An Application of Fuzzy Logic and ANFIS in Intelligent Fault Detection and Localization in Medium-Voltage Networks

Accurate and timely fault diagnosis in medium-voltage networks is essential for enhancing the reliability and operational resilience of power systems. Traditional protection schemes, while fast, often struggle with accuracy in the presence of high fault resistance, dynamic load variations, and unsynchronized measurements. This paper presents a hybrid intelligent framework that integrates a fuzzy logic-based module for fault detection and classification with an Adaptive Neuro-Fuzzy Inference System (ANFIS) for fault location estimation. The proposed method utilizes simple voltage and current indicators as input features and does not rely on GPS synchronization or extensive training datasets. A rule-based fuzzy inference system ensures interpretability and robustness, whereas ANFIS provides accurate fault distance estimation. The system is implemented in MATLAB/Simulink and validated under various fault scenarios. Simulation results demonstrate that the proposed approach can accurately detect, classify, and localize different types of faults, making it suitable for real-time protection in conventional substations and resource-constrained environments.