Discrimination of faulty compressors and healthy one using Kohonen maps [Kohonen hari·ta kullanarak hatali ve saglikli kompresörleri·n ayriştirilmasi]

Abstract

Kohonen's Self Organizing Map (SOM) is a very usable method for the classification problems. In this work SOM is used to classify the possible faults of Hermetic Compressors which are used in the refrigerators. In refrigerators the main cause of the annoying noise/sound is the compressor sound and different faults have different noise types. Classification of the types of the sounds produces valuable information both in the R/D phase and after the production chain. In order to make this classification, the sound pressure values of different compressors with definitive faults have been collected and the healthy one and the faulty compressors are identified using SOM