Development of Accident Prediction Model for Low Frequency and High Severity (LFHS) Industrial Accidents

Abstract

Back Ground: Industrial safety is a multi-disciplinary subject involving interaction of human, equipment and work environment. It is difficult to design equipment and work environment for each individual as each individual is unique. Therefore, accident control measures should be implemented by safety through design, supplemented by administrative controls and personal protective equipment. In order to implement an effective accident control program, all accidents should be investigated and analyzed to identify significant contributors. Methods: Regression analysis is one of the tools for identification of significant contributors through development of accident prediction models. The paper illustrates multiple linear regression analysis of Low Frequency High Severity (LFHS) accidents occurred in a Public Sector Power Company in India during the period of 10 years from 2006 to 2015. The independent variables of the model are ‘Types of LFHS Accidents’ and dependent variable is man-days loss attributed to LFHS accidents. Results: The significant contributors of the LFHS accidents are ‘Exposure to or Contact with Electric Current’, ‘Fall of Persons from Height’ and ‘Stepping on Striking Against or Struck by Object’. The paper also gives accident control measures to be taken by ‘Safety through Design’, ‘Administrative Controls’ and ‘Personal Protective Equipment’. Conclusion: Effective implementation of these measures will support organization in accident prevention.