Re: operator +: nonconformant arguments (op1 is 10x10, op2 is 10x9)

[Doug Stewart]

On Thu, Aug 9, 2018 at 8:31 PM, Doug Stewart <address@hidden> wrote:

On Thu, Aug 9, 2018 at 1:53 PM, Mike Miller <address@hidden> wrote:

On Thu, Aug 09, 2018 at 13:09:45 -0400, Doug Stewart wrote:
> I modified the end of your file to:
>
>  n_cycle
>  if (n_cycle ==1)
>     Acl=sys.A + sys.B*ones(16,9);
>  else
>      Acl=sys.A + sys.B*ones(16,10);
>   endif
>
> now it runs 15 times.
> What this shows is that the first tome through you have a 9x9
> and after that you have 10x10  sys.A

Yes, that is one obvious workaround. I thought you were interested in
investigating why that is.

I think Beginner1 is attempting to claim that sys.A should always be
10×10, and sys.B should always be 10×16, and that there is a bug in here
somewhere. I don't yet know whether that is true.

That's why I asked for a bug report, we need a concise explanation of
what is passed in, what is returned, and how that differs from what is
expected.

--
mike

I have figured out what is happening.

The loop uses kv from 0 to 15

the code  forms:

sys_ol

sys_il

sys_nb

sys_vsc

The code then uses the connect comand

sys=connect(sys_ol,sys_il,sys_nb,sys_vsc,inname_group,outname_group);

to connect them all together into one big sys.

On the 1st pas kv=0

and this makes one row all zeroes, and connect then eliminates that row!!!!

if we set the 15 voltages to not start at 0 but close to zero then all works as expected.

cycle_vector = linspace(0.0000001,15,15);

you probably want

cycle_vector = linspace(0.0000001,15,16);

That is a work around that will get you going.

We will have to look at connect and see why it is dropping the 0 vector.

Doug

--

DAS

Before I spend time looking at connect, I would like to know what matlab does.

would someone with matlab run the attached code and see if there is a size mismatch reported,please.

--

DAS

test444124.m
Description: Text Data