You are on page 1of 12

Data Analysis of Inspection, Diagnosis,

and Rehabilitation of Flat Roofs


J. Conceição 1; B. Poça 2; J. de Brito 3; I. Flores-Colen 4; and A. Castelo 5

Abstract: This paper presents the data collected from the inspection of 105 Portuguese flat roofs. The inspections were based on a system of
inspection, diagnosis, and rehabilitation created and validated previously for flat roofs. The use of this system increases the objectivity and
effectiveness of an inspection and supports the inspection of flat roofs. The use of flat roofs is gaining momentum in the construction sector
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

worldwide, and there is little information concerning the correlation among the most common types of anomalies, their most probable causes,
applicable diagnosis tests, and the most suitable rehabilitation techniques. The information gathered during the field inspection of 105 flat
roofs allowed the validation of the proposed inspection system and the determination of the most common features that can occur in this type
of roofing system, as well as their correlation with the materials, characteristics, location, and position on the roofing system. The data
analysis of the information gathered aims at helping designers, end users, and inspectors as a decision-basis tool regarding this roofing
system, and is presented in various charts in order to facilitate their interpretation and help identify the most conditioning parameters
in the performance and durability of this type of nonstructural system. DOI: 10.1061/(ASCE)CF.1943-5509.0001252. © 2018 American
Society of Civil Engineers.
Author keywords: Flat roofs; Anomalies; Causes; Diagnosis; Rehabilitation; Data analysis.

Introduction Despite the existence of new and better-performing waterproof-


ing materials, the poor training of the waterproofing layer appliers
A roof consists of a set of load-bearing and protective elements is a constant source of problems. Due to these deficiencies, there
intended to protect the interior of a building from meteorological are water seepages that affect the underlying layers of the top floor,
phenomena such as rain, heat, cold, wind, and snow, sheltering the causing damages that translate into costs, not only from the repair
building’s structure and ensuring high standards of living condi- but also due to the fact that the referred areas become unusable.
tions (Gomes 1968). In order to reduce the lifetime costs of the roof, it is fundamental
The use of flat roofs, considered to have a maximum slope of 7°, to have a solid knowledge of state-of-the-art technology, execution
is gaining momentum throughout the world, and every region has techniques, good implementation principles, and appropriate diagno-
its own tradition or preference in terms of the materials used. In sis and repair methods, which will increase the roof’s durability and
Portugal, according to data from the National Institute of Statistics service life and slow down the appearance of degradation signs.
(INE), the ratio of flat roofs to pitched roofs has been steadily Also, the studied maintenance/rehabilitation techniques allow a nor-
increasing, and between 2006 and 2011, the use of mixed and flat malization of the maintenance procedures and contribute to a cost
roofs represented 12.5% of those used in Portugal. Therefore, and reduction because more efficient interventions can be carried out.
given the increased importance of these structures, it was important The inspection and diagnosis system (Conceição et al. 2017)
to develop a system that supports the repair of anomalous situations validated in the paper is innovative and scientifically corroborated
in roof operations (INE 2012). by other studies. The system was created and validated with data
collected from the inspection of 105 flat roofs throughout mainland
Portugal and is based on a methodology that has been applied and
1
Lieutenant, Portuguese Army, Military Academy, Rua Gomes Freire, validated by the same research team to other nonstructural building
Lisbon 1169-203, Portugal. Email: joao.melro@tecnico.ulisboa.pt materials: waterproofing systems of flat roofs (Walter et al. 2005),
2
Lieutenant, Portuguese Army, Military Academy, Rua Gomes Freire,
External Thermal Insulation Composite System (ETICS) (Amaro
Lisbon 1169-203, Portugal. Email: bruno.poca@tecnico.ulisboa.pt
3
Full Professor, Civil Engineering Research and Innovation for
et al. 2013, 2014), pitched roofs (Garcez et al. 2012a, b), painted
Sustainability, Dept. of Civil Engineering, Architecture and Georresources, surfaces (Pires et al. 2015a, b), and window and door frames
Instituto Superior Técnico, Univ. of Lisbon, Ave. Rovisco Pais, Lisbon (Santos et al. 2017a, b).
1049-001, Portugal (corresponding author). Email: jb@civil.ist.utl.pt
4
Associate Professor, Civil Engineering Research and Innovation for
Sustainability, Dept. of Civil Engineering, Architecture and Georresources, Inspection Plan for Flat Roofs
Instituto Superior Técnico, Univ. of Lisbon, Ave. Rovisco Pais, Lisbon
1049-001, Portugal. Email: ines.flores.colen@tecnico.ulisboa.pt The inspections intend to identify the causes and precondition sig-
5
Ph.D. Student, Civil Engineering Research and Innovation for nals of each anomaly, allowing, during the use phase, an increased
Sustainability, Instituto Superior Técnico, Univ. of Lisbon, Ave. Rovisco
ability to detect the need for intervention and thus reduce the risk
Pais, Lisbon 1049-001, Portugal. Email: andre.castelo@tecnico.ulisboa.pt
Note. This manuscript was submitted on February 2, 2018; approved on of unexpected anomalies. For the implementation of a proper diag-
August 6, 2018; published online on November 27, 2018. Discussion per- nosis by the inspection team, inspection and validation sheets
iod open until April 27, 2019; separate discussions must be submitted for should be prepared, as well as the necessary material for the
individual papers. This paper is part of the Journal of Performance of inspection (binoculars, markers, rulers, measuring equipment, pho-
Constructed Facilities, © ASCE, ISSN 0887-3828. tographic equipment, and other necessary accessories).

© ASCE 04018100-1 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Table 1. Proposed classification system for anomalies in flat roofs mapping determines the incidence of each anomaly, as well as their
Code Anomaly severity, allowing the creation of their repair project and of other
constructive elements (Silvestre and de Brito 2007). Also, in terms
General
of suitable repairs, the manufacturers should provide detailed guid-
A-G1 Surface wear
A-G2 Fracture/failure
ance and specific requirements regarding installation and use of
A-G3 Detachment/peeling their membrane and flashing products.
A-G4 Creasing/bulging With the inspection campaign, it was possible to validate the
A-G5 Cracking proposed classification systems and the corresponding correlation
A-G6 Puncture matrices. Based on the information collected, a data analysis was
A-G7 Absence/inadequate layer positioning made.
A-G8 Debris accumulation
A-G9 Inadequate slope/ponding
A-G10 Biological growth Characterization of the Materials, Roofing Systems,
A-G11 Corrosion Sample, and Elements
A-G12 Moisture stains of condensation/leaks
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

Specific Characterization of the Materials and Roofing Systems


A-S1 Inadequate expansion joint design
A-S2 Inadequate downspout design In order to take into account the main aspects to consider in the
A-S3 Inadequate overflow drainage tube design design and execution of flat roofs, the latter can be classified
A-S4 Inadequate gutter design according to accessibility, type of superficial protection, type of
A-S5 Inadequate lap-joint design waterproofing membrane, type of constructive system, slope, and
A-S6 Defects in the fastenings supporting structure.
A-S7 Defects in the coping When applying a flat roof insulation system, several aspects
A-S8 Defects in the tail ends should be taken into consideration to maximize the life span of
the applied system. One of these aspects is the protection of the
insulation system against environmental agents and surface wear.
The inspection sheets condense the information gathered in the The protective layer can be categorized in three classes: unpro-
inspection of flat roofs, including the building characteristics and tected, light protection, and heavy protection, as exemplified in
other relevant information necessary for a proper characterization Fig. 1. The unprotected class has no type of protection over the
of the anomalies. Thus, the inspection sheets should include the waterproofing membrane. In the light protection class, the protec-
following fields: for general information, a header with the inspec- tion is usually applied at the production stage and contains sand or
tion record number, date of inspection, person in charge, and gravel encrusted in the superficial layer of the exposed waterproof-
purpose of the inspection; for each building, the location, type of ing membrane. In the heavy protection class, the type of protection
to apply depends on the accessibility of the roof. Light concrete and
dominant use, later interventions, number of floors above ground,
small concrete slabs are the more usual types of heavy protection
type of environment, and proximity to the sea; for each flat roof,
solutions.
whether there is access to the interior, the roof’s total area and
The type of waterproof membrane can classify a flat roof system
slope, type of insulation system, thermal insulation material, type
as traditional or nontraditional, depending on the material applied.
of use (e.g., inaccessible, terrace, garden), type of protection, exist-
The most common types of waterproofing membrane materials
ence of mechanical fasteners, existence of circulation paths, and
found during the inspections were, as exemplified in Fig. 1, liquid
existing singularities (chimneys, pipes, downspouts); for any on-
membranes, PVC membranes, and bituminous membranes. The
going maintenance operations, implemented type and frequency
liquid membrane is an adherent membrane that consists on the ap-
of inspections, type and frequency of operations, characteristics
plication (usually through painting) of a liquid material on the
of operations, date of execution, technique used and its materials,
thermal insulation or other waterproofing membranes and usually
existing means of access to the building roof for the realization of
requires superficial protection. The PVC membrane is a nonadherent
inspection/intervention (e.g., ladders, skylights), and registry of membrane that is normally mechanically fixed to the underlying
other observations of inspections. layers of the flat roof system. The bituminous membrane is the most
In turn, the validation sheets are designed to validate the inspec- commonly used type of membrane and can be used in adherent
tion system and the proposed diagnosis, so for each inspected roof, (through heat) and nonadherent (through mechanical fixings)
a validation sheet was created. In these sheets, every anomaly de- systems.
tected was registered and analyzed in accordance with the anomaly The position of the waterproofing membrane divides systems
classification system presented in Table 1. The anomalies were into traditional when the waterproofing membranes are located over
also characterized in accordance with several parameters (where the thermal insulation and inverted when the waterproofing mem-
applicable): conditions for the anomaly progression; percentage branes are located under the thermal insulation, as shown in Fig. 1.
of affected area; aesthetic value of the affected areas; state of deg-
radation; humidity stains; occurrence of leaks; correct fixation ex-
ecution; propensity of the anomaly to affect in the short medium Characterizations of the Sample and Elements
term the state of the waterproofing coating; suitable materials; The 105 inspections performed were carried out in buildings
adequate tail-end execution; suitable coatings over the tail ends; from the Portuguese Army (26 buildings), the Portuguese Air Force
existence of the item in question; protruding joints; proper execu- (31), the National Republican Guard (GNR) (8), and civilian
tion of lap joints; sufficient overlap width; existence of drains; buildings (40).
convenient location; insufficient slope; and severity level. The inspected roofs are mostly (approximately 85%) located in
One of the recommended methods during inspections is map- the district of Lisbon. The remaining inspections were carried out
ping the anomalies in addition to inspection sheets, since these do in the Santarém (8%) and Braga (5%) districts.
not allow an accurate location definition of the anomalies. The Table 2 presents the full sample and element characteristics.

© ASCE 04018100-2 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Type of superficial protection

No protection Light protection Heavy protection


Type of waterproofing membrane material
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

Liquid membrane PVC membrane Bituminous membrane


Type of constructive system
Traditional system Inverted system

1 - supporting structure; 2 - vapour barrier; 3 - thermal insulation; 4 - geotextile; 5 - waterproofing membrane; 6 - protection layer

Fig. 1. Classification of flat-roof waterproofing systems according to the type of superficial protection, type of waterproofing membrane, and type of
constructive system.

Data Analysis each occurrence of the anomaly. As for A-G10, it was identified in
62% of the sample (65 flat roofs), to which 216 probable causes
The data analysis presented in this study follows the classification were allocated, resulting in an average of 3.3 causes for each
system used in the inspection, diagnosis, and repair system occurrence of the anomaly.
previously developed (Conceição et al. 2017). All anomalies, The causes of the A-G8 (debris accumulation) anomaly are
causes, diagnosis methods, and repair techniques are thus men- mainly from the C-U group (maintenance/use errors, responsible
tioned according to their designation in the system and the relevant for 60% of cases), mainly C-U3 (absent/inadequate inspection,
abbreviation. 21% of all causes of this anomaly), C-U1 (lack of cleaning debris
As shown in Table 1, the anomalies are classified into two in the current zone, 19%), C-U2 (accumulation/obstruction of
groups: A-G (general anomalies, comprising 12 anomalies) and debris in downspouts or gutter entrances, 12%), C-U5 (vandalism,
A-S (specific anomalies, comprising 8 anomalies). The most 9%), and C-A1 (wind, 17%) of the C-A group (19%).
common anomalies detected during the inspections are presented Biological growth (A-G10) mainly results from causes of the
in Fig. 2. Table 3 presents the classification system for the causes, C-A group (environmental actions, corresponding to 59% of all
which are divided into five groups: C-P (project/design errors, causes of this anomaly), including C-A2 (ultraviolet/solar radia-
comprising 11 causes), C-E (execution errors, comprising 12 tion, 12%), C-A3 (moisture, 25%), C-A4 (prolonged presence of
causes), C-A (environmental actions, comprising 7 causes), C-U water/rain, 6%), and C-A5 (vegetation, 16%).
(maintenance/use errors, comprising 5 causes), and C-M (external The more common anomalies are, in terms of frequency, A-G1
mechanical actions, comprising 2 causes). (surface wear), A-S2 (inadequate downspout design), and A-G4
(creasing/bulging), registering 50%, 48%, and 46%, respectively.
The group of causes with the highest frequency in A-G1 is C-A
Frequency Analysis of the Anomalies and Causes
(65%), particularly due to the contribution of C-A2 (solar radiation)
Fig. 3 shows the relative contribution of each anomaly to the total and C-A6 (natural aging), both detected in 30% of the cases of this
sample of anomalies detected in flat roofs. In turn, Fig. 4 shows the anomaly.
relative frequency of anomalies identified at least once in each of The groups of causes that are more related with the appearance
the 105 inspections. Debris accumulation (A-G8) and biological of A-S2 are C-E (execution errors, 48%) and C-P (project/design
growth (A-G10) were the most common anomalies. A-G8 was errors, 36%). In the execution errors, there is usually poor labor
identified in 74% of the sample (78 flat roofs), to which 313 pos- quality (C-E1) or poor execution of the downspouts (C-E7), or
sible causes were allocated, resulting in an average of 4.0 causes for both. This problem becomes relevant when sinking or protruding

© ASCE 04018100-3 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Table 2. Sample and elements characteristics
Characteristic Description
Sample
Environment 90% urban and 10% rural.
Typology 60% services, 10% housing, and 30% others.
Rehabilitation 67% of the sample presents no rehabilitation/intervention, 33% were rehabilitated/intervened.
Number of floors 53 buildings have one floor; 14 buildings, 2 floors; 12 buildings, 3 floors; 3 buildings, 4 floors; 2 buildings, 5 floors; 9
buildings, 6 floors; and 12 buildings, 7 or more floors.
Circulation paths 10% presented circulation paths and 90% presented no circulation paths.
Area 19% of the sample has an area lower than 50 m2 ; 25% between 50 m2 and 100 m2 ; 17% between 100 m2 and 150 m2 ;
18% between 150 m2 and 400 m2 ; 11% between 400 m2 and 650 m2 ; and 10% higher than 650 m2 .
Slope 4% of the sample has a slope lower than 1°; 42% between 1° and 2°; 42% between 2° and 3°; 5% between 3° and 4°; 1%
higher than 4°; and in 7% of the sample, the slope was not measured.
Elements
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

Type of constructive system 23% of the roofs have a traditional waterproofing system; 23% have an inverted system (i.e., the waterproofing is below
the insulation), and in 54% of the cases, the position of the insulation layer could not be determined or there was no
insulation layer.
Type of protection 24% of the elements present a heavy protection system; 62% a light protection system; and 14% of the elements are
unprotected.
Type of membrane 82% of the elements have a bituminous waterproofing membrane; 7% a PVC waterproofing membrane; and 11% a liquid
waterproofing membrane.
Type of thermal insulation 73% of the elements could not be classified according to the type of thermal insulation applied; 22% of the elements had a
synthetic insulation; and 5% had a mineral insulation material.
Type of fastening Only 14% of the elements presented mechanical fastening.

COMMON ANOMALIES DETECTED

A-G1 Surface wear A-G4 Creasing/bulging A-G8 Debris accumulation

A-G10 Biological growth A-S2 Inadequate downspout’s design

Fig. 2. Most common anomalies detected during the inspections.

areas are created adjacent to the overlapping layers of the insu- of the inspected flat roofs. The anomalies classified as singular
lation, affecting the flow of water. In the project/design errors anomalies represent 33% of the total anomalies; however, the clas-
group, there is often a deficient design/detailing of the down- sified general anomalies that can also occur at singular points must
spouts (C-P4). be added to this value.
The groups that most contribute to the appearance of A-G4 When making a comparison between the most common anoma-
(creasing/bulging) are C-E (execution errors, 57%) and C-A lies in flat roofs and in pitched roofs (Garcez et al. 2012a), the
(environmental actions, 32%). In the first group, C-E1 (poor labor following conclusion could be taken:
quality) and C-E11 (poor placement of the waterproofing mem- • Anomaly A-G8 (debris accumulation) has a much greater
brane) are the most frequent causes of this anomaly. Both causes impact in flat roofs (74.3%) when compared to pitched roofs
are intrinsically linked to labor quality. As for the environmental (28%), mainly due to its relative position and increased ten-
actions, the causes that most contribute to the appearance of this dency to accumulate debris. So, it is even more important for
anomaly are C-A2 (ultraviolet/solar radiation), C-A3 (moisture), flat roofs to have routine maintenance/cleaning in current zones,
and C-A7 (cycles of wetting-drying), detected in 8%, 13%, and singular points, and drainage spots; and
11% of the cases, respectively. • Anomaly A-G10 (biological colonization) was detected in simi-
On the other hand, A-S3 (inadequate overflow drainage tube lar percentages in flat roofs (61.9%) and pitched roofs (72.5%),
design) and A-S6 (defects in the fastenings) have the lowest prob- reflecting the greater ease of colonization of flat roofs and the
ability of occurrence, appearing in only 4% and 7%, respectively, less frequent maintenance of pitched roofs.

© ASCE 04018100-4 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Table 3. Classification system of possible causes of anomalies in flat roofs Regarding the causes of these anomalies, the following conclu-
Code Cause sions were drawn:
• The accumulation of debris in pitched roofs is associated with
Project/design errors
an incorrect slope or incorrect/insufficient cleaning of singular
C-P1 Deficient design/detailing of the coatings
points (e.g., gutters). In flat roofs, incorrect slope is not consid-
C-P2 Deficient design/detailing of the expansion joints
ered so relevant to this anomaly because the slope is too low to
C-P3 Deficient design/detailing of the protruding elements
C-P4 Deficient design/detailing of the downspouts
influence the accumulation of debris in these areas. The main
C-P5 Deficient design/detailing of the gutters
reasons are (1) wind displacement of debris, which promotes
C-P6 Deficient design/detailing of the lap joints the concentration of debris, especially in the singular points
C-P7 Deficient design/detailing of the pitch (corners, gutters); (2) a lack of maintenance/cleaning action,
C-P8 Deficient design/detailing of the tail ends and associated as in pitched roofs; and (3) vandalism from the concentration
protection elements of waste—in particular, cigarette butts. In flat roofs, cleaning
C-P9 Deficient design/detailing of the overflow drainage tubes of specific points such as the drainage system is particularly
C-P10 Difficult/impossible access to the surfaces relevant to allow water to flow.
• Biological colonization in pitched roofs is due to solar radiation,
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

C-P11 Inadequate choice of materials


biological action, insufficient slope, and lack of or insufficient
Execution errors maintenance actions (detected in 50% of the cases of the anom-
C-E1 Poor labor quality
aly). In flat roofs, the slope of the roof has no direct or indirect
C-E2 Application in humid/rainy weather
relationship with this anomaly. Solar radiation, the presence of
C-E3 Poor materials storage/transport
moisture, and the appearance of vegetation have similar rele-
C-E4 Poor execution of tail ends and associated protection
elements
vance when compared to pitched roofs, occurring on average
C-E5 Poor execution of expansion joints
58% of the times the anomaly occurs. Regarding maintenance
C-E6 Poor execution of lap joints actions, their occurrence is more important in flat roofs, being
C-E7 Poor execution of downspouts registered in 83% of the occurrences of the anomaly. When com-
C-E8 Poor execution of gutters pared to pitched roofs, the drainage system has higher relevance
C-E9 Poor execution of overflow drainage tubes in flat roofs because 30% of the cases of this anomaly come from
C-E10 Deficient fastenings a lack of maintenance/cleaning of the drainage system.
C-E11 Poor placement of the waterproofing membrane Fig. 5 shows the distribution of anomalies according to the
C-E12 Deficient pitch material used in the waterproofing membrane of the flat roof
system.
Environmental actions For A-G1 (surface wear), bituminous membranes present a
C-A1 Wind
higher value frequency. This result is primarily due to the fact that
C-A2 Ultraviolet/solar radiation
this material can be used both with heavy and light protection,
C-A3 Moisture
which can lead to a faster and more intense surface wear. On
C-A4 Prolonged presence of water/rain
C-A5 Vegetation
the other hand, the same situation was not detected on the PVC
C-A6 Natural aging
membranes because all the PVC membranes had heavy protection.
C-A7 Wet-dry cycles The anomalies considered typical of this type of material [A-G1,
A-G3, A-G4, A-G5, and A-G12 (surface wear, detachment/peeling,
Maintenance/use errors creasing/bulging, cracking, and moisture stains, respectively)] were
C-U1 Lack of cleaning debris in current zone not detected in PVC membranes. This absence is mainly due to the
C-U2 Accumulation/obstruction of debris in downspouts or gutter applied constructive solution that always uses heavy superficial pro-
entrances
tection, which makes this material more durable and results in less
C-U3 Absent/inadequate inspection
problems. The main anomalies detected in this type of membranes
C-U4 Change of the type of use of the roof
C-U5 Vandalism
are due to a lack of maintenance, resulting in the development of
A-G8 (debris accumulation) and A-G10 (biological growth).
External mechanical actions A-G10 (biological growth) is more recurrent in PVC membranes
C-M1 Dynamic loads than in other waterproofing membranes. This is justified by the over-
C-M2 Static loads exposure of the protective layer and the waterproofing membrane to

Fig. 3. Relative contribution of each anomaly to the total sample. (See Table 1 for anomalies.)

© ASCE 04018100-5 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Fig. 4. Frequency of anomalies detected at least once in each of the 105 inspections. (See Table 1 for anomalies.)
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. Relative frequency of anomalies in different waterproofing membrane materials. (See Table 1 for anomalies.)

high-moisture conditions, creating a favorable environment for the high correlation with the lack of quality and poor labor, particularly
development of vegetation. This condition can also be justified in the preparation of bonding of the surface.
by the higher occurrence of A-G9 (inadequate slope/ponding) in Fig. 4 also shows that bituminous membranes show a higher
PVC membranes when compared to other materials (Fig. 5). frequency of A-G5 (cracking), which may be explained by their
The predominance of A-G7 (absence/inadequate layer position- darker color, which absorbs more energy in the form of heat,
ing) in PVC membranes relative to other materials is due to the relative to other lighter materials.
absence, in the inspected roofs, of a separation layer between Liquid membranes have the lowest frequency of A-S8 (defects
the waterproofing membrane and the heavy protection. in the tail ends), which can be explained by the ease of the system to
From the analysis of Fig. 5, one can also observe that the PVC adhere to any material and be applied to any type of surface geom-
membrane is the material with more occurrences of A-G6 (punc- etry. The types of membranes that present the lowest frequency of
ture), which can be explained by the fact that the system consists of A-G10 (biological growth) also present a lower frequency of A-S8
a single layer only. (defects in the tail ends). These data support the conclusion that a
The high occurrence of A-G3 (detachment peeling) and A-G4 correct tail-end execution prevents the early development of
(creasing/bulging) in the bituminous and liquid membranes has a A-G10, due to a lower probability of accumulation of water and

© ASCE 04018100-6 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Relative frequency of anomalies in bituminous membranes of different superficial protections. (See Table 1 for anomalies.)

small debris (dust, sand, earth), which can accelerate the develop- operations, whereas the latter is due to bad detailing or execution
ment of this anomaly. of downspouts, or both.
The different anomalies detected in bituminous membranes, In PVC membranes, regarding the integrity of the waterproofing
according to the different types of superficial protection, are pre- membrane, A-G6 (punctures) has a high rate of occurrence
sented in relative frequency in Fig. 6. (13.3%), which can jeopardize the integrity of the membrane. In
As expected, unprotected membranes and those with light the case of inverted systems, where PVC membranes are adopted,
protection have an increased frequency of A-G1 (surface wear) A-G7 (absence/inadequate layer positioning) was detected in
compared to those with heavy protection. 13.3% of the cases. This anomaly occurred mainly in the thermal
The frequency of detection of A-G4 (creasing/bulging) in insulation, which sometimes was damaged due to wind, vandalism,
different protection materials was expected because bituminous and biological growth. A-S8 (defects in the tail ends) occurred with
membranes without protection have the highest exposure to a frequency of 6.7%, which shows the small importance that is
solar radiation and temperature variation when compared given to detailing and execution of singular points.
to light protection, and even more when compared to heavy Regarding the use of liquid membranes, the most common
protection. anomalies detected, besides the ones described previously, are
There is a higher frequency of A-G10 (biological growth) in A-G4 (creasing/bulging, 9.4%) and A-G7 (absence/inadequate
membranes with heavy protection due to the accumulation of mois- layer positioning, 9.4%). The first is associated with poor labor
ture between the protection and the membrane, creating a favorable quality, particularly when placing the waterproofing membrane
condition for the development of the anomaly. on a moist substrate. The latter, in turn, relates to the lack of a glass
The average number of anomalies detected in the different types fiber net layer or the extra liquid membrane layer in some specific
of waterproofing membranes is calculated by dividing the total locations.
number of anomalies of a waterproofing membrane by the total As for bituminous membranes, besides the anomalies previ-
number of occurrences of that waterproofing system. Thus, the ously described, those that have a higher occurrence are A-G1
average number of anomalies by the waterproofing system is (surface wear, 9.6%) and A-G4 (creasing/bulging, 8.2%). In the
5.82 for PVC membranes, 6.27 for bituminous membranes, and first case, it is the action of solar radiation that predominantly pro-
4.29 for liquid membranes. motes the aging of bituminous membranes. In the second case, the
The number of inspected liquid membranes (11) is much lower anomaly is related to an inefficient fastening of the membrane
than the number of bituminous membranes (86), but the average along with the presence of moisture in the substrate.
number of anomalies is relatively close. The level of severity/repair emergency of each anomaly was
The anomalies that occur more often, on average, in the three also characterized in the fieldwork. In this inspection system, three
solutions are A-G10 (biological growth), A-G8 (debris accumula- levels of severity were considered: 0 (immediate intervention,
tion), and A-S2 (inadequate downspout design). The first two are up to 6 months); 1 (medium-term intervention by 1 year); and
due, once again, to the absence of cleaning and maintenance 2 (monitoring of the anomaly).

© ASCE 04018100-7 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


studies continue to identify the design phase as one of the main
reasons for lack of quality in flat roofs, with a weight of up to
50% (Cnudde 1991; Josephson and Hammarlund 1991). However,
the validation of this inspection system showed a lower percentage
of anomalies with design causes. This value is expected to continu-
ally decrease because the materials have an increasingly better
quality, are easier to apply, and are submitted to detailed tests
for quality control in many countries (Silva and Gonçalves 2001).
Fig. 7. Relative contribution of each group of causes. As shown in Figs. 8 and 9, C-E1 (poor labor quality) is the most
common cause and is responsible for most of the execution errors
detected. C-A2 (ultraviolet/solar radiation) is the environmental ac-
The level of severity for each anomaly was based on the infor- tion with the biggest impact because it faster deteriorates the layers
mation gathered in the inspection and validation sheets and respec- of the system. C-U3 (absent/inadequate inspection) is the third
tive levels of severity/repair urgency referred to in the anomaly most common cause related to incorrect cleaning and maintenance
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

sheets. Level of severity 2 was allocated to 50.3% of all anomalies, of the system.
whereas the level of severity that requires immediate intervention
was allocated to 19.2% of all anomalies, in particular to A-G12
(moisture stains), A-S2 (inadequate downspout design), and Frequency Analysis of the Diagnosis Methods
A-S6 (defects in the fastenings). The highest level of severity Along with the development of the inspection system, a total of
was attributed to these anomalies in more than 50% of the seven diagnosis methods, as presented in Table 4, were divided into
cases. For A-G11 (corrosion), A-G10 (biological growth), A-G9 five groups: D-A (visual analysis), D-B (electric methods), D-C
(inadequate slope), A-G8 (debris accumulation), A-G7 (absence/ (thermohygrometric methods, D-D (nuclear methods), and D-E
inadequate layer positioning), A-G5 (cracking), AG-3 (detachment/ (load tests).
peeling), and A-G1 (surface wear), the medium-term intervention In the 105 inspections, a total of 936 diagnosis methods
classing was attributed in more than 30% of the cases. were attributed to 608 detected anomalies, resulting, on average,
In Fig. 7, the relative contribution of each group of causes is in 1.54 diagnosis methods per anomaly. The total number of
presented. It is apparent that the C-E (execution errors) group recommended diagnosis techniques is higher than the number
presents the highest rate of incidence of all groups, being the of anomalies, because for many of the anomalies, a combination
one that most contributes to the appearance of anomalies, followed of several techniques may be necessary for a proper diagnosis.
by the C-A (environmental), C-P (project/design), C-U (maintenance/ The first conclusion to be drawn from Fig. 10 is that all diag-
use), and C-M (external mechanical) causes groups, with the lowest nosis techniques proposed in the classificatory system are likely to
number of occurrences. be used in flat roofs because they have been registered at least once
Two of the most recognized studies mentioned previously indi- during the inspections.
cated that project and design errors represented 40% of the causes As shown in Fig. 10, D-A1 (visual inspection) is the only
of the anomalies detected (Bureau Securitas 1984a, b). More recent method common to all anomalies because it is the primary method

Fig. 8. Relative frequency of the causes from the C-P and C-E groups. (See Table 3 for causes.)

Fig. 9. Relative frequency of the causes from the C-A, C-U, and C-M groups. (See Table 3 for causes.)

© ASCE 04018100-8 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Table 4. Proposed classification system of diagnosis methods of anomalies Table 5. Proposed classification of rehabilitation techniques of anomalies
in flat roofs in flat roofs
Code Diagnosis method Code Rehabilitation technique
Visual analysis General
D-A1 Visual inspection R.1 Cleaning of the exterior coating of the flat roof
D-A2 Measurement of slope R.2 Application/repair/substitution of the waterproofing system
Electric methods R.3 Application/repair/substitution of the thermal insulation
D-B1 Electric leak detection R.4 Application/repair/substitution of the separation layer
D-B2 Capacitance test R.5 Application/repair/substitution of the vapor barrier
Thermohygrometric methods R.6 Application/substitution of the shaping layer
D-C1 Infrared thermography R.7 Creation of circulation routes
Nuclear methods
D-D1 Nuclear method Singularities
Load tests R.8 Application/repair/substitution of expansion joints
D-E1 Flood test R.9 Repair of joints and associated protective elements
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

R.10 Application/repair/substitution of the drainage system


R.11 Application/repair/substitution of fastening elements

of diagnosis of any anomaly, followed by more specific methods


if available and necessary. The second most indicated method is A-G8 (debris accumulation) and A-G10 (biological growth) can
D-C1 (infrared thermography), with 12.2% of all recommended be resolved by using the R.1 technique due to its swiftness, easy
methods, followed by D-B1 (electric leak detection), registered implementation, and low cost. The results for this technique were
in 11.2% of cases. D-B2 (capacitance test), D-D1 (nuclear method), expected because of the absence of routine maintenance, along with
and D-E1 (flood test) have very close contributions, considering the proliferation of vegetation.
their individual constraints. The high frequency of R.2 is mainly due to the different types
The uniform distribution among the diagnosis methods (with the of anomalies that can occur in the waterproofing membrane. The
exception of the D-A1 technique, visual inspection) can be related application of this technique does not imply that the waterproofing
with high frequency of two anomalies that can be diagnosed by all system is not working, but that the membrane is not in perfect con-
the methods presented: the A-G2 (fracture/failure) and A-G12 dition, affecting its proper functioning and the durability of the
(moisture stains) anomalies, detected in 21.9% (A-G2) and remaining materials of the system.
19.0% (A-G12) of all the flat roofs. The use of R.9 (repair of joints and associated protective ele-
ments) is related to the detection of A-S7 (defects in the coping)
and A-S8 (defects in the tail ends) and is the third most recom-
Frequency Analysis of the Rehabilitation Techniques
mended technique, in more than half of the roofs inspected.
Along with the development of the inspection system, a total of 11 As shown in Fig. 12, PVC membranes are the ones that overall
diagnosis methods, as presented in Table 5, was divided into two require less intervention, particularly using R.2, i.e., they have
groups: for general anomalies (comprising 7 methods), and for sin- greater durability. Still, it is not possible to say that PVC mem-
gular anomalies (comprising 4 methods). branes are more durable than the other materials because in the
In the 105 inspections, a total of 385 rehabilitation techniques inspections performed, all PVC membranes had heavy protection.
were attributed to 608 detected anomalies, resulting, on average,
in 0.63 techniques per anomaly and 3.67 techniques per roof.
An average below 1 demonstrates that the rehabilitation techniques Lessons Learned and Conclusions
are intended to eliminate the anomaly and not the cause.
As shown in Fig. 11, the most common rehabilitation techniques With the 105 inspections carried out, it was possible to validate
are R.1 (cleaning of the exterior coating of the flat roof, 93%) and and calibrate the proposed classification systems, perform a data
R.2 (application/repair/substitution of the waterproofing system, analysis of the collected information, and even change some as-
95%). These values demonstrate the lack of maintenance that most sumptions that improved the application of the classification sys-
roofs presented when inspected. tem to real situations; however, the calibration of the system should

Fig. 10. Relative frequency of each of the diagnosis methods recommended for the anomalies detected. (See Table 4 for diagnosis methods.)

© ASCE 04018100-9 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Fig. 11. Relative frequency of the rehabilitation techniques. (See Table 5 for rehabilitation techniques.)
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 12. Relative frequency of the rehabilitation techniques adequate for different waterproofing materials. (See Table 5 for rehabilitation techniques.)

be continued in terms of more inspections in order to increase the (A-G11) (100%), and moisture stains of condensation/leaks
number and diversity of the sample. (A-G12) (35%); and
The data analysis of the anomalies concluded that A-G8 (debris • a correct design/detailing of the tail ends and associated
accumulation) and A-G10 (biological growth) are the most likely protection elements (C-P8) to prevent detachment/peeling
anomalies to de detected during an inspection in 74% and 62% (A-G3) (this cause is responsible for 39% of the occurrences
of the cases, respectively. As expected, the relevance of the causes of this anomaly), moisture stains of condensation/leaks
associated with errors of use/maintenance and environmental ac- (A-G12) (35%), defects in the coping (A-S7) (88%), and
tions was demonstrated. defects in the tail ends (A-S8) (100%).
In bituminous membranes with heavy protection, there was 2. Concerning execution errors (C-E), it is essential to improve the
a higher incidence of A-G10, about 86%, than in membranes that quality of labor (C-E1) in order to avoid/minimize the occur-
have no protection (33%) or in those with light protection (61%). rence of detachment/peeling (A-G3) (this cause is responsible
This result was predictable because the protection in the water- for 58% of the occurrences of this anomaly), creasing/bulging
proofing system slows the evaporation of water, creating a prosper- (A-G4) (85%), absence/inadequate layer positioning (A-G7)
ous environment for the proliferation of vegetation. (62%; for example, in the downspouts when the waterproofing
Roofs without heavy protection had a higher number of cases layer is not extended into the interior of the tube), inadequate
with A-G1 (surface wear), with values close to 65%. These results slope/ponding (A-G9) (79%), inadequate design of the expan-
show the importance of selecting the best type of protection be- sion joint (A-S1) (85%), inadequate design of expansion joints
cause it protects the membrane from mechanical strain and UV (A-S1) (86%), defects in the fastenings (A-S6) (43%), and
and thermal variations. defects in the tail ends (A-S8) (71%).
Comparing the results obtained for flat roofs with those from a 3. Concerning environmental actions (C-A), the main causes of the
pitched-roofs study (Garcez et al. 2012a, b), it can be concluded occurrence of anomalies are
that, as expected, the environmental agents have a higher influence • ultraviolet/solar radiation (C-A2): 75% of surface wear (A-G1),
in the occurrence of anomalies in pitched roofs, in comparison to 65% of fractures/failure (A-G2), 29% of creasing/bulging
flat roofs, which can be justified by their geometry and position. (A-G4), and 78% of cracks (A-G5) are due to this cause; and
It is then possible to conclude the following: • presence of moisture (C-A3): responsible for 100% of the
1. Concerning project/design errors (C-P), it is essential to take occurrences of corrosion and 48% of creasing/bulging
into account (A-G4).
• the use of appropriate materials (C-P11) to minimize surface 4. Concerning maintenance/use errors (C-U), it is essential to
wear (A-G1) (the choice of inappropriate materials is respon- improve the following aspects:
sible for 57% of the occurrences of this anomaly), fracture/ • Absent/inadequate inspection (C-U3) greatly contributes to
failure (A-G2) (26%), cracking (A-G5) (57%), corrosion the accumulation of debris (A-G8) (83%) and for biological

© ASCE 04018100-10 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


colonization (A-G10) (83%); because they are the two most growth), each of which was observed in 20% of the inspected PVC
frequent anomalies, it becomes essential to have cleaning/ membranes.
maintenance operations in flat roofs. Of the three types of membranes analyzed, liquid membranes
• A lack of cleaning debris in the current zone (C-U1) clearly were those that had a lower occurrence of A-G10 (biological
contributes to the accumulation of debris (A-G8) (74%), in growth) and A-S8 (defects in tail ends).
particular near the drainage system where the wind action is The objective of this study was to perform an analysis of the
preponderant. data gathered in field inspections of flat roofs. This data allowed
5. Concerning external mechanical actions (C-M), it is noted that the calibration of the inspection, diagnosis, and rehabilitation of
• static loads (C-M2) are responsible for 30% of the occur- flat roofs, and it was properly calibrated and validated through
rences of fracture/failure (A-G2) and 37% of that of punctu- an extensive campaign of inspections. The system is therefore con-
res (A-G6); and sidered a good tool to aid in future inspections. The data presented
• dynamic loads (C-M1) are responsible for 53% of the occur- in this paper help review the design choices according to the types
rences of punctures (A-G6). of materials and systems to use and the possible problems that may
In comparison to Freitas and Sousa (2003) in terms of execution occur in the service life of the flat roof. The use of a computerized
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

errors, there was a significance decrease (from 80% to 30%); how- system that allows the condensation of the information gathered is
ever, it is still the group that most contributes to the occurrence of important. The use of this system facilitates the standardization of
anomalies in flat roofs. Thus, it can be said that since 2001 there has maintenance operations and can reduce the costs of the operations
been a significant improvement in waterproofing execution, par- by achieving more efficient choices and effective interventions.
ticularly in terms of skilled labor. A similar result is presented
in Garcez et al. (2012b), in which 31% of the causes are execution
errors. In comparison with Silva and Gonçalves (2001), the results Acknowledgments
obtained in this study are similar, but the design errors decrease
from 38% to 22%. In short, it is concluded that the results are gen- The authors gratefully acknowledge the support of the CERIS Re-
erally in line with other studies, even though a broader universe of search Institute, Instituto Superior Técnico—University of Lisbon
causes was considered. and Foundation for Science and Technology (FCT).
When comparing the occurrence of anomalies in different
materials of the waterproofing membranes, it can be concluded
that the highest number of anomalies can be detected in bitumi- References
nous and liquid membranes. This should be considered in a
more demanding specification of materials when choosing a bi- Amaro, B., D. Saraiva, J. de Brito, and I. Flores-Colen. 2013. “Inspection
and diagnosis system of ETICS on walls.” Constr. Build. Mater.
tuminous or liquid membrane solution, with particular attention
47 (Oct): 1257–1267. https://doi.org/10.1016/j.conbuildmat.2013.06
to the strength and durability of the material when exposed to .024.
the sun. Amaro, B., D. Saraiva, J. de Brito, and I. Flores-Colen. 2014. “Statistical
Level 2 of severity/repair emergency was the most often re- survey of the pathology, diagnosis and rehabilitation of ETICS in
corded, with 50.3%, showing that anomalies of this gravity should walls.” J. Civ. Eng. Manage. 20 (4): 511–526. https://doi.org/10
be monitored in order to check their evolution. Anomalies with a .3846/13923730.2013.801923.
repair emergency of 0 (19.6%), because their occurrence affects Bureau Securitas. 1984a. “Étude statistique de 10000 dossiers de sinistres,
one of the main functional requirements of a roofing system—its [Statistical study of 10.000 claim files].” [In French.] In Annales
water tightness—should lead to an immediate intervention. The se- L’institut Technique du Bâtiment et des Travaux Publics, 378. France:
verest anomalies detected were moisture stains of condensation/leaks L’institut Technique du Bâtiment et des Travaux Publics.
Bureau Securitas. 1984b. “Étude statistique de 12200 dossiers de sinistres
(A-G12), inadequate downspout design (A-S2), defects in the fasten-
en 1982 [Statistical study of 12.200 claims files in 1982].” [In French.]
ings (A-S6), and fracture/failure (A-G2). In Annales L’institut Technique du Bâtiment et des Travaux Publics,
Visual inspection (D-A1) is recommended in an initial diagnosis 426. France: L’institut Technique du Bâtiment et des Travaux Publics.
of any anomaly, being easy to use and not requiring any special Cnudde, M. 1991. “Lack of quality in construction.” In Proc., European
equipment, apart from a skilled experienced professional; however, Symp. on Management, Quality and Economics in Housing. Quality
where possible, it is recommended to use auxiliary means in order and Economics in Housing, 508–515. London: E&FN Spon.
to obtain a more complete diagnosis. Conceição, J., B. Poça, J. Brito, I. Flores-Colen, and A. Castelo. 2017. “In-
Many of the inspected roofs show no signs of an implemented spection, diagnosis, and rehabilitation inspection system for flat roofs.”
maintenance plan. This demonstrates the need to use R.1 (cleaning J. Perform. Constr. Facil. 31 (6): 04017100. https://doi.org/10.1061
/(ASCE)CF.1943-5509.0001094.
of the exterior coating of the flat roof) in 93% of the sample because
Freitas, V., and M. Sousa. 2003. “Building pathology: A catalogue.” In
it is designed to repair A-G8 (debris accumulation) and A-G10 Proc., 2nd Int. Symp. on Building Pathology, Durability and Rehabili-
(biological growth). tation, 477–488. Lisbon, Portugal: Portuguese National Laboratory of
The most recommended techniques for bituminous membrane, Civil Engineering.
with a recommendation frequency higher than 50%, were clean- Garcez, N., N. Lopes, J. de Brito, and G. Sá. 2012a. “Pathology, diagnosis
ing of the exterior coating of the flat roof (R.1), repair of joints and repair of pitched roofs with ceramic tiles: Statistical characterisation
and associated protective elements (R.9), and application/repair/ and lessons learned from inspections.” Constr. Build. Mater. 36 (Nov):
substitution of the waterproofing system (R.2), given the high 807–819. https://doi.org/10.1016/j.conbuildmat.2012.06.049.
frequency of anomalies in the membrane, such as surface wear Garcez, N., N. Lopes, J. de Brito, and J. D. Silvestre. 2012b. “System
of inspection, diagnosis and repair of external claddings of pitched
(A-G1), detachment/peeling (A-G3), and creasing/bulging (A-G4).
roofs.” Constr. Build. Mater. 35 (Oct): 1034–1044. https://doi.org/10
The most recommended techniques for PVC membranes, with .1016/j.conbuildmat.2012.06.047.
a recommendation frequency higher than 70%, were cleaning of Gomes, R. 1968. Flat roof. Technical information buildings. Lisbon,
the exterior coating of the flat roof (R.1) and application/repair/ Portugal: National Laboratory of Civil Engineering.
substitution of the waterproofing system (R.2). The use of R.1 is INE (National Statistics Institute). 2012. Census 2011: Definitive results:
mainly due to A-G8 (debris accumulation) and A-G10 (biological Portugal. [In Portuguese.] Lisbon, Portugal: INE.

© ASCE 04018100-11 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100


Josephson, P., and Y. Hammarlund. 1991. “Sources of quality failures in Santos, A., M. Vicente, J. de Brito, I. Flores-Colen, and A. Castelo. 2017b.
building.” In Proc., European Symposium on Management, Quality and “Inspection, diagnosis, and rehabilitation system of door and window
Economics in Housing, 671–680. London: E&FN Spon. frames.” J. Perfor. Constr. Facil. 31 (3): 04016118. https://doi.org/10
Pires, R., J. de Brito, and B. Amaro. 2015a. “Inspection, diagnosis, and .1061/(ASCE)CF.1943-5509.0000992.
rehabilitation system of painted rendered façades.” J. Perform. Constr. Silva, J., and P. Gonçalves. 2001. “Frequent pathologies in flat roofs in
Facil. 29 (2): 04014062. https://doi.org/10.1061/(ASCE)CF.1943-5509 Portugal.” [In Portuguese.] In Proc., National Construction Congress.
.0000534. Lisbon, Portugal: Instituto Superior Técnico.
Pires, R., J. de Brito, and B. Amaro. 2015b. “Statistical survey of the inspec- Silvestre, J. D., and J. de Brito. 2007. “Technical note: Statistical analysis of
tion, diagnosis and repair of painted rendered façades.” Struct. Infrastruct. defects of tiles’ joints.” [In Spanish.] Materiales de Construcción
Eng. 11 (5): 605–618. https://doi.org/10.1080/15732479.2014.890233. 57 (285): 85–92. https://doi.org/10.3989/mc.2007.v57.i285.42.
Santos, A., M. Vicente, J. de Brito, I. Flores-Colen, and A. Castelo. 2017a. Walter, A., J. de Brito, and J. Lopes. 2005. “Current flat roof bituminous
“Analysis of the inspection, diagnosis, and repair of door and window membranes waterproofing systems: Inspection, diagnosis and pathol-
frames.” J. Perform. Constr. Facil. 31 (3): 04017098. https://doi.org/10 ogy classification.” Constr. Build. Mater. 19 (3): 233–242. https://doi
.1061/(ASCE)CF.1943-5509.0001095. .org/10.1016/j.conbuildmat.2004.05.008.
Downloaded from ascelibrary.org by URI LIBRARIES on 12/01/18. Copyright ASCE. For personal use only; all rights reserved.

© ASCE 04018100-12 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2019, 33(1): 04018100

You might also like