Development of a fuzzy qualitative risk assessment model applied to construction industry

Dissertação para obtenção do Grau de Doutor em Engenharia Industrial

The construction industry is plagued by occupational risky situations and poor working conditions. Risk Assessment for Occupational Safety (RAOS) is the first and key step to achieve adequate safety levels, particularly to support decision-making in safety programs. Most construction safety efforts are applied informally under the premise that simply allocating more resources to safety management will improve safety on site. Moreover, there are many traditional methods to address RAOS, but few have been adapted and validated for use in the construction industry, thus producing poor results. The contribution of this dissertation is a qualitative fuzzy RAOS model, tailored for the construction industry, named QRAM (Qualitative Risk Assessment Model). QRAM is based on four dimensions: Safety Climate Adequacy, (work accidents) Severity Factors, (work accidents) Possibility Factors and Safety Barriers Effectiveness. The risk assessment is based on real data collected by observation of reality, interviews with workers, foreman and engineers and consultation of site documents (working procedures, reports of work accident investigation, etc.), avoiding the use of data obtained by statistical tecnhiques. To rating each parameter it was defined qualitative evaluators - linguistic variables - which allow to perform a user-friendly knowledge elicitation. QRAM was, firstly evaluated by “peer” review, with 12 safety experts from Brazil (2), Bulgaria (1), Greece (3), Turkey (3) and Portugal (3), and then, evaluated by comparing QRAM with other RAOS tecnhiques and methods. The safety experts , concluded that: a) QRAM is a versatile tool for occupational safety risk assessment on construction sites; b) the specific checklists for knowledge elicitation are a good decision aid and, c) the use of linguistic variables is a better way to make the risk assessments process more objective and reliable.

Fundação para a Ciência e Tecnologia - PhD Scholarship SFRH/BD/39610/2007