Predicción del hurto de energía eléctrica a través del uso de la inteligencia artificial mediante algoritmos de machine Learning para CNEL EP unidad de negocios Santo Domingo

Electrical energy losses are a problem, which have not been completely resolved in electric power systems. These losses may happen during the stages of generation, transmission and distribution of electrical energy. This research project was developed in the area of electrical distribution, with the aim of analyzing the reasons and possible solutions that generate non-technical losses, through the implementation of intelligent software. Technical losses occur from generation and they represent the energy that is not used or that is lost during transmission, sub-transmission and distribution, caused by the Joule effect, Eddy currents (Foucault’s currents) or hysteresis. However, this type of loss cannot be completely eliminated, due to the physicochemical phenomena that occur in ferromagnetic nuclei, heating in the conductors, which cannot be eliminated. Non-technical losses in the distribution of energy are generated due to the following reasons: 1.- Administrative: Users without metering, erroneous readings, inaccurate computer systems and a culture of “non-payment”. 2.- Poor operation: Poor maintenance, poor accuracy of measurement equipment and phase unbalance. 3.- Fraudulent: Unauthorized connections, intervention in the database, meter intervened or targeted, being as such "energy theft". Reducing technical losses implies a large economic investment since a resizing of conductors and transformers should be carried out. Electricity distribution companies have sought to eliminate non-technical losses by implementing anti-theft connections, replacing electromechanical meters with digital meters, but people use any method, device or mechanism to violate those connections and measurement devices, to seek their personal good that is, not paying in full monthly electricity consumption. Currently Artificial Intelligence (AI) is used in a wide range of areas, one of them is the electricity sector. Through AI, a large amount of data obtained in the stages of generation, transmission, sub-transmission and distribution of electric power can be analyzed, and with help of AI algorithms an optimal solution to problems can be found in a logical and reasonable way. xiv This research project was applied in the distribution area, to analyze through Machine Learning (ML) algorithms, the behavior of non-technical losses in the different types of users of the Santo Domingo electrical company with the aim of finding possible offenders (users who steal energy).

• Main Author: Macao Sánchez, Richard Alex (author)
• Other Authors: Pujota Cuasapaz, Edison Javier (author)
• Format: bachelorThesis
• Language: spa
• Published: 2022
• Subjects: Pérdidas técnicas y no técnicas; Inteligencia artificial; Machine learning; Hurto de energía; Algoritmos; Aprendizaje de máquina
• Online Access: https://repositorio.uteq.edu.ec/handle/43000/6618