Simple bar rotating

Typically: "How do I... ", "How can I... " questions
Post Reply
hamzaali
Posts: 2
Joined: 18 Jan 2018, 13:41

Simple bar rotating

Post by hamzaali » 19 Jan 2018, 17:04

Hi Guys,
I am beginner to V-rep and this forum. So sorry in advance for etiquette.

Model:
i am rotating a simple bar, undamped and zero friction.

Problem:
After running the simulation, i found out that i leave bar at -90 degree. But the free vibration could not reach -90 degree during oscillation it was -83 degree.

https://mega.nz/#!IGBXEDYS!ETVNMkVUof0U ... KrSyypgBKo

As it is basic physics, energy is conserved, how can i reach -90 degree during oscillation?

coppelia
Site Admin
Posts: 6361
Joined: 14 Dec 2012, 00:25

Re: Simple bar rotating

Post by coppelia » 22 Jan 2018, 09:47

Hello,

this is directly dependent on the physics engine you use. You can notice that with the Newton engine, the oscillation amplitude is growing. Each physics engine will use different parameters and perform different calculations. This is sometimes difficult to influence. You will have to look at the dynamic parameters of joints, shapes, and also at the overall dynamic settings for the physics engine (such as dynamic simulation step size, constraint solving iterations, damping, kinetic loss, etc.).

Best in your situation is to reduce the dynamic simulation step size to 1ms: you can do this in the dynamics engine general properties. The reason why this has such a large influence is because physics engines have difficulties handling large rotational velocities.

Cheers

hamzaali
Posts: 2
Joined: 18 Jan 2018, 13:41

Re: Simple bar rotating

Post by hamzaali » 25 Jan 2018, 18:44

coppelia wrote:
22 Jan 2018, 09:47
Hello,

this is directly dependent on the physics engine you use. You can notice that with the Newton engine, the oscillation amplitude is growing. Each physics engine will use different parameters and perform different calculations. This is sometimes difficult to influence. You will have to look at the dynamic parameters of joints, shapes, and also at the overall dynamic settings for the physics engine (such as dynamic simulation step size, constraint solving iterations, damping, kinetic loss, etc.).

Best in your situation is to reduce the dynamic simulation step size to 1ms: you can do this in the dynamics engine general properties. The reason why this has such a large influence is because physics engines have difficulties handling large rotational velocities.

Cheers
Hey I reduced 5e-3 to 1 e-4 and the problem was almost solved.
Can you please explain, what is the reliability and credibility of my simulation in V-rep? because from what i think it can't be used without knowing the results, because it will mess up even in basic simulation!!!

coppelia
Site Admin
Posts: 6361
Joined: 14 Dec 2012, 00:25

Re: Simple bar rotating

Post by coppelia » 27 Jan 2018, 15:44

As I said, physics engines have difficulties with larger angular velocities. Some more, others less. And a dynamic engine will not be able to handle every aspect of a simulation perfectly. It is your task to identify the focus of your simulation (e.g. conservation of energy in the system) and to adjust the physics engines' parameters accordingly. For details, you should look at the physics engines' documentation.
But for the simply fact that you cannot know how accurate or realistic a dynamic simulation is, we have added 4 different physics engines to V-REP: this way you can switch from one engine to another and compare the results. Or you can select the engine that performs best for your specific use case.

Cheers

Post Reply