Attitude control with magnetorquers

[schaubh]

Hi Basilisk people,

I'm working on a cubesat model, which ADCS is solely magnetorquers and magnetometers. Among the BSK tutorials there are ones describing how to use magnetic torque bars (MTB) along with reaction wheels to drive the net angular momentum of the RWs to zero for example. I am wondering is it possible to use only MTBs for attitude control?

Thank you!
Regards, Daria

[schaubh]

Hello Daria,

MTBs can generate a magnetic moment vector in any direction in the spacecraft body-frame, with a maximum magnitude dictated by the saturation capacity of the individual coils or rods. This magnetic moment m interacts with the Earth magnetic field B and subjects the spacecraft to an external torque given by cross(m,B). As a result, it is impossible to produce a magnetic torque in a direction that is colinear to the magnetic field.

This is why you cannot achieve generic 3-axis control with MTQs alone, because there always is a direction about which it is impossible to produce an instantaneous torque.

Cheers,
Ben

[schaubh]

As Ben says, you can do 2D attitude control. There are solutions that converge over many orbits as the 2D plane where you can apply torques varies over the orbit with the local magnetic field direction. We don't have such a module in BSK yet. You would have to create your own.

[schaubh]

Thank you for your response Ben!

[schaubh]

Thank you prof Schaub! Will try to do so

[leogeo13]

Hello,

For attitude control with magnetorquers, wouldn't it be possible to use the mtbeffector to exert the torque on the body?

[schaubh]

Yes, this BSK effector mimics the magnetic torque bar behavior.

[r-u-n-e]

I'm trying to simulate a cubesat with magnetorquers and no reaction wheels.

The documentation for MtbEffector says the output (cmdTorqueOutMsg) is of msg type "CmdTorqueBodyMsgPayload" but that variable doesn't exist. The only output variable i see is (mtbOutMsg) of msg type "MTBMsgPayload". So, because of the this i'm getting a msg error.

How should I connect the torques computed from MtbEffector to the the spacecraft body?

[ephraim271]

@r-u-n-e Checkout basilisk/examples/scenarioMtbMomentumManagement.py
Line 350 :

    mtbEff = MtbEffector.MtbEffector()
    mtbEff.ModelTag = "MtbEff"
    scObject.addDynamicEffector(mtbEff)
    scSim.AddModelToTask(simTaskName, mtbEff)

[r-u-n-e]

Oh that makes sense ! Thank you

[r-u-n-e]

I have made a diagram of how I connected modules to achieve attitude pointing with magnetorquers. I would like to make sure this is the proper way of going about this before moving on. I hope this is clear enough.

Thanks

[schaubh]

MRPFeedback produces a 3D control torque, but the MTBs can only produce the torque, at best, along 2 axis. With this setup you are not guaranteed full controllability.

[IvelinPenchev]

In src/simulation/dynamics/MtbEffector/MtbEffector.i, line 43, shouldn't the payload include be "MTBMsgPayload.h"? The ConfigMsgPayload is already included in line 41.

I believe part of the reason you are getting an error is because that. We just ran across the bug, so I believe this is a good place to report it.

%include "architecture/msgPayloadDefC/MTBMsgPayload.h"

[schaubh]

I believe you are correct, that msg definition include is missing from the swig interface file. I'll create a PR to fix this.

[schaubh]

Just pushed a fix to this on develop

[IvelinPenchev]

great thanks for the speedy action, we pulled and confirm it works :) I believe it also needs to be fixed in the documentation for Mtb Effector:

[schaubh]

The documentation is updated on this pushed branch as well. Note that the online documentation is only updated when we release a tagged release. If you want the documentation of the current develop branch, then please see

http://hanspeterschaub.info/basilisk/Support/User/createHtmlDocumentation.html