6 Replies Latest reply on Aug 12, 2014 2:08 AM by Toni Pintar

    Vessel modelling (beam vs plate)

    Toni Pintar


      I am considering the best way to model a large section of a vessel. Since it's a fairly large model I plan to use surface elements for all plating, major beams and brackets. And than for modelling stiffeners I would like to use beams, hopefully that would reduce computing time. 


      This approach however has issues:

      1. If I use beams attached to plates, will that play nicely with the rest of the model (see image 01)

           - my preliminary assumption is that if I don't connect longitudinal stiffeners to transversal beam it shouldn't significantly impact results.


      2. If using beams, is there any way to model connection of horizontal to vertical stiffeners and also bracket connections (see image 02)

           - these are easily modelled by plates, but if I model stiffeners as beams there will be a large gap between them, or between stiffener and bracket.


      Thanks in advance for any assistance.

        • Re: Vessel modelling (beam vs plate)
          Nicholas Luyster

          Hi Toni,


          I'm having trouble visualizing what is going on from the screen shots and the description.  Would it be possible to send a jpg of the whole model with arrows pointed to the sections referred to in the pictures?


          With this said, you can get decent results through mixed meshing.  However, mixed meshing does have limitations and if you want detailed stress results on some of the brackets and structural beams, you may have to switch over to solids at those locations.  Often times, I'll start out with a mixed mesh and then I'll duplicate the study and transition to solid elements at locations which I believe to be concerning.  This may result in several simulations. However, it will be the best approach for getting reliable results if the simulation is indeed big. 


          All the best,



          SolidWorks Simulation Training

          • Re: Vessel modelling (beam vs plate)
            Jared Conway

            what is the goal of the analysis? like nick said, if you're interested in the interfaces, you have to model them. if you just want overall structure behavior, go with mixed mesh. this might be worth looking into sub modeling.

            • Re: Vessel modelling (beam vs plate)
              Toni Pintar

              So to further clarify..


              The scope of the model will be full width of the vessel and about 70% of frames, this means the analysis targets a more global effect, that is why I am even considering mixed mesh. For local effects off course switching to a solid mesh would be more reliable, however the detail would have to be specially tiny seeing as how vessels are purely plate constructs.


              I don't have the whole model unfortunately, but as mentioned vessel.. standard construction skin plates, frames longitudinal stiffeners. Image 01 shows a typical detail of vessel frame, and image 02 shows a typical connection detail between stiffeners or two beams. Sketch in image 03 shows isometric view of 02.


              Now back to principal question:

              Image 03: left side is rough reality sketch, right side is assumed model method (plates are surface, beams stiffen surface.).



              1. force transfer between horizontal and vertical beam,

              2. force transfer from beams to bracket (If I model bracket connected to top& Side plate directly the results will be different)


              Is this scenario even possible, or can this type of mixed mesh only be used with an open ended stiffened plate.

                • Re: Vessel modelling (beam vs plate)
                  Shaun Densberger

                  It's hard to say without getting a good look at the model, but based on 03.png and how I understand it, I'd just use all shell elements (and skip the beam elements). As both Jared and Nick have pointed out, results near the interface between two different types of elements are fairly inaccurate. Again, this may or may not be a problem depending on what you're after results wise.

                  • Re: Vessel modelling (beam vs plate)
                    Jared Conway

                    "Is this scenario even possible, or can this type of mixed mesh only be used with an open ended stiffened plate."


                    are you asking if this can be done with mixed meshing? the answer is yes


                    if you're asking how, i'd suggest building a really simplified model and start trying some things. i don't fully follow your graphic. but it should be possible. you may need to use element types that aren't ideal to get a connection that you want (for example the beam to gusset), or you may need to build a model from scratch for analysis to help you, for example sheetmetal to speed up the creation of shells.

                  • Re: Vessel modelling (beam vs plate)
                    Toni Pintar

                    OK guys, thanks for trying.


                    The end conclusion on my side model all in plate elements except for stiffeners connected only to parent plate. Those are to be modelled as beam elements, stiffening the plate.