[Equest-users] [Bldg-sim] Roof Conduction Load

[Aaron Powers]

To try to add to that, the sequential calculation of load then system
reaction is more a result of the calculation engines preferring to use only
linear solvers than the fundamentals of the procedure used.  The main issue
arises because the load generally depends on the action of the HVAC system
(e.g. surface temperatures depend on convection coefficients) and the HVAC
system action depends on the load, which is a nonlinear relationship.

The Heat Balance Method (weird name because there is no such thing as a
conservation of heat) simply enforces an energy conservation equation at
each discrete temperature/enthalpy node.  For N nodal temperatures, this
gives a set of N simultaneous equations to solve at each time step.  If all
of these equations are linear, the calculation engine can use more
efficient linear solvers to resolve the temperatures.  Nonlinear solvers
would require some sub-timestep iterations, which adds calculation time,
and there's also the problem that nonlinear solvers do not always converge
and can be sensitive to initial guesses.  The preference for linear solvers
is also the reason that radiation is usually linearized.

Strictly speaking, Energy Plus is an "integrative" modeling tool and does
solve the loads and system response "simultaneously".  The use of previous
time step temperatures are used because the energy balance equations
general involve transience which requires a finite difference approximation
of heat storage.  As Joe pointed out, this can lead to errors especially
when large time steps are used in the vicinity of large temperature swings.
 The integrative technique of Energy Plus is to first assume the thermostat
setpoint is exactly met, then calculate the system response to meet this
load, then to recalculate the zone temperature caused by this system
action.  I'm not sure how many of these iterations are made each timestep,
but there is some information passing back from the system reaction to the
zonal temperatures.  Similarly, this method (they call it Predictor
Corrector method) is used to model system controllers.  Again, I'm not sure
if the information loop stops with one iteration or continues until a
convergence criterion is met.

The weight factor method used by DOE-2 relies on the convolution principle
of linear systems, so there is no way to obtain a simultaneous solution
(unless there is some way to first deconvolve the weight factors).  On the
other hand, network or quasi-network solutions like the Heat Balance Method
have the ability to solve for space temperature and HVAC reactions
simultaneously, but this is not a requirement of the method, and true
simultaneous HBM solvers are either rare or do not exist because of the
computational expense.

Aaron



Jeremiah,

Your  question is not quite on the mark because what I've described is
downstream from the calculation of room loads where techniques such as the
Weighting Factor or the Heat Balance methods come into play.   Before I get
to that issue, though, I want to point out that the "Adjust Loads"
procedure I described is done automatically by DOE-2 and totally invisible
to the User;  I did so only to explain why Bhartendu (the original poster)
was not seeing any conduction loads through the ceiling.

Now, back to this "Adjust Loads" procedure, there is no simulation program
I know of that solves simultaneously for the heat flows and the zone
temperature.  DOE-2 does it in the two-step way that others and I have just
described. When you said "compared to the Heat Balance Method", I assume
you're thinking of EnergyPlus.  So, how does EnergyPlus meet this problem?
Whereas DOE-2 does one year of LOADS and then one year of SYSTEMS,
EnergyPlus does them in the same time step, but also sequentially.
Therefore,  EnergyPlus calculates the heat flows using the zone temperature
from the previous time step which are then used in its HVAC simulation
without adjustment. This should be okay when zone temperatures don't
change, but what about at morning start-up when temperatures could change
by 5-10 F within one hour?  I've always wondered whether that's why
EnergyPlus gives much lower heating loads in mild California climates than
does DOE-2?  As far as which procedure is "more realistic", I can't say.

Joe

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 108D
Moraga CA 94556yjhuang at whiteboxtechnologies.comhttp://weather.whiteboxtechnologies.com
for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"


On 8/19/2013 9:57 PM, Jeremiah Crossett wrote:

Joe, Just wondering,  given your expertise how would you compare this to
the heat balance method.., and what do you think is more realistic?

*
*
*
*
*Jeremiah D. Crossett  | Senior Analyst  |  Phase Change Energy Solutions*
*
120 E. Pritchard St.  | Asheboro, NC 27203  | Mobile 503-688-8951
www.phasechange.com
*

*
*
*
*



On Mon, Aug 19, 2013 at 6:14 PM, Joe Huang <yjhuang at whiteboxtechnologies.com
> wrote:

> Just to repeat what David and Rodrigo have already sad, DOE-2 calculates
> conduction loads in a two step process. In LOADS, a constant reference
> temperature is specified, which DOE-2 uses to calculate the amount of heat
> flow through the surfaces. This is the LOADS loads summarized in the LS-E
> and LS-F reports.  In SYSTEMS, DOE-2 calculates the true temperature of the
> zone based on the loads passed from LOADS, as well as the actions of the
> HVAC system.  In the course of this calculation, the heat flows ("loads")
> from LOADS are adjusted, but this is not done surface by surface, but in
> total using the so-called "Zone Conductance" times the difference between
> the actual zone temperature and the reference temperature used in LOADS.
> At this point, heat flows through internal walls are also considered, using
> the  difference between this hour's zone temperature and the previous
> hour's zone temperature of the adjacent space, e.g., the Attic or Plenum.
> If you are interested in this number, I believe it can be extracted in an
> hourly report.
>
> In reference to Rodrigo's question, I don't see any benefit in forcing a
> heat flow between the room and the attic by using different reference
> temperatures.   You will simply get a constant heat flow at that
> temperature difference throughout the simulation. In reality, attics are
> generally colder than the room in the winter and warmer than the room in
> the summer, so the derived heat flow with this temperature difference
> doesn't get you any closer to reality, in my opinion.
>
> Joe
>
> Joe Huang
> White Box Technologies, Inc.
> 346 Rheem Blvd., Suite 108D
> Moraga CA 94556yjhuang at whiteboxtechnologies.comhttp://weather.whiteboxtechnologies.com for simulation-ready weather data
> (o) (925)388-0265
> (c) (510)928-2683
> "building energy simulations at your fingertips"
>
>
> On 8/19/2013 12:05 PM, David Wood wrote:
>
> Hi All,
>
> The LS-C report comes from the LOADS simulation portion of DOE 2.2. The
> plenum reference temperature used in LOADS is adjusted in the SYSTEMS
> portion of the program and the loads in the spaces adjacent to the plenum
> are corrected
> correspondingly.
>
> Dave Wood
>
>
> On Thu, Aug 8, 2013 at 8:58 AM, Rodrigo Cerqueira <
> rodrigo.cerqueira at mra.qc.ca> wrote:
>
> Hi,
>
> This happens because the load calculation is made with a fixed temperature
> (see picture). All the spaces, include plenum zones, have the same
> temperature by default (70°F). So, all roof load appears in the plenum
> spaces. To considerate the roof load, you have to change this temperature
> in plenum spaces and a part of roof load will appear like "internal surface
> conduction". But I'm not sure if this is a good practice. Maybe someone
> else has an opinion about that.
>
>
>
>
>
> *Rodrigo Cerqueira, *ing. jr, M.Ing.**
>
> *Simulation énergétique*
>
>
>
> [image: signatureCouriel]
>
>
>
>
>
>
>
> *De :* equest-users-bounces at lists.onebuilding.org [mailto:
> equest-users-bounces at lists.onebuilding.org] *De la part de* Bhartendu
> Awasthi
> *Envoyé :* 8 août 2013 06:35
> *À :* equest-users at lists.onebuilding.org; bldg-sim at lists.onebuilding.org
> *Objet :* [Equest-users] Roof Conduction Load
>
>
>
> Hi All,
>
>
>
> I am doing an energy simulation model in eQuest. I noticed that the load through roof conduction in LS-C is zero where as in LS-B (for plenum space) it is showing some load.
>
>
>
> As per my understanding this load should appear in LS-C also.
>
> Also, if I reduce the U value of the roof, the load through roof reduces and can be accounted in LS-B only (plenum space).
>
>
>
> Refer attached LS-C and LS-B reports.
>
> Any help will be appreciated.
>
>
>
>
>
> Thanks & Regards
>
>
>
> Bhartendu Awasthi
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/equest-users-onebuilding.org/attachments/20130822/cf71fe75/attachment-0002.htm>