3 Replies Latest reply on Mar 26, 2014 3:31 AM by Jared Conway

    Flow Simulation Of A Saloon Car Model

    Guy Davis

      Hi Guys,

       

      For a college project I am testing the affect aerodyanmic devices have on a saloon car model.

      To analyse the affect ride height and aero devices have on the model i am using gobal and surface goal plots to record the levels of lift/downforce. The testing so far is going well, however i have a few questions for the users more experienced with flow simualtion than me.

       

      Firstly, as mentioned, i am currently using surface goals to measure the downforce or lift generated from certain ride heights/ aerodyanmic devices. The surface plot values do chnage slightly when the car is lowered, however i dont think the values are correct. I have also selected surface plots on the aero devices they seem to  be producing a realistic downforce value.

       

      However, the aero devices ( the added parts into the assembly) arent interacting with the surface goal plots at the wheels. Is it possible to get the spoiler and splitter to interact with the car model created, so that i can then detect the increase in downforce at the surface plots located on the wheels. I have mated the parts, but maybe there is an advanced mate that is needed ?

       

      What is the correct way to test the downforce of the model ? Thanks

       

      Secondly, im not able to simulate roating wheels. I saw a guide on the internet where you have to produce a solid extrusion over the wheel and make sure that this wasnt merged. However i wasnt able to select the extrusion seperately. I have a feeling a certain modification was missed out from the guide or i have mis-understood so was wondering whether someone could point me in the right direction to simulate wheel rotation.

       

      Lastly, to calculate the drag generated i was using a global goal in the the direction of flow over the model. This was working and was producing around 600N of drag. This value was deemed to be around usual for the model and the drag coefficient also seems to be correct. However after my modifications, i re-ran a test and the model is now supposedly producing around 1800000 N of drag which is obviously very wrong regardless of the modification. I cant figure out what would cause this reading. I have re-ran several tests and still get this value, all of the other goals are fine.

       

      Any help/ ideas would be appreciated

       

      Many thanks, Guy

        • Re: Flow Simulation Of A Saloon Car Model
          Jared Conway

          have you looked through the forum? lots of a good discussions about this and how challenging this is going to be with flow simulation or any other CFD tool. if you can get first-order type accuracy (results that can be compared relative to each other), i would accept that as a successful project.

           

          for measuring lift or drag, you need to select a closed surface. the top and the bottom of the surface. if you try and choose an element or a specific face, your results aren't going ot be good. i'd recommend playing with a block in an airstream to see how this behaves. and look the how/why of this in the help. so basically you have to choose the whole car all the time. all of the faces that touch fluid. if you don't, you have to make sure your reference pressure is close to what is on the non selected side.

           

          rotating wheels, again, would start with a super simple model. you're looking for wall motion. insert boundary condition > wall.

           

          and there are no mates that are needed to connect things together. the software doesn't know anything about how they are connected, it is just integrating pressure over the surface. a dip into the help will show you the equations.

           

          to get really good and accurate results, you will need a decent sized comp domain and mesh. take a look at a 2d wing simulation to get a grip on how big and how much and then see how that scales to your problem and you'll understand what i said in the first paragraph and also why wind tunnels still exist today.

            • Re: Flow Simulation Of A Saloon Car Model
              Guy Davis

              Thanks for the reply Jared

               

              Yes, i did spend some time looking through the forum for posts about surface plots and how to measure downforce but the search bar didnt  bring anything up that could help , mostly brought up posts with measuring in. Even when i was searching for 2d wing simulation i cant seem to find anything?- maybe im searching wrong.

               

              Ok, well after testing a cube in a Comp domain. It can be seen that a global goal in the Z and Y direction gives the same results as using a surface plot and selecting all 6 sides of the square in the Z and Y direction.

               

              Now unfortunately i am unable to do a surface goal on the whole model, it takes around 2 hours to mesh and then i run out of memory. Technically, due to the results from the cube test, this should mean that a global goal will give the same results as using a surface plot on every face of the car/aero devices.

               

              Which takes me to where i am currently.

               

              Roughly of the top of my head. At an air speed of 70 mps, using a surface goal with the top and bottom face of the wing selected, i was getting a value of -330N, and when the bottom faces of the splitter, undertray and diffuser were selected i was getting a value of around -900N, now these figures seem about right.

               

              Now when i do a simulation on the bog standard car, standard ride height, with cut outs underneath to disturb airflow. I get a drag reading of 877N in the Z direction and a value of 382N in the Y direction. When i do a test on the fully modified car with optimized ride height, splitter, undertray, diffuser and spoiler i get 319N in the Y direction and 820 in the Z direction.

               

              Now i may be terribly wrong and i have just noticed on a video online that i am supposed to subtract the vehicle weight from that value. But regardless of that. How has the overall downforce only increased by -80N  ?

               

              Standard Car Results.pngModified Car results.png

              Ps. Not to sure whether the images will be any good to see the results. But  the results are listed in the text anyway.

               

              Thanks

                • Re: Flow Simulation Of A Saloon Car Model
                  Jared Conway

                  1. you cant' select specific faces or elements and report them if they aren't closed entities. it will result in erroneous results. focus on the global goals for now and do some research on surface goals and how they might work for you based on some isolated tests. remember to change the reference pressure and see how it affects these results if you aren't selecting all the faces of the vehicle.

                  2. potential reasons for your downforce not being your expectation: 1. maybe it isn't as good as you think it should be 2. numerically you only have 1 data point, you need to check mesh and comp domain for convergence on a result. if your computer can't handle it, make the scope of the problem smaller. a 2d section of your vehicle will probably tell you a lot about what is happening and then you can see more about the effect of mesh and comp domain on results in a reasonable amount of time.

                  3.i'm not following the logic about subtracting the forces. do a freebody diagram here.

                  4.not finding the flow articles. try google search instead of the forum search or browse through some of the pages of articles. you can also look at the technical referenes for some examples as well.