O uso de redes bayesianas para modelagem e simulação dos fatores de risco no transporte rodoviário de cargas

Abstract

In recent years, quantitative risk analysis in Road Freight Transport (TFR) has been successfully applied in studies to assess the risks to which the chemical industries and other segments give rise. Recently, relevant studies have shown that transporting different modes (on roads, tracks, pipelines and inland waterways) of hazardous materials plays an important role in determining risk. In particular, with regard to road transport carried out by trucks, as it is an important modality for economic development and very common for the handling of various types of cargo, to assess the level of risk of a given activity, it is necessary to determine the severity of index of that risk for each situation that may occur during TFR. In this context, the purpose of this research is to investigate and calculate the risks related to the activity of road freight transport through Bayesian Networks. These models can estimate different risk scenarios in the TFR activity, with a view to allowing greater assertiveness in measuring the level of risk. The computational models were implemented in software from Bayesian Networks and data entry was performed in Excel® spreadsheets establishing a simplified use interface. The methodology adopted for this study is field research based on the participation of specialists and academic sources, as well as the use of a systematic literature review, the application of the Delphi technique and, finally, a Survey. The results showed that through the use of the proposed model, it was possible to have greater assertiveness in choosing the best scenario for carrying out the TFR activity, since it is also possible to identify whether a given scenario can be classified as low, medium or high degree of risk. Thus, the risk prediction method for CRT makes it possible to assess the probability of the occurrence of one or more risk factors during its activity. In the end, the proposed approach contributes to a better understanding of the probability of the most recurrent risk factors in TFR.