Adams - issues with reaction wheel implementation

[JRsen]

Hi,

I have some problems with modelling reaction wheel in Adams. I need to analyze behaviour of separation mechanism that set both linear velocity along Z axis and angular velocity around Z axis to fly-away mass. In order to stabilize fly-away mass that is separated from a base and ejected, I want to use one reaction wheel, that will be rotating around axis aligned to Z axis. Reaction wheel is rotating with angular velocity of 23000 deg/s. Mechanism is analyzed in no gravity condition (separation and latter movement will take place in space). Unfortunately I am encountering some issues during analysis. First of all, angular velocity magnitude of reaction wheel and fly-away mass changes throughout the time and constantly decreasing, same is happening with angular momentum magnitude of fly-away mass and reaction wheel.

Picture 1,

https://files.engineering.com/getfile.aspx?folder=33accb13-789d-45c2-b92b-a836d130ef8d&file=1.png

Picture 2

https://files.engineering.com/getfile.aspx?folder=c611b7b4-d4b7-4c3a-8c02-d691ddd8a97c&file=2.png

They should not be changing, because after separation there are no forces acting on components. I am using timestep that is set at 1e-3s. Characteristics of the curves are also unsettling:

Picture 3

https://files.engineering.com/getfile.aspx?folder=7581e335-5d55-454b-b7e9-74953e634a31&file=3.png

With smaller time step i.e. 1e-5 magnitudes of angular velocity and angular momentum are constant, but while plotting components of them, this is what I get (some peaks occur randomly):

Picture 4

https://files.engineering.com/getfile.aspx?folder=38817d49-2673-4313-ae1a-be88c49130ed&file=4.png

To clarify, model without reaction wheel was also implemented and results have been validated. I assume that, issues are connected with reaction wheel.
In order to connect reaction wheel with fly-away mass I am using revolute joint, motion is assigned to the revolute joint (23000 deg/s around Z axis), initial angular velocity of fly-away mass is set at 0 deg/s around Z axis. Could you help me with finding out where the problem may lay?

Thanks!

 

[GregLocock]

So you have two parts connected by a revolute joint. One is spinning at high speed at T=0? are your plots axes correctly labelled? What happens at T=8?


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[GregLocock]

Also you should investigate why the speed spikes up in the first time step in image 3.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[JRsen]

Thanks for answering,
Fly-away mass is not rotating at T=0, ejection system sets it in motion during first 0.15 second (Ejection system consists of compressed springs). As a result of ejection fly-away mass rotates with angular velocity of around 85 deg/s in 0.15s. To be honest I have no idea, what is happening in T=8, no forces act on bodies after ~0.15s. Reaction wheel shall rotate with angular velocity of 20000 deg/s from T=0 till the end of simulation. In first time steps (0-0.15s) there is a speed spike because of ejection of fly-away mass.

 

[GregLocock]

Is all this rotating in 3d? It could be a numerical problem as the local coordinate systems go into lock. Perusal of the msg file will probably give some hints. So you understand the initial jump in angular velocity at T=0, to 85 deg/s, the real puzzle is why the reaction wheel deg/s is so noisy, and whta the fly awya mass is up to at T=8. The inital jump on plot 3 from 20000 to 20070 is presumably the reaction wheel accelerating up to the same angular speed as the FAM, but that doesn't make sense to me if the only connection is a revolute joint.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[JRsen]

It is rotating in 3D. Regarding reaction wheel angular velocity, I presume that plots are created in reference to global CS, while motion of reaction wheel (20000 deg/s) is assigned in reference to fly-away mass (~85 deg/s), so angular velocities may add up? I am also afraid that this may be numerical issue and contacted support as well.