Discrete element problem

[ptc-4921095]

Hi, I am trying to solve the attached problem. It includes a system of equations (with deltaC and J) and also a discretized relationship that marches forward from i to i+1 (with Q, C, Y, deltaC and J). Right now I am getting a 'this value must be an array' error. Any help is appreciated.

[nas0k]

Remember that there are two subscripts: Literal "." and Array "[". Your error may be in your use of the incorrect subscript.

Norm

[ptc-4921095]

Thanks Norm, I have double checked the subscripts and I don't think there's anything there. Unless it's really slipping my sight...

[Werner_E]

The function calls to soln are missing the argument DeltaP or better yet DeltaP should be deleted in the argument list of the soln-definition, I guess.

BTW, for lower x (I tried 48) you get (of course) the same problems as with the prior version of that file. In your new file you would only see that the graphs end prematurely (as of the wrong values going complex and not being plotted), but if you change the expression (Q.r[i)^(1/3) to the nth-root notation you will see.

[AndyWesterman]

I think that the approach of limiting the Q.r value was the right move but not quite enough.

The equation for Y.r looks to be empirical in nature & probably has assumptions built in.

Not knowing the appilcation means that we have no way to understand what the intent of the equation is.

(Worth checking the limits that this equation is valid for.)

However, when Q.r tends to 0 the value of Y.r must become infinite & this can cause mathcad serious problems for the numeric processor.

I have added a line to the program loop to keep the value positive.

If the minimum value is held at 0.001 then it seems to work around the singularity.

Reduce the limit to 10^-5 & S1 shows a spike

Reduce further & it seems to get better again ...

Given this I'm not sure that this is the whole answer but it may be another step on the way.

Regards

Andy

[Werner_E]

This problem is part of this thread http://communities.ptc.com/message/200165#200165

The leftover DeltaP which was the cause of the initial error stems from there.

And JM stated there, that Qr and Qs are supposed to be linear over the whole range which is definitely not the case if we limit Qr.

I am not sure if its really an adding of small error or rather that the solveblock get inaccurate very quickly for low values of x.

I guess that it would be hard to produce a solution without changing the constraint/equations in the solve block. For low values of x that solve block is very sensible and will quickly fail to find a solution if TOL and/or CTOL are set to smaller values than the defaul 10^-3.