A Study of Multilayer Perceptron Networks Applied to Classification of Ceramic Insulators Using Ultrasound

Interruptions in the supply of electricity cause numerous losses to consumers, whether residential or industrial and may result fines being imposed on regulatory agency’s concessionaire. In Brazil, electrical transmission distribution systems cover a large territorial area, because they are usually outdoors, exposed environmental variations. this context, periodic inspections carried out networks, ultrasound equipment is widely used, due non-destructive analysis characteristics. Ultrasonic inspection allows identification defective insulators based signal interpreted by an operator. This task fundamentally depends operator’s experience interpretation. way, it intended test machine learning applications interpret signals obtained from grid insulators, distribution, class 25 kV. Currently, research area uses several models artificial intelligence for various types evaluation. paper studies Multilayer Perceptron networks’ application classification different conditions ceramic restricted database ultrasonic recorded laboratory.