When a failure occurs in a system, the first essential questions concern the location and severity of the fault, as these factors determine appropriate corrective actions and help maintain continuous operation. Although numerous studies address this topic, most rely heavily on deep system knowledge and are tailored to a limited set of predefined faults.

In this research, we propose a generalizable fault-diagnosis model that reduces dependence on expert knowledge and complex analytical procedures, thereby improving system resilience through fast fault localization, severity estimation, and timely intervention. The methodological contribution of this work lies in integrating a fuzzy inference layer with the Random Forest algorithm, enabling the model to combine operator experience with data-driven learning. This hybrid fuzzy–ML structure enhances interpretability through rule-based reasoning while improving robustness in uncertain environments—an approach aligned with recent advances in fuzzy–ML fusion for fault diagnosis. The proposed framework is validated on two quadcopter fault scenarios. The experimental results show that the model provides rapid processing, straightforward implementation, and stable performance under uncertain conditions. These characteristics make the method a practical and accessible tool for both researchers and industrial practitioners.