Developing a risk management framework for agricultural water systems using Fuzzy Dynamic Bayesian Networks and decision-making models

Given the various natural and human-caused hazards that threaten the agricultural water distribution process from the main source to farms, establishing a framework to analyze these risks is crucial. This study aims to develop an intelligent risk management framework to help stakeholders devise long-term and sustainable solutions for managing agricultural water systems. First, we developed a Fuzzy Dynamic Bayesian Network (FDBN) model for multi-hazard risk assessment, taking into account the temporal causal interactions between parameters and incorporating fuzzy theory. Next, we defined several risk management scenarios across structural, non-structural, automated control, and integrated methods. These scenarios were implemented in the FDBN model to mitigate the risks associated with the system. Various economic, social, environmental, and technical criteria were considered, and scenarios were ranked using the WASPAS, TOPSIS, and MultiMoora methods. The Copeland approach was used to combine the ranking results. The results showed that automated scenarios, specifically Model Predictive Control (MPC) and Proportional-Integral (PI) controllers, could reduce the system's risk by 11.4% and 9.8%, respectively, and were ranked the highest. The findings of this study and the proposed framework can assist operators in the sustainable planning and management of water systems in light of anticipated threats.