hexagon logo
Register Sign in
  • State Not Answered
  • Replies 4 replies
  • Subscribers 11 subscribers
  • Views 110 views
  • Users 0 members are here
Options
Related

Spike on the graph of reaction force exerted by the ground on the rear wheel at the bump

asheshshrestha
Hi,
My two wheeler Adams model is subjected to run through the bump of height 10 cm approx. And I am interested in the ground reaction on the rear wheel at the bump when it moves with a velocity 25 km/hr. So I have plotted the ground reaction force on the wheel for a 6-second time span. I have made two graphs with different steps ( 400 steps and 600 steps ). I am pretty confused about which curve should I really trust. The curve of finner steps ( greater than 500) have spikes that show 25 KN ground reaction at the bump while in course one (less than 500 steps), it is smoother than the graph of finner steps ( no spikes at bump ) but of course, It has the effect of bump that shows 2500 N ground reaction force on rear wheel. The ground reaction at rear wheel is 1400 N at rest condition. Any logical suggestion might be helpful for me.
Thanks in advance
errortoask
errortoask2
!
Parents
  • 0
    You are surprised that you don't see spikes when you have longer time steps (and please talk time steps, not number of steps)? The solver might simply have taken a step over that event with the longer time step.
    ADAMS is a numeric solver. No numeric solver is exact. So there is a process to tune the solver settings.
    1. Do a simulation with default solver setting (ERROR and HMAX). Use regular GSTIFF I3. Save the results.
    2. Decrease ERROR an order of magnitude and run again.
    3. Plot some typical displacements from these two runs. (in your case, for example front fork stroke, swing arm angle, body heave and pitch). (ERROR mainly control accuracy of displacements)
    4. If they differ too much, go back to 2.
    5. Decrease HMAX an order of magnitude. Simulate.
    6. Plot some typical accelerations and forces from the last two iterations. In your case, tire forces, wheel and body accelerations. HMAX does control accuracy of accelerations.
    7. If they differ too much, go to 5.
    8. Now repeat this for different solver settings.
      1. GSTIFF SI2 is usually much more accurate for the same ERROR and HMAX. But also quite a bit slower.
      2. HHT is usually faster, if you have lots of contacts, much faster. Would not expect that to be the case here.
Reply
  • 0
    You are surprised that you don't see spikes when you have longer time steps (and please talk time steps, not number of steps)? The solver might simply have taken a step over that event with the longer time step.
    ADAMS is a numeric solver. No numeric solver is exact. So there is a process to tune the solver settings.
    1. Do a simulation with default solver setting (ERROR and HMAX). Use regular GSTIFF I3. Save the results.
    2. Decrease ERROR an order of magnitude and run again.
    3. Plot some typical displacements from these two runs. (in your case, for example front fork stroke, swing arm angle, body heave and pitch). (ERROR mainly control accuracy of displacements)
    4. If they differ too much, go back to 2.
    5. Decrease HMAX an order of magnitude. Simulate.
    6. Plot some typical accelerations and forces from the last two iterations. In your case, tire forces, wheel and body accelerations. HMAX does control accuracy of accelerations.
    7. If they differ too much, go to 5.
    8. Now repeat this for different solver settings.
      1. GSTIFF SI2 is usually much more accurate for the same ERROR and HMAX. But also quite a bit slower.
      2. HHT is usually faster, if you have lots of contacts, much faster. Would not expect that to be the case here.
Children
No Data
© 2025 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement