[mongya]

I am planning to measure velocity of wind made with DIY pitot tube because I don’t have any available flow measurement. But what I concern about my plan is that premise of Bernoulli equation is fluid should be inviscid and that as far as I studied pitot tube should be used in the enclosed pipe. Is it theoretically valid to measure wind with as I said DIY pitot tube?

[Gerry Kan]

Originally Posted by mongya

I am planning to measure velocity of wind made with DIY pitot tube because I don’t have any available flow measurement. But what I concern about my plan is that premise of Bernoulli equation is fluid should be inviscid and that as far as I studied pitot tube should be used in the enclosed pipe. Is it theoretically valid to measure wind with as I said DIY pitot tube?

Dear Jaemin:

Just to be clear, you are using an existing pitot tube, correct? I don't think making your own pitot tube is a good idea because, as simple as it looks, it is still a precision instrument for measuring flow velocity.

In theory, you can get very reasonable mean wind speed. But because it only measures flow in the direction facing the tube opening, it is only suitable when you know the flow come predominantly in that direction. This is why you often see them in such applications (i.e., in a pipe, or on vehicles). For outdoor usage, unless you know the wind is predominantly coming from a single direction, it is probably not very useful.

Whether a pitot tube is useful for your purposes, depends on what you would like to get. From your previous posts, you are looking for validation data for an exercise CFD code you are writing for a class. For such cases, validation data should be fairly easy to come by. I imagine you will be applying your code on relatively established test geometries. I would suggest this instead of trying to obtain your own measurement, which adds an additional layer of complexity and uncertainty.

Unless your course also demands you do this in addition to your coding, but I think this is very likely because it simply takes much more time and effort to accomplish from scratch.

Hope that helps, Gerry.

[mongya]

Originally Posted by Gerry Kan

Dear Jaemin:



Just to be clear, you are using an existing pitot tube, correct? I don't think making your own pitot tube is a good idea because, as simple as it looks, it is still a precision instrument for measuring flow velocity.



In theory, you can get very reasonable mean wind speed. But because it only measures flow in the direction facing the tube opening, it is only suitable when you know the flow come predominantly in that direction. This is why you often see them in such applications (i.e., in a pipe, or on vehicles). For outdoor usage, unless you know the wind is predominantly coming from a single direction, it is probably not very useful.



Whether a pitot tube is useful for your purposes, depends on what you would like to get. From your previous posts, you are looking for validation data for an exercise CFD code you are writing for a class. For such cases, validation data should be fairly easy to come by. I imagine you will be applying your code on relatively established test geometries. I would suggest this instead of trying to obtain your own measurement, which adds an additional layer of complexity and uncertainty.



Unless your course also demands you do this in addition to your coding, but I think this is very likely because it simply takes much more time and effort to accomplish from scratch.



Hope that helps, Gerry.

Dear Gerry
Firstly, I apologize for late reply. I didn’t know that you answered for my question.
I am actually trying to make pitot tube to measure velocity of wind made by fan. And the reason why I am trying to measure the velocity is to decide the velocity boundary condition for simple algorithm for 2D convection-diffusion heat transfer equation.
The picture below is blueprint of pitot tube that I designed. Because the fan wind has one predominant velocity component, I think the outcome measured by hand-made pitot tube will be pretty valid. Also, I am doing just self-fulfilling project for deeper understanding to fluid dynamics and CFD using FVM, so such a thing, I guess, gonna be experience of study rather than be scratch.

Lastly, I really thank you for your such wonderful detailed reply and sincerely appreciate your thoughtful advice.
IMG_0939.jpg

[Gerry Kan]

Originally Posted by mongya

Dear Gerry
Firstly, I apologize for late reply. I didn’t know that you answered for my question.
I am actually trying to make pitot tube to measure velocity of wind made by fan. And the reason why I am trying to measure the velocity is to decide the velocity boundary condition for simple algorithm for 2D convection-diffusion heat transfer equation.
The picture below is blueprint of pitot tube that I designed. Because the fan wind has one predominant velocity component, I think the outcome measured by hand-made pitot tube will be pretty valid. Also, I am doing just self-fulfilling project for deeper understanding to fluid dynamics and CFD using FVM, so such a thing, I guess, gonna be experience of study rather than be scratch.

Lastly, I really thank you for your such wonderful detailed reply and sincerely appreciate your thoughtful advice.
Attachment 100899

Dear Jaemin:

What you showed me here is not a pitot tube. A pitot tube measures the difference between the stagnation pressure and static pressure to obtain the flow velocity in the direction of measurement.

What you have here is a device that measures the drop in static pressure across two points in an enclosed area, which gives you the mean flow between the two points. If this is your setup, and all you need is the mean flow speed, something like this should be reasonably simple to construct. You can find these differential pressure flowmeters in wind tunnels.

As for a CFD validation exercise, you will need much more than that. Your setup might be good enough for determining boundary conditions for your problem (i.e., volumetric / mass flow rate), you will need at least point measurement for validation. Since it is an enclosed area, you might want to consider obtaining real pitot tubes for measure, say, the flow velocity at the center of your wind tunnel, or even a flow profile, to see how closely your CFD code compares with these measurements.

Hope that helps, Gerry.

[mongya]

Originally Posted by Gerry Kan

Dear Jaemin:



What you showed me here is not a pitot tube. A pitot tube measures the difference between the stagnation pressure and static pressure to obtain the flow velocity in the direction of measurement.



What you have here is a device that measures the drop in static pressure across two points in an enclosed area, which gives you the mean flow between the two points. If this is your setup, and all you need is the mean flow speed, something like this should be reasonably simple to construct. You can find these differential pressure flowmeters in wind tunnels.



As for a CFD validation exercise, you will need much more than that. Your setup might be good enough for determining boundary conditions for your problem (i.e., volumetric / mass flow rate), you will need at least point measurement for validation. Since it is an enclosed area, you might want to consider obtaining real pitot tubes for measure, say, the flow velocity at the center of your wind tunnel, or even a flow profile, to see how closely your CFD code compares with these measurements.



Hope that helps, Gerry.

Dear Gerry Kan
Oh I should have given more details of the tube. Actually the tube is filled with some water and the blue surface of the tube means the free surface of the water.
And what I am trying to do is not validation exercise. It’s to decide boundary condition of velocity at wall.
Also, the reasons why I doing this with that tube is because if I had any flow meter, I could have measured the wind made by fan, but unfortunately as a student since I have no money to purchase the flow meter, all I can do is making pitot tube to determine the velocity BC instead of buying flow meter and measure the magnitude of velocity.
Here is a thing also. Of course I could have assumed the wind velocity for like 3~5m/s. But, I decided to choose former method which I mentioned above because I think it’s better in terms of study experience.

I am so touched with your detailed reply. You must be big guy in this field. I hope there will be only happy events in your life.

Sincerely, Jaemin.

[Gerry Kan]

Dear Jaemin:

I understood what you have presented here. It sounds to me you are trying to force your experimental setup to measure something it cannot, just because it is relatively cheap and easy to construct.

Before you invest any more time and resources on this, you should discuss your course project thoroughly with your course instructor. He is a much better judge on what is feasible given your technical competence and available resources.

It also strikes me that you have some issues understanding the technical fundamentals based on what you wrote. Your instructors will be able to give you the proper guidance.

Gerry.

[mongya]

Originally Posted by Gerry Kan

Dear Jaemin:



I understood what you have presented here. It sounds to me you are trying to force your experimental setup to measure something it cannot, just because it is relatively cheap and easy to construct.



Before you invest any more time and resources on this, you should discuss your course project thoroughly with your course instructor. He is a much better judge on what is feasible given your technical competence and available resources.



It also strikes me that you have some issues understanding the technical fundamentals based on what you wrote. Your instructors will be able to give you the proper guidance.



Gerry.

Dear Gerry Kan

Your replies inspired me keep in mind that scientific measurements should be more tight. Always reminding your replies, I will try to have more strict attitude dealing with experiments.

Actually, I have no instructor. Unfortunately, the high school that I go to is not sort of science school so, any of my school teachers cannot give advice neither regarding my study or experiments. So I have been studying by myself related theories through internet or YouTube videos or some textbooks which I feel hard to understand fully.

I think I don’t have good understanding of fundamentals of each related theories. I will try to improve theoretical backgrounds as possible as I can

Sincerely, Jaemin

[mongya]

Originally Posted by mongya

Dear Gerry Kan

Your replies inspired me keep in mind that scientific measurements should be more tight. Always reminding your replies, I will try to have more strict attitude dealing with experiments.

Actually, I have no instructor. Unfortunately, the high school that I go to is not sort of science school so, any of my school teachers cannot give advice neither regarding my study or experiments. So I have been studying by myself related theories through internet or YouTube videos or some textbooks which I feel hard to understand fully.

I think I don’t have good understanding of fundamentals of each related theories. I will try to improve theoretical backgrounds as possible as I can

Sincerely, Jaemin

Also, there was obvious my misunderstanding of your replies. So i want to apologize for that. Since my first language is not English, please understand me.

[Gerry Kan]

Dear Jaemin:

If you are still in high school (and by that you mean secondary school) and this is your science project, it's a very different story.

I assume you are looking to build a wind tunnel and try to measure the flow velocity through it, and then build a CFD model representation. Is that correct? If not you might need to give me more details.

The next question is: What equipment do you have at the moment (i.e., do you already have the wind tunnel, or do you have to build it), and how much time do you have to get this done?

I advise against building your own CFD code, or use existing commercial code to do this. It is not exactly an easy thing to code quickly (my personal record is about one week for a 2D code, and I wouldn't want to do it again). Instead, download ready-to-use wind tunnel simulators that work very nicely on a PC.

You let me know what you want to do and I will do my best to help you.

Gerry.

[mongya]

Originally Posted by Gerry Kan

Dear Jaemin:



If you are still in high school (and by that you mean secondary school) and this is your science project, it's a very different story.



I assume you are looking to build a wind tunnel and try to measure the flow velocity through it, and then build a CFD model representation. Is that correct? If not you might need to give me more details.



The next question is: What equipment do you have at the moment (i.e., do you already have the wind tunnel, or do you have to build it), and how much time do you have to get this done?



I advise against building your own CFD code, or use existing commercial code to do this. It is not exactly an easy thing to code quickly (my personal record is about one week for a 2D code, and I wouldn't want to do it again). Instead, download ready-to-use wind tunnel simulators that work very nicely on a PC.



You let me know what you want to do and I will do my best to help you.



Gerry.

Dear Gerry Kan

This is everything about my project

Theme : Thermal measurements and analysis of my classroom through 2D convection-diffusion discretized equation through simple algorithm using MATLAB (uFVM environment)

Research motivation : since air conditioner in my classroom makes only certain directions of wind, some classmates who are in the directions of wind say they’re cold and some who aren’t say they’re still hot even if the air conditioner is on. So through thermal analysis of classroom, I am trying to suggest probable solution. (since there are two fans each opposite side of classroom, if I can take advantage of convection occurred by the flow accelerated by the two fans, by making temperature of whole classroom evenly, I think I can suggest some decent solution for the problem)

The obstacles that I faced from on now :
1. Difficulty in making mesh ( uFVM add-ons in MATLAB is based on openFOAM but, some functions like blockMesh is impossible. So all I can do to make mesh is writing every node myself which is I think massively inefficient. )
For example,
(0.5 0 0)
(1 0 0)
(1.5 0 0)
…
(0.5 0.5 0)
(0.5 1 0)
…

2. setting boundary conditions :
As I said above, I am trying to make the most of effects of convection through two fans installed at the opposite side of each walls in classroom. So to get results reflecting how the fans have an effect on the thermal state in classroom I have to set the velocity boundary condition at wall. But, I don’t know the exact value of velocities made by fans. Even if I can assume the velocity value as certain well-known value on the internet, I wanted to set the velocity more similar with the velocity made by fan installed in my classroom. So as to measure the magnitude of velocity of fan, I came up with the idea measuring velocity with some device imitating the principle of pitot tube which I thought easy to make. After composing the device, I wanted to get advice on whether there is theoretical errors or issues with the device. So that’s why I uploaded the initial post.


And, unfortunately what I have regarding equipments is nothing but laptop. (That’s sort of main reasons why I try to do everything with theories and CFD.)
The time left to me from on now is about 2 weeks. ( the better if the project finishes sooner )

What I want to do is to get thermal contours on my classroom using FVM simple algorithm with CFD program which is not commercial software. (I have used Autodesk CFD and Ansys fluent before and now I really want to do CFD with programs like MATLAB or openFOAM. Now I think it’s better to use openFOAM rather than MATLAB with uFVM add-ons due to the struggles in making grid.)

If I can code grid at files including points file, faces file, neighbors file and so on, I want to use MATLAB because I think it’s way easier to run CFD with MATLAB with uFVM add-ons than openFOAM.

[Gerry Kan]

Dear Jaemin:

Thanks for giving me a better idea of what you want to do, and I (at your age) would not have thought of something like this for a science project. Your idea is, indeed, interesting and ambitious.

On the other hand, I think this kind of work, for a two week project, might be a little too much for you and your laptop. Here are a few things I would suggest to make this much easier.

1) How experienced are you with OpenFOAM at all? I would recommend a code that is bit easier to use. I found this semi-commercial code from Japan called Flowsquare+ (I am not affiliated with the product). This CFD code is designed to run on a Windows laptop (unlike OpenFOAM). Their trial version is limited, but you can request a student license that will give you all the necessary features for one year. You might want to spend a few hours exploring this before falling back on Matlab / OpenFOAM, which is much more complex to use and requires much more computational power.

Here is the link for the software: https://fsp.norasci.com/en/index.html

2) For flow measurement, a simple wind gauge will be a very cheap, reliable and much more accurate option. Search Amazon (or similar) for "wind gauge" (or "풍속계") and you will find a lot of options that are below 30,000 won (about 20 Euros). You might need to come up with a way to mount this in a fixed location, or you can hold it by hand at different locations, just make sure you mark down the coordinates beforehand. Since you are looking for steady state calculation, the point measurements should be fairly quick and simple.

Here is an example I found from a Korean site. It cots only 12,500 won (not even 10 euros): https://prod.danawa.com/info/?pcode=4683013. This will do much more than your U-tube setup.

3) Given the time you have, I would just model your room as a rectangular box. The AC will be your cold air inlet (you will have to guess the temperature). Fans are a little tricky to model. I would not model them for now. Just identify the cold spots and suggest how they can be made more even. You need an outlet so that your room is not continuously accumulating air. You might need to "leave the door open" so to speak to do that. Otherwise, introduce crevices around doors and windows, but this can be very tricky to setup.

For a secondary school project there should be plenty of material to go through. Good luck with this, and let me know if you have any more questions.

Gerry.

[mongya]

Originally Posted by Gerry Kan

Dear Jaemin:

Thanks for giving me a better idea of what you want to do, and I (at your age) would not have thought of something like this for a science project. Your idea is, indeed, interesting and ambitious.

On the other hand, I think this kind of work, for a two week project, might be a little too much for you and your laptop. Here are a few things I would suggest to make this much easier.

1) How experienced are you with OpenFOAM at all? I would recommend a code that is bit easier to use. I found this semi-commercial code from Japan called Flowsquare+ (I am not affiliated with the product). This CFD code is designed to run on a Windows laptop (unlike OpenFOAM). Their trial version is limited, but you can request a student license that will give you all the necessary features for one year. You might want to spend a few hours exploring this before falling back on Matlab / OpenFOAM, which is much more complex to use and requires much more computational power.

Here is the link for the software: https://fsp.norasci.com/en/index.html

2) For flow measurement, a simple wind gauge will be a very cheap, reliable and much more accurate option. Search Amazon (or similar) for "wind gauge" (or "풍속계") and you will find a lot of options that are below 30,000 won (about 20 Euros). You might need to come up with a way to mount this in a fixed location, or you can hold it by hand at different locations, just make sure you mark down the coordinates beforehand. Since you are looking for steady state calculation, the point measurements should be fairly quick and simple.

Here is an example I found from a Korean site. It cots only 12,500 won (not even 10 euros): https://prod.danawa.com/info/?pcode=4683013. This will do much more than your U-tube setup.

3) Given the time you have, I would just model your room as a rectangular box. The AC will be your cold air inlet (you will have to guess the temperature). Fans are a little tricky to model. I would not model them for now. Just identify the cold spots and suggest how they can be made more even. You need an outlet so that your room is not continuously accumulating air. You might need to "leave the door open" so to speak to do that. Otherwise, introduce crevices around doors and windows, but this can be very tricky to setup.

For a secondary school project there should be plenty of material to go through. Good luck with this, and let me know if you have any more questions.

Gerry.

Dear Gerry Kan

Firstly, you will never know how big gratitude I am feeling right now. Thank you for your detailed advice. I am touched a lot especially for suggesting to purchase wind gauge with Korean-translated word.

And, I am trying my best to lead my project as you advised. Thank you.

And, also I am planning another project regarding adapting machine learning to acceleration of the speed of calculation of CFD. As I searched, many researchers have been researching about adaption of ML to RANS or LES. But, I have only shallow superficial understanding of RANS and all I know is just transport equation (especially, convection-diffusion)

So regarding doing something with adaption of ML to CFD, is there any possibility to research acceleration of transport equation ?

And when doing ML is it done with FVM CFD?

[Gerry Kan]

Originally Posted by mongya

And, also I am planning another project regarding adapting machine learning to acceleration of the speed of calculation of CFD. As I searched, many researchers have been researching about adaption of ML to RANS or LES. But, I have only shallow superficial understanding of RANS and all I know is just transport equation (especially, convection-diffusion)

So regarding doing something with adaption of ML to CFD, is there any possibility to research acceleration of transport equation ?

And when doing ML is it done with FVM CFD?

Dear Jaemin:

We can talk about your next idea once you have done what you have here. It is better to focus on one thing at a time, when, in my opinion, this is not a trivial task, given the time your have and the amount of work that still remains.

Gerry.