Re: Question about Solar Radiation Processor.

["Michaël Kummert" <kummert@xxxxxxxxxxxxxxx>]

Dear Mr Jong,

I think you could have a look to hourly values (or values for the time step
of your simulation). Sometimes, if the measured horizontal radiation  is too
high just after sunrise or just before sunset, it can give huge errors on
the computed radiation for tilted surfaces. And I guess the behaviour of
different models for tilted surface radiation can be very different in these
"extreme" situations.
But I must say the numbers you got for daily sum are impressive!

Furthermore, there is a strange point in TYPE 16.
If I am right, if the sunrise/sunset occurs during a time step, the outputs
of TYPE 16 give average radiation on the period for which the sun is up. In
this case, the integrated radiation for one day is not exact.
I give a short example :

Let's say the sunrise happens at 8:30 and your simulation time step is 1h.
Between 8h and 8:30,  radiation is 0
Between 8:30 and 9:00, say the average radiation is 50 kJ/h/m2

Then, if I am right, the output of TYPE 16 for the time step between 8:00
and 9:00 will be 50.
In my opinion, it should be 25 (the real average on 1h)

This, combined with an error on measured radiation which is often important
when the sun altitude  is low, could lead to strange results for these
periods.
But normally, these errors have a small influence on daily values (as far as
I know!)

I would be happy to have your (and "TRNSYS people") comments about this.


Michael Kummert



______________________________________________

Michakl Kummert

Fondation Universitaire Luxembourgeoise

Avenue de Longwy, 185
B 6700 Arlon
Belgium

Phone : +32 (0)63 23 08 59
Fax      : +32 (0)63 23 08 00

kummert@ful.ac.be


-----Original Message-----
From: Jong, M.J.M. <jong@ecn.nl>
To: TRNSYS E-mail group (E-mail) <trnsys@relay.doit.wisc.edu>
Date: Thursday, March 02, 2000 11:38
Subject: Question about Solar Radiation Processor.


Dear TRNSYS Users,

I often use TRNSYS for buildingsimulations. For the Solar Radiation
Processor I always use the Perez skymodel (parameter 3=4). For a special
case I had to use the Isotropic skymodel (parameter 3=1). I remarked that
there was a great difference in the diffuse radiation on a square that is
South East orientated (tilt = 35 0). With the Perez skymodel, the annual
total diffuse radiation on this square (tilt 35 0, orientation South East)
is 3,0 GJ/m2 a year. With the Isotropic skymodel, this radiation is "only"
2,4 GJ/m2 a year. This is a difference of 0.6 GJ/m2 a year. When I do the
same exercise for a sqaure that is South orientated, then the result for the
Perez Model is 2.6 GJ/m2 a year and for the Isotropic skymodel 2.4 GJ/m2 a
year. The difference is only 0.2 GJ/m2 a year.

To explain this, I made a graph of the day-to-day total diffuse radiation
for the different skymodels and orientations (see attached file Graph.zip
(zipped Excel-file)). You can see that for the South East orientation with
the Perez sky model there is a "sky-high" diffuse radiation on day 299. I
think this cannot be correct.

This could be caused by incorrect input for Total Horizontal Radiation, the
Diffuse Horizontal Radiation or the Beam Radiation. To check this, I plotted
in the same graph these radiations (also day-to-day totals). In the graph
you can see that the input is totally normal on day 299, so the input can
not cause the "sky-high" diffuse radiation.

Does anybody has equal experiences and does anybody know an explenation for
these strange results ?


With kind regards,


Maurice Jong
<<Graph.zip>>

===========================================================
E-Mail : jong@ecn.nl - Maurice Jong.
ECN - Energy Research Foundation      Department: Wind, Group DEGO
Telephone:    (+31) 224 564938             Fax:       (+31) 224 563214
P.O. box 1    1755 ZG  Petten                Country:   The Netherlands
===========================================================