[thedal]

Can you tell me if the characteristic dimension D will change for flow over a multiple cylinders in tandem?

Let two square cylinders of side length 1m is separated by say, 2m distance is subjected to a flow velocity of 1m/s air. Will the characteristic dimension used to find Re is 1m that is D of the single cylinder, Or will it change? And if so, What is its Reynolds Number?

[FMDenaro]

Originally Posted by thedal

Can you tell me if the characteristic dimension D will change for flow over a multiple cylinders in tandem?

Let two square cylinders of side length 1m is separated by say, 2m distance is subjected to a flow velocity of 1m/s air. Will the characteristic dimension used to find Re is 1m that is D of the single cylinder, Or will it change? And if so, What is its Reynolds Number?

You have to define a lenght for the Reynolds number but in some cases this lenght can be differently chosen for the same flow problem. In your case, if the bluff body is in a unconfined flow, you have only D as characteristic lenght. In a confined flow, you could also use the height of the channel even if such a choice does not characterize exactly the non-dimensional parameter for the bluff-body flow.

[flotus1]

There are two ways of looking at the problem of choosing the "correct" characteristic length.
1) You are trying to re-create an experiment or a simulation. In this case, choose the same characteristic length as the work you are referring to
2) You are doing your own simulation. In this case choose whatever you think is adequate. And most importantly: write down your definition if you publish your work.
You can even define multiple characteristic lengths and thereby multiple Reynolds numbers. Re_D for the cylinder and Re_S for the separation between the cylinders.
It is just a definition that does not change physics.

[piu58]

Osborne Reynolds experimented with tube flow. His charactersitic length was the diameter of the tube.

The concept (similarity of flow) can be adapted and is adapted to other geometries. In many cases there doesn't exist such a thing like "the" Reynolds number. You may chose any of the geometric values, in your case the diameter or the distance are appropriate. It is not correct to say "my Reynolds number is beyond 2300, that means ...". It means nothing this way. What you can say is If I establish another, but geometric similar geometry and adapt the flow values getting the same Reynold number, my flow characteristic is similar, or in case of a simulation, equal.

[FMDenaro]

The key is that even if we can define differently the Reynolds number (depending on our choices), generally we want that the non-dimensional equations are characterized in the relevance of each term by the non-dimensional numbers. That means that we not only want to make non-dimensional the set of equation but we want to "normalize" the variables (that is getting O(1)). In such a way the values of the non dimensional numbers really characterizes the flow problem. And this is obtained only chosing a proper lenght, velocity, time, etc.

[thedal]

Thank you all for the replies.
I am taking "D" as the characteristic dimension for single cylinder. I thought it will keep increasing as the number of cylinders increase (say 2D,3D...), which I feel is wrong now. As you all said, Characteristic dimension is chosen such that it non-dimensionalize the governing equations correctly. Hence, It depends all about my flow physics. Hence, D is sufficient in my case for non-dimensionalizing for all multiple cylinder cases.

Always
Thedal