Reliability of single office simulation in early building load estimation

Reliability of single office simulation in early building load
estimation
Grahovac, M.1, Liedl, P.2, Frisch, J.3, Tzscheutschler, P.4
International Graduate School of Science and Engineering (IGSSE), Project Team 2.08
Technische Universität München, 80290 München, Germany,
eMail: mgrahovac@ewk.ei.tum.de
Abstract
The tool we are developing provides users with recommendations for the construction of building envelope and
energy supply system, taking climate and location into account. The main goal is to increase energy efficiency of
buildings by including the subject of energy consumption in the very early stage of design. The envelope
evaluation is based on single office simulation – the energy plant requires building loads. To link the envelope
and plant selection, we performed test simulations to show if the loads obtained during single office simulation
can be used to approximate total building load, just by scaling the results. Comparisons of building simulation
results with scaled results of single office simulation have confirmed that for early cooling load estimation of an
office building, planed to be built in Shanghai, it is enough to simulate just one office for all necessary
orientations and scale the result. The scaled result and the loads of the internal zone should be summed up to
obtain the building load. This approach ensures fast and flexible method for estimation of energy demand and
plant capacity. In addition, it is possible change the planed building size without the need for additional
simulations.
1. Introduction
The goal of our project is to create an easy-to-use tool for architects, whose application
insures implementation of energy efficient solutions into design process, from the earliest
stages of office building design [1]. The in-development tool provides recommendations for
the construction of envelope and energy supply system, taking the climate and location into
account. While working on the project, we encountered the following problem: to test the
envelope, it is enough to simulate just a single office, considering all the walls except the
external one, adiabatical. On the other hand, to be able to recommend an air conditioning or
heating plant, the loads for the whole building are required. The question was, can those loads
- obtained during single office simulation - be used to approximate the building load. If so, the
proper connection between envelope and energy plant selection within the tool was ensured.
We have chosen to use Shanghai climate for this research, since it is one of our project’s goal
climates.
Four simulations of one office room were performed, each for one cardinal direction. After
choosing the appropriate geometry for the building, we scaled the office results and compared
them to results of the building simulation. The building zone and the scaled office, to which
the zone has been compared, have had the same orientation. To assess the influence of
building's interior - obviously not considered during office simulation - we investigated a
building model with and without heat exchange between the interior and periphery offices.
Roof and floor influences had been investigated similarly. Additionally, since we decided to
use a three zone model, we checked the influence of corner offices, with two external walls.
End results are agreeable, especially for cooling, being the main energy consumer
in Shanghai office buildings. It is possible to make a solid estimation, both for yearly and

1 Lehrstuhl für Energiewirtschaft und Anwendungstechnik, mgrahovac@ewk.ei.tum.de
2 Lehrstuhl für Bauklimatik und Haustechnik, petra.liedl@lrz.tum.de
3 Lehrstuhl für Computation in Engineering, frisch@bv.tum.de
4 Project Team Leader, Lehrstuhl für Energiewirtschaft und Anwendungstechnik, ptzscheu@tum.de

---

monthly energy demand, just by simulating a single office. This short and simple one-zone
office simulation can help estimate the fuel consumption and encourage utilisation of
renewable energy. Furthermore, load profiles show good correspondences as well. We can
trust the peak load, obtained by scaling the office simulation results. The given information
helps us determine the exact plant size, and decide if base and peak load plant should be used.
2. Methods
Parameters for the office room simulated are based on the data from [2] and are given in the
Table 1. The same parameters have been used in building simulation. We designed the
building to have an elongated shape with six floors, illustrated in Fig. 1. There is a string of 15
offices on each of the two opposite sides of a floor. Due to the shape of our test building we
divided it into three zones.
Table 1. Envelope properties
Wall Layers
Thickness
Internal Gypsum
Mineral Wool
126 mm
Gypsum
External Gypsum
Wood
133 mm
Insulation
Wood
Ceiling/Floor Carpet
Screed
Mineral Wool
295 mm
Concrete
Air
Mineral Wool
Window
Interpane IPASOL natura 6634 6/16/4

Two basic orientations have been simulated; the first was the North – South after which the
building was rotated for 90°, to face the East – West orientation. Since the majority of the
results presented in this paper are from the North – South orientation, we will refer to the
zones of the building as North, South and Internal zone. The Internal zone refers to a zone
between the North and the South.

Fig. 1. Scheme of the building and the office with basic dimensions

---

Internal gains used are represented in Table 2. In early design stages the precise function of a
particular part of the building is unknown, so we used a constant gain of 50 W/m2 inside the
internal zone. No shading has been applied in this test scenario.
Table 2. Internal gains
Internal gains
Power
No. per Office
North and South zone
Person
30W radiation, 45W convective
2
PC 230W
2
Lights 5,77W
1
Internal zone
50 W/m2 x 1000 m2=50 kW

To compare the result of building simulation with accordingly scaled result from the office
simulation, four building variations have been simulated, see Fig. 2. The first variation,
Building 1, takes heat exchanges between the zones into account, as well as those through the
northern and southern external wall. All other walls, except the walls inside one zone are
considered to be adiabatical. Building 2 is equivalent to the office simulation and allows the
heat exchange only through south and north external wall. The third version is more realistic
and adds heat exchange through east and west walls of zones S and N to Building 1. Building
4 is the version closest to reality, keeping only vertical external walls of Internal zone
adiabatical.
In both building and office simulation we introduced schedules to control the conditioning
operations. Heating, cooling and ventilation (1/h air changes) were operating during office
hours between 08:00 and 18:00. Infiltration of (0.1)/h was allowed from 00:00 to 24:00.
Due to the Shanghai climate and the fact that office buildings have high internal loads that
coincident with the sun radiation, we emphasized the cooling loads result analysis. The East –
West orientation gave results similar to those from South – North, when it comes to
accordance with the office simulation results.

Fig. 2. Definition of 4 building variations, which vary from complete equivalence to office simulations
(Building 2) to the option closest to reality (Building 4). N – North zone, I – Internal zone, S – South zone
The office simulation results representing cooling and heating energy demand, as well as
sensible energy gain from ventilation and infiltration, have been scaled to meet the size of our
building. The scaled office results have been compared to all of the building versions.
3. Results
We present the results divided into four chapters: cooling, ventilation, heating and infiltration.
The results consist of comparisons of the load profiles and monthly energy amounts. Those
presented below refer to south – north building orientation.

---

3.1. Cooling
140

Scaled Office
120
kW
Building 1
100
r
t
h,
Building 2
Building 3

No
80
Building 4
ad
60
g Lo
i
n
40
ol
Co
20
0
Transition season (23.-29. Apr) Typical week (16.-27. Jul) Peak week (27. Aug - 03. Sep)

Fig. 3. Cooling load profile for North zone during three representative weeks within one year. The bigger
the load the better the curves agree.
Cooling load profiles of building simulation show good agreement with the scaled office
profile during cooling season. During the transition season error is more significant due to the
influence of Internal zone.
140

120
Scaled Office
kW
h,
Building 1
100
Building 2
out
S
Building 3
80
Building 4
60
Load
i
ng
40
ol
o
20
C
0
Transition season (23.-29. Apr) Typical week (16.-27. Jul) Peak week (27. Aug - 03. Sep)

Fig. 4. Cooling load profile for South zone during three representative weeks within one year. The bigger
the load the better the curves agree.
North, 1566 cooling operation hours
South, 1795 cooling operation hours
100
100
80

Building 1
80

Building 1
Building 2
%
Building 2
r
,
%
r,
r
o 60
Building 3
60
Building 3
r
r
o
er
Building 4
e
e
Building 4
t
i
ve 40
tiv
a
40
la
e
Rel
R
20
20
0
0
0
20
40
60
80
100
0
20
40
60
80
100
Cooling operation time, %
Cooling operation time, %

Fig. 5. Relative error of building simulation versions compared to scaled room for North and South. Only
the results during operation time have been taken into account.
The version with the biggest relative error is, as expected, Building 4, see Fig. 5. It is
important to know that the error is increased during the transition season, as shown in Fig. 4,
while the loads are small, so this error won’t have a significant influence on the energy
consumption estimation.

---

North South
26000
24000
Scaled Office
h
W 22000
Building 1
K 20000
Building 2
nd,
a 18000
Building 3
16000
Building 4

dem 14000
gy 12000
10000
8000
l
i
ng ener
o
6000
o
C
4000
2000
0
Jan Feb Mar Apr May Jun
Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun
Jul Aug Sep Oct Nov Dec

Fig. 6. Comparison of monthly energy demands for North and South - scaled office vs. building
simulations.
When it comes to monthly energy demand, see Fig. 6, the differences between the building
options are almost insignificant, but the absolute error between the office and building
simulation stays almost constant during the year.
14000
Building 1

h
Building 2
W
12000
Building 3
,
K
Building 4
nd
10000
a
8000
gy dem
6000
g ener
4000
i
n
ol
o
2000
C
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec

Fig. 7. Comparison of monthly energy demands for Internal zone.
In Building 2 the internal zone does not exchange the heat between the zones and thus has
constant cooling load during the year, which equals the internal gains of the zone. Other three
versions have very similar monthly demand values during the main cooling period of the year,
as shown in Fig. 7, whereas those during heating season are reduced.
3.2. Ventilation
Ventilation has shown very good correspondence, which can be seen from Fig. 8 and Fig. 9.

20
W

k
r
t
h,
0
o
N
d
Scaled Room
a -20
Building 1
n Lo
Building 2
-40
t
i
o
Building 3
t
i
l
a
Building 4
n -60
Ve

During heating peak (15 - 21. Jan) During cooling peak (27. Aug - 03. Sept)

Fig. 8. Ventilation load profile for North zone during heating and cooling peek weeks

---

W
20

k
h,
ut
0
So
ad
Scaled Room
o
-20
Building 1
L
n
Building 2
t
i
o
-40
Building 3
i
l
a
Building 4
-60
Vent

During heating peak (15 - 21. Jan) During cooling peak (27. Aug - 03. Sept)

Fig. 9. Ventilation load profile for South zone during heating and cooling peek weeks
North, 2609 ventilation operation hours
South, 2609 ventilation operation hours
100
100
90
Building 1
90
Building 1
80

Building 2
80

Building 2
70
70
r
,
%
Building 3
r
,
%
Building 3
r
o
60
r
o
60
er
Building 4
Building 4
50

er
50
t
i
ve
40
t
i
ve
40
l
a
l
a
e
30
e
R
30
R
20
20
10
10
0
0
0
10
20
30
40
50
60
70
80
90 100
0
10
20
30
40
50
60
70
80
90 100
Ventilation operation time, %
Ventilation operation time, %

Fig. 10. Relative error of building simulation versions compared to scaled room for North and South. Only
the results during operation time have been taken into account.
North South
Jan Feb Mar Apr May Jun
Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun
Jul Aug Sep Oct Nov Dec
h
W 2000
,

K 1000
n
0
-1000
gy gai
er -2000
n

e -3000
t
i
on -4000
i
l
a
Scaled Room
-5000
nt
Building 1
-6000
ve
Building 2
l
e -7000
Building 3
i
b
-8000
i
s
Building 4
n
e -9000
S

Fig. 11. Comparison of monthly ventilation energy gains for North and South - scaled office vs. building
simulations.
Ventilation profiles for the Internal zone are almost identical for all four building simulation
versions, Fig. 7.
3.3. Heating
Heating loads are relatively low and do not exceed 140 kW. The load profiles of the building
follow the scaled room load profile, but do not show such a good coincidence as in the case of
cooling.

---

140
Scaled Room
120
Building 1
,
kW
Building 2
100
r
t
h
Building 3
Building 4

No
80
ad
60
g Lo
t
i
n
40
a
He
20
0
Peak week (15 - 21. Jan) Typical week (12-18. Feb) Transition season (05-11. March)

Fig. 12. Heating load profile for North zone during three representative weeks within one year.
For Building 4 only 10% of the operation time is the relative error under 20%. We have seen
the same problem with small cooling loads. However, the profile shapes do correspond.
Heating is not needed in the Internal zone due to internal gains and heating in the offices on
the periphery.
3.4. Infiltration
As expected, the infiltration for all the versions shows less then 10% error during 90% of the
year.
W
2

k
Scaled Room
1
d,
Building 1
0
Building 2
-1
Building 3
-2
t
i
on Loa
Building 4
-3
-4
f
i
l
t
r
a
I
n
-5
Durin heating peak (15 - 21. Jan) During cooling peak (27. Aug - 03. Sept)

Fig. 13. Infiltration load profile for South zone during heating and cooling peek weeks.
4. Conclusion
We have confirmed that for early cooling load estimation during a conventional office
building designed in a standard way, planed to be built in Shanghai, it is enough to simulate
just one office in two orientations and scale the result. Afterwards, this scaled result and the
loads of the internal zone should be summed up. Additional office orientation simulations are
required only if the shape of the building demands it. This approach ensures fast and flexible
method for estimation of energy demand and plant capacity. In addition, we can change the
building size without the need for additional simulations.
There were problems during transition season, where the profile shape of the building does
agree with the scaled office profile, but the load values have relative errors larger than 50%.
Further steps would include performing additional comparisons with other building and office
parameters, such as envelope properties and climate. Synchronization of radiation distribution
in the two models requires more attention, as well as improvement of transition season
results. Based on this research, it would be possible to introduce correction coefficients for
better approximation of the real load.

References
[1] Pfaffinger, M., Liedl, P., Egger, M., van Treeck, C. Tzscheutschler, P., Grahovac, M., Rank, E., Hausladen,
G., Wagner, U.: Zusammenspiel zwischen Gebäuden, Nutzern, Klima und Energieeffizienz, BauSIM 2008;
[2] Hausladen, G., de Saldanha, M., Liedl, P.: ClimateSkin, Basel-Boston-Berlin 2006.

---