Learning Forum

[Jiezougt]

What would be the requirement in order to get the System Coupling to work? To be specific for my questions:nIf one-to-one mapping is not needed between the meshed model in Fluent and Mechanical, is there any other requirement I need to be careful?nDo the geometry (the geometry from SpaceClaim) need to be exactly same?nThe walls/wall shadow in Fluent, which are autogenerated between regions, can the results at the wall location be transferred into Mechanical? This wall is a steel structural plate separating two air regions.nThe wall/wall shadow pair is between two fluid regions (both are air) in FluentnThe wall is suppose to be a thin steel plate, which in fluent is just a wall BC with a thickness assigned and layer conduction activated.nIn mechanical, if this wall is a shell-meshed area (not solid). Is this meshed area mappable from fluent wall results? (Fluent has this wall as a wall between solids, ie an area between fluids, not a dedicated area for wall simulation)nAdditional question : I am trying to model steel bridge box girder (thin steel plate wrapping as a box), the air outside of the girder, and the air inside of the girder in Fluent. nShould I use area and shell element for the steel? nOr should I use really thin solid element for the steel in Fluent? (steel box is about 30m long, 2m high, 3m wide, and 10mm thickness). nI am having a hard time determining which way is the best - (a) thin plate sounds like a good fit for area and shell, but Fluent is giving me hard time when using water-tight geometry workflow for meshing.nnThanks very much!nBrandonn

[Karthik R]

Hello,nHere are some answers.nThe nodes on the fluid and structural volumes do not need an exact one-to-one mapping. However, for best results, it would be a good practice to ensure that the overall surface areas are the same. Also, when the data is interpolated from one surface to another, it uses the resolution on the surface for this. You might also want to keep a similar mesh resolution on both.nIt generally is part of the same geometry. When you obtain your volumes for Structural and CFD analyses, they come from the same geometry.nI'll defer the other two questions to other FSI experts. nKarthikn

[Stephen Orlando]

nFor 3. & 4, in Mechanical you'll need to either ncreate a thin body with solid shells, so that you can select each side of the platenor model the plate with two surface bodies that are connected with a bonded contact.nCan you provide some more details on the physics and goals of this simulation? If the deflection of the bridge doesn't have a significant affect on the flow, then this can be modelled with a 1-way transfer of pressure from Fluent (steady) to Mechanical (static structural) with a direct connection in Workbench. This 1-way method doesn't use System Coupling. If the deflection does affect the flow, you'll need to use a 2-way transfer with System Coupling.n

[Jiezougt]

Hi @Jiezougt ,For 3. & 4, in Mechanical you'll need to either create a thin body with solid shells, so that you can select each side of the plateor model the plate with two surface bodies that are connected with a bonded contact.Can you provide some more details on the physics and goals of this simulation? If the deflection of the bridge doesn't have a significant affect on the flow, then this can be modelled with a 1-way transfer of pressure from Fluent (steady) to Mechanical (static structural) with a direct connection in Workbench. This 1-way method doesn't use System Coupling. If the deflection does affect the flow, you'll need to use a 2-way transfer with System Coupling.https://forum.ansys.com/discussion/comment/106980#Comment_106980

Thanks Karthik and Steve!Here are more details on the goal of this simulation:nThe model is a bridge structure - Concrete deck and thin steel box girders.nSolar load with ray tracing, wind for convection, and conduction as loads for thermal.nTransient thermal analysis in fluent - The solar load will change (both direction and magnitude) with time. I am trying to add wind temperature (inlet temperature) to vary with flow time. (import .txt files if possible).nThe thermal analysis result for each step from Fluent need to be transferred to Mechanical (structural transient) to simulate the stress change with time due to the thermal change.nThe bridge is about 40m long and 10m wide 2m deep, the deflection is expected to be less than 3 cm max. It doesn't have significant affect on the flow. 1-way transfer of thermal is great! (not pressure, not force, but thermal)nnWith using System Coupling, I only found the data transfer from Fluent to Mechanical could only be 'Force' as 'Force' is the only option under 'Target, Mechanical' (not exact text...) . It doesn't sound a good fit for my simulation purpose.nnI tried to use Fluent in Workbench, drag the 'solution' to Transient Structural 'setup'. In this case, there is an 'imported body temperature' could be added into Mechanical. However, I am not sure what kind of 'Temperature' is this - In Fluent, if you plot contour, there are several temperature options: Total temperature(solid effect + dynamic fluid effect), Static temperature (only solid effect with no fluid), Wall temperature (not sure about this one even after reading the definition in Help... is it same as Total Temperature?) and so on. This question also holds true for CFD-post - CFD-post only imports one temperature by default, and it is called 'Temperature'. I am not sure which temperature it imports into it. nAny input/advice is highly appreciated!nBrando