at SciVerse ScienceDirect

Building and Environment 57 (2012) 194e204

Contents lists available

Building and Environment

journal homepage: www.elsevier .com/locate/bui ldenv

Building façade design for daylighting quality in typical governmentoffice building

Yaik-Wah Lim a,*, Mohd Zin Kandar a, Mohd Hamdan Ahmad b, Dilshan Remaz Ossen a,Aminatuzuhariah Megat Abdullah a

aDepartment of Architecture, Faculty of Built Environment, Universiti Teknologi Malaysia, Malaysiab Institute Sultan Iskandar of Urban Habitat and Highrise, Universiti Teknologi Malaysia, Malaysia

a r t i c l e i n f o

Article history:Received 14 February 2012Received in revised form23 April 2012Accepted 24 April 2012

Keywords:BlindsField measurementGlareGlazingLight shelfSimulation

* Corresponding author. Tel.: þ60 167592058.E-mail address: limyaikwah@gmail.com (Y.-W. Lim

0360-1323/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.buildenv.2012.04.015

a b s t r a c t

Daylighting is one of the potential passive strategies to improve energy performance and users’ visualcomfort in existing offices without expensive installation and operational cost. For the first time,empirical study of daylighting performance was conducted for an existing typical government officebuilding designed by Public Work Department (PWD) in Malaysia. Daylight field measurement ofexternal illuminance and internal work plane illuminance was carried out. The findings demonstratedthat internal daylight level in the building was insufficient despite the abundance of external daylightavailability in the tropics which can achieve as high as 130 klx. Validation of simulation using Radiance-based software against the measurement under actual tropical sky indicated significant Pearson corre-lation of 0.709. Then, modification of window glazing and shading devices was experimented withsimulation. The findings evidenced that light shelf increased daylight distribution uniformity, but failedto reduce glare on vertical plane when direct sunlight patches occurred. Integration of light shelf withpartial blinds tilted at 45� showed improved performances of mean Guth Visual Comfort Probability(VCP) and CIE Glare Index. This paper concludes that simple modification of window glazing and shadingdevice is able to provide significant improvement in tropical daylighting quantity and quality for visualcomfort.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Since 2000’s, there are arising awareness and concerns towardsenergy efficiency in office buildings among the government andpolicy makers in Malaysia. A few prominent energy efficientbuildings were designed and constructed by the government asshowcase buildings to the building profession as well as the public.These include the Low Energy Office (LEO), Putrajaya in year 2004,Green Energy Office (GEO), Bangi in year 2007 and EnergyCommission (EC) Diamond Building, Putrajaya in year 2010. Allthese buildings have successfully achieved significant reduction inenergy consumption (with BEI 65 to 135 kWh/m2/yr) in compar-ison with the typical office buildings (with BEI 250 kWh/m2/yr) inMalaysia [1,2].

To promote energy efficiency in Malaysian government officebuildings, PublicWork Department (PWD) plays an essential role asthe driving forces on development and construction of government

).

All rights reserved.

buildings. PWD has designed standard office building for federaland states governments since 1960’s. However, many of thesestandard designs were developed for cost and space efficiencywithout considering energy efficiency. This has led to high energyconsumption among the existing typical government office build-ings. Hence, extensive efforts are urgently needed to study andimprove the energy performances in the existing typical govern-ment offices. Existing buildings are more constrained in compar-ison with new buildings in applying energy efficient designstrategies. Thus, feasible retrofit is needed with less cost effects.Thereby this paper suggests means of energy efficiency measuresby modifying the glazing types and solar shading to improve thebuilding daylighting and visual performances.

Daylight is more efficient in providing light in comparison withelectric light while producing less heat for the same amount of light[3]. Zain-Ahmed et al. [4] mentioned that minimum energy savingsof 10% can be achieved by using daylighting strategies alone.Besides, daylight is the full-spectrum of light that matches closestto human visual response. Quality of daylight is much better thanelectric light. Lesser amount of daylight is needed to perform a taskthan performing the same task under electric light. Apart from that,

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204 195

daylight also gives better colour rendering and visual environmentthat enables the occupant to see objects properly in a room [3].

As the use of computers and video display terminals (VDT) hasproliferated, visual comfort has become an important issue in officedesign [5]. Galasiu and Veitch [6] reviewed several surveys andconcluded that most of the responders believed daylight is superiorto electric light in its effects on people. Besides, human has greatertolerance to daylight uniformity in comparison with electric light[7]. Daylighting gives better visual comfort to the occupants.

The major drawback to harvest daylight in tropical climate ishigh intensity of solar radiation which will result unwanted solarheat gain. Record from Subang Jaya Meteorological Station showedthat the annual maximum intensity of solar radiation falling onhorizontal is about 1000 W/m2 and on vertical surface is about850 W/m2 for east and west facing surfaces [8]. Apart from that,uncontrolled daylight can lead to excessive glare especially whenworking with computer task [9]. Dubois [7] suggested that absolutework plane illuminance (WPI) higher than 500 lx is too bright forcomputer work. However, the global illuminance in tropical sky canbe higher than 100 klx [10]. Thus, high tropical daylight availabilitywill cause discomfort glare to the occupants.

As tropical sky is predominant by intermediate sky with clouds[10,11], the sky can be bright with patches around the edges of theclouds. Direct glare problem occurs when the occupants look to thehigh sky luminance through the windows. While the sun is notobscured by the clouds, direct sunlight patches falling on theinterior spaces also cause non-uniform daylight distribution,resulting great contrast and visual discomfort. Consequently, thechallenge in tropical daylighting is not about achieving the daylightquantity, but to control the daylight quality. Occupants need toavoid direct exposure to the sky component and direct sunlightpatches by using shading devices [12,13].

Inappropriate use of the shading devices will sacrifice naturaldaylighting and create discomfort glare. Many building users closeup their windows totally with blinds in order to avoid glare. Asa result, the abundance of daylight in tropical climate is yet to beutilised [14]. Therefore, the balance among the prevention of heatgains, control of discomfort glare and harvesting of daylight is verycrucial in order to develop energy efficient and visual comfortabledesign solutions [13,15e17].

1.1. Typical government office buildings by PWD

The design of typical government office buildings in Malaysia isconstrained by the organisational requirement as stated in

Fig. 1. Wisma Persekutuan Johor

Economy Planning Unit (EPU) Standard in Malaysia [18]. Accordingto Malaysian Ministry of Work, there are total of 64 federalgovernment office buildings designed by PWD and located atdifferent states of the country [19]. The conventional space plan-ning locates the individual officer’s rooms at the perimeter of thebuilding, thus limiting the potential of daylight utilisation. Tintedglazing and overhangs are commonly applied for all buildingfaçades without considering the orientation. Internal blinds orcurtain are widely used to control glare problem. This kind ofstandard design is continuously duplicated for many governmentoffice buildings for more than 50 years. However, there is lack ofdaylighting research on these buildings.

A recent survey of lighting conditions in 5 Malaysian govern-ment office buildings with different design typologies had beenconducted by the authors [20,21]. The findings evidenced that allthe selected buildings showed full dependency on electric lightingalthough there was sufficient external daylight availability in thetropical region. When the electric lights were switched off, all the 5buildings failed to meet MS 1525:2007 lighting recommendation[22]. The research concluded that all the 5 office buildings were notdesigned for daylight utilisation, with average daylight factor lowerthan 1.5% and poor daylight distribution uniformity. Therefore,investigation on the daylighting performance in typical govern-ment office buildings to improve energy efficiency and visualcomfort is critically demanded. This paper presents an empiricaldaylighting study for the first time in an existing typical govern-ment office building designed by PWD. Building façade modifica-tion was proposed to improve the daylighting performance.

1.2. Study case e Wisma Persekutuan Johor Bahru (WPJB)

WPJB is an existing typical government office building designedby PWD and completed in year 1978. This building is located at thecity centre of Johor Bahru with latitude 1�27’37"N and longitude103�45030"E. This building was selected for study because it showsthe design typology (floor layout, building form, building heightand façade design) which is commonly employed by PWD. It isorientated with linear façades facing north-east and south-west.The building is 10-storey high with total lettable area of19,694 m2 and ceiling height of 2450 mm. It is free from shadowcasting from adjacent buildings and vegetation.

As shown in Fig. 1, the building façade is clothed with extensiveuse of external shading device with overhang and vertical screenfor all orientations. This kind of shading device is a commonelement in many typical government office buildings in order to

Bahru by PWD in year 1978.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204196

avoid direct solar heat gain and glare problem. Survey by theauthors had identified at least 18 typical government office build-ings located throughout Malaysia with similar building façadedesign as shown in Fig. 2. However, the potential of this shadingdevice for daylighting demands further investigation. Apart fromthat, the common glazing types employed in these buildings aretinted glazing with low visible transmittance (VT).

2. Method

2.1. Field measurement

As shown in Fig. 3, a typical individual officer’s room in WPJBfacing south-west (220�N) was selected for daylight measurement.The selected room is at the top floor and the blockage from adjacentbuilding or vegetation is minimised. This room has deep plangeometry with room depth (D) to width (W) ratio 2.2 (as shown inFig. 4), where the diminishing effect of side-lighting is significant.The floor area of the room is representative to the office room sizefor all government staffs with Grade 41e46 (18 m2) according toEPU Standard [18].

One illuminance meter Delta OHM LP-PHOT 02, Probe E (PE)with data logger was installed on the roof-top to measure theexternal illuminance (Fig. 5a). The measurement range of theequipment was 0e150 klx. Concurrently, 2 illuminance metersDelta OHM LP-PHOT 01 with data loggers, Probe 1 (P1) and Probe 2(P2) were installed inside the room to measure internal WPI asshown in Fig. 5b. P1 was located almost at the centre of the roomwhile P2 was situated near to the rear wall, at the work planeheight (750 mm) as indicated in Fig. 4. All the measurements weretaken from 26 March until 8 April 2011 (9 days), with intervals of10 mins. However, this office room was occupied by user andelectric lighting was used during the working days. Thereby onlythe data during 2 weekends (4 days; 26e27 March and 2e3 April)when the electric lights were off was taken for daylighting analysis.

A few assumptions were made to control the design variables tobe studied. During the measurement, all the transparent openingsapart from the external window were covered with opaque card-boards so that only one side-lighting effect was considered. Thiswas to eliminate influences from other lighting sources apart fromthe daylight. Thus, the diminishing effect of side-lighting can bedistinctly observed. Besides, the internal shading (vertical blinds)was totally opened in order to ignored its effect on daylighting. This

Fig. 2. Examples of typical government office buildings with similar plan shape and façadKompleks Pejabat Kerajaan Jalan Duta, Kuala Lumpur in 1978; (c) Wisma Persekutuan Kota BTaiping, Perak in 1991; and (f) Wisma Persekutuan Tawau, Sabah in1984.

was because the user’s adjustment on the internal shading wassubjective and unable to be generalised. Hence, in this study, onlythe effect of the external shading was investigated. The details ofthe characteristics of the office room are stated in Table 1.

2.2. Computer simulation with radiance

This paper employed Radiance-based computer simulation tocompare with the field measurement results. Radiance is devel-oped by Greg Ward at Lawrence Berkeley National Laboratories,widely recognised and validated by the lighting professionals. It canhandle virtually any space geometry, as well as non-diffusereflectance [23,24]. Previous research has demonstrated the reli-ability of this software in simulating daylighting performances forvarious design variables and sky conditions [7,13,25e28].

In modelling the officer’s room, a few assumptions were made.Previous research showed that the relative differences of modellingwith furniture and empty room ranged from 0 to 35%, varied asa function of sun angle, distance from window and also furniturearrangement [7]. Thus, any single furniture arrangement will givea specific illuminance distribution. It is unpractical to generalise thefurniture arrangement for daylighting study. Therefore, in thisstudy, office furniture was ignored in order to simplify themodelling.

In order to imitate the condition of the field measurement, onlyone side external window was modelled without internal shading.The window frame was ignored in the modelling since the effectwas not significant. All the surfaces reflectance values weremodelled according to the actual room. Intermediate sky wasemployed for simulation as previous research proved that tropicalsky is predominated by this sky condition [10,11]. The fieldmeasurement recorded unstable external illuminance levels. Thisindicated dynamic formation of cloudswhich further supported theprevious finding that tropical sky is predominantly intermediate.The locations of the measurement points (P1 and P2) were exactlysame with the field measurement while the external illuminancevalues were taken from the summary of sky condition in thesimulation software. The date and time of simulations were 27March 2011, 800 he1700 h. For further analysis on the effect ofsurface colours on simulation results, 2 models were developed asshown in Fig. 6: (a) with colour and (b) with grey scale. Both of themodelling imitated surface reflectance values as stated in Table 1.Then, the simulation results were compared with the mean

e design: (a) Bangunan Tun Datuk Patinggi Tuanku Haji Bujang, Sarawak in 1975; (b)ahru, Kelantan in 1979; (d) Wisma Persekutuan Melaka in 1978; (e) Wisma Persekutuan

Fig. 3. Individual officer’s room selected for daylight measurement.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204 197

measurement results during the 2 weekends (4 days) when electriclights were off.

Further simulation tests were carried out by modifying theexisting glazing and shading device in order to study the impact ofthese design variables on improving the daylighting performance.Firstly, the existing window was replaced with a bared windowwithout any shading device and using clear glazing with VT 75.0%.Then, an external and internal light shelf was installed at 1130 mmheight from window sill in order to enhance the daylight unifor-mity. Finally, integrations of the light shelf and partial blinds titledat 45� were developed to eliminate glare problem. All the test casesare shown in Fig. 7.

Fig. 4. Geometry of the officer’s room and locations o

Previous research demonstrated that the application of CIE(International Commission on Illumination) skies to simulatetropical daylighting will result in underestimation of absolutedaylight level [29]. This is because the external illuminance in thetropics can be as high as 130 klx while the simulationwith CIE skiesgives external illuminance below 20 klx. Thus employing relativeratios is more reliable in comparison with the absolute values.

Daylight ratio has been employed by previous daylightingresearch in tropical climate. In this paper, daylight ratio wasemployed to determine the percentage ratio of internal WPI levelwith its simultaneous external illuminance level as indicated inEquation (1). This measurementmethod is viable in regions close tothe equator to study the indoor daylight availability [30]. Besides,WPI uniformity ratio as shown in Equation (2) was applied toinvestigate the daylight distribution. Higher uniformity ratio indi-cates better daylight distribution [7,13,31]. All simulations weredone by employing intermediate sky as tropical sky is predomi-nantly intermediate [10,11]. The dates and times selected for theseanalyses were 21st March, 22nd June and 22nd December, 900 h,1200 h and 1500 h to indicate the different solar angles throughoutthe day when the Sun is in equinoxes, north solstice and southsolstice respectively.

Daylight ratio ¼ internal WPI=external illuminance�100 (1)

WPI uniformity ratio ¼ minimum WPI=average WPI (2)

For glare analysis, Guth Visual Comfort Probability (VCP) and CIEGlare Index were employed. Guth VCP is to measure percentage ofpeople who would be satisfied with a particular view; while CIE

f WPI measurement points: (a) Plan; (b) Section.

Fig. 5. Instrumentation: (a) Delta OHM LP-PHOT 02 with data logger for external illuminance and (b) OHM LP-PHOT 01 with data loggers for internal illuminance.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204198

Glare Index is a unitless index of visual comfort which is used as aninternational standard. Guth VCP value of 65% and CIE Glare Indexvalue of 19 are acceptable for sedentary conditions; Guth VCP valueof 52% and CIE Glare Index value of 22 are acceptable for transientconditions [32].

It is not practical to simulate all possible views, dates and timesfor glare analysis [7,13,32]. Subsequently, the view to the externalwindow from the centre of the room at eye-level with sittingposition was selected for the analysis as the main daylight glaresource was from the external window. Since this room was facingsouth-west (220�N), the most critical date and time for glare wasduring Decemberwhen the Sun is in south solstice and at afternoonwhen the Sun is at low angle in the west. Thus, the date and timeselected were 22 December 1500 h with solar altitude 52.3� andazimuth 226.8�.

3. Result

3.1. Field measurement

As shown in Fig. 8, the internal WPI was influenced by theexternal illuminance. Generally, 27 March 2011 and 3 April 2011showed higher external illuminance in comparison with the other

Table 1Characteristics of the officer’s room.

Element Sub-element Detail

Window Window-to-wallratio

64.5%

Glazing type TintedVisual transmittance(VT)

37.0%

Shadingdevice

External Overhang with verticalscreen

Internal Vertical blind (light green) eIgnored

Surfacereflectance

Wall 49.5% (green)Floor 8.5% (blue carpet)Ceiling 78.6% (white)Door 11.4% (brown)

Electriclighting

Fluorescent (2 tubes) 3 quantity

days. The highest recorded external illuminance was about 130 klxduring noon time, while the lowest during office working hour wasapproximately 20 klx. This demonstrated that there is abundance ofdaylight in Malaysian tropical sky.

The high quantity of external illuminance failed to illuminatethe officer’s room. The results of P1 and P2 showed that the WPIwas constantly below 70 lx, which was much lower in comparisonwith the minimum requirement of 300 lx in general office as statedin MS 1525 [22]. This indicated that the room had insufficientbrightness for any office task performance without electric light. P2was consistently lower than P1, showing the diminishing effect ofside-lighting.

3.2. Simulation validation

Fig. 9 shows that the simulatedWPI was consistently lower thanthe measured result. The highest difference of simulated WPI wasapproximately 80% lower than the measurement. This indicatedagreement with the previous research [29] that simulation tendedto underestimate the absolute WPI under tropical sky as theexternal illuminance in simulated CIE sky was much lower.Nevertheless, the result demonstrated that daylighting simulationwith colour and simulation with grey scale gave similar perfor-mance results.

A clearer comparisonwas made by calculating the daylight ratioof the measured and simulated results. Fig. 10 demonstrated thatthe measured and simulated daylight ratios gave similar perfor-mance pattern. The average difference between the measured andsimulated results for P1 and P2 were approximately 12.4% and65.1% respectively for modelling with colour. This showed that asthe distance from the external window increased, the simulationtended to underestimate the daylight level. This could be due to thelimitation of the modelling to imitate the internal surfaces reflec-tance and office furniture in the actual room. However, themodelling with colour and with grey scale recorded identicalperformance results.

Statistical analysis was employed to compute the Pearsoncorrelation among the daylight ratio results of the measurement,simulation with colour and simulation with grey scale as shown inTable 2. The analysis indicated significant correlation between the

Fig. 6. Computer modelling: (a) with colour and (b) with grey scale.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204 199

measured and simulated results with Pearson correlation of 0.658and 0.709 for modelling with colour and modelling with grey scale,respectively. The Pearson correlation between the modelling withcolour and modelling with grey scale was as high as 0.937. Theresults justified that the Radiance-based simulation software wasable to accurately predict daylight ratio under tropical sky. Besides,simplification of modelling can be made by modelling the surfacereflectance values with grey scale rather than imitating the actualcolour.

3.3. Illuminance analysis

Table 3 indicates the daylighting performance results of theexisting test case and modification of the building façade. The

Fig. 7. Test case: (1) Existing; (2) Bared window with clear glazing; (3) Clear glazing with ligand blind 2.

existing case showed that the mean daylight ratio was low. Thelowest mean daylight ratio was 0.09% at 1200 h on 21 March whilethe highest was 0.83% at 1500 h on 22 December. This was becausethe external windowwas facing south-west which received highestamount of daylight during afternoon in December when the Sunwas at south solstice. Apart from that, the existing test case recor-ded low daylight uniformity ratio which ranged from 0.03 to 0.07.In overall, the existing test case did not show potential of daylightutilisation with mean daylight ratio constantly below 1.00% due tothe extensive use of external overhang and tinted glazing.

Modification of the façade demonstrated evident improvementin daylight availability and distribution uniformity. By simplyomitting the existing external overhang and replacing the tintedglazing with clear glazing, the bared window showed as much as

ht shelf; (4) Clear glazing with light shelf and blind 1; (5) Clear glazing with light shelf

Fig. 8. WPI (lx) and external illuminance (klx) measurement results on 26e27 March 2011 and 2e4 April 2011.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204200

2176% increment of mean daylight ratio at 1500 h on 22 December.Similar with the existing test case, the bared window test case alsoyielded lowest daylight ratio at 1200 h while the highest was at1500 h on 22 December. The mean daylight ratios were higher than6.00% at 1500 h on 21 March (10.04%) and at 1500 h on 22December (18.89%). According to MS 1525:2007, daylight factor(DF) above 6.00% will cause glare and thermal problem [22]. On theother hand, the bared window gave minor improvement inuniformity ratio which ranged from 0.04 to 0.09. This was becausethe bared window allowed deeper daylight penetration incomparison with the existing design.

Installation of light shelf achieved success in both reducingexcessively high mean daylight ratio and increasing uniformityratio. The highest mean daylight ratio at 1500 h on 22 Decemberwas reduced as much as 54.05% from 18.89% (bared window) to8.68%. While for 1500 h on 21 March, the mean daylight ratio wasbelow 6.00% which avoided glare and thermal problem. However,during 1200 h on 21 March and 22 June, the mean daylight ratioswere only 0.60%, slightly below the MS 1525:2007 recommenda-tion of DF 1.00% for average daylighting. On the other hand, therewas obvious improvement in the daylight uniformity ratio. Thewindowwith light shelf recorded uniformity ratio from0.07 to 0.13,approximately 85.19% increment from the bared window. The

Fig. 9. Comparison of measured and simulated (with colour and grey scale) WPI (lx).

lowest uniformity ratio was recorded at 1500 h on 22 Decemberwhen the daylight availability was the highest.

3.4. Luminance and glare analysis

Table 4 summarises the illuminance, luminance and glareanalysis for all the test cases at 1500 h on 22 December, which wasthe most critical time and date for glare in the south-west facingroom. For the existing case, as the indoor daylight availability waslow (mean daylight ratio of 0.83%), the glare problem was notcritical. The mean Guth VCP was 29.56% and mean CIE Glare Indexwas 23.37 (slightly higher than 22 which is acceptable for transientcondition). The glare threshold of 67.95 cd/m2 indicated thatgenerally the room was dark.

The results indicated glare problemwas critical for clear glazingwithout external overhang. The bared window showed 34.72%decrement of mean Guth VCP (19.29%) and 19.47% increment ofmean CIE Glare Index (27.92) in comparison with the existingdesign. The glare threshold was increased to 576.25 cd/m2. Intro-ducing light shelf improved the daylight uniformity but actually

Fig. 10. Comparison of measured and simulated (with colour and grey scale) daylightratio (%).

Table 2Correlation analysis among daylight ratios of measurement, simulation with colourand simulation with grey scale.

Measured Simulated(with colour)

Simulated(grey)

Measured PearsonCorrelation

1 0.658a 0.709a

Sig. (2-tailed) 0.002 0.000N 20 20 20

Simulated(With Colour)

PearsonCorrelation

0.658a 1 0.937a

Sig. (2-tailed) 0.002 0.000N 20 20 20

Simulated(Grey)

PearsonCorrelation

0.709a 0.937a 1

Sig. (2-tailed) 0.000 0.000N 20 20 20

a Correlation is significant at the 0.01 level (2-tailed).

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204 201

worsen the glare problem. The test case with light shelf furtherreduced the mean Guth VCP to 15.22% whereas the mean CIE GlareIndex and glare threshold remained similar with the baredwindow.

As previous research showed that blinds were good in reducingglare [7,13], 2 types of blinds (Blind 1 and Blind 2) were installed asshown in Fig. 7. The simulation results showed that blind at themiddle part of the window (Blind 1) gave more significantimprovement in reducing glare compared to blind at the bottompart of the window (Blind 2). For test case with Blind 1, the meanGuth VCP was 31.96% while the mean CIE Glare Index was 25.17.Meanwhile, test case with Blind 2 recorded mean Guth VCP of

Table 3False colour contour, mean daylight ratio and uniformity ratio for various test cases.

17.04% and mean CIE Glare Index of 26.54. The use of Blind 1 alsoslightly reduced the high mean daylight ratio to 7.11%, while Blind 2evidently reduced it to 2.21%. Unfortunately, Blind 1 reduced theuniformity ratio from 0.07 (with light shelf) to 0.03. This was due tothe dark shade created by the blind at the middle part of thewindow. On the contrary, Blind 2 improved the uniformity ratio to0.23. The glare thresholds for test case with Blind 1 and Blind 2were 377.24 cd/m2 and 438.55 cd/m2, respectively.

4. Discussion

The validation of simulation against field measurement sug-gested that adaptive application of the Radiance-based simulationsoftware is needed in order to accurately predict the tropicaldaylighting performance. As shown by previous research [29],simulation performance criteria using relative ratio (such asdaylight factor) was more reliable than using the absolute value(such as WPI). In this paper, the Pearson correlation between thesimulated and measured daylight ratio was significant. Moreover,simulation of modelling with colour and modelling without colourrecorded similar results (Pearson correlation of 0.937). Subse-quently, this paper recommends the use of daylight ratio as theperformance criteria for tropical daylighting simulation. Surfacecolours can be ignored in order to generalise the surface reflectancevalues.

Previous research by Sharifah and Sia [15] studied the impact ofsolar shading on various orientations of a government mixed-usecommercial building, Kompleks Tun Razak (KOMTAR) in Penang,Malaysia. However, KOMTAR has 12 similar-sided polygons floor

Table 4Summary of illuminance and glare analysis for various test cases on 22 December; 1500 h.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204 203

plan that looked almost circular from afar; is not a conventionalgovernment office building. WPJB demonstrated the most commontypologies of floor plan, room height and façade design for theexisting government office buildings in Malaysia designed by PWD.These kinds of design typologies were initially developed for costand space efficiency. They had been widely replicated in manygovernment office buildings. Nevertheless, the findings of the fieldmeasurement evidenced that these kinds of design typologies didnot show daylighting efficiency. In spite of the external illuminancein tropical climate was peaked with 130 klx, the indoor daylightlevel in WPJB was always below 70 lx.

Common daylighting design rule of thumb recommended 2.5 H(height of the top of external window from work plane height) asthe maximum depth of the daylit area from external window [33].However, WPJB showed much lower daylit area depth. P1 withapproximately 2.0 H (3.5 m from external window) recordedmaximum of 70 lx only. Previous research by Shahriar and Mohit[34] in tropical climate stated that the depth of daylighting zone for300 lx WPI was 3.5 m. A study by Lim [13] proved that daylit areawith WPI above 100 lx for an office with bared window in tropicalclimate under 20 klx overcast sky can be 3.6 H. This justifies that thelow daylight penetration in WPJB was not due to the deep roomplanning. Therefore, inappropriate building façade design was thereason for poor daylighting.

The simulation results further supported that modification offaçade design in the existing government office building wasurgently demanded to achieve daylighting efficiency. Simplemodification such as omitting the existing overhang and changingthe glazing type demonstrated substantial increment in indoordaylight availability. Nevertheless, the findings suggested thata bared window in tropical climate will cause glare and thermalproblem (with maximum daylight ratio of 18.89%). Lim [13]expressed that this kind of design, although providing sufficientdaylight level, will eventually result total ‘rejection’ of daylight bythe users through closing up the window with blinds. This was dueto users’ visual and thermal discomfort. Thereby light shelf andblind were recommended to control the daylight quality for visualand thermal comfort [7,13,35].

Light shelf successfully reduced the excessively high work planedaylight level under tropical sky. Besides, it also improved thedaylight distribution uniformity by reflecting the daylight into thedeeper room area. Subsequently, light shelf is one of the mostpotential passive strategies for tropical daylighting. However, thesimulation on the most critical day for glare (22 December) showedthat light shelf was not good in reducing glare on vertical planewhen the direct sunlight patches penetrated into the room. In fact,the installation of light shelf reduced the mean Guth VCP incomparison with the bared window. This was because light shelfcreated dark shade that contrastedwith the bright sunlight patches.

The integrations of light shelf and blinds resulted successfulreduction in glare on vertical plane. However, the simulationresults showed that proper placement of blinds was as critical asdemonstrated by previous research [7,13,36]. The placement ofblind at the middle part of the window (Blind 1) yielded 87.56%increment on mean Guth VCP in comparisonwith the placement ofblind at the bottom part of the window (Blind 2). Therefore,building designers shall take into consideration of the properinstallation of blinds for effective daylighting. In this paper, inte-gration of light shelf with Blind 1 (tilted 45�) was the best option interm of reducing glare on vertical plane when there was directsunlight patch. However, it reduced the daylight distributionuniformity on horizontal plane. Thus, the use of blinds shall beflexible. When there is no direct sunlight patch (varied time fordifferent orientations), the blinds shall be rolled up to allow betterdaylight penetration and distribution uniformity.

5. Conclusion

This paper has presented empirical study of daylightingperformance for the first time in an existing typical governmentoffice designed by PWD in Malaysia. The findings urged reconsid-eration on the typical façade design that was in fact not appropriatefor tropical daylighting. This paper demonstrates that simplemodification of the external shading device and glazing type couldprovide significant improvement in the indoor daylight quantityand quality. However, flexible internal shading device such as blindwas necessary to control the direct sunlight patches to avoid glareproblem.

This paper focuses only on the daylighting performance forvisual comfort. Providing comfortable daylighting condition will ofcourse reduce the use of electric lighting thus contributing tobuilding energy saving. Besides, avoiding direct sunlight patch willalso decrease the heat gain and thermal discomfort. Nevertheless,further research is needed on the thermal performance of theproposed modification of building façade as tropical daylighting isalways associated with intensive solar heat gain.

Acknowledgement

This research is funded by the Public Works DepartmentMalaysia (PWD) and Universiti Teknologi Malaysia (UTM) througha collaborative research grant, Vote 73751. The authors would alsolike to acknowledge the assistance from the PWD and UTMresearch teams in conducting the research.

References

[1] Chan SA. Green building Index - MS1525. In: PAM CPD Seminar: applyingMS1525:2007 code of practice on energy efficiency and use of renewableenergy for non-residential buildings. Kuala Lumpur, Malaysia; 14 Feb 2009.p. 1e22.

[2] Mohd Zin K. Energy efficient building design towards sustainable develop-ment. Symposium International Urbanization, Architecture and Developmentin Nusantara. Selectra Building. Universitas Sumatera Medan Indonesia; 18Jun 2005 [Malay].

[3] Robbins CL. Daylighting: design and analysis. New York: Van NostrandReinhold Company; 1986.

[4] Zain-Ahmed A, Sopian K, Othman MYH, Sayigh AAM, Surendran PN.Daylighting as a passive solar design strategy in tropical buildings: a casestudy of Malaysia. Energy Conversion Manag 2002;43:1725e36.

[5] Guzowski M. Daylighting for sustainable design. New York: McGraw-Hill;1999.

[6] Galasiu AD, Veitch JA. Occupant preferences and satisfaction with the lumi-nous environment and control systems in daylit offices: a literature review.Energ Building 2006;38:728e42.

[7] Dubois MC. Impact of solar shading devices on daylight quality: simulationswith radiance. Research Report. Lund, Sweden: Lund University; 2001.

[8] Ossen DR. Optimum overhang geometry for high rise office building energysaving in tropical climates. Doctor of Philosophy, Faculty of Built Environment.Universiti Teknologi Malaysia; 2005.

[9] Carmody J, Selkowitz S, Lee SE, Arasteh D, Willmert T. Window systems forhigh performance buildings. 1st ed. New York: W.W.Norton & Company, Inc.;2004.

[10] Zain-Ahmed A, Sopian K, Othman MYH, Abidin ZZ. The availability of daylightfrom tropical skies: a case study of Malaysia. Renew Energ 2002;25:21e30.

[11] Djamila H, Ming CC, Kumaresan S. Estimation of exterior vertical daylight forthe humid tropic of Kota Kinabalu city in East Malaysia. Renew Energ 2011;36:9e15.

[12] Bülow-Hübe H. Energy-efficient window systems. Doctor of Philosophy. LundInstitute of Technology, Lund University; 2001.

[13] Lim YW. Internal shading for efficient tropical daylighting in high-rise openplan office. Doctor of Philosophy, Faculty of Built Environment. UniversitiTeknologi Malaysia; 2011.

[14] Lim YW, Mohd Hamdan A. Daylight and users’ response in high rise open planoffice: a case study of Malaysia. 3rd International Graduate Conference onEngineering, Science, and Humanities. Skudai, Johor, Malaysia: UniversitiTeknologi Malaysia; 2e4 Nov 2010.

[15] Sharifah FSF, Sia SJ. Sunlight control and daylight distribution analysis: theKOMTAR case study. Building Environ 2004;39:713e7.

[16] Tzempelikos A, Athienitis AK. The impact of shading design and control onbuilding cooling and lighting demand. Solar Energ 2007;81:369e82.

Y.-W. Lim et al. / Building and Environment 57 (2012) 194e204204

[17] Lim YW, Mohd Hamdan A, Ossen DR. Effect of solar shading, fenestration androom geometry on office building energy efficiency in hot humid tropic. 3rdinternational Conference on Built Environment in Developing Countries.Penang, Malaysia: Universiti Sains Malaysia; 2-3 Dec 2009. p. 174e185.

[18] EPU. Garis panduan dan peraturan bagi perancangan bangunan. Putrajaya,Malaysia: Economy Planning Unit (EPU), Department of Prime Minister; 2008.

[19] http://www.kkr.gov.my.[20] Mohd Zin K, Yong Razidah R, Mohd Sabere S, Ossen DR, Mohd Hamdan A,

Lim YW, et al. Survey of surface reflectance, work plane illuminance level anduniformity in Malaysian government office buildings. ICEESD 2011, Interna-tional Conference on Ecosystems. Venice, Italy: Environment and SustainableDevelopment; 28-30 November 2011. p. 3330e3335.

[21] Mohd Zin K, Mohd Sabere S, Yong Razidah R, Ossen DR, Aminatuzuhariah MA,Lim YW, et al. Investigating daylight quality in Malaysian government officebuildings through daylight factor and surface luminance. ICEESD 2011,International Conference on Ecosystems. Venice, Italy: Environment andSustainable Development; 28-30 November 2011. p. 1793e1798.

[22] Department of Standards Malaysia. MS 1525: 2007. Malaysian standard: codeof practice on energy efficiency and use of renewable energy for non-residential buildings. 1st revision; 2007. Malaysia.

[23] Mistrick RG. Desktop radiance overview. US: Pacific gas and Electric; 2000.[24] Reinhart C, Fitz A. Findings from a survey on the current use of daylight

simulations in building design. Energ Buildings 2006;38:824e35.[25] Reinhart CF, Walkenhorst O. Validation of dynamic RADIANCE-based daylight

simulations for a test officewithexternal blinds. EnergBuildings2001;33:683e97.[26] Kim SY, Song KD. Determining Photosensor conditions of a daylight Dimming

control System using different Double-skin Envelope Configurations. IndoorBuilt Environ 2007;16(5):411e25.

[27] Li DHW, Cheung GHW, Cheung KL, Lam JC. Evaluation of a simple method fordetermining the vertical daylight factor against full-scale measured data.Indoor Built Environ 2009;18(6):477e84.

[28] Iversen A, Nielsen TR, Svendsen SH. Illuminance level in the urban fabric andin the room. Indoor Built Environ 2011;20(4):456e63.

[29] Lim YW, Mohd Hamdan A, Ossen DR. Empirical validation of daylight simu-lation tool with physical model measurement. Am J Appl Sci 2010;7(10):1412e9.

[30] Dahlan ND, Jones PJ, Alexander DK, Salleh E, Alias J. Daylight ratio, luminance,and visual comfort assessments in typical Malaysian hostels. Indoor BuiltEnviron 2009;18(4):319e35.

[31] Ho MC, Chaing CM, Chou PC, Chang KF, Lee CY. Optimal sun-shading designfor enhanced daylight illumination of subtropical classrooms. Energ Buildings2008;40:1844e55.

[32] Murray V, O’Flynn CJ, Beattie K. Advanced building services simulationsoftware providing design solutions in Dublin and Boston. 7th Inter-nation IBPSA Conference. Brazil: Rio de Janeiro; 13-15 August 2001.p. 499e504.

[33] Egan MD, Olgyay V. Architectural lighting. 2nd ed. New York: McGraw-HillHigher Education; 2002.

[34] Shahriar ANM, Mohit MA. Estimating depth of daylight zone and PSALI forside lit office spaces using the CIE standard general sky. Building Environ2007;42:2850e9.

[35] Mansour A, Sopian K, Zain-Ahmed A, Reimann G. Daylight distribution ofa new design for future commercial office building in Malaysia. ISESCO SciTechnol Vis 2006;2(1):53e6.

[36] Koo SY, Yeo MS, Kim KW. Automated blind control to maximize the benefitsof daylight in buildings. Building Environ 2010;45:1508e20.