Three-Phase Electric Motor Isolation Fault Detection Modeling Caused Unbalance Voltage Using Radial Basis Function Network
(a) Alfin Sahrin* (b) Mauridhi Hery Purnomo
(a) Politeknik Energi dan Mineral Akamigas + alfin.sahrin[at]esdm.go.id
(b) Institut Teknologi Sepuluh Nopember + hery[at]ee.its.ac.id
Abstract
One of the fault detection of three-phase electric motor isolation is caused by unbalanced voltage. The National Electrical Manufacturers Association (NEMA) standard for unbalanced voltage is 1%. Fault detection of electric motor isolation is too complex to be modeled in mathematical form. Therefore, this paper proposes to model the electric motor isolation caused by unbalanced voltage using the Radial Basis Function Network (RBFN). RBFN is implemented for training and testing data with input variations number of unbalanced voltage, rotation speed, running hours and output variations number of insulation resistance, index polarization. The modeling process is comparing the number of neurons and the learning rate to get accuracy and speed of time. The results of modeling using RBFN obtained Mean Square Error (MSE) of 2.11 e-04 and this method is very well applied for fault detection of electric motor isolation.
Keywords: electric motor fault detection; unbalance voltage; radial basis function network
Topic: Electrical Engineering