[TRNSYS-users] Errors when using Powell's method for iteration

[David Bradley]

Chun Kwong,
  You have to be a bit careful when using the Powell's Method solver to 
make sure that the controller models that you are using (Type2, etc.) 
are set up to accomodate that solver and are in the appropriate mode for 
that solver. The idea of the Powell's Method solver is that if the 
system that you are trying to control has very nearly no capacitance (in 
other words no ability to store energy anywhere) then the system is 
going to react so quickly to changes in controller output that the 
Successive Substitution solver will not be able to converge. A good 
example is a PV panel and a resistive load. If the thermal capacitance 
of the PV panel is neglected then the combination of those two 
components may have difficulty converging on an operating point. The 
Powell's Method solver seeks to remedy the problem; my limited 
understanding of it is that it wiggles the inputs to each component by a 
little bit and it looks to see if the outputs wiggle a lot or a little. 
If they wiggle only a little, the system is said to be stable and 
converged. If they wiggle a lot, the solver tries to find a different 
discrete control state that is a better solution.  99 times out of 100, 
convergence problems in a successive substitution simulation are the 
result of oversimplification of the system (neglecting the piping, 
choosing models that have no thermal capacitance, tolerances too loose, 
timestep too big) and will not really be helped at all by switching to 
the Powell's Method controller.
Kind regards,
 David


a8304506 at graduate.hku.hk wrote:
> Hello, Trnsys users.
>
> I try to perform a simulation using the Powell's method for iteration, but with
> errors "inconsistency in the input data" found for some components.  However,
> when I switch back to the successive method, there is no error.  Do anyone know
> how to rectify the problems?  Your reply is appreciated.  Thank you!
>
> Best Regards
> LEE, Chun Kwong
> Research assistant
> Division of Building Science and Technology
> City University of Hong Kong
>
>
>
>
>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> TRNSYS-users mailing list
> TRNSYS-users at cae.wisc.edu
> https://www-old.cae.wisc.edu/mailman/listinfo/trnsys-users
>   

-- 
***********************************************************************
Thermal Energy System Specialists (TESS), LLC 
David BRADLEY                      22 N. Carroll Street - Suite 370 
Partner                            Madison, WI 53703 
                                   USA
P: +1.608.274.2577 
F: +1.608.278.1475
E-mail: bradley at tess-inc.com 
Web Pages:  http://www.tess-inc.com    and      http://www.trnsys.com

***********************************************************************

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/trnsys-users-onebuilding.org/attachments/20090113/a7e082cc/attachment-0001.htm>