# forces calculation

 Register Blogs Members List Search Today's Posts Mark Forums Read

 June 17, 2005, 00:28 forces calculation #1 Chien Guest   Posts: n/a Hi, I am seeking advices on force calculation in CFX-Post, in particular on a 2D location. I am running a simulation with transient/laminar flow over a cylinder. In comparing the following: 1) force_x() @ cylinder 2) areaInt_x(Pressure) @ cylinder + areaInt_x(Wall Shear) @ cylinder there's always a slight discrepancies. I was wondering if anyone how why? I have also monitored the "Normal Force on cylinder (X)" in the CFX-Solver, again it's also different from 1)and 2). At the end of the day, I want to plot force in x-direction of the cylinder over time. I noticed that the data obtained using force_x() function is only available at Full backup results. That's why I want to which force calculation is the right one. I appreciate any inputs and comments. Chien

 June 17, 2005, 04:46 Re: forces calculation #2 James Date Guest   Posts: n/a Yep, i've also noticed that. Check my previous post. "Cf and Cvp Contributions to Cd Using CEL" James

 June 17, 2005, 13:50 Re: forces calculation #3 Rui Guest   Posts: n/a Hi, I donīt know the answer.But I have checked in CFX-Post that areaInt_x(phi)@Surface is different from areaInt(phi*Normal X)@Surface. They are equal only if the surface is a Plane. The documentation says about areaInt: "If a direction is selected, the result is an integration over the projected area of each face onto a plane normal to that direction". I have checked that areaInt_x(1)@Surface corresponds exactly to the projected area. But I have checked that areaInt(Normal X)@Surface only corresponds to the projected area if Surface is a plane.Any curved surface is represented by a set of planes, and Normal X should correspond to each of these planes, but probably it is evaluated at each node, and as each node is part o more than 1 plane some average (with some error) is done.From this, I think one may conclude that areaInt_x(phi)@Surface is more accurate that areaInt(phi*Normal X)@Surface. About areaInt_x(Wall Shear)@Surface. Remember that Wall Shear is a vector (Pressure is not), and that CFX evaluates Wall Shear as the norm of the vector. Let's imagine your surface has an area of 1m2 and is perpendicular to the x-axis, and the Wall Shear vector all over the surface is (0,10,0) Pa (its norm is 10). If you calculate areaInt_x(Wall Shear)@Surface you'll obtain 10 N. However, the Wall Shear vector has no component on the x-direction and thus doesn't produces any force on that direction.I think the correct way to evaluate the shear force on the x-direction is areaInt(Wall Shear X)@Surface. However I have compared (on a surface normal to the y-axis, where areaInt_x(Pressure)@surface = 0) force_x@Surface with areaInt(Wall Shear X)@Surface and the results are different. I don't know why. I think this should be answered by someone from CFX Regards, Rui

 June 17, 2005, 16:09 Re: forces calculation #4 Chien Guest   Posts: n/a Hi James: Yes, I have read your posting on the "Cd and cvp..", they were helpful. I was hoping someone from CFX would provide an answer to that. CFX's documentation mentioned the inclusion of momentum flux in the calculation for force, but i'm still confused about the discrepancies. Rui: Thanks for the advices. They were clear and helpful as well. However, the areaInt(phi*Normal X) doesn't seem to work for the surface on my cylinder. CFX-Post was saying "No data exist for that...". I have tried that with an abritary plane, didn't work as well. Any ideas? Anyway, my main concerns are still: 1) discrepancies between force_x().. and areaInt(Pressure).. 2) force() only available with full backup results

 June 20, 2005, 11:53 Re: forces calculation #5 Rui Guest   Posts: n/a Hi, It's strange that areaInt(phi*Normal X) doesn't work. Normal X is allways available in CFX-Post, even when you load a .def or .gtm file. What is exactly the error message, and which version of CFX are you using? About the force () function being only available with full backup files, try to set the expert parameter Output Eq Flows to true. But I'm not sure if this will let you calculate force () with any transient file (I think the documentation is not very clear in this point). About the discrepancies, try to post a message in the CFX-Community forum, or ask CFX support how the force () function is calculated. Regards, Rui

 June 21, 2005, 03:16 Re: forces calculation #6 Chien Guest   Posts: n/a Rui: When i tried to do areaInt(Pressure*Normal X)@ cylinder, I received the following message: No data exists for variable 'Pressure * Normal X' specified in object 'cylinder'[/i]. I am using CFX 5.7 at the moment. thanks, CNg

 June 21, 2005, 07:35 Re: forces calculation #7 Rui Guest   Posts: n/a Hi, I've checked it in CFX-5.7 and I also received that message.You have 2 options to work it around:1- Create an expression and define it as Pressure*Normal X, create a variable (called var, for example) and set it to that expression. In the calculator, you may then obtain areaInt(var)@cylinder2- Use CFX-5.7.1 (with SP1) and it will let automatically obtain from the calculator areaInt(Pressure*Normal X)@cylinder I'm using CFX-5.7.1 and that's why I found strange that you couldn't obtain areaInt(phi*Normal X)@surface. Apparently it was a bug of CFX-5.7 which wouldn't let you calculate function(variable1*variable2)@locator , but it has been fixed in CFX-5.7.1.I think you should update to 5.7.1 and install the SP1 (it's available from the CFX-Community web site). Have you tryed to set the expert parameter Output Eq Flows to true? Can you calculate force () with .trn files? Regards, Rui

 June 22, 2005, 13:58 Re: forces calculation #8 Chien Guest   Posts: n/a Rui, Apparently, we are using CFX 5.7 SP1 at the moment. I don't suppose there's a 5.7.1 upgrade available for download, is there? CFX-Pre documentation states that the Output Boundary Flow = All[/b] in Output Control replaces the expert parameter output eq flows = f. I tried re-running my simulation with both methods, but still unable to get force() with transient data (partial). Chien

 June 23, 2005, 07:27 Re: forces calculation #9 Rui Guest   Posts: n/a Hi I meant that the SP1 for CFX-5.7.1 is available for download. We received the standard version of CFX-5.7.1 in CDs. The documentation states:"When Output Boundary Flows = All is set you get all equation flows written to the file you have setup. The internal default setting for full backup/results and transient files is All. The default setting for Selected Variable and Minimal transient files is None. This CCL parameter replaces the solver expert parameter output eq flows.""The force on a boundary is calculated using momentum flow data from the Results file, if it is available. The result can be positive or negative indicating the Direction of the force. To include the required momentum flow data in your Results file to calculate forces on boundaries, you should set the expert parameter Output Equation Flows to ON on the Expert Parameters form in CFX-Pre""Force calculations on boundaries require additional momentum flow data. This can be included in the Results file by adding the Expert Control Parameter output eq flows and setting its value to T." That's why I thought that setting the Expert Paremeter output eq flows to true (or the Output Boundary Flow to All) would you let you to calculate force () with partial transient files (.trn). I think you should ask this to CFX support. Regards, Rui

 June 23, 2005, 11:45 Re: forces calculation #10 Chien Guest   Posts: n/a Hi, For those who wanted to know the outcome, I found out that you have to specify the [/b]Output Boundary Flows = All[/b] in the Transient Results section in order to get force () for partial transient (.trn). However, the expert parameter output eq flows = t does not do the trick for this case. Thanks Rui and James's advices on this. When i get a response from CFX regarding the force discrepancies, i'll let you know. CE