[esp-r] Re: Slab-on-grade for thermally activated slab btw zones

[Marco Manzan]

Alle mercoledì 5 maggio 2010, esp-r-request at lists.strath.ac.uk ha scritto:

>
> Message: 1
> Date: Wed, 5 May 2010 10:33:52 +0200
> From: "Achim.geissler" <achim.geissler at intergga.ch>
> Subject: [esp-r]  Slab-on-grade for thermally activated slab btw.
> 	zones
> To: esp-r at lists.strath.ac.uk
> Message-ID: <9e81925a2f99b36452f085f029d0b6c9 at 212.203.77.246>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Dear all
>
> has anyone used the Slab-on-grade hydronic floor plant component
> successfully for a "normal" thermally activated floor slab between two
> offices? Results seem to imply that the "other" zone does not know about
> the thermal activation. The internal surface temperature of the slab in the
> "other" zone agrees with what one would expect for a heavy concrete slab
> without thermal activation (i.e. approx. the mean temperature of the daily
> temperature swing of the zone).
>
> The setup was this: Two stacked zones, 320 mm concrete floor slab split in
> three layers as symmetric MLC (80 - 160 - 80). A Slab-on-grade component is
> linked to node 6 (centre of inner 80 mm layer) of the floor of the upper
> zone. The connection between the zones seems correct, both .geo and the
> .cnn read correctly. Temperature for the floor surface of the upper zone
> are plausible. Temperatures for the ceiling of the lower zone as described
> are not.
>
> Thanks for any ideas.
> Best
> Achim
>
>
>

dear Achim

I've used extensively the Slab-on-grade component and there are some problems, 
the first is that the underlying zone does not know about the thermally 
activated floor, since the routine injects the heat flux only to the floor 
nodes partaining to the upper zone, second problem the simulation time 
increase substantially. The second problem arise partly from the computations 
performed in the coefficeint generator routine and partially in routine  
RadFloor_store.
I solved the first problem by adding some lines of code to the routine 
RadFloor_CoeffGen. in the following I include the diff file for 
radiant_floor.F
essentially I injects the same amount of flux to the slab floor and to the 
linked structure if exists, you can see the difference by plotting the 
Xsection temperatures for the two structures.
As far as the second problem is concerned, I solved the problem of 
RadFloor_store by performing the copy instructions for the affected 
structures only, I will submit the modified version to Jon Hand. But using 
the component requires still a huge computing time, and pheraps a new 
component should be developed.

I'm working on this components:
1) a component that behaves as a radiator and injects a flux lineary dependent 
with the meanlog temperature from component and slab layer
2) a component that computes the heat flux in a static manner using meanlog 
temperature from component and slab layer (as EnergyPlus does)
I will appreciate any suggestions

Be aware that if you use my changes your are using them at your own risk since 
they have not been checked by esp-r developers.

bye

marco

Index: radiant_floor.F
===================================================================
--- radiant_floor.F	(revisione 4957)
+++ radiant_floor.F	(copia locale)
@@ -212,6 +212,15 @@
      &     ConDp(MCOM,MS), TetaMf(MCOM,MS), TetaMp(MCOM,MS),
      &     Tpipef(MCOM,MS), Tpipep(MCOM,MS)
       REAL    COUT(MPCOE), J11, J12, J21, J22
+
+      COMMON/GR1D01/NNDS,NNDZ(MCOM),NNDC(MCOM,MS),NNDL(MCOM,MS,ME)
+
+C Common saving coupled zones/surfaces
+      common/Zone_Couplings/ bCoupled_Zone (MCOM,MS),
+     &                       iCoupled_Zone (MCOM,MS),
+     &                       iCoupled_Surf (MCOM,MS)
+
+      logical bCoupled_Zone
       
       character outs*124
 C----------------------------------------------------------------------
@@ -577,7 +586,16 @@
                Qboiler = ADATA(IPCOMP,1) * FLWCp * (Tin - Tout)
                Qnode   = ADATA(IPCOMP,1) * Qcir
                GENFLXF(Iz,Is,In) = Qnode / AreaR
+C.....store the same flux for adiacent surface
+               if ( bCoupled_Zone( Iz, Is ) ) then
 
+               Iza = iCoupled_Zone ( Iz, Is)
+               Isa = iCoupled_Surf ( Iz, Is)
+               Ina = NNDC(Iz, Is) - (In - 1)
+C  
+               GENFLXF(Iza,Isa,Ina) = Qnode / AreaR
+ 
+         endif 
 C     Establish additional output variables
 C     Minimum inside surface temperature
                PCAOUT(IPCOMP,1) = Tmin

-- 
 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _        
|Marco Manzan               Email: manzan at units.it            |        
|Tel:   (+39) 040 5583506   Fax  : (+39) 040  572033          |        
|Dipartimento di Ingegneria Navale, del Mare e per l'Ambiente |        
|Universita' degli Studi di Trieste                           |        
|piazzale Europa 1, 34127 TRIESTE                             |        
|_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _|