This thesis proposes a methodology for fault location in distribution systems, focusing on the stages of segmentation and detection, location, and cause characterization. The methodology aims to improve the efficiency and reliability of fault location, which is crucial for maintaining service quality and reducing maintenance costs. The thesis is structured into several chapters: 1. **Generalities of Fault Analysis**: This chapter introduces the characteristics of faults in distribution systems and the process of network operation during a fault, including detection, location, and cause analysis. 2. **State of the Art**: This chapter reviews existing methods for fault detection, location, and cause characterization in distribution systems. It covers techniques such as Fourier transform, wavelet transform, Kalman filter, Hilbert-Huang transform, and vectorial transformation. 3. **Fault Segmentation**: This chapter presents a methodology for segmenting faults, identifying transient and stable states from instantaneous voltage and current values. The proposed method is validated with rectangular waveforms and events with fast and slow transitions. 4. **Fault Location**: This chapter describes a metaheuristic optimization method for fault location, using measured voltage and current values to compare with calculated values. The method is validated through simulations on an industrial distribution network with different sources. 5. **Cause Characterization**: This chapter explores the characterization of fault causes, aiming to classify fault events based on contact with trees, animals, or lightning discharges. The analysis is based on real event records and oscillograms, using machine learning techniques for classification. The thesis concludes with a discussion of the results and future research directions, emphasizing the importance of adapting fault location methods to the evolving nature of distribution networks, including the integration of distributed generation and smart grid technologies.This thesis proposes a methodology for fault location in distribution systems, focusing on the stages of segmentation and detection, location, and cause characterization. The methodology aims to improve the efficiency and reliability of fault location, which is crucial for maintaining service quality and reducing maintenance costs. The thesis is structured into several chapters: 1. **Generalities of Fault Analysis**: This chapter introduces the characteristics of faults in distribution systems and the process of network operation during a fault, including detection, location, and cause analysis. 2. **State of the Art**: This chapter reviews existing methods for fault detection, location, and cause characterization in distribution systems. It covers techniques such as Fourier transform, wavelet transform, Kalman filter, Hilbert-Huang transform, and vectorial transformation. 3. **Fault Segmentation**: This chapter presents a methodology for segmenting faults, identifying transient and stable states from instantaneous voltage and current values. The proposed method is validated with rectangular waveforms and events with fast and slow transitions. 4. **Fault Location**: This chapter describes a metaheuristic optimization method for fault location, using measured voltage and current values to compare with calculated values. The method is validated through simulations on an industrial distribution network with different sources. 5. **Cause Characterization**: This chapter explores the characterization of fault causes, aiming to classify fault events based on contact with trees, animals, or lightning discharges. The analysis is based on real event records and oscillograms, using machine learning techniques for classification. The thesis concludes with a discussion of the results and future research directions, emphasizing the importance of adapting fault location methods to the evolving nature of distribution networks, including the integration of distributed generation and smart grid technologies.