study of factors related to accidents occuring during the construction phase of oil, gas and petrochemical projects

introduction: construction phase in industries is a dynamic process that is naturally and intrinsically dangerous and as it becomes more complicated, the accidents rate also increases. one should note that without considering a model, one could not obtain useful and reliable information and method to prevent accidents. therefore, to achieve useful methods for preventing accidents, it is desirable to consider a model. the general goal of this study was presentation of a model. a model is the reflection of a fact. in other words, it should be said that the model represents a system or process whose behavior can be predicted. models are therefore used for understanding the behavior of actual terminals and show a theory in the way that covers important variables for describing phenomena and instead, ignore factors of low importance in the expression of those phenomena. methods: this study was a research article conducted in 2004-2005 in the assaluyeh region. data was gathered from accident reports present in security and health records of the projects and also statistics present at the treatment centers. in this study, an analytical model (multi-regression) was presented to describe the impact of effective and deep factors on the possibility of an increase in accidents leading to death, through measurement of the effects of independent variables on the dependent variables. for this purpose, the structure of 50 accidents that led to death were studied along with another 2700 accidents, and after studying the accident reports and related documents, observing operations and equipment, counseling with accident observers and an expert team of managers, supervisors and engineers, and simulation of some accidents, unsafe conditions and functions, mismanagement and use of worn out and defective tools, equipment, devices and machinery were considered as the four independent variables and the job accidents leading to death were considered as dependent variables. the relationship between independent and dependent variables, evaluation of regression coefficients and the test of different models were based on multi-regression analytical model and analyzed using eviews software program. results: the final findings of this study, while specifying the possibility of occurrence of accidents leading to death with the existence of any of the independent variables showed that among the independent variables, unsafe conditions and mismanagement have relatively more important roles to play in the occurrence of accidents leading to death such that in conclusion, these factors have been defined as root causes in the model. conclusion: since the adjusted coefficient determined for the model in this research was 0.99 that is, the specified model could describe 99 percent of changes related to the number of job accidents leading to death and it is only for one percent of other accidents that there was no justified answer. in another words, those causes were not seen in the model. therefore, it could be concluded that this research as compared to similar researches gained more useful results. thus, by omitting or lowering unsafe conditions and mismanagement factors, accidents resulting in deaths can be reduced.