Application of Bayesian networks for geotechnical risk assessment: case of the Ghazaouet highway viaduct (Algeria)

This paper analyses the subsoil-related uncertainties for a better management of geotechnical risks in the framework of the viaduct project of Ghazaouet (Algeria). To achieve this, we used Bayesian networks as they integrate a systemic approach that facilitates risks identification and relations between different components of the system. From a methodological perspective, we first defined the context of our analysis. Subsequently, our work was decomposed using a systemic approach supported by a functional analysis. The steps followed aimed at identifying the hazardous elements and analysing them and then designing the undesired event logic trees and building the Bayesian networks, while associating the probabilities in each node. Thus, we were able to highlight the relevance of the use of Bayesian networks in the risk management of viaduct projects. The proposed method makes it possible to treat the uncertainties related to the geotechnical risk, by representing the latter as a probability scenario. This representation of the random variables and the relationships between them makes it possible to visualise the different scenarios and determine which ones are critical. The application of this method enabled us to highlight two major scenarios, namely stop activities and the accident/injury/death. The probability of these two scenarios is high, and this is due to the weak or non-existent risk management system within the project. Thus, we were able to highlight the relevance of the use of Bayesian networks in the risk management of viaduct projects in Algeria, although further studies and work are underway for validation.