Abstract

MURILLO-SOTO, Luis D.; CALDERON-ARCE, Cindy  and  FIGUEROA-MATA, Geovanni. Electric motor failure detection based on indexes of powers and neural networks. Tecnología en Marcha [online]. 2018, vol.31, n.1, pp.81-97. ISSN 0379-3982.  http://dx.doi.org/10.18845/tm.v31i1.3499.

We present an alternative method to determine anomalous operating conditions in a threephase induction motor, based on a set of dimensionless indices, they calculated from their power losses. These indices are sensitive to variations in the conditions under which the motor operates and will therefore be assessed if, on the whole, they are capable of discriminating between an unbalance of voltage, under-voltage or a thermal variation, among others. In this work the equations necessary to calculate the electrical losses that allow the calculation of the proposed indexes are developed. It is important to note that these indices unify the interpretation of electrical, mechanical and thermal variables with the intention of using them in the design of an automatic fault detection system that operates in real time, through a prototype based on neural networks. In addition, experiments are carried out on a three - phase two - horsepower induction motor, 230 Volt - 60 Hertz, the testbed constructed and the final part of the work graphically depicts the variability of the indices under different conditions Abnormal, which makes it possible to assess their sensitivity and relation to the associated faults.

Keywords : Loss indices; fault detection; induction motor; neural networks.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )