solve block question

[tanfolio]

When solving a system of simultaneous equations, do the variables need to be unitless?

I have a system of six equations with six unknowns in a solve block. At the end of the solve block {i.e. find u,v,w,x,y,z) = }One of the arguments highlights and I get an error message that says "The units in this expression do not match those of the others in the list".

 

[IRstuff]

They don't have to be unitless, but Mathcad, to date, does not allow mixed units in a matrix.

If you are returning more than 1 variable from a Find(), then it'll wind up in a matrix at the output, where the units must all be the same.

TTFN

 

[tanfolio]

Thanks for the reply. Since the variables I wish to solve for do not have the same units, is there a clever work around to get the solve block to work or do I need to adjust the equations so that they are dimensionless?

Steve

 

[Bung]

I'm curious, and I could kick myself when I see the answer, but is there a real physical situation where you would have mixed units in a matrix?

I can see some problems with matrix multiplication, for example, if units are not consistent in each matrix.

Unless of course your matrix is actually just a big table or array of numbers treated as one block for convenient indexing.


Bung
Life is non-linear...

 

[IRstuff]

Tanfolio,

AFAIK, that's about the only way, i.e., normalizing the units


Bung,

AFAIK, there's no obvious physical problem with that sort of solution. It's basically an artifact of the way Mathcad outputs results from a Solve block and a limitation that Mathcad imposed on arrays.

TTFN

 

[GregLocock]

A real case I can think of is when describing a 6dof rigid body say, where the natural units are length and number for the x and rx components respectively.

Cheers

Greg Locock