Hi, Ruochun
Thank you for your reply, I wiil try it. By the way, can you share
me the method how to debug the code step-by-step in linux? In addition, if
I use the DEMTracker to set the location and rotation of geometry at each
time step, will this affect computational efficiency? And how to set the
location of geometry at each time step? From the
demo, DoDynamics(frameTime) is many times longer than timestep, and if we
set the frameTime by timestep, it will result in a huge data file.
Best regards,
Wenxuan
在2024年6月27日星期四 UTC+8 23:06:58<Ruochun Zhang> 写道：
> Hi Wenxuan,
>
> 1. When you load a mesh or an analytical object, it's by default in family
> 255, a reserved and automatically fixed family. You probably changed your
> mesh into a family that does not have any motion prescription, then in that
> case it follows the "normal physics" like all the particles you load in,
> including falling under gravity. If you wish to fix that family, just use
> *SetFamilyFixed*.
>
> 2. This is more complicated. First, we have to understand that in DEME,
> strictly speaking, objects work with their principal frames of inertia
> (centered at CoM, having a diagonal inertia matrix). This is a current
> limitation due to how the inertia matrix-defining APIs are designed: you
> can only input the principal moments of inertia. But this also means that
> setting the frame correctly is not necessary if you completely prescribe
> the motion of that object (a.k.a. this object does not respond to physics),
> and you don't care about getting a torque reading. However, in your case, I
> suggest you do so, since you'll prescribe a complex motion, and
> standardizing the frame to use at least the CoM frame can save you some
> head-scratching. I suggest that you make it such that in the mesh file you
> load in, the (0,0,0) point is right at the CoM of this mesh; but if not,
> you can adjust the CoM and principal axes with *InformCentroidPrincipal*
> after loading in the mesh.
>
> After that is taken care of, you need to combine the two rotations into a
> rotation about this object's own CoM and a linear translation of this
> object's CoM (because family motion prescription allows for setting these
> two motions only). Use your MBD knowledge to do that. I don't guarantee the
> correctness of this, but some back-of-the-envelop calculation seems to put
> the angular velocity to be (w2×cos(w1×t), w2×sin(w1×t), w1), and the linear
> velocity to be 5×(w2×cos(w2×t)×sin(w1×t)+sin(w2×t)×w1×cos(w1×t),
> -w2×cos(w2×t)×cos(w1×t)+sin(w2×t)×w1×sin(w1×t), 0). Again, please verify
> yourself. After that, prescribe the object's family's motion using it.
>
> But if this is too complicated and you can write your code to calculate
> the location and quaternion of this object at any given *t *(which for
> some people is easier than mathematically deriving the analytical
> expression for the linear and rotational motions), then you can set this
> object fixed, then use the tracker methods *SetPos *and *SetOriQ *to
> enforce the location and rotation of this object *at each time step*. It
> should achieve the same assuming this object is not moving at a very high
> velocity.
>
> Thank you,
> Ruochun
>
> On Thursday, June 27, 2024 at 2:49:43 PM UTC+8 [email protected] wrote:
>
>> Hi,
>> I want to achieve a composite rotational motion of a geometric body,
>> as follows
>> Motion 1: Rotate at a speed of w1 rad/s, with the rotation axis being the
>> z-axis
>> Motion 2: Rotate at a speed of w2 rad/s, with a rotation vector of [1 *
>> cos (w1 * time), 1 * sin (w1 * time), 0], and an action point of [0, 0, 5],
>> as shown in attached figure.
>> [image: rotating_Chute.png]
>>
>>
>> However, the following issues were encountered:
>> 1. After using DEMTracker to track geometry, even if I don't add any
>> motion, the geometry will continue to drop down. How can I fix the geometry?
>> 2. How to achieve the composite rotational motion mentioned above?
>>
>> Best regards,
>> Wenxuan Xu
>>
>
--
You received this message because you are subscribed to the Google Groups
"ProjectChrono" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/projectchrono/72d33399-446f-4f57-9a0c-d3b8b06c1bcan%40googlegroups.com.
