5 Replies Latest reply on May 2, 2010 10:55 AM by Scott Pratt

    Simulation Advice

    Scott Pratt

      Hello All,

      I am planning out my first Simulation analysis and would like to get some feedback to make sure I am going about it the right way. The frame consists of two welded frames that are bolted together. I will go through and add welds constraints to the welded frames and then add bolt connections that link the two frames together. Is there a standard bolt preload I can use? I suppose I should put something in instead of having no preload, right? I have a frame that has a central axis about which it rotates. Do I need to add the part that constrains this post to be able to create the right sleeve constraint? The frame sits on four rollers, which are not shown but they are represented by the black arrows. I suppose I need to add these to the model and provide the proper mating conditions between two round parts, no? A set of forces are applied to the frame, which are represented by the red arrows. We manually calculated these loads by using civil engineering calculations that gave us the loads from our specified wind load. The two load points are connected together by the structure that mounts to this frame. Do I need to add this structure or something that represents this structure? Or can I just constrain the movement of these points? Is there anything else I need to be cognizant of when doing this analysis? BTW, would it be inappropriate in post my results of the analysis and get people's feedback? I don't want to abuse this forum, but would love to have some hand holding as I am doing this. Thanks in advance for your help.

       

      -Scott

       

      Frame FEA.jpg

        • Re: Simulation Advice
          Joe Grant

          Scott - Congratulations on your first analysis. Alot of questions there. I'll do my best to answer a few. You provided a picture so you'll probably get some good feedback.

          Some things to think about before you start any analysis -

          1. What are the requirements. Is the analysis performed to satisfy internal, customer, or government regulations or requirements?

          2. What type of analysis are you performing and why (back to requirements). Static analysis, fatigue analysis, vibration? I would be concerned about the natural frequency of the structure if it is subjected to wind loading.

           

          IN GENERAL -  I usually approach an analysis in a multi-staged approach. I start with a relatively simple model and apply loads and boundary conditions (pre-processing), I then let it solve and look at the results (post-processing). Looking at the results and animation gives me a good idea if the structure is behaving the way I expect it to. Based on that review I then may choose to add more complexity to the model. This in turn will increase the computer computation time and resources. At some point in time I reach a point that I conclude that the model sufficiently represents reality and then I write a report detailing the findings.

           

          SPECIFIC TO YOUR ANALYSIS - You will have to decide if the added complexity of weld constraints and bolted connections are pertinent to your model. You may choose to simply have contacting bonded faces. As far as bolt pre-load goes, the short answer is yes. There are general tables that give recommended torques for various screws. That said though, you as the engineer should be designing for a pre-load based on your loading conditions. How much force is being applied to the structure and what resultant force do the bolts see when you create a simple free body diagram on paper? Your bolt selection should be such that the force applied during loading never exceeds your pre-load. A simple formula for calculating torque is T = KxDxF where: T = torque, K= experimental factor (ranging between .2 for bare steel to .14 with a moly grease coating), D= nominal diameter of the fastener, F=force (along axis of fastener).

          You don't have to add another structure at the red arrows but you may want to consider how that structure will react in a separate analysis. I would definitely consider appropriate restraints at the four bottom locations as this may have a big impact on the result versus having the structure completely supported underneath.

           

          I suggest you post your results and no you are being inappropriate to ask for this level of feedback. The forum is open an no one is obligated to respond. I would be cognizant of who may see these results as it relates to your customer and competition. Good luck. Looking forward to seeing some results.

            • Re: Simulation Advice
              Scott Pratt

              Hey Joe,

              Thanks for the advice. I am trying to design the frame to be able to survive the right wind conditions. This is an internal spec. Yes, I would like to do a static analysis and then do another one to look at the natural frequency.

               

              Simplifing the model for an initial analysis makes a lot of sense. The locations where the load is applied, should I just create a split line that represents the mounting holes for the mating component that subjects this load? What about the sides of the frame. I created those parts as a weldment. Do I need to add a weld connection at each joint? If so, this could be pretty time consuming. In this case, should I just recreate the frame and simplify the design more. For example, make the weldment side one part that is shelled? Should I do this anyway?

              -Scott

                • Re: Simulation Advice
                  Joe Grant

                  It appears from your picture that you have two sets of lugs or eyelets that will be used to mount another structure by way of a pin or bolt. I would simply apply the load to the inside face of the hole surface. Not sure I understand your question about sides of the frame.

                   

                  If your model faces are all touching or coincident and your analysis default is to have all contacting faces bonded then your shouldn't have to recreate anything. With regards to shell elements, I would stay away from those. That's just a personal preference. My reason is two fold, shells by definition only account for reactions in two directions. It's not always easy to tell if that is important to an analysis or not, so I just leave everything as solid elements and take the penality of increased computation time. Also, with shells you have to worry about whether the shells are created at the mid-plane of a part or the surface and how that impacts the parts contact with a mating part. Again, I just skip shells.

                   

                  Questions back to you.

                  1. Have you run the analysis yet. If not, try it. Does it run or crash. If it runs then great if not you have to back up a few steps and figure out why.

                  2. Is your VAR in the loop. If they are doing their job you should be able to make a cut at teh analysis and forward the files to them for some additional expert input.

                    • Re: Simulation Advice
                      Scott Pratt
                      Hey Joe,
                      Applying loads to the inside surfaces of the mounting holes makes sense and I will do this. I will use my current model as is and make sure that the program treats the contact surfaces properly.
                      Since my truss frame, which is not shown, is mounted to the eyelets, don't I need to add a rotational constraint? The truss frame connects to both eyelets together. It seems I will need to restrain the rotation on about the vertical axis. Does this make sense?
                      Also, the ring rest on one drive wheel and three idler wheels. Do I need to add these to my assemly for the analysis or can I just create a round split line that represents the contact surfaces?
                      I actually have not had time to run the analysis yet. I am out of town in training all week and thought this would be a good time to get advice about how to do things before I actaully do it. BTW, the main reason for my analysis is to compare a prototype frame that we built to some design changes that were recommended to reduce costs. Like to see how well we designed our prototype and finds ways to make it better and reduce the costs. I will be running some analyses this weekend and will post my results.
                      No, my VAR is not in the loop. I originally wanted a mentoring session where an experienced simulation guy would hold my hand through a set of analyses to show me how to do it. They wanted to be paid to do this and my company doesn't have the funds right now to do this. It was hard enough to get the Simulation license. It seemed to me that they would not assist me unless I paid them. Thought the forum would be a great way to get free assistance, which is greatly appreciated by the way.    

                      -Scott
                        • Re: Simulation Advice
                          Scott Pratt

                          I went ahead and tried to run an analysis. I wanted to simplify the analysis as suggested by Joe and I connected all of the components that are in contact. I created split lines to represent the contact areas between the ring and the support shafts. I added fixed constraints on the contact areas and then added the gravity load. I tried to apply a load to my frame but I kept getting as error (see attached pic). I first tried to apply the load on a bolt hole, then on a flat surface, and finally on an edge. Neither worked. Does anyone know why this won't work? Thanks.

                           

                          -Scott