The assessment of a building with damage has often been compared to the approach to a patient in health care. In both professions, the most important, but at the same time most difficult phase is the diagnosis. Despite the resemblances with medicine practice, the field of building pathology still lacks an 'encyclopaedia of diseases'. Although some atlases and expert systems are available (e.g. MDDS), especially for structural damage we still miss a consistent way of describing the processes leading to this damage, including a coherent use of a well-defined terminology.
Damage is the manifestation of a lack of performance of a building. Thus, the assessment of damage can be regarded as an evaluation of the present and expected loss of performance, combined with proposals for improvement or repair. Diagnosis is an essential part of this assessment: its aim is not only to indicate what has caused the damage, but, moreover, to try and understand what initiated the damaging process (root cause) and how this process has developed in time. In medicine, this understanding of the course of a disease is indispensable for finding ways to treat symptoms, to cure a patient of his illness, and, if possible, to prevent a disease from affecting other people as well.
The process of diagnosing damage is a deductive process. It is based on the evaluation of the available data on symptoms and context of the damage, and of the investigator's knowledge on damaging processes. This leads to the formulation of one or more hypotheses, which, subsequently, are verified until the cause of damage is diagnosed with a certain accuracy. For a sound diagnosis, the investigator needs thorough insight into all processes possibly leading to damage, and also the ability to distinguish between these processes. This is complicated by the fact that damage often results from several causes, rather than from just one. In its Recommendations, the ICOMOS International Scientific Committee for Analysis and Restoration of Structures of Architectural Heritage (ISCARSAH) (2005) states: 'This is why intuition and experience are essential components in the diagnostic process.'
A correct diagnosis of the damaging process and its root cause is necessary for the proper assessment of damage, and the importance of this correctness has even more increased since we aim at minimal interventions. Therefore, the best way to improve the assessment of damage is by improving the accuracy of the diagnosis. There are two ways to achieve this: either by improving the quantity and/or quality of the information available to the surveyor; or by improving the qualities of the surveyor in processing (analysing, interpreting, and evaluating) this information. Please note that the word 'information' is used here for data and knowledge in two domains. On the one hand, it refers to data on a specific case: the symptoms and the context of the damage. On the other hand, it also concerns the more general knowledge of damaging processes, which is essential for the formulation of hypotheses.
In the past decades, many studies have contributed to the improvement and facilitation of the diagnosis of damage. We can classify these studies into five groups, according to their main goal:
- Improvement of the data on damage;
- Improvement of the knowledge on damaging processes;
- Improvement of the analysis of information;
- Improvement of the evaluation of information;
- Improvement of the exchange of information.
Many studies belong to the first group. Their aim is to improve the completeness and quality of data on the symptoms or the context of damage. These studies have resulted in new and better techniques for on-site investigation, for on-site and laboratory testing, and for monitoring. Special attention has been paid to the development of a range of non-destructive tests (Binda et al. 2000, Binda & Saisi 2002). A good overview of the techniques currently available is given in the documentation of the Onsiteformasonry project (2005).
Tragedies such as the sudden collapse of the civic tower of Pavia in 1989, the Umbria-Marche earthquake of 1997 and the Molise earthquake of 2002 have led to a second group of research projects. This group has aimed at extending our knowledge on specific causes of damage. Especially the creep behaviour of masonry and the effects of seismic actions on buildings have been investigated (Ignoul et al. 2006, Binda etal. 1992).
The third group of studies has focused on the development of techniques to help the investigator to analyse the present behaviour of a structure and interpret the way damage has occurred. The growing possibilities in computation have led to a range of methods and strategies such as various methods for numerical modelling (Lourengo 2002). For the analysis of collapse mechanisms due to seismic action, the macro-element approach has commonly been used (Doglioni et al. 1994). One of the applications of this approach can be found in the digital, didactic handbook MEDEA (Manuale di Esercitazioni sul Danno Ed Agibilita), which gives an overview of damage patterns in masonry and concrete structures that have suffered from an earthquake (Papa & Zuccaro 2004).
Furthermore, a fourth group has proposed methods to facilitate the evaluation of information. One major aspect in this is the evaluation of the vulnerability of structures, with the aim to predict their future behaviour and the related risks (Augusti et al. 2001, Speranza et al. 2006). The Multi-Hazard Assessment of Vulnerability method (MHAV (historic buildings)) uses the macro-element approach to link building typologies to their characteristic vulnerabilities to natural hazards such as earthquakes, storms, and floods. The method evaluates the expected loss of both material and cultural significance (D'Ayala et al. 2006).
Finally, a fifth, smaller group of studies has focused on the improvement of information exchange. These studies point out that a basic need is still to be fulfilled: the need for better communication between all parties involved in a restoration project. Suggestions have been made to improve the exchange of data by using a consistent terminology and standardised survey forms (Van Balen 2001, Kelley & Sparks 2006).
The difficulties in information exchange mentioned above are part of a larger problem: they show the limitations that are inherent to our traditional diagnostic process. The development of new techniques and methods for data collection, analysis and evaluation has certainly contributed to the improvement of diagnosis. However, less attention has been paid to the improvement of the diagnostic process itself and to a more efficient, practicable distribution of knowledge. This paper discusses the limitations of the current intuitive approach, and suggests a method to increase its quality and efficiency.
2 DIAGNOSING DAMAGE: LIMITATIONS OF EXPERIENCE AND INTUITION
The current approach to diagnosing damage is, for most experts, an intuitive one. The deduction is a mental process that strongly depends on the intuition and the experience of the investigator. These abilities are highly valued, but the subjectivity of the approach also brings about some inherent limitations.
Our brains are fast processors, and they are well-trained in finding references. When diagnosing damage, we use this ability to evaluate and compare a new situation with our personal experience. It helps us to judge and weigh different types of information, and it makes us capable of readily formulating the most probable hypothesis.
However, despite its benefits, this method has some disadvantages. First of all, we select our references more or less at random. This is a quick way to retrieve our knowledge, but it is also subjective, irreproducible, and sometimes inconsistent.
Then, experience can also make us biased. It could make us jump to a conclusion by paying more attention to facts that support our initial ideas, while we overlook facts that are contrary. In this way, we would fail to notice alternative hypotheses that could explain the type of damage as well.
Moreover, our intuitive approach may keep us unaware of the assumptions we make during the deduction, wittingly and unwittingly. Because communicating this approach is difficult, the facts and suppositions on which the diagnosis is based may remain veiled to others. This makes it hard to judge the reliability of a diagnosis, since the conditions under which the deduction is valid are insufficiently clear.
To support and improve the diagnostic process, we need to understand how the process of diagnosing works. Therefore, a questionnaire was held among experts. This questionnaire contained pictures of four buildings with visible damage, and for each case the experts were asked to answer the following questions:
- What could be the cause of this damage?
- On what information do you base this hypothesis?
- What data would you need to be able to verify your hypothesis?
- If no initial hypothesis could be established on the basis of the photograph(s), what would be the next step(s) in the investigation process?
The results of this questionnaire were surprising. On the basis of a first impression, the experts have found different clues, made different assumptions, and suggested a range of possible causes. In the next paragraphs, the results of two of these cases are discussed (Figs 1-5).
The damage in the sidewall of building A, as shown in Figure 1, forms a typical damage pattern. Most
experts immediately referred to the classic 'back-of-an-envelope' yield line pattern, which is the failure mode of a rectangular slab, supported around its perimeter, and uniformly loaded perpendicular to its plane. The crack pattern, thus, gives a clear indication of the collapse mechanism. In accordance to the plate bending mode, the loading must have been horizontal, out-of-plane, and directed outwards. In this way, the symptoms of the damage led to hypotheses such as wind load (wind suction) and horizontal impact load (explosion), but also hindered volume changes of a concrete floor incorporated into the wall.
The experts used the context of the damage to exclude some specific causes. The local character of the damage and its distance to the ground led them to omit the possibility of settlement as initiator. Furthermore, some experts (correctly) assumed the building to be located in the Netherlands, which made them conclude that an earthquake would be a less likely cause.
Figure 2 gives an overview of the different hypotheses and anti-hypotheses that were brought forward by the participants. It also shows the arguments and the assumptions that the experts used to frame and support their hypotheses. It appears that, despite the typicality of the damage pattern, the cause of damage is not so obvious. Therefore, in response to the third question, experts asked for more data on the properties of the building, on its construction history, on the nature of the location, and on the characteristics and distribution of the damage.
The damage in building B is less evident than the damage in building A. It can be described as an inplane deformation of the façade (Fig. 3). The skewing
of the windows at the first floor, visible in Figure 4, is accompanied by severe crack development in the lintels and sills of the windows, and in the arch above the door opening. The edges of the cracks show displacements and demonstrate that the parts on either side of the fracture have undergone a clockwise rotation.
Concerning the information on the context of the damage, the participants particularly focused on the construction of the building. For example, they pointed out the difference between the timber-framed construction of the sidewall, and the stonework façade. In addition, the relatively large openings in the façade, the apparent age of the building, and speculations on the adjacent building on the left side and the open area on the right side led them to the assumption that building B may have little or no lateral stability. This assumption helped them to explain the lateral displacements in the façade.
Figure 5 shows that, in this case, both the symptoms and the context of the damage were used to formulate hypotheses. These hypotheses vary from environment-related processes such as differential settlement, to overloading due to a lack of horizontal restraint, and hindered volume changes due to corrosion of the anchors. The hypothesis of differential settlement appears to be supported by many arguments and assumptions. Nevertheless, its root cause remains unclear, although several options were expressed.
Summarising, the results of this questionnaire demonstrate the benefits of our current approach. Even on the basis of only some photographs, experts have shown that their skill and experience make them able to formulate a hypothesis that could explain the damage.
The use of photographs in this questionnaire did hinder the evaluation of damage and context. Nevertheless, this method has been chosen with precisely this goal: It forced the participants to be more aware of the facts on which they based their hypotheses. In this way, the responses also show the limitations of the intuitive approach.
One of these limitations is that experience can sometimes make experts biased. This could be overcome by joining our experience, as the range of
alternative hypotheses resulting from the combination of the responses shows.
Furthermore, the results emphasize that our line of reasoning is influenced by unconscious assumptions. For example, most experts interpreted the damage in building A as caused by out-of-plane loading of a slab supported around its perimeter. Apparently, they assumed that the perimeter of the crack pattern matches with the borders of a room behind the wall. Yet, in reality the floor lies at the level of the horizontal crack, which omits the possibility of wind suction and explosions as a cause, and increases the likeliness of hindered volume changes of the floor itself as a hypothesis. It is clear that assumptions like these should be mentioned explicitly when formulating a hypothesis.
Based on the results of the questionnaire, we can now make some suggestions to improve the current approach to diagnosing:
- Pay more attention to alternative hypotheses;
- Place more emphasis on explicitly mentioning the facts and assumptions on which the diagnosis is based;
- Find ways to join and share our experience; and
- To obtain these goals, our knowledge needs to be accessible, retrievable, and communicable.
Was this article helpful?