Keywords: fuzzy regression model, regression analysis, cause-effect relation, building collapse, failure identification, sample data, environmental hazards.

Building collapse is of concern throughout the lifecycle of constructed facilities. The occupants of the building, whether they are construction workers, office workers, residents, clerks, shoppers, and so forth, are most severely affected by building collapse. Numerous accounts are given of loss of life due to building collapse including, most notably, that of the collapse of the World Trade Centre twin towers in New York City.

The occupants and owners of surrounding property also are adversely affected independent of whether the collapse takes place early, during the construction phase of the building, during the service life of the building, or during demolition or deconstruction.

Much about building collapse can be learned from studies of controlled demolition. Controlled demolition by implosion is similar to building collapse. Researchers conducted air quality studies immediately after the demolition of a 22-story building by controlled implosion in Baltimore, Maryland. At sampling sites within a four block radius the concentration of airborne dust particles was as much as 3,000 times higher than they had been prior to the collapse. At seven and a half blocks away the concentration was 20 times higher than normal. The elevated conditions only existed for approximately 15-20 minutes. At locations upwind from the site and at indoor sampling locations the dust concentrations were at normal levels. Dust, which is defined as microscopic particles of solid material, can contain many hazardous and carcinogenic substances. For instance, the collapse of a concrete structure can introduce bits of crystalline silica into the air. Inhalation of crystalline silica can lead to the non-reversible lung disease silicosis, which scars the lungs and reduces oxygen capacity. Depending on the age of the structure, asbestos can be released into the atmosphere, possibly resulting in asbestosis, another lung scarring disease. Other items in dust might include pulverized glass, fragments of sprayed-on fire resistive material, paint (possibly lead containing), and lime from gypsum wallboard.

Dust eventually settles out of the atmosphere to the earth and surface waters. The potential now exists for the once airborne particulates to enter the food chain by deposition in agricultural soil and in drinking water sources. The deposition of lead, a highly toxic substance, has proven especially persistent in urban areas. Lead poisoning is especially harmful to children under 6 years-old, causing, among other ailments, learning disabilities, attention deficit disorder, and kidney damage. Dust deposition also upsets the balance of wildlife, plants, flora and fauna when introduced into ecological environments.

The repercussions of building collapse at each stage of the building lifecycle present their own unique concerns and safety challenges. This paper examines building collapse during construction and describes software aimed at assessing the risk to contractors of building collapse.

It is first necessary to investigate all the possible causes contributing to building collapse and interrelationships among these possible causes. Methods of regression analysis are commonly used to build a model using collected data containing uncertainties and to obtain a prediction equation for the building collapse.

This paper describes a fuzzy regression model to examine the variables contributing to building collapses and investigate their cause effect relation based on historical building accident data in Taiwan.

Fuzzy regression is used because most of the information regarding the causes of building collapse come from expert opinion and are inherently qualitative rather than quantitative. In conventional regression analysis, the errors between observed data and a regression model are regarded as a random variable, whereas the errors are viewed as the fuzziness of the model structure in fuzzy regression analysis. Ordinary regression can only fit crisp data, whereas fuzzy regression can be used to fit both fuzzy data and crisp data. Uncertainties in building collapse data can be mainly attributed to ambiguity and vagueness in defining the variables and their relations.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £95 +P&P)