Design Optimization of Glazing Façade by

Using the GPSPSOCCHJ Algorithm

Mona Khatami1, M.Sc. Maria Kordjamshidi2, PhD Behrouz Mohammad Kari 3, PhD
[University of Ilam] [University of Ilam] [University of Tehran]
Mona.khatami@yahoo.com m.kordjamshidi@ilam.ac.ir kari@workmail.com
Alireza Zolfaghari4, PhD
[University of Birjand]
zolfaghari@birjand.ac.ir

ABSTRACT
Engineering design is a process to find the best solution to satisfy various design criteria. This work aims
to optimize the glazing façade performance and the window size by minimizing the heating, cooling and electric
lighting demand of office buildings. Accordingly, this paper presents a comprehensive analysis in order to study
the balance between daylighting benefits and energy requirements in perimeter office spaces taking into account
glazing properties control with window size, orientation and climatic conditions. The glazing area and
thermophysical properties of the window were taken as the main variables. The optimization was carried out by
using a combination of Energy plus7.0.0 and GenOpt softwares. The energy consumption can significantly
change affected by geometric parameters, materials properties and types of window glass, orientation and
climatic conditions. Optimum range of each parameter was calculated in order to minimize annual energy
consumption with a hybrid multidimensional optimization algorithm: GPSPSOCCHJ algorithm. Furthermore,
since the annual energy consumption effectively depends on the type of air conditioning system, the optimization
process was carried out individually with both evaporative cooling system and compression cooling system. The
results indicated that using the evaporative cooling system compared is more appropriate and economical in
comparison with the compression cooling system. Also, investigations indicated that reflective double glass and
low-e double glazed with argon layer glass is appropriate for Tehran office building and can respectively allocate
the maximum level of window area and the minimum of energy consumption.
Keywords: Glazing façade, Optimization, GPSPSOCCHJ optimization algorithm , Energy consumption

INTRUDUCTION
Window is considered as one of the most important components influencing the thermal performance of
buildings. T heir shape, size, optical and thermal properties, orientation and shading/daylighting attachments
determine the interior daylighting conditions as well as the visual and thermal comfort for the occupants. T he
balance between daylight provision and reduction in energy consumption or demand through appropriate control
of solar has been investigated in a few studies by several researchers (Lee et al., 1995; Citherlet et al., 2001;
Franzetti et al., 2004; Hviid et al., 2008; Tzempelikos et al., 2010). Coupling between daylighting and thermal
simulation is necessary for a comprehensive an alysis. In 1998, Clarke et al. compared the annual energy
consumption of three different types of glazing system using ESP -r and found reductions of about 4.5%, 10.9%
and 6% in maximum heating capacity, maximum cooling capacity and total energy consumption respectively.

30th INTERNATIONAL PLEA CONFERENCE 1

16-18 December 2014, CEPT University, Ahmedabad
 Optimized glass facade design may improve exploitation of daylight and result in significant savings in
electricity consumption for lighting. Reinhart (2002) calculated the daylight availability for several Canadian
locations considering the effects of climate, external shading, facade orientation, glazing type and occupancy
schedules. T he study showed that location, orientation and blind slat angle all have a significan t impact on
daylight autonomy while external objects and glazing type were less important. Optimized glass facade design
may improve exploitation of daylight and result in significant savings in electricity consumption for lighting.
Reinhart (2002) calculated the daylight availability for several Canadian locations considering the effects of
climate, external shading, facade orientation, glazing type and occupancy schedules. T he study showed that
location, orientation and blind slat angle all have a significant impact on daylight autonomy while external objects
and glazing type were less important.
T he sophisticated characterizations of window and shading systems sparked a large amount of studies on this
topic (Reinhart and Walkenhorst, 2001; Walkenhorst et al., 2002; Robinson and Stone, 2006; Loutzenhiser et al.,
2007), and various calculation models that predict illuminance on the interior surfaces of a building as well as on
the work plane level are available (Mardaljevic, 2001; Fakra et al., 2011). T he different models have some
limitations; for example, some models use constant glass transmittance, some others use limited evaluation
metrics such as daylight factors (Ghisi and Tinker, 2005); and some have limitation in sky luminance inputs.
Moreover, it is complicated to modify existing software codes to adapt specific necessities or to present results
using different measures. As to the latter, advanced daylighting metrics may be properly used in daylight
performance evaluation (Nabil and Mardaljevic, 200 6; Reinhart et al., 2006). Finally, the significant
computational time, the complex calculation procedure and the inability to interpret simulation results are all
factors preventing the design community from picking up the advanced design analysis schemes with very few
exceptions (Reinhart and Wienold, 2011).
T his study has been tried to optimize the window size and glass type with the objective of minimization of
annual energy consumption function. In such a way, while reducing energy consumption, occupants’ thermal
comfort and the brightness level of each space remain in the acceptable range. For this purpose the modeling of
thermal and visual performance of building’s transparent façade is performed by EnergyPlus software and the
results are optimized by GenOpt software and GPSPSOCCHJ algorithm and the effect of all parameters among
solar heat gain coefficient (SHGC), thermal transmittance of window (U value), and visual transmittance is
considered. Finally, calculations and evaluations will lead to provide window design recommendations due to
climate.

METHODS
In this paper, as shown in Figure 1 , a case room is considered in accordance with the case No. 600 in
3
ASHRAE 140 standard. Accordingly, this sample space is an office with the dimensions 6 × 8 × 2.7 m in the
middle of a tall building which only a wall with 8m width and 2.7m height is in contact with outdoor climatic
conditions of Tehran. According to Table1 the wall adjacent to the outdoor, specified by common materials for
office buildings that respectively from in to out includes veneer plaster, insulation, concrete block, stucco and
stone.
Table1. The wall adjacent to the outdoor construction
field units obj1 obj2 obj3 obj4 obj5
name 25mm concrete 50mm
stone stucco block insulation plaster(light)
roughness medium medium medium medium
rough Smooth rough rough smooth
thickness m 0.03 0.0254 0.2 0.0508 0.01
conductivity W/m.K 3.17 0.72 0.33 0.03 0.16
density kg/m3 2560 1856 1380 43 600
specific heat J/kg.K 790 840 880 1210 1000

30th INTERNATIONAL PLEA CONFERENCE 2

16-18 December 2014, CEPT University, Ahmedabad
 As shown in Table2 , eight types of window glass have been considered: 6mm clear single glazed, clear
double glazed with argon layer, clear double glazed with air layer, low-e clear single glazed, reflective clear single
glazed, low-e clear double glazed with argon layer, low-e clear double glazed with air layer, reflective clear
double glazed with air layer. Dimming of overhead electric lighting is determined from interior daylight
3
illuminance calculated at one or two reference points. T wo reference points in coordinates 3 ×1.6×1 m and
3
3×6.4×1 m toward the wall adjacent to the outdoor are considered as lighting evaluation criterion. Also, in order
to simulate the thermal and lighting energy demands, the EnergyPlus software is used.

Table2. Types of window glazing construction

Field Obj1 Obj2 Obj3 Obj4 Obj5 Obj6 Obj7 Obj8
Name 6mm clear clear low-e reflective low-e clear low-e clear reflective
clear double double clear clear double double clear double
single glazed glazed single single glazed glazed with glazed with
glazed with argon with air glazed glazed with argon air layer air layer
layer layer layer
Outside Clear Clear 3mm Clear PYR B REF a PYR B PYR B clear REF a clear
layer 6mm 3mm clear clear mid clear 6mm 6mm mid 6mm
6mm 6mm
Layer2 Argon Air Argon Air 13mm Argon
13mm 13mm 13mm 13mm
Layer3 Clear 6mm Clear Clear 6mm Clear 6mm Clear 6mm
6mm

EnergyPlus is one of the most comprehensive whole-building energy simulation tools that are capable of
modeling several features including solar irradiance and illuminance under different sky conditions, advanced
fenestration systems, blind controls, indoor illuminance maps, lamp controls, and heating/cooling energy impact
associated with daylighting controls (Seo et al., 2011). Building model, location and Climatic conditions design in
Software environment. EnergyPlus weatherdata file is used for energy performance calculations and indoor
climate analysis. Hourly based outdoor climate data (dry-bulb air temperature, relative humidity, wind speed,
direct solar radiation and diffuse radiation on horizontal surfaces for 8784 hours) was used to create the model for
calculation. Comparability of current study results for other climatic areas can be done through monthly and
yearly average parameters which are indicated in Table3 (Hanni.al et al., 2012).

Table 3. Reference year parameters

Air Relative Wind Direct solar Diffuse
temperature humidity speed radiation radiation
Month ˚C % m/s Wh/m2 Wh/m2
Jan 2.4 63 1.7 3014 1176
Feb 4.8 55 2.5 3506 1604
Mar 10.2 44 2.9 3820 1923
Apr 16.2 36 3.3 4735 2343
May 22.3 30 3.3 5859 2396
Jun 27.5 24 3.1 7640 2319
Jul 30.9 24 2.8 7632 2032
Aug 29.5 24 2.2 7234 2049
Sep 25 25 2.3 6687 1642
Oct 18.2 33 2.1 5238 1488
Nov 11 45 1.8 3959 1169
Dec 5 59 1.5 2992 1085
Avg 16.9 38.5 2.5 5193.0 1768.8

30th INTERNATIONAL PLEA CONFERENCE 3

16-18 December 2014, CEPT University, Ahmedabad
 By Considering Constraints that describe below the building annual energy consumption with a focus on
providing residents thermal comfort is calculated by EnergyPlus.
1. -0.5 ≤ PMV ≤ 0.5
2. T he minimum illuminance required by international standards on the desk: 500 Lux
3. Heat generated within the space caused by a computer, printer and other accessories available: 800 W
4. Number of people: 4
5. Hours due to the discontinuous use: 7:00 to16:00
6. People with the metabolic rate of 100 W while seated
7. People Clothing thermal resistance, 0.6, 0.5, 0.7, 0.9 clo respectively for spring, summer, autumn and
winter conditions.
8. Constant heating set point: 23.5˚C
9. Constant cooling set point: 26˚C
10. Internal gains for lights: Lighting level calculation method is used to create t he maximum amount of
lights to this set of attribute choices : 400 W

Figure 1 Office space in the initial position study

A key part of using optimization tools with artificial intelligence-based algorithms for optimal design is
defining an appropriate objective function and constraints. In the issue examined in this article inside light level,
inside temperature, sunshade dimensions, wall thermal resistance, energy consumption In order to provide
lighting, heating and cooling, All are measurable quantities that can be offered based on the objective function
and constraints. On the other hand, the purpose of this study was the amount of illuminanace inside the building
and its thermal behavior which is obtained by minimizing the building's annual energy consumption influenced by
the optimal size of the window and its different types. For daylighting control types available in EnergyPlus,
optimization algorithms must support discrete (on-off or 2 or 3 steps controls) and continuous (dimming cont rol)
variables. In addition, the selected algorithm should support intrinsic approximation problems. Detailed buildings
energy simulation tools such as EnergyPlus, T RNSYS, and DOE-2 involve solving a series of systems of partial
and ordinary differential equations that are coupled to algebraic equations. Therefore, an optimal solution for a
continuous cost function may be difficult to obtain without using a heuristic approach (Wetter et al., 2003). Wetter
(2008) recommends hybrid algorithms using the General Pattern Search (GPS) method coupled with the
HookeeJeeves algorithm with multiple starting points or the Particle Swarm Optimization (PSO) algorithm. Using
this algorithm in GenOpt, with Energyplus output as input of the optimization problem, can be found to answer
issue.
As previously noted, the objective function of this issue is the total annual energy consumption which is
minimized by determining the coefficients for the efficiency and production cost of the energy. The above issue is
optimized and analyzed for two efficiency, compression cooling system and evaporative cooling system in Tehran
climate.

30th INTERNATIONAL PLEA CONFERENCE 4

16-18 December 2014, CEPT University, Ahmedabad
RESULTS
Order to determine and Analyzed the optimum dimensions of the window at the four main directions, for 8
types of glass studied by using either compression or evaporative cooling system, after performing optimization
for 64-state results were as follows. Figures 2, 3, 4 and 5 show the optimal window area respectively for the
North, South, East and West orientations, for evaporative cooling system in comparison with compression cooling
system. As can be seen, the use of evaporative cooling system in the same condition the optimum window size
will be larger in all four directions. According to Figure 2 , in the north, reflective double glass in both cases is an
2
option and its size in evaporative cooling system is 4.5×2.25 m (47% of surface) and in compression cooling
2
system is 4×2 m (37% of surface). As shown in Figure 3 in the south, reflective double glass is the best and its
2
optimum size in evaporative cooling system is 5×2.5 m (58% of surface) and in compression cooling system is
2
4.31×2.15 m (43% of surface).

Figure 2 Comparison of optimal window sizes for eight types in North, for systems, evaporative cooling and
compression cooling

Figure 3 Comparison of optimal window sizes for eight types in South, for systems, evaporative cooling and
compression cooling

30th INTERNATIONAL PLEA CONFERENCE 5

16-18 December 2014, CEPT University, Ahmedabad
 14
Window Area with Compress ion cooling system

12 Window Area with Evaporative cooling system

East Window Area (m2)

10

0
6mm clear clear double clear double low-e clear low-e clear low-e clear reflective reflective
single glazed glaz ed with glaz ed with double double single glazed clear double clear single
air layer argon layer glaz ed with glaz ed with glaz ed with glaz ed
air layer argon layer air layer

Figure 4 Comparison of optimal window sizes for eight types in East, for systems, evaporative cooling and
compression cooling

Figure 5 Comparison of optimal window sizes for eight types in West, for systems, evaporative cooling and
compression cooling

Data in Figure 4 indicate that East is a good choice for reflective double glass and the dimensions of
2 2
evaporative cooling system is 2.23×4.46 m (46% surface) and of compression cooling system is 4×2 m (37% of
surface). As well as shown in Figure 5 in west direction, If using compression cooling system reflective double
glass, and if using evaporative cooling system Low-emissivity double glass with argon layer and reflective double

30th INTERNATIONAL PLEA CONFERENCE 6

16-18 December 2014, CEPT University, Ahmedabad
 2
glass are better choice and Their dimensions are respectively 1.98 × 3.96 m (36% surface) in compression
2
cooling system and 2.23 × 4.46 m (46% surface) for both glasses in evaporative cooling system.
T he Figure 6 shows that the optimum surface area for eight types of glass windows in the four cardinal
directions, by using Evaporative cooling system instead of compression cooling system, increases in various
glasses from 13 to 300 percent. As seen in Figure 6 , using an evaporative cooling system the window size can be
significantly increased with the objective to minimize the energy consumption. Among this low-emissivity double
glass with argon layer with more than 300% increase in the south had the highest and reflective double glass with
air layer with up to 34%, had the lowest increase. T his difference is due to the Low-emissivity double glass is
greater than reflective double glass amount of solar energy.
Also, Figures 7 and 8 show the window optimum area for eight types of glasses in four different directions
with each of the desired cooling system. It is observed that with evaporative cooling system, the optimum amount
of window area in the south is higher than the other main directions and the area of reflective Double glazed
window in both systems is higher than the other window.

Figure 6 Percent increase in the optimal value of the window area for eight types in four directions, for use
of the evaporative cooling system for comparing compression cooling system

Figure 7 The optimal value of the window for the evaporative cooling system efficiency

30th INTERNATIONAL PLEA CONFERENCE 7

16-18 December 2014, CEPT University, Ahmedabad
 Figure 8 The optimal value of the window for the compression cooling system efficiency

DISCUSSION AND CONCLUSION

In this study, the effects of location, material and size of the windows were investigated in order to minimize
the annual energy consumption of administrative units with emphasis on the effect of fenestration surface. T he
results indicate that using the evaporative cooling system compared to compression cooling system is more
appropriate and more economical in T ehran climate and if office window has been placed in the optimum
orientation, glass area can allocate up to 50% of façade surface. Also, reflective double glass and low-e double
glazed with argon layer glass are appropriate for Tehran office building units and can respectively allocate the
maximum level of window area and the minimum of energy consumption. Moreover, in the same conditions,
using the evaporative cooling system, window optimum size in the four main directions can be varied from 19%
for single-glazed window to 58% for reflective double-glazed window in the South direction.

REFERANCES
Lee, E., Selkowitz, S., 1995. “The design and evaluation of integrated envelope and lighting control strategies for
commercial buildings”. ASHRAE Transactions, 101 (1), 326-342.
T zempelikos, A., Bessoudo, M., Athienitis, A.K, Zmeureanu, R., 2010. “Indoor thermal environmental conditions
near glazed facades with shading devices – Part II: Thermal comfort simulation and impact of glazing and
shading properties”. Building and Environment, 45, pp. 2517-2525.
Hviid, C., Nielsen, T.R., Svendsen, S. 2008. “Simple tool to evaluate the impact of daylight on building energy
consumption”. Solar Energy, 82 (9), pp. 787-798.
Franzetti, C., Fraisse, G., Achard, G., 2004. “Influence of the coupling between daylight and artificial lighting on
thermal loads in office buildings”. Energy and Buildings, 36, pp. 117-126.
Citherlet, S., Clarke, J.A., Hand, J., 2001. “Integration in building physics simulation”. Energy and Buildings, 33,
pp. 451-461.
Walkenhorst, O., Luther, J., Reinhart, C., Timmer, J., 2002. “Dynamic annual daylight simulations based on one-
hour and one-minute means of irradiance data”. Solar Energy, 72(5), pp. 385-395.
T zempelikos, A., Athienitis, A., 2007. “The impact of shading design and control on building cooling and lighting
demand”. Solar Energy, 81, pp. 369-382.
Reinhart, C., Walkenhorst, O., 2001. “Validation of dynamic Radiance-based daylight simulations for a test office
with external blinds”. Energy and Buildings, 33 (7), pp. 683-697.
Loutzenhiser, P., Manz, H., Felsmann, C., Strachan, P.A., Maxwell, G.M., 2007. “An empirical validation of
modeling solar gain through a glazing unit with external and internal shading screens”. Applied Thermal
Enginneering, 22 (2), pp. 528-538.

30th INTERNATIONAL PLEA CONFERENCE 8

16-18 December 2014, CEPT University, Ahmedabad
Robinson, D., Stone, A., 2006. “ Internal illumination prediction based on a simplified radiosity algorithm”. Solar
Energy, 80 (3) , pp. 260-267.
Walkenhorst, O., Luther, J., Reinhart, C., Timmer, J., 2002. “Dynamic annual daylight simulations based on one-
hour and one-minute means of irradiance data”. Solar Energy, 72(5), pp. 385-395.
Mardaljevic, J., 2004. “Spatio-temporal dynamics of solar shading for a parametrically defined roof system”.
Energy and Buildings, 36 (8), pp. 815-823.
Fakra, A.F., Miranville, F., Boyer, H., Guichard, S., 2011. “ Development of a new model to predict indoor
daylighting: Integration in CODYRUN software and validation”. Energy Conversion and Management, 52
(7), pp. 2724-2734.
Ghisi, E., T inker, J., 2005. “An Ideal Window Area concept for energy efficient integration of daylight and
artificial light in buildings”. Energy and Buildings, 40 (1), pp. 51-61.
Reinhart, C., Selkowitz, S., 2006. “Daylighting—Light, form, and people”. Energy and Buildings, 38(7), pp. 515-
517.
Nabil, A., Mardaljevic, J., 2006. “Useful daylight illuminances: A replacement for daylight factors”. Energy and
Buildings, 38(7), pp. 905-913.
Reinhart, C., Wienold, J., 2011. “T he daylighting dashboard – A simulation-based design analysis for daylit
spaces”. Buildings and Environment, 46(2), pp. 386-396.
Seo, D., Ihm, P., and Krarti, M., 2011. “Development of an optimal daylighting controller”, Building and
Environment, 46, pp. 1011-22.
Hani, A., and Koiv, T.A., 2012. “Optimization of office building facades in a warm summer continental climate”,
Smart Grid and Renewable Energy, pp. 222-230.
Wetter, M., Polak, E., 2003. “A convergent optimization method using pattern search algorithms with adaptive
precision simulation”, IBPSA Conference, Eindhoven, Netherland.
Wetter, M., Wright, J., 2003. “Comparison of a generalized pattern search and a genetic algorithm opt imization
method”, IBPSA Conference, Eindhoven, Netherland.
Wetter, M., 2008. “GenOpt generic optimization program”, User manual version 2.1.1., Technical report LBNL -
54199, Building Technologies Program, Simulation Research Group, Lawrence Berkeley National Laborat

30th INTERNATIONAL PLEA CONFERENCE 9

16-18 December 2014, CEPT University, Ahmedabad