Occupational health, safety and environmental risk assessment in textile production industry through a Bayesian BWM-VIKOR approach

Abstract

Occupational risk assessment (ORA) is a process that consists of evaluating, ranking, and classifying the hazards and associated risks arising in any workplace from the viewpoint of occupational health and safety. Many ORA methods have been proposed in the literature, from a single independent expert to participatory methodologies made by group decision and simple to complex ones. In this paper, a holistic ORA is presented, which uses two important multi-attribute decision methods named Bayesian Best-Worst Method (Bayesian BWM) and VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR). Bayesian BWM is used to determine the importance weights of six different assessment criteria, which are the probability of hazardous event (P), frequency (F), severity (S), detectability (D), cost (C) and sensitivity not to use personal protective equipment (SNP). Since the classical BWM finds solution to the weights of a number of criteria from only one expert’s judgment, Bayesian BWM is preferred in this paper (1) to enable participation of a group of experts, (2) to aggregate the preferences of these multiple experts into consensus without loss of information and (3) to follow a probabilistic way for solving the ORA problem. The hazards are then ranked by VIKOR. The approach is implemented in the ORA process of a textile production plant. Results of risk analysis showed that electricity hazard and associated risks constitute the highest risk ratings. These hazards arise from the product, process, human and working environment. The associated risks are evaluated, prioritized, and detailed control measures are proposed. This study made comparisons with the classical BWM-VIKOR approach to demonstrate the proposed approach’s difference and practicality. Results can also help practitioners and risk analysts in formulating the improvement measures to increase the overall safety of the working environment further.