Defect characterization, diagnosis and repair of wood flooring based on a field survey

A. Delgado^a, C. Pereira^b,^*, J. de Brito^b, J. D. Silvestre,^b

^a. Instituto Superior Técnico (IST) - Universidade de Lisboa, (Lisbon, Portugal)

^b. CERIS-ICIST, Dept. of Civil Engineering, Architecture and Georresources, Instituto Superior Técnico (IST) - Universidade de Lisboa, (Lisbon, Portugal)

* clareira@sapo.pt

1. INTRODUCTIONTOP

Wood floorings are a long-established flooring solution in every continent. Wood flooring (WF) comprises both solid wood and composite wood solutions. WF is known for its easy maintenance, durability, low cost, quality control of the production process and good environmental performance (1). Functionally, wood floorings comply with today’s comfort, habitability and aesthetic requirements, and additionally are quick to install.

The characteristics of the substrate, the hygrothermal conditions and the availability of skilled labor are determining factors during the installation of WF and influence their performance over time, both in indoor and outdoor areas. Defects occurrence and the degradation level of WF are highly dependent on these early-stage factors, which may lead to repairs and/or refurbishment before time. With a system that impartially identifies, records, and classifies defects in WF and their causes, the task would be easier. Prescribing the most appropriate diagnostic methods and repair solutions to eliminate the defects and corresponding causes is another useful asset of the system (2). The work of Delgado et al. (1) proposes such an inspection system for WF, based on a methodology also tested on gypsum plasterboard walls (3), gypsum plasters (4), renderings (5), ceramic tiling (6), natural stone cladding (7), external thermal insulation composite systems (8), flat roof waterproofing systems (9), pitched roof claddings (10), industrial floor coatings (11), architectural concrete surfaces (12), and door and window frames (13). This methodology was adapted to the specific WF technology and the study of floorings’ pathological processes. To the best of the authors’ knowledge, it remains the only such system in the literature.

Delgado et al. (1) describe in detail this WF inspection system. This paper presents a detailed statistical analysis of the results of the inspection program that validated the WF inspection system. It is a systematic approach to defect, causes, diagnostic methods and repair solutions, based on a set of 98 WF case studies, similar to the analysis of Gaião et al. (14).

2. INSPECTION PROGRAMTOP

An inspection campaign was planned for buildings with defects in WF, including the pathological characterization of 98 WF representative case studies, analyzing the most suitable diagnostic and repair methods available for each specific circumstance. This field work made it possible to validate the expert system proposed for WF and its procedures, such as the defect classification system, defect rating, classification of probable causes, classification of diagnostic methods and classification of repair techniques.

The inspected WF included softwood and hardwood floors, and laminated and engineered wood floors. Nevertheless, in the statistical survey the type of flooring was not considered, as the sample was not varied enough. It is focused mainly on cork tiles and traditional timber tiles. The inspected buildings were intended to be representative of the potential market for future inspections and rehabilitation. The standard inspections consisted mainly of a visual observation of the WF, which was documented through a standard inspection and validation file. Furthermore, in situ tests were recommended for each defect.

All the buildings are located in the center-south region of Portugal, and 48% are in Lisbon. As for the distribution of the age of the buildings in the sample, most buildings had been built in the previous five years (37%), followed by buildings 5 to 20 years old (26%), and buildings at least 100 years old (17%). Buildings 50 to 100 years old represent 11% of the sample, while only 9% of the buildings were 20 to 50 years old. The sample of 98 WF comprised 35 different buildings. The most significant part of the inspected WF was in commercial buildings (30%), heritage buildings (29%) and residential buildings (19%). Heritage buildings are those considered historical heritage, functioning as public or cultural buildings, as opposed to ordinary buildings. Figure 1 shows that heritage buildings have a higher ratio of defects per pavement. For every building each continuous WF was described in an inspection file.

2.1. Inspection and validation filesTOP

The inspection files contain all relevant information on the buildings and WF, so that different inspections can be compared. An inspection file was filled describing each inspected WF. First, the building location, use, construction date, area, percentage of wood floor covering, constructive typology, and any contact information are indicated. Next, all the wood floorings are described stating the installation date, detailed constructive characterization, type of cleaning undertaken, and relevant location specific data.

While the inspection file allows a constructive characterization of the building and of each inspected WF, the validation file provides information on the detected pathological phenomena, for statistical purposes. One validation file was filled for each WF, identifying and recording the main WF defects detected. These records were then used to validate the inspection system, namely the classification of defects, causes, diagnostic methods and repair techniques, as well as the correlation matrices between these parameters. Each validation file identifies: the kind of defects found, location and size of every defect, conditions for the defect to progress, substrate cracking, evidences of expansion of the parts, aesthetical value of the affected area and severity level, probable causes, diagnostic methods proposed, and repair techniques proposed. The content of the inspection and validation files is fully described by Delgado et al. (1). The discussion of the data in the inspection and validation files is the basis of the present statistical analysis.

2.2. Diagnosis and statistical characterization of WF pathologyTOP

The WF inspection system includes a classification system for defects in WF and their direct and indirect causes (1). The system groups defects (D) according to location on the surface (S) of the WF or in joints and interfaces (JI), within the WF, or between the WF and other floor coverings or protruding elements. In total, 16 types of defects were defined, according to scientific references (15–22). Defects are coded with a D for defect, hyphen and the group reference, namely S for surface, and JI for joints and interfaces, followed by sequential numbering. D‑S group defects are then divided into two subgroups: aesthetic defects (D.1‑S) and functional defects (D.2‑S). Probable causes, diagnostic methods and repair techniques are coded using a similar method (1).

The inspection system considers 38 probable causes of WF defects, which were classified in chronological order, according to the following groups: production errors (C‑A), design errors (C‑B), execution errors (C‑C), exterior mechanical actions (C‑D), biological actions (C-E), environmental actions (C-F), and maintenance errors (C‑G).

For each detected defect, further diagnostic methods were recommended from a set of 11 proposed appropriate methods. These methods were divided in four groups. Three are nondestructive tests (T-ND): T.1-ND Aided visual observation, T.2-ND Traditional nondestructive tests and T.3-ND Nontraditional nondestructive tests. The fourth group is T-D Destructive tests.

The inspection system, presented by Delgado et al. (1), also associates a set of 18 preventive (RP) and curative (RC) repair techniques with the defects. These two main groups of repair techniques were divided in four subgroups corresponding to intervention in WF covering (RP-A and RC-A), WF surface (RP-B and RC-B), WF substrate (RP-C and RC-C) or in joints between pieces of WF or with protruding elements (RP-D and RC-D).

3. RESULTS AND DISCUSSIONTOP

3.1. Defects observed in the sampleTOP

In all, 239 defects were detected in the sample of 98 WFs (only one event per defect type was recorded for each WF), resulting on average in 2.44 defects per pavement. Figure 2 shows the relative frequency of detected defects (number of records divided by 98, the number of inspected WF). Defects are described in detail by Delgado et al. (1). D.1‑S3 Scratches or wrinkles (84.7%) was by far the most common defect detected in the sample. According to Baar et al. (23), the occurrence of scratches on the wood surface protection layer may influence the susceptibility of wood to the occurrence of other defects, when exposed to water. Hence, the incidence of scratches in this sample may be significant in terms of degradation modelling. D.1-S3 was followed by D‑JI3 Change of joint size (28.6%) and D.1‑S4 Wearing or detachment of the finishing layer (25.5%). Considering some defect classification differences, the results on Neto and de Brito’s (24) color change defects, referring to natural stone claddings, are similar to those on WF, as D.1-S1 Staining or color change and D.1-S2 Cigarette marks represent about 21% of the sample, while Neto and de Brito’s (24) color change defects represent about 27% of all defects. However, in Neto and de Brito’s (24) sample, joint defects were as relevant as color change defects (28% of all defects), while in WF all D-JI Joints and interfaces defects represent about 41% of the sample. This is a sign of the vulnerability of joints in WF, as they are subject to variations in wood moisture content and, consequently, to size variations according to climate changes, as verified by the high frequency of the defect D-JI3 Change of joint size in WF.

Five of the defect types were only detected up to two times, namely D.1‑S2 Cigarette marks, D.2‑S4 Detachment of elements from substrate, D.2‑S5 Rot, D.2‑S8 Delamination and D.2-S9 Crumbling. Nevertheless, these defects were kept in the analysis.

These general results, specifically on WF, differ from those of Chong and Low (25), referring to general floors defects. In that sample, the most frequent defects found in floorings under use were cracks, water seepage and delaminated tiles.

Considering the groups of defects, aesthetic defects (D.1-S) are the kind of defects most frequently detected: on average each pavement had 1.4 aesthetic defects.

Figure 3 shows the relationship between the average number of defects of each type and the age of the pavement it was detected in. Defect D.1-S3 Scratches or wrinkles, as the most frequent defect, has the highest average number of defects in WF of every age. Contrary to expectations, older buildings in this sample do not show greater frequency of this anomaly than the others, probably due to the type of solution, materials used and previous repairs. Defects D.1-S3 represent an average of 1.02 defects in buildings 20 to 50 years old, showing lack of maintenance. As for other defects, D.2‑S7 Xylophage attack is the defect most related with the age of the WF, as an average of 0.55 defects occur in WF more than 100 years old. Although some of the inspected floorings had possibly undergone repair operations to eliminate previous xylophage attacks, it is significant that one in every two floorings more than 100 years old shows this type of defect, even if on a limited area, requiring immediate action. Despite these results, it is important to highlight that, in the whole sample, defects D.2‑S5 Rot, D.2‑S6 Moisture and D.2‑S7 Xylophage attack were not representative. As general literature usually refers to mould, fungi and insects as the main origin of defects on wood (26), it is relevant that the inspected WF were not substantially affected by the effects of these organisms, whether or not they had already been subject to repairs. Buildings less than 5 years old have greater incidence of defect D.1-S4 Wearing or detachment of the finishing layer, which may show that the choice of materials was not adequate to use conditions. According to the results obtained, age does not appear to be a determining factor for the occurrence of defects, which contradicts the 42 years of estimated service life of hardwood floors calculated by Minne and Crittenden (27). The maintenance that WF are subjected to seems to play a more relevant role than age.

Every defect was rated in terms of repair urgency (1):

Repair urgency depends on the severity of the defect when it is detected. For each defect the criteria to rate its repair urgency were clearly defined, supporting the inspection file. For example, the defect file for D.2‑S4 Detachment of elements from substrate indicates that this defect is classified by the percentage of affected area and by the aesthetic value of the affected area (high, medium or low). When this defect is detected, if users’ safety is compromised, or if the affected area is over 5%, then the repair urgency level is 0. If the defect presents progression features, the repair urgency level is 1. For the remaining instances of D.2‑S4 defect, the repair urgency level is 2.

The commonest repair urgency level is “1” (42%) and the highest severity level (0) represents 23% of the sample. Observing Figure 4, a repair urgency per defect analysis leads to different conclusions. Defects D.2‑S3 Broken or splintered elements, D.2‑S4 Detachment of elements from substrate, D.2‑S5 Rot, D.2‑S7 Xylophage attack, D.2‑S8 Delamination, D.2-S9 Crumbling and D-JI2 Detachment or loosening of mass of joint material have the highest repair urgency (0) in more than 33% of the sample. They can all compromise the flooring’s integrity, even D-JI2, whose urgency levels reveal how important joint maintenance is to extend the WF’s service life. Considering that D.2‑S3 and D.2‑S7 affect a wider sample (14 and 10 defects respectively), their urgency frequency may be more significant. Although aesthetical defects require less demanding works, and do not compromise the flooring’s integrity, they should not be less considered, especially according to the flooring’s function.

As for the area of flooring affected by each defect, Figure 5 shows the data registered at the inspections. Defect D.1‑S3 Scratches or wrinkles, which is the one that affects the widest sample (83 defects), is also the one with the widest affected areas, since in 5% of the cases the D.1‑S3 defects affect 25% to 50% of the WF’s area. In the sample, defect D.2‑S3 Broken or splintered elements affects a large flooring area (from 5% to 10%) in 50% of the cases. Comparing these results with those on natural stone floorings (24), some differences are observed. In natural stone floorings, color changes are the type of defect that affects wider areas, while in WF, the corresponding defects (D.1‑S1 Staining or color change and D.1-S2 Cigarette marks) mainly affect areas of less than 5% of the flooring. Neto and de Brito (24) attributed this result to infrequent cleaning, but that may not be the case in WF, as staining is most probably associated with humidity or fungi.

3.2. Direct or indirect probable causes of defectsTOP

In total, 1319 probable causes (of 38 different types) were identified in the inspection campaign, as direct, or close, causes, or as indirect, or initial, causes (1). Therefore, an average of 5.5 probable causes were assigned to each recorded defect. It must be noted that no in situ or laboratory tests were actually performed, as inspections consisted solely on visual observation. So, the registered probable causes point out several possibilities.

C‑D Exterior mechanical actions account for 32% of the causes in the sample, followed by C-C Execution errors, which represent 25% of the sample. It is important to disseminate detailed information on WF construction systems, so that solutions are chosen according to the use of the floor, as well as to improve the selection of techniques and installation materials. Exterior mechanical actions are not accounted for in the Chong and Low’s (25) study, but poor workmanship is also a relevant cause for flooring defects. It is responsible for 35% of defects in floorings under use.

As each defect is confronted with the groups of probable indirect causes, some issues become evident. Figure 6 (left) shows that 76% of aesthetic defects (D.1‑S) were indirectly caused by exterior mechanical actions (C‑D). Aesthetic defects are also highly associated (56%) with execution errors (C‑C). Aesthetic defects area not indirectly caused by any production errors. As for functional defects (D.2‑S), 52% of defects are indirectly caused by one environmental action (C-F) and 48% by maintenance errors (C‑G). 18% of defects on joints and interfaces (D-JI) are indirectly caused by at least one execution error (C‑C).

Observing Figure 6 (right), it is interesting to notice that all aesthetic defects (D.1-S) are directly caused by at least one exterior mechanical action (C‑D). Biological direct causes (C‑E) are also moderately associated with aesthetic defects (33%). As for functional defects (D.2‑S), 59% of these kind of defects are directly caused by biological actions (C‑E), the same amount of functional defects (D.2‑S) directly associated with maintenance errors (C‑G). Defects on joints and interfaces (D-JI) are directly associated with execution errors (C‑C) in 24% of the D-JI defects registered.

Production errors (C‑A) are not directly associated with any specific defect. On the one hand, design errors (C-B) and execution errors (C-C) are only relevant for D-JI defects. On the other hand, environmental actions (C‑F) and maintenance errors (C‑G) are only directly associated with functional defects (D.2‑S).

Additionally, the results reveal other features. Aesthetic defects (D.1‑S) are not associated with production errors. Execution errors (C-C) are the main contributors to the occurrence of D.1-S Aesthetic defects (100%), but C-D Exterior mechanical actions (85%) and C-B Design errors (79%) also play a main role. Functional defects (D.2-S) are evenly associated with C-B Design errors (61%) and C-F Environmental actions (60%). As for D‑JI Defects on joints and interfaces, there are four groups of causes that stand out, as their association with these group of defects is above 70%, namely C-D Exterior mechanical actions, C-C Execution errors, C-B Design errors and C-A Production errors. C-E Biological actions are the group of causes less associated with WF defects in this sample.

3.3. Probable causes contribution to each defectTOP

To understand the contribution of each individual cause to the occurrence of each defect, it is important to analyze them individually. For this analysis, only defects that were observed more than 10 times were considered. Causes that were observed at least once for each defect were considered. Considering these criteria, only D.1-S1 Staining or color change, D.1-S3 Scratches or wrinkles, D.1-S4 Wearing or detachment of the finishing layer, D.2-S1 Warping, swelling, or other flatness deficiencies, D.2-S2 Cracking of elements and/or joints (possibly form substrate cracking), D.2-S3 Broken or splintered elements, D.2‑S6 Moisture and D-JI3 Change of joint size are analyzed in detail.

Figure 7 shows the data on causes probably associated with defect D.1-S1 Staining or color change. This defect was identified 19 times, to which 5.2 probable direct and indirect causes were attributed on average. All these defects can be associated with a finishing layer unsuited to use (C-C10) and natural ageing (C-F4). Covering not complying with use requirements (C-B3) and poor or insufficient maintenance/cleaning works (C-G2) are associated with over 80% of these defects, leading to being considered highly probable causes. This defect is associated with causes from groups C‑B, C‑C, C‑E, C-F and C-G. They are not related with production errors nor exterior mechanical actions.

Defect D.1-S3 Scratches or wrinkles was identified in 83 WF cases. To each an average of 5.3 probable causes (direct and indirect) were considered. All the sample was associated with a covering not complying with use requirements (C-B3), a finishing layer unsuited to use (C-C10), user or car circulation (C-D7) and with actions not foreseen at design stage (C-D8), as shown in Figure 7. Exterior mechanical actions (C-D group of causes) are highly associated with this defect.

In Figure 7, the data on causes of defect D.1-S4 Wearing or detachment of the finishing layer are also available. It was identified in 25 pavements and associated with 4.1 probable causes (direct and indirect) per defect. All these defects were associated with a finishing layer unsuited to use (C-C10), impacts (C-D5), user or car circulation (C-D7) and with poor or insufficient maintenance/cleaning works (C-G2). Once again, C‑D Exterior mechanical actions are the group of probable causes most associated with this defect.

All the surface aesthetic defects (D.1-S) were associated with cause C-C10 Finishing layer unsuited to use, an execution error that reveals unfamiliarity of the project team with materials’ technical features and limitations and requirements of the flooring system.

As for defect D.2-S1 Warping, swelling, or other flatness deficiencies, data on probable causes are available in Figure 8. This defect was identified in 15 of the inspected pavements, to which 7.9 direct or indirect probable causes were attributed on average. Only one probable cause is associated with all detected defects, which is inadequate wood drying (C-A1). But the wrong design of joints between elements or with protruding elements (C‑B1), and the use of excessively dry or wet wood (C-C3), both associated with 93% of these defects, are also highly probable causes. Causes from the group C-C Execution errors are highly represented in this sample of defects.

Defect D.2-S2 Cracking of elements and/or joints (possibly form substrate cracking) was present in 13 of the inspected pavements. An average of 5.5 probable causes (direct or indirect) were attributed to each of these defects. As Figure 8 shows, this defect was always considered to be associated with an inadequate fixing/fastening system (C-B2), differential substrate-WF movements (C-D1) and with the initial shrinkage and/or expansion of joint material (C-D4).

In Figure 8, data on probable causes of defect D.2-S3 Broken or splintered elements are also shown. This defect was detected in 14 floors, each associated with an average of 6.1 probable causes for their occurrence. No probable direct or indirect cause is associated with all these defects, and the sample presents a set of causes with an equal contribution to its appearance. Voids between wood pieces and bedding material (C‑C7) and shrinkage or expansion of bedding material (C-D3) are only sporadic causes. The main contributor is the group of causes C-D Exterior mechanical actions.

Defect D.2-S6 Moisture was identified in 12 WFs. An average of 2.2 probable causes were attributed to each. Environmental actions (C-F) and maintenance errors (C‑G) generate moisture. The most significant are WF humidification/rising damp (C-F3) and pipe ruptures near or over WF (C-G3), as shown in Figure 8. Poor or insufficient maintenance/cleaning works (C‑G2) are a less relevant cause.

There are 25 causes associated with more than 10% of surface functional defects (D.2-S). The cause most associated with these defects was C-B2 Inadequate fixing/fastening system, associated with 47% of defects.

Analyzing the causes for defect D-JI3 Change of joint size, it is evident in Figure 9 that causes belong to various groups of causes (C-A, C-B, C-C, C-D and C-F). The following four probable causes are associated with all these defects: inadequate wood drying (C‑A1), wood cutting not following wood fiber direction (C-A3), wrong design of joints between elements or with protruding elements (C-B1), and differential substrate-WF movements (C-D1). These causes are the most representative in the whole sample of defects on joints and interfaces (D-JI). D-JI3 defect was identified in 18 pavements, with an average of 7.6 probable causes attributed to each one.

Design errors (C-B) are associated with some of the main defects found in the sample. It agrees with the results of Ishak et al. (28), referring to the implications of design deficiency on the building’s service life. Inefficient detailing, inadequate material selection and poor design for maintenance access, along with thermal movement, were considered the main implications of design fault on building maintenance. As mentioned, inefficient detailing affects mainly D.2-S1 Warping, swelling, or other flatness deficiencies, D.2‑S2 Cracking of elements and/or joints and D-JI3 Change of joint size. As for inadequate material selection, it is associated with D.1-S1 Staining or color change and D.1-S3 Scratches or wrinkles. These defects are highly frequent in the sample (D.1‑S3, D‑JI3 and D.1‑S1), which proves the importance of careful design, considering the dire consequences.

3.4. WF diagnostic methodsTOP

The inspection system of WF includes a proposal of inspection methods to better diagnose the detected defects (1). In some cases, these methods may be useful to describe the severity, extension and evolution of defects, which otherwise could not be assessed. They may also help to confirm the causes of defects. Eleven in situ diagnostic methods were considered, organized in four groups. In the inspection program, for each detected defect, an average of 1.7 diagnostic methods were recommended, taking the classification list of defects into account. Only 7% of the diagnostic methods indicated correspond to destructive tests, which is an acceptable number. Non-destructive diagnostic methods are preferred as they preserve the integrity of the cladding and, generally, have a lower cost, although sometimes only qualitative results are obtained.

The most prescribed method was T.1-ND3 Evaluation of other singularities. It was recommended for more than 50% of the detected defects (Figure 10). This diagnostic method associates the analyzed defect with other defects or with temporary environmental circumstances that may help to understand the pathological phenomena. Technically and financially, T.1-ND3 is a useful low-demanding method.

The diagnostic methods T.2-ND3 Moisture measurement (surface and/or environmental) and T.2-ND4 Temperature measurement (surface and/or environmental) were prescribed for more than 20% of the defects. As wood is a raw material sensitive to environmental conditions, it makes sense that temperature and moisture conditions are evaluated to diagnose the origin of an anomalous symptom.

3.5. WF repair techniquesTOP

For each defect identified in the inspection campaign an average of 1.5 repair techniques were prescribed. There are situations in which more than one technique is viable, the choice depending on costs and efficiency. Furthermore, the combination of repair techniques can be the best procedure in some cases, namely when preventive and curative actions should be considered. Moreover, 42% of the repair techniques were preventive (RP), i.e. they intend to eliminate the causes, and 58% were curative (RC), i.e. they intend to eliminate the defects.

Considering the sample of 98 wood floors, each floor should be subject to works considering an average of 3.6 repair techniques, since each floor presents an average of 2.44 defects. Repairs should be planned considering the floor as a whole, analyzing all defects and causes, and combining all necessary operations in a multiple-step intervention.

Observing Figure 11, there are two repair techniques that stand out. RC-B1 Removal/replacement of finishing layer was prescribed for 45% of the defects and RP-B1 Application of suitable finishing layer was prescribed for 36% of the defects. Both techniques refer to the WF surface, specifically to the finishing layer. So, as expected, in 31% of defects, both techniques were recommended, as their combination seems to be the correct approach. They were prescribed for all defects identified as D.1-S4 Wearing or detachment of the finishing layer, and they were also considered appropriate for 74% of D.1-S1 Staining or color change defects. RC-B1 was prescribed to 81% of D.1-S3 Scratches or wrinkles defects.

Nejad et al. (29) tested coated heat-treated and untreated wood samples used in flooring solutions. Scratch resistance was better in heat-treated samples, probably due to a shallow penetration of the coating, leading to a higher dry film thickness. Heat-treated coated wood samples also had a significantly lower color change after house-hold chemical tests. The application of an adequate finishing layer, combined with a proper wood treatment, may prevent scratches defects on WF, as well as stains or color changes.

As for other representative defects in the sample, RC-A6 Partial or total WF replacement and RC-C1 Partial or total replacement of bedding material were prescribed for all D.2-S4 Detachment of elements from substrate defects.

RC-A6 Partial or total WF replacement was considered suitable for all D.2-S3 Broken or splintered elements defects. Repair techniques RC-A3 Reinforcement with metal elements and RC-B2 Removal/replacement of corroded metal elements are also considered appropriate for 86% of D.2-S3 defects.

As for D.2-S6 Moisture, repair techniques RP-A2 Protection against physical-chemical agents (waterproofing and fire protection products) and RP-C1 Repair of defects in pipes are considered the most suitable, as they were prescribed for 75% and 83% of these defects, respectively.

An example of a relationship between defects and repair techniques in WF is shown in Table 1. This table includes the curative and preventive repair techniques that may be used in interventions in defect D.2-S5 - Rot (the techniques with a high correlation are underlined).

4. CONCLUSIONTOP

The analysis presented allowed outlining the degradation profile of a statistically significant sample of WF in existing buildings. Such an analysis is very important for suppliers, designers, installers and users, as all need to learn from previous mistakes in order to improve the construction quality of WF.

Different expert systems for the evaluation of defects in buildings already exist, but, in terms of scope and aim, no similar research work has been found in the specialized literature. Thus, the data analysis presented in this paper is innovative and robust, as it is supported by a statistically significant validation work. The inspection program included 98 wood floorings (WF) with defects, a sample diverse in age, building typology and WF system. It was essential to validate the inspection procedure and confirm the initial theoretical presuppositions, while developing the inspection and diagnosis system. It is expected that the systematic approach used in the presented inspection campaign can help future inspection procedures to be more objective and, at the same time, standardized, based on a consistent working database for the prevention and repair of similar defects.

The analysis of the collected data highlighted the most frequent defects and its causes. Scratches and wrinkles were the most detected defect, highly associated with external mechanical actions. In fact, external mechanical actions and execution errors were the main causes of detected defects. In this context, it is clear that more attention needs to be paid to a good choice of materials, namely finishes need to be adequate to use requirements and the fixing/fastening system needs to be carefully considered at an early stage. At the production phase, wood needs to be dried properly, and while in use, WF need regular specific maintenance.

Further collection of data on WF defects will enable building a wider sample and precise degradation models.

ACKNOWLEDGEMENTSTOP

The authors gratefully acknowledge the support of CERIS-ICIST from IST, University of Lisbon, and FCT, Foundation for Science and Technology.

REFERENCESTOP