelling assumptions without losing sight of your study objective. And based on my experience I can assure you that this is the real fun while doing any CFD project!
Can I handle the Geometry?
Every CFD project required CAD geometry and its 3D equivalent to be used in CFD simulation. The CAD geometry of the actual phenomenon can be a real bottleneck in the CFD process. Many times rigorous CAD clean-up process is required to make it possible to do any CFD analysis on the process under consideration. Focus should also be there on how complex the CAD model is while doing CFD feasibility study
Catch 22 between meshing method and mesh count
Meshing is among the most intensive work during execution of any CFD project. Any CFD analysis is not possible without a good quality mesh and there is a real catch 22 situation here! You can go on making a very fine and complex mesh which will satisfy all physics requirement but then such a fine mesh can be unsolvable or computationally very expensive to solve. So you need to decide the optimum point here! This should also be considered while doing any feasibility study and taking decision on timeline of the CFD project.
There are always the solver unknowns
Just when you think that you have considered everything in the feasibility check, comes the difficulties in solver! There are many unknown problem that might be encountered during solver execution part of the project but they are equally difficult to predict beforehand. The best way to estimate these solver unknowns is to talk with CFD experts who have handled such physics before.
Following is a list of some points that you should consider during your feasibility study. These points will help you to design your own feasibility study approach.
- Understand the fundamentals of CFD
- Understand the theory behind CFD
- Get an overview of CFD software
- Get an overview of application areas of CFD
- Based on your interest and knowledge select an application area of CFD
- Get a thorough understanding of different area where CFD is applied in your field of interest
- Select a real world phenomenon from that field of application which you are interested to study or design
- Define the objective of using CFD in your study
- Whether CFD is at all necessary
- Which software are you going to use?
- What size of CFD model are you going to generate
- How much mesh are you going to generate?
- What is the computational power required
- What all physics is involved
- Are there models available in the selected CFD software which can capture the real world phenomenon?
- What approximations are necessary?
- Is the end objective met?
What is the best time to conduct feasibility study?
The ideal time to conduct feasibility study is right at the start of your project selection stage. You should ideally select 3-4 project topics. Conduct feasibility study of each and then decide based on your interest and comparative analysis.
CFD is also a bit about compromise
CFD is also a bit about compromise. It is very rare that you can perform CFD simulation without any constraints. Similar it is also equally rare that you will be able to replicate the extract phenomenon within a CFD model. Hence it is important to understand that CFD process can also be an exercise with compromise but you as an engineer performing the CFD simulation are the decision makers with an effort to reduce the number of compromises.
Feasibility can save you in future
Feasibility study done at the right time during a CFD project can save you in the long term. It can be a differentiator between a successful and technical sound CFD project and a somehow wrapper up CFD exercise with heavy compromise!