Comprehensive maintainability scoring system (COMASS) for commercial buidings in tropical climate of singapore 3

Table 5.1 Maintainability factors for basement and their RWs Component Maintainability factors Critical defect mitigated Non-critical defect mitigated Wt... 42, 26 and 10 number of fa

Chapter 5 Maintainability Scoring for Building Subsystems

5.1 Introduction

Maintainability is a design parameter and hence should be addressed right from the design phase Traditionally the maintainability guidelines are focussed on design and to a certain extent on construction and installation But Chapter 4 has highlighted that to achieve high maintainability, like design or construction, maintenance should also be ‘designed’ or planned Each subsystem was divided into top-down hierarchy in as many numbers of components as possible and the related defects were graded in terms of cause and criticality These findings were extensively used in current section to develop the best maintainability practices These guidelines in the form of maintainability checklist factors are linked with the defects they can mitigate and such linkage is expressed through the factors’ relative weights Hence this chapter addresses the second objective of this research: ‘Setting benchmark for good design, construction and maintenance practices and provide guidelines for optimum selection of maintenance strategies by developing individual maintainability scoring framework for major building systems’

5.2 General format of maintainability scoring

5.2.1 Mathematical principle

The details can be referred back to the research methodology (Section 3.3) Briefly, guidelines for design, construction and maintenance were developed in the form of a checklist The checklist factors are grouped under life-cycle phases (design, construction and maintenance) and further sub-grouped under components For each factor a design scheme can score 5 for compliance and 1 for violation of the given guideline This score is adjusted if for the same factor, different standards are followed in different sections of the building Upon weighted summation, the total score for the subsystem is determined Hence the scoring

contains two parts: (1) development of guidelines for the maintainability factors and (2) derivation of their relative weights For derivation of relative weights, details of defects and criticality values should be referred back to the corresponding tables in Chapter 4 Each of the nine subsystems follows the same format

5.2.2 Maintainability Handbook

As mentioned, this is the collation of proposed maintainability guidelines and the second main deliverable of this research project It spans about 150 pages and is an extract of more than 400 various sources of literature and practical knowledge elicited from site visits and interviews Apart from its academic contribution of integrating causes and effects of building defects, it aspires to be a benchmark for industry practices Hence it is presented as a standalone document in Appendix C (Section 1-9)

5.3 Maintainability scoring for basement

Basements are subjected to a permanent hydrostatic pressure and probable aggressive soil conditions Water-tightness is a critical issue for basement maintainability It depends primarily on the system selection, structural concrete, detailing of waterproofing, drainage system to reduce hydrostatic pressure around basement wall, adequacy of waterproofing over penetrations, projections or joints and coordination with other services located in the basement Additionally the flooring and finishes add to the maintainability as they have direct influence on the ease of usage A basement once constructed has limited scope of repair or replacement upon occurrence of cracks or subsurface seepage In fact maintenance involves only floor, walls and drainage Hence maintainability in terms of water-tightness should be high Sequence of construction and quality control are also critical issues (Chew, 2000)

Maintainability guidelines (Appendix C.1) was developed for the major components, namely, water proofing, structural elements, drainage system, finishes on floor and wall and ancillary facilities Out of total 59 maintainability factors, 28 are for design, 21 are for construction and

rest are for maintenance and external factors The summary of the factors and their relative

weights are presented in Table 5.1 Details of the defects and their criticality index can be

referred back to Table 4.4

Table 5.1 Maintainability factors for basement and their RWs

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW Design phase

a1 Selection /usage A3, A5 0.454 0.021 a2 Application feasibility A3, A5 0.454 0.021 W/P system

a4 System / construction method selection

a5 Design (mix & rebar) A3, A4 A5 A1, A111, A17 1.117 0.052 a6 Joint details A1, A12, A17 0.333 0.016 Concrete

a13 Pipe penetrations A19 0.289 0.014

Waterstop (in

Type B

a17 Cavity wall design A6, A11 0.065 0.003 a18 Masonry block and

a20 Cavity floor design A20 A6 0.311 0.015

Drainage

Flooring

a26 Paint selection A8, A10 0.230 0.011 Wall finishes

a27 Tile selection A12, A13 0.061 0.003 Ancillary

a31 Material quality A20 A11, A17 0.057 0.003 a32 Casting & curing A3, A4, A5 A1, A12, A17 1.110 0.052 RCC

a34 Substrate & material quality

a36 Inspect & test A3 A1 0.541 0.025 W/P install

a39 Material quality &

condition

Cavity

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW

a45 Substrate & material quality

Paint &

plaster

a47 Screed & tile prep A12, A14 0.056 0.003

Tiling

Construction subtotal 0.285

Maintenance phase

General a50 Inspection A4, A5, A15, A19 A2, 18 1.569 0.074

Drainage

External

factors

a56 Soil permeability A1, A3, A5 0.541 0.025 a57 Aggressive chemical A1, A3 0.541 0.025

a59 Building age A1, A3, A4 0.865 0.041

Total 1.000

5.4 Maintainability scoring for facade

Facade as building envelope needs to meet the primary requirement of weather exclusion

Hence water tightness is the key concern for facade maintainability followed by aesthetics

Various facade options ranging from traditional brick or block masonry to modern metal or

glass curtain wall vary largely in these two aspects Each system should be carefully detailed

to minimize problems during the service life Apart from system selection, wall shape, grid

and joint details should be considered Additionally, complexity in building profile affects

accessibility and exposure condition influences facade durability Considering these factors,

materials are usually chosen that can facilitate cleaning and partial removal However facade

maintainability remains incomplete without addressing the issues of window They are weak

points in facade fabric allowing a path for seepage and control run-off profile Moreover,

windows require more frequent cleaning compared to the rest part of the facade Hence it is

necessary to have safe and easy cleaning provisions These aspects were taken into

considerations while developing maintainability scoring for facade (Table 5.2, Appendix

C.2) Corresponding defect information can be found in Table 4.6 Total 78 factors were

identified for 8 various types of facades 42, 26 and 10 number of factors were attributed to

design, construction and maintenance phase respectively,

Table 5.2 Maintainability factors for facade and their RWs

Compo-nent

Maintainability factors

Critical defect mitigated

Compo-nent

Maintainability factors

Critical defect mitigated

0.159 0.007 b64 Setting fixings B44 B32, B37, B40 0.275 0.013

Metal

b65 Panel erection B44 B32, B34, B35,B36,B40,

B41

0.318 0.015 b66 Substrate prep B44 B6, B42, B45 0.274 0.013

Total 1.000

5.5 Maintainability scoring for wet area

Wet area needs to fulfil the requirement of internal water tightness to prevent leakage and

keep a safe and healthy environment Water-tightness relies mainly on the adequacy of

waterproofing over the floor and wall surfaces punctured by unavoidable penetrations,

projections and joints Therefore, proper selection, detailing of a waterproofing system and

sound workmanship should be in focus The finishes on wall and floors are the first barrier to

water infiltration and hence should be durable Additionally, the material should allow easy

maintenance in terms of resistance against stain, chipping, cracking and cleaning material or

cleaning methods Wet area contains many elements of sanitary-plumbing system Efficient

plumbing design not only reduces number of penetrations, but also allows adequate space for

regular inspection and cleaning Considering these issues, the maintainability factors of wet

area were identified (Table 5.3) The detailed guidelines are presented in Appendix C.3 and

background information of defects is documented in Table 4.8

Table 5.3 Maintainability factors for wet area and their RWs

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW Design phase

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW Construction phase

c33 Slab casting C2, C4 C1, C3, C6, C11,C12 0.352 0.023

c34 Embedding of

services

C4, C15, C16 C11 0.547 0.035 Floor

5.6 Maintainability scoring for roof

Among various forms of roofing systems, accessible type reinforced concrete flat roof is most

commonly used in commercial buildings of Singapore Flat roof is defined as a roof with

pitch less than 5° to horizontal (SS CP 82) The regular components are: deck, waterproofing,

insulation, and protective surface (Baskaran, 1996) Various types of roof have various

arrangements of these elements selected according to the usage and exposure condition Flat

roofs can either be (1) inaccessible i.e access for cleaning and repair only or (2) accessible

i.e can host various human activities (BS 6399-3) Flat roofs not only should carry both the

live load and abundant rainfall, but also withstand the exposure condition Roof components should be compatible with each other to form a durable structure together Except the drainage or finished surface other roof elements are beyond the scope of regular inspection or maintenance and hence should be constructed well to avoid costly re-roofing Maintainability aspects of roof are described in Table 5.4 along with detailed guidelines in Appendix C.4 Relative weights (RW) of the factors are determined based on the associated defect criticality (Table 4.10) A total of 64 factors were identified out of which 35, 21, and rest were dedicated

to design, construction and maintenance respectively

Table 5.4 Maintainability factors for roof and their RWs

Component Maintainability factors Critical defect

mitigated

Non-critical defects mitigated Wt RW Design

Component Maintainability factors Critical defect

mitigated

Non-critical defects mitigated Wt RW Construction phase

d42 Finished surface D3, D5, D21 D2,D4, D6,D9, D11,D12 0.817 0.049 d43 Storage & prep D6, D9, D10, D11 0.117 0.007 d44 LAM apply D6, D9, D12, D14, D15 0.141 0.008 d45 Membrane

application

D21 D6, D9, D10, D13, D14,

D15

0.429 0.025 d46 Joint & penetration D3, D5,,D21 D4,D11, D14,D24 0.757 0.045

d56 Testing D5, D21, D22 D24 0.745 0.044

Construction subtotal 0.320

Maintenance phase

d57 Check D3, D19 D2, D7, D8, D20 0.425 0.025 Roof

5.7 Maintainability scoring for sanitary-plumbing system

Main functions of sanitary and plumbing system are hot and cold water supply along with removal of liquid waste or liquid borne solid waste (Heerwagen, 2004) Health and convenience are the two major main focus of this system Outbreak of SARS brought into limelight the importance of a well-maintained, hygienic sanitary-plumbing system as the disease spreads as a result of infected water droplets (Watts, 2003) Additionally in Singapore due to limited land of water catchments, efficiency of the system has been emphasized by Ministry of Environment and Water Resources (MEWR) As sanitary-plumbing system consists of different types of pipes (cold & hot water supply and waste disposal) passing in

the close proximity and often concealed from direct access, chances are higher for a defect to remain unnoticed unless the damage is significant Hence it is important that design, construction and maintenance guidelines are aimed for a zero-defect scenario

Maintainability guidelines (Appendix C.5) was developed for the major components, namely, water supply system, storage tank, pumps (general and sewage), sanitary appliances, sanitary piping, sewage ejector & solid diverter tank and finally sewer drains There are total 93 maintainability factors (49 for design, 25 for construction and rest for maintenance and external factors) The summary of the factors and their relative weights are presented in Table 5.5 Details of the defects and their criticality index can be referred to Table 4.12

Table 5.5 Maintainability factors for sanitary-plumbing system and their RWs

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW Design phase

e15 Cleaning facility E3, E5 0.446 0.012

Cold water

supply

piping

Hot water

supply

e20 Temp control E3, E11, E12 0.628 0.016

e25 Circulation E3 E9, E10 0.226 0.006 Storage tank

Pumps

e34 Floor waste & trap E19, E20 0.712 0.018

Sanitary

appliances

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW

e36 System select &

venting

e37 Pipe designE18 E23, E24, E25 0.691 0.018

Sanitary

pipe

e41 Pit design E27, E28, E29 0.478 0.012

Sewage

ejector &

solid

diverter tank

e50 Storage & handling E2, E11, E12 0.728 0.019

e51 Laying E3, E5, E6, E14 E7 0.898 0.023

e52 Joint & penetration E3, E4, E5 E9 0.744 0.019

Water

supply

e56 Achieve water

Sanitary

appliance

e66 Storage & handling E23, E24 0.515 0.013

Sanitary

pipe

Swg eject &

Hot water

Storage

Pumps

e84 Check E19, E21, E24, E27 E22 1.103 0.028

Sanitary

appliances

Sanitary

Component Maintainability factors Critical defect

mitigated

Non-critical defect mitigated Wt RW

Sewer

Ext factor e93 Building age E2, E12, E22 0.659 0.017

Maintenance subtotal 0.234

Total 1.000

5.8 Maintainability scoring for HVAC system

Among various types, central chilled water HVAC system is commonly used in commercial buildings of Singapore Operation and maintenance cost for HVAC is very high as it consumes more than 50% of the total energy (Lu, 2005) Apart from energy efficiency the standard of internal environmental quality should be met in order to enhance the productivity and ensure user comfort Hence design issues are focussed on these two parameters

Table 5.6 Maintainability factors for HVAC system and their RWs

Component Maintainability factors Critical defect mitigated Non- crit

defects Wt RW Design phase

Component Maintainability factors Critical defect mitigated Non- crit

Component Maintainability factors Critical defect mitigated Non- crit

distribution f91 Clean F32, F34, F39 0.692 0.027 Ext factor f92 Building age F2, F5, F12, F19, F23 F3 0.870 0.034

Maintenance subtotal 0.383

Total 1.000

Air handling unit (AHU) and fan coil units (FCU) are two options for circulating, cleaning, heating / cooling, (de)humidifying and mixing of air A choice or a combination of both controls the efficiency of the system A chiller comprises of evaporator, condenser, compressor and various control devices for cooling the refrigerant, while at cooling tower the water from chiller rejects heat to the atmosphere All these thermodynamic cycle should work

in tandem A HVAC system if not properly maintained can raise alarm duet to Legionella and mould growth in cooling tower and return air plenum respectively The maintainability factors for HVAC system are discussed briefly in Table 5.6 and illustrated in Appendix C.6 A total

of 92 factors were developed Design, construction and maintenance carry 54, 12, 26 % of relative weights derived from associated defect criticality (Table 4.14)

5.9 Maintainability scoring for elevators

Elevators or lifts form the vertical transportation spine of high-rise buildings It is a permanent lifting equipment serving two or more landing levels, including a car for transportation of passengers and/or other loads, running between rigid guide rails, either vertical or inclined to the vertical by less than 15° (Janovsky, 1999) Hydraulic elevators are generally used for buildings with less than seven storeys high while traction elevators are used in taller building For this research, traction elevators were considered which is more common in Singapore commercial buildings, complex than hydraulic type and hence difficult

to maintain Apart from the safety against free fall, the main issue for regular function is control Machine room equipments control the speed of travel, response against a call and

need to work in tandem with the other components attached to the car, car door and lobby

door Though machine room, hoistway or pit is not in direct contact with building occupants,

these areas and the equipment they house have crucial role in travel performance of elevator

and safety of maintenance personnel These aspects were considered for deriving the

maintainability factors (Table 5.7) Corresponding detailed guidelines and defect data can be

obtained from Appendix C.7 and Table 4.16 respectively A total of 84 checklist factors were

formulated for the major components, out of which design involves 44 factors, construction

involves 10 and rest are attributed to maintenance

Table 5.7 Maintainability factors for elevator and their RWs

Compone

nt Maintainability factors

Critical defect mitigated

Non-critical defect mitigated Wt RW Design phase

nt Maintainability factors

Critical defect mitigated

Non-critical defect mitigated Wt RW

1.140 0.051 g60 Clean & lubricate G15 G4, G17, G18 0.286 0.013

Traction

machine

& motor

g61 Check & record G15 G1 0.222 0.010

g62 Clean & lubricate G19 G5, G20 0.204 0.009

Brake

assembly g63 Check &record G5, G20 0.089 0.004

g64 Clean & lubricate G22 G21 0.129 0.006

Governor

machine g65 Check & record G22 0.102 0.005

g66 clean & lubricate G27, G28 G25 0.219 0.010

g79 Clean & lubricate G57 G8, G59 0.363 0.016

g80 Check & record G55, G58 G4, G59 0.599 0.027

Total 1.000

5.10 Maintainability scoring for electrical system

Electrical system in a commercial building is vast as it forms the backbone of the whole M&E services It differs significantly based on the power rating For a typical commercial building, the electrical energy is drawn from service provider at the main services entrance and goes through a series of voltage step down processes and gets distributed to the end utilization devices The usual components are: transformer, wiring (cable and busbars), distribution equipment, protective device, lighting, standby / emergency power supply, grounding and lightning protection system (SS CP 97-2) Schematically electrical system for a building is illustrated in Fig 5.1 From maintainability point of view the functions of these elements were considered (IEEE & ANSI, 1991; Halt, 1994; Holt, 1998; Miller & NFPA, 1999) in order to define the maintainability factors (Table 5.8) For detailed discussion Appendix C.8 should be referred along with Table 4.18 for defect classification A total 107 maintainability factors can

be grouped under design (57), construction (23) and maintenance (27) phases

Fig 5.1 Schematic diagram of electrical system of commercial building (Siemens, 2008)

Table 5.8 Maintainability factors for electrical system and their RWs

Component Maintainability

factors

Critical defect mitigated

Non-critical defect mitigated Wt RW Design phase

Component Maintainability

factors

Critical defect mitigated

Non-critical defect mitigated Wt RW

Component Maintainability

factors

Critical defect mitigated

Non-critical defect mitigated Wt RW Construction phase

5.11 Maintainability scoring for fire protection system

The prime objective of fire protection system is to save life and property by prevention, detection, control and suppression of fire For this purpose the main components falling under fire protection system are detector, communicator, sprinkler, hydrant, hose and portable fire extinguishers For a safe and panic-free evacuation the escape route should be fortified with fire door, standby power supply and pressurization system for smoke control Active fire protection system is covered by the scope of this research As these elements are used only during emergency, they bear a high probability of being overlooked unless the mishap occurs Design and installation of the system is strictly standardized by the statutory body Hence the maintainability guidelines (Table 5.9, Appendix C.9) are primarily based on local SCDF code (2007) To mitigate defects (Table 4.20) in fire protection system, a total of 90 factors were identified: 39, 20 and 31 are related to design, construction and maintenance respectively

Table 5.9 Maintainability factors for fire protection system and their RWs

Component Maintainability factors Critical defect

mitigated

Non-critical dfct

mitigated Wt RW Design phase

Component Maintainability factors Critical defect

j75 Check J21,J22,J23, J24,J25 0.149 0.009 Valves

j77 Check J26, J28, J29 J27, J30 0.679 0.040 j78 Replace / recharge J28, J29 0.374 0.022

Component Maintainability factors Critical defect

5.12 Overview of scoring of all subsystems

For the entire building a total of 731 guidelines were proposed (Table 5.10) Design related guidelines were undoubtedly maximum in number (52.26% in average) But construction had 75.6% higher emphasis for C&A subsystems while maintenance guidelines had 104.4% greater influence for M&E subsystems

Table 5.10 No of maintainability factors

Design Construction Maintenance System Sub-system

Chapter 6 Comprehensive Maintainability Scoring System 6.1 Introduction

Comprehensive maintainability scoring system or COMASS is the third phase of this research addressing the third objective and together with criticality analysis provides answer for the second research question It is the integration of nine major building subsystems Their relative weights (more precisely global weights or GW w.r.t overall building maintainability) were generated by analyzing the AHP (analytic hierarchy process) questionnaires The local weights (LW) at each level of hierarchy reflected the grading logic of the decision makers (DM) and were compared with the subjective knowledge shared by them during interview

6.2 Data analysis

Data analysis in AHP includes: data processing, dealing with inconsistency, aggregation of results into group decision making (GDM) and derivation of GWs Their methodology has been discussed in detail in Section 3.4.6 It also illustrates how aggregation of individual priorities (AIP) using geometric mean method (GMM) was found most suitable for this research Fig 6.1 shows the demography of the DMs who provided valid responses

Fig 6.1 Demography of the respondents

6.2.1 Data processing in ExpertChoice (EC 11.5)

Compared to other available softwares, EC works in tandem with AHP (Bodin & Gass, 2003)

In EC, it is easy to construct the hierarchy and instantaneously check the important information such as inconsistency ratio (IR) and LW First the master template of hierarchy

a Hierarchy for building maintainability

b Auto generation of Inconsistency Ratio (IR)

c Local weights (LW) generated in EC

Fig 6.2 Data processing in Expert Choice (EC)

was created (Fig 6.2a) in which judgments obtained from individual experts were keyed in Instantly the IR appeared at the bottom left corner of the matrix (Fig 6.2b) Upon the completion of data-entry, the LWs were automatically reflected (Fig 6.2c) A sample print- out of such report is presented in Appendix D.1

6.2.2 Dealing with inconsistency

After considering the criticism and possible improvement of inconsistency, it was decided to explore data falling beyond Saaty’s 10% cut-off limit The idea was to include acceptable data

to maximum possible extent so that large amount of information was not lost Apostolou and Hassell (1993) have reported such ‘inconsistent’ data does not vary significantly from the data with ≤ 10% IR

There were 37 sets of complete questionnaires Previously no significant difference among groups from various work profile was found (Section 4.2.2) Hence instead of work profile,

IR was used for grouping the questionnaires They were classified into three major categories, namely (1) IR ≤ 10% or 0.1; (2) IR ≤20% and (3) IR ≤ 30% and the remaining responses were

discarded Under each, two sub-categories were defined as by DM or respondent and by comparison matrix (Table 6.1) This process allowed six ways of analyzing the data (Table 6.2) whereby patterns could be observed to select the best data set for further analysis Individual priorities by DMs who provided all matrixes with IR ≤ 10%, 20% and 30% were

grouped under Dataset 1, 2 and 3 respectively In the other method, irrespective of source, the matrixes with IR ≤ 10%, 20% and 30% were grouped under Dataset 4, 5 and 6 respectively It

should be noted that by definition, Dataset 3 and 6 were the same

Table 6.1 Six classifications of data set

Classification type IR< 10% IR< 20% IR< 30%

Table 6.2 Illustrative example of IR chart

Inconsistency ratio (IR) Matrix

I1 (5⋅5) Among 5 M&E subsystems for low rise bldg

I2 (5⋅5) Among 5 M&E subsystems for mid rise bldg

I3 (5⋅5) Among 5 M&E subsystems for high rise bldg

It was found that among 37 completed survey questionnaires, 11, 5 and 7 responses showed

IR of 10%, 10-20% and 20-30% respectively Rest 14 responses were discarded for further analysis Finally after addition, the groups, namely, IR≤ 10%, IR≤20% and IR≤30% were

having 11, 16 and 23 sets of data respectively It was observed that with experience of the DMs, their consistency (IR ≤ 10%) of answer improved almost exponentially (Fig 6.3)

1

2

3

5 5

6

0 1 2 3 4 5 6 7 8 9

5-10 Yrs 10-15 Yrs 15-20 Yrs > 20 Yrs.

Fig 6.3 Comparison of years of experience and consistency of results

6.2.3 Selection of threshold limit of IR and the best dataset

As mentioned earlier this particular AHP model contains mutual exclusiveness of location and height Hence for each of 12 location-height combinations, there can be (12x6) = 72 sets

of GW for nine subsystems Before proceeding further it was essential to set the threshold limit for IR and select the best dataset among six accordingly For this purpose, the mutual exclusiveness of locations and heights were set aside temporarily and GWs for 9 subsystems from 6 datasets were derived in a usual manner For selection, three tests were conducted:

● Preliminary test to visualize graphically any abrupt change in GW with rise in IR

● Statistically test any such difference by using Wilcoxon signed rank test for two datasets at

a time and Friedman test for all six datasets

● Checking for any rank reversals

6.2.3.1 Preliminary test result

For various datasets, no visually noticeable different trend was observed (Fig 6.4) With every 10% increase in IR, change in GW among six datasets was less then 2% Hence no firm conclusion could be drawn from the graphical comparison and further testing was required

Set 5

Fig 6.4 GW for nine –subsystems

6.2.3.2 Statistical test result

Such aggregated data may not be normally distributed (Apostolou & Hassell, 1993) Hence Wilcoxon signed rank test was used This is a nonparametric test equivalent to the parametric t-test and is useful when data does not meet the assumptions and requirements of the t-test (Siegel & Castellan, 1988) A confidence interval of 95% (α= 0.05) was used to determine whether the null hypothesis of no significant difference could be rejected There was no significant difference (p < 0.05) found between two consecutive datasets (Appendix D.2) However difference (p=0.016) among all six datasets were found in Freidman test indicating that all datasets are not of equal quality Hence at this point also the threshold point of IR could not be set and the data was treated for next step of analysis through rank reversal

6.2.3.3 Rank reversal

The main causes of rank reversal are inconsistency in pair-wise comparison or change in hierarchical structure or both (Tamura et al, 1998) In this case effect of inconsistency was checked by comparing ranks for all the six datasets (Table 6.3) It was observed that rank reversals occurred when IR increased beyond 10% As AHP assigns relative weights or helps

to select the best alternative by ranking, phenomena of rank reversal can not be ignored

Table 6.3 Rank reversal with change in IR

Classification by respondent Classification by matrix Dataset 1 Dataset 2 Dataset 3 Dataset 4 Dataset 5 Dataset 6 Sub-systems

Basement 0.028 9 0.030 9 0.036 9 0.034 9 0.035 9 0.036 9 Facade 0.119 3 0.110 3 0.098 4* 0.112 g 4 0.103 4 0.098 4 Wet area 0.077 5 0.067 7* 0.068 6* 0.088 5 0.073 6* 0.068 6 Roof 0.060 7 0.046 8* 0.043 8 0.053 g 8 0.043 8 0.043 8 San-plumb 0.060 8 0.070 6* 0.064 7* 0.067 g 7 0.061 7 0.064 7 HVAC 0.278 1 0.271 1 0.281 1 0.249 1 0.279 1 0.281 1 Elevator 0.201 2 0.221 2 0.210 2 0.209 2 0.219 2 0.210 2 Electrical 0.112 4 0.110 4 0.119 3* 0.117 g 3 0.110 3 0.119 3 Fire prot 0.064 6 0.076 5* 0.081 5 0.071 6 0.077 5* 0.081 5 Note: R=Rank * Any consecutive rank reversal g Rank reversal between Dataset 1 and 4

6.2.3.4 Selection between Dataset 1 and 4

At this stage, it was affirmed that IR> 10% caused rank reversal and data beyond that limit may not be very consistent It was checked if the threshold can be set at a more stringent value of IR ≤ 5%, but not a single set of response could achieve such standard Hence it was concluded that Saaty’s recommendation of IR ≤ 10% was the optimum threshold It led to selection between Dataset 1 and 4 – both with IR ≤ 10%, but classified by person and by cell respectively

Rank reversal was also identified between Dataset 1 and 4 in spite of the same IR (≤10%)

This could be explained using the fact that Dataset 4 contains many comparison matrixes from DMs who might have failed to provide consistent judgment for the whole hierarchy The inconsistency of all such matrixes indirectly affected the reliability of result in group-decision making and rank reversal occurred Hence at this point Dataset 1 was considered the most suitable and henceforth was used for derivation of GW and drawing conclusions

LWA , = local weight of m-th alternatives w.r.t p-th criteria

Apart from this general equation an additional normalisation was required to handle the mutually exclusive options of location and height in hierarchy model (Fig 3.3) For each possible combination of location and height, the relative weights were re-normalized by ignoring all non-applicable options Hence four options for locations and three options of height yielded (4 x 3) =12 sets of re-normalized weighted priority As for a particular building, 1 out of the 12 values will be used; they were named as effective LW Table 6.4 illustrates the normalized results

Table 6.4 Renormalized LW of for all zone-height combinations

Normalized LW for options of zone & height Weighted Priority (Re-normalized LW) Effective LW Zone (z) 0.199 Height (h) 0.801 Zone Height Zone (z’) Height (h’)

L h R z

R z

×+

×

×

)()(768

L h R z

L h

×+

generic rules of AHP Table 6.5 illustrates the case of first combination of location=

residential zone with height=low The equations are self explanatory In this case weighted

priority of location and height are z’= 0.232 and h’=0.768 respectively (from Table 6.4) GWs

for all 12 combinations are presented in Table 6.6 The calculations can be obtained from

Section 3 of Appendix D

Table 6.5 Derivation of GW for location = residential zone and height = low

Zone / location priority z’ = 0.232 Height priority h’ = 0.768

Priority

of sub- system

Wghtd

priority

Priority

of system

Priority

of subsystem

Wghtd

priority

GW of subsystem w.r.t both zone &

Note: R = Residential, C = Commercial, I = Industrial, Mx = Mixed,

L = Low, M = Medium and H = High

6.3 Results and discussion

COMASS is the integration of building subsystems based on both objective and subjective parameters of maintainability GWs for nine subsystems can represent meaningfully their relative contribution in overall maintainability if their numerical values can be synchronized with the subjective thinking of DMs For this purpose the sources of these numbers i.e the intermediate LWs which are directly provided by the DMs were compared with their subjective interviews in the following sections

6.3.1 Influence of criteria: location and height

In general, DMs reported that, compared to height, location has less influence unless it hinders the availability and accessibility for services The M&E system which is in general costlier to maintain usually remains unaffected by a building’s location The same logic is more or less applicable to internal elements namely, basement and wet area Numerical values

of LWs 0.199 and 0.801 for location and height respectively reaffirmed this fact

Moreover when GWs of subsystems were compared for various locations keeping the height fixed and vice versa (Fig 6.5 and 6.6), it was observed that GW for different locations with the same height almost overlapped each other while it is not true for the other case Such results tallied with the comments brought forth by the experts whereby maintainability relies less on external factor such as aesthetics and more on functional performance Hence if a building changes its location, there will be no significant change in maintainability profile as the function and services remain same

6.3.2 Influence of sub-criteria: location

However, when experts had to compare among various locations, it was noted that location controls maintainability to a certain extent in terms of working hours for maintenance, accessibility and user expectation Commercial buildings in residential, commercial and mixed zones, as opposed to industrial area, have time restrictions for maintenance works to be

carried out in order to minimize disturbance to the neighbourhood Secondly, location affects accessibility for maintenance For example, in CBD (central business district), space constraints may restrict the installation of gondola system and result in a costlier alternative

Fig 6.5 Variation of GWs with location, keeping the height constant

Location = residential zone

Location = mixed zone Height = variable

Finally, user expectation varies with location (Fig 6.7a) Shopping malls and community centres of residential area mostly cater for the regular customers from the neighbourhood Hence compared to CBD, maintenance priority can be kept lower Commercial areas are dotted with government offices, MNC headquarters, big shopping malls and five star hotels

In order to retain the global tenants and reflect the image of the country to locals and foreigners, buildings are kept at their best condition Industrial area is infamous for pollutants and the users are mostly skilled or unskilled labourers or people involved in manufacturing industry, hence user expectation is low However, the emphasis lies in between the first two locations so as to cope up with the detrimental environmental conditions Mixed area has very high degree of uncertainty To accommodate the requirements of all the three above- mentioned options, the standard becomes a superset of all requirements hence the stress is highest among all

Low Medium High

Fig 6.7 RWs of various (a) locations and (b) heights

6.3.3 Influence of sub-criteria: height

Exponential hike in maintenance need with height was reported (Fig 6.7b) as M&E systems become complex rather than same simple facility of low-rise building multiplied by the number of floors (Section 3.4.2.4) Additionally greater height means more difficult access for equipment For example, a transformer located at load centre may be at intermediate floor Its travel route for replacement should be heavy duty and can cause disruptions to other services leading to a high cost Usually, maintenance of a tower is significantly higher than a low-rise sprawling building of same gross floor area

6.3.4 Relative importance of C&A and M&E systems

Irrespective of location and height, M&E systems were found to have higher emphasis (Fig 6.8) By comparing the objective results with the experts’ comments, reasons for such trend was noted These arguments exactly matched with the comparison of both systems as obtained from literature (Section 2.2):

● C&A systems usually last for minimum 10 years due to avail of warranty E.g water

proofing membrane

● C&A systems, except wet area do not have much direct impact on end-users as compared

to M&E systems

● M&E systems are complex and have higher number of components compared to C&A

systems Hence their frequency of break down is also higher HVAC is the most complex and expensive among all

Fig 6.8 RWs of C&A and M&E systems for various locations and heights

6.3.4.1 Influence of location on C&A and M&E systems

Keeping height constant, the locations were compared (Fig 6.9a) RW of C&A and M&E system remained somewhat unchanged for same location Maintenance of external elements

in residential zone is difficult as working at night is prohibited by law and working in day affects the commercial activities of the building As a result, C&A elements (except wet area) are prioritized in this zone Business parks mostly located in industrial area are usually equipped with cargo lifts and require an uninterrupted power and water supply Poor

performance of M&E systems will heavily affect the production and hence it explains why M&E systems in that zone require higher maintainability Another reason of lower importance of C&A systems in industrial zone might be the fact that users prioritise the facilities within the building (usually their work place) that has direct relation with productivity over the aesthetics i.e on facade – the prime member of C&A system

Fig 6.9 RW of C&A and M&E systems with variance in (a) location and (b) height

6.3.4.2 Influence of height on C&A and M&E systems

Keeping the location constant when the height was varied, a uniform trend was obtained for

RW of C&A and M&E systems (Fig 6.9b) With height, importance of M&E constantly increased as the overall maintenance (Fig 6.7b), but the curve is flatter It suggests, though both M&E and C&A contribute to the overall maintainability the former controls the trend This fact is justifiable because with height, the complexity and intensity of M&E systems

increase significantly The effect of height on C&A might not be as much as M&E system, for instance, the roof area and depth of basement would be the same regardless of the building height The C&A components affected by height would be the facade and wet area, which would increase in capacity rather than complexity

6.3.5 Relative importance of all nine subsystems

Unanimously HAVC was selected as the most important contributing element in maintainability Second position was assigned to elevator system (in 10 cases) and to facade (only two cases –both for low rise building) The third most influential subsystem was facade (8 out of 10) though electrical system was counted twice in third rank for medium rise and high-rise buildings in industrial area (Table 6.7)

Building heights

Fig 6.10 Relative weights of C&A systems for various locations and heights

Building heights

Fig 6.11 Relative weights of M&E systems for various locations and heights

Influence of criteria and sub-criteria on the ranking of subsystems was analysed Facade, among all C&A systems, got the highest emphasis under various locations and heights as it

needs to perform in terms of noise attenuation and water resistance (Fig 6.10) Moreover, cleaning could be cumbersome when it is not easily accessible A gradual increase of emphasis therefore occurs with increase in height In general, facade had a high rank in commercial area as in CBD especially aesthetics is important for business growth Also there lie many administrative buildings whose facade gets a higher emphasis to reflect the country’s image Among all five M&E systems, HVAC showed high prominence as it has many components and next was elevator system (Fig 6.11) These rankings are discussed in details

in the following sections

Table 6.7 Ranking of nine subsystems for all locations and heights

Height: low (L) Height: Medium (M) Height: High (H) Area

Subsystem GW Subsystem GW Subsystem GW Rank

Residential (R)

Commercial (C)

Industrial (I)

Mixed (Mx)

6.3.5.1 Rank 1: HVAC system

HVAC system comprising of many complex and expensive components, controls the indoor air quality, in turn the productivity For instance, chiller in HVAC system needs to stop operation during the annual maintenance; which would incur high cost In addition, the professional fee to maintain the chiller is very high Another component of the HVAC system is the cooling tower where Legionnaires’ disease needs to be taken care through proper treatment of water Furthermore, the filters in the fan coil units need to be cleaned in order to ensure the quality of air Also, in the event when the water outlet is not cleared, jelly forming could happen In tropical climate mould growth in ducts is a serious issue As such HVAC requires much maintenance attention and hence being the most expensive system for all zones and heights Probably presence of HVAC in M&E group increased its importance over C&A system

6.3.5.2 Rank 2: Elevators

Elevator system was considered as the second most influential contributor to building maintainability As commented by the experts, faults rarely occur in elevators, but a breakdown can affect business heavily and an accident can be fatal This therefore explains the trend for its high importance Interestingly, there were two exceptions that low-rise buildings in residential and commercial zone had elevator system at third position Such exception is not unlikely as those buildings might have neither cargo lifts like industrial areas nor high speed elevator of high rises Probably escalator serves the main purpose of vertical transportation in those two building groups Commercial buildings in industrial zones are mostly business parks and have production related activities, hence large capacity and unquestionable reliability of elevator system is a mandate This is the reason why buildings in industrial area impose high weightage on elevators even for a low-rise building