Irahis Rodríguez, Universidad de Carabobo. email : firstname.lastname@example.org, Roberto Alves Universidad Simon Bolivar. email: email@example.com, Victor Barrios, Universidad de Carabobo, Rafael Ortega Universidad de Carabobo
Abstract. This paper deals with the application of Neural Network Techniques forthe detection of rolling-element bearing damage in induction motor monitoring electric power. Three-phase induction motor are the “workhorses” of industry and are the most widely used electrical machines. In an industrialized nation, 70% of industrial process use induction motor. For this reason is very important incipient motor failures detection, avoiding production lost and reducingoperational costs. The main idea about motor damage detection is “motor variables, current, flux, torque, electric power, vary (in particular the spectrum) with respect to the time-varying normal operating conditions of the motor”. Motor Current Signature Analysis (MCSA) is a noninvasive, on line monitoring technique to diagnose faults in three-phase induction motor drives, detecting differences in thecurrent spectrum. Frequency Spectrum Analysis can be an imprecise technique due to introduction of harmonics of the voltage supply, noise, and load variation. Neural Network is a successful technique to solve problems of pattern recognition, because the learning way is independent of noise, load variations, and voltage supply harmonics, and the problem of motor damage detection can be classified likea problem of pattern recognition. For this reason, it decided utilize Neural Network Technique to detect Induction Motor failures, monitoring electric power. It chooses monitoring electric power for increasing the effect of the failure in data training of the neural network.
Motor Induction can operate with asymmetries, like:
• Stator Windings Failures like,interturns short circuit.
• Conexiones anormales del devanado del estator.
• Broken Rotor Bar and End Ring Faults.
• Mechanical Failures, like bearing damage, motor shaft failures, air gap eccentricity.
Operation Asymmetric Motor Induction brings consequences like asymmetric flux and current, to increase loss and pulsations torque. This produce poor efficiency and excessive increasetemperatura, producing failure on the marine. Then it is very important detection incipient failures.
Detecting variations in entrance current, torque, flux, when a failure appear in the la machine, you can establish, if it has been a failure in the motor or not. Bearings problems are one major cause for drive failures. Their detection is posible by noise, vibration and temperature monitoring. Theimplementation of these measuring systems is expensive and proves only to be economical in the case of large motors or critical applications. During the past 20 years, there has been a substantial amount of research into the development of new condition monitoring techniques for induction motors. One successful technique is Motor Current Signature Analysis (MCSA), but we make instant electricalpower monitoring, getting satisfactory results.
Failure Diagnosis deals recognition patterns problem, and solution in problems with this characteristics has been gotten in a successful way using Neural Networks. This paper explains a method, based in instant electrical power monitoring, using neural network to establish if it is producing a failure bearing. Neural Network makeincipient failure motor detection since normal conditions and failure conditions training data, detecting changes in the instant electrical power frequency spectrum patterns. Neural Networks learns entrance power specific patterns and after monitoring in several operating conditions (normal and failures conditions), can determine which condition is.
Later we are going to describe Power Equipment,...