Faults classification of induction machine using an improved ant clustering technique

In this paper, a new approach is applied to solve classification problems for the diagnosis of faults in induction motors. This new method finds its origins in works on the unsupervised classification algorithms based on ant clustering and the heuristic principles of the K-means algorithm and the principal components analysis (PCA). The main advantage is that requires no information about the system or about a possible number of classes. The proposed algorithm is evaluated in the Benchmark data set (IRIS) and applied to the diagnosis of a squirrel-cage induction motor of 5.5 kW in order to clustering data sets and verify the fault detection capability. The obtained results prove the efficiency of this method for the monitoring of electrical machines.