Levy Ely de Lacerda de Oliveira

Detection of Localized Bearing Faults in Induction Machines by Spectral Kurtosis and Envelope Analysis of Stator Current

Early detection of faults in electrical machines, particularly in induction motors, has become necessary and critical in reducing costs by avoiding unexpected and unnecessary maintenance and outages in industrial applications. Additionally, most of these faults are due to problems in bearings. Thus, in this paper, experimental bearing fault detection of a three-phase induction motor is performed by analyzing the squared envelope spectrum of the stator current. Spectral kurtosis-based algorithms, namely, the fast kurtogram and the wavelet kurtogram, are also applied to improve the envelope analysis. Experimental tests are performed, considering outer bearing faults at different stages, and the results are promising.

Predictive Maintenance by Electrical Signature Analysis to Induction Motors

Industries always try to increase the reliability of their productive process. In this context, predictive maintenance performs a fundamental role in order to reach high availability and reliability concerning their pieces of equipment. Predictive maintenance can be understood as the action on the equipment, system or installations based on the previous knowledge about the operation condition or performance, obtained by means of parameters previously determined (Bonaldi et al, 2007).

Induction Motor Efficiency Evaluation Using a New Concept of Stator Resistance

This paper presents an air-gap torque (AGT)-based method for efficiency estimation of induction motors. A new concept of stator resistance that includes the mechanical losses effect is proposed. This new stator resistance is estimated through a particle swarm optimization approach based on the stator flux equations and minimization of torque error at the rated operation point. Then, the obtained stator resistance is used in the AGT equations to estimate the shaft torque and then the efficiency. Moreover, the rotor speed is estimated using induction motor current signature analysis. Thus, the proposed methodology for induction motor efficiency estimation relies only on line currents, line voltages, and nameplate data, being appropriate for in-service applications. Finally, the simulation and experimental results are presented to validate the proposed method at different load conditions.

A stator flux synthesis approach for torque estimation of induction motors using a modified stator resistance considering the losses effect

This work proposes a methodology for induction motor torque estimation based on the stator flux model. The presented methodology just relies on line voltages, line currents and nameplate data; and it adopts a modified stator resistance, which already comprises the mechanical losses effect. This modified stator resistance is estimated through a Particle Swarm Optimization (PSO) algorithm, which is a kind of artificial intelligence applied to optimization problems. The PSO presented objective function aims to minimize the torque error at the rated operation point. Simulation and experimental results validate the effectiveness of the presented methodology.

A system for turbogenerator predictive maintenance based on Electrical Signature Analysis

This paper proposes a prototype to detect incipient faults in non-salient pole synchronous generators based on the Electrical Signature Analysis (ESA) technique. The methodology consists in taking measurements of the machine stator voltages and currents and processing the acquired signals. By analyzing the voltages and currents spectra, it is possible to distinguish a faulty condition from a healthy one and to infer the failure type of the generator. This is summarized in a set of ESA failure patterns, which were proved through tests conducted in a custom made scale model laboratory. The custom laboratory setup consists of a small fault injection capable two-pole synchronous generator driven by an inverter-fed induction motor. The proposed prototype is operating in four turbogenerators of a Brazilian power plant. The main advantages of the presented system are the low intrusiveness, the ease of installation and it is economically viable. Moreover, the ESA failure patterns are based on defined frequencies and the structural features of the machine.

Differential Evolution based Air-Gap Torque method approach for induction motor efficiency estimation

This paper proposes a methodology for torque and efficiency estimation of in-service induction motors based on the Air-Gap Torque (AGT) method. The proposed methodology only relies on induction motor stator voltages, currents and nameplate data. A modified stator resistance is estimated including the mechanical losses effect. This modified stator resistance is estimated by using a Differential Evolution (DE) based algorithm. The DE procedure estimates the modified stator resistance by using the stator flux equations, minimizing the torque error at the rated operation point. The estimated stator resistance is applied in the AGT method equations to compute the induction motor torque and efficiency. Simulation tests were performed to validate the proposed methodology.

An air-gap torque based method for efficiency evaluation using pso to estimate a new concept of stator resistance including the losses effect

This work proposes an air-gap torque (AGT) based method for efficiency and torque estimation of induction motors. A new concept of stator resistance including the mechanical losses effect is presented. This modified stator resistance is estimated through a Particle Swarm Optimization (PSO) approach. The proposed methodology only relies on line currents, line voltages and nameplate data, which is appropriate for in-service applications. Finally, simulation and experimental tests were conducted to validate the proposed method at different load conditions.

Discrimination of Synchronous Machines Rotor Faults in Electrical Signature Analysis Based on Symmetrical Components

Electrical signature analysis (ESA) has been successfully applied to predictive maintenance of synchronous machines. The fault diagnosis is performed by analyzing failure patterns in the current or voltage spectra, which allow the discrimination of a healthy and a faulty condition on the monitored machine. Generally, the failure patterns are a function of the line frequency, rotor rotation frequency, and some structure features of the machine. The rotor rotation frequency pattern, for instance, is indicative of rotor mechanical problems and rotor winding interturn short-circuit. An increase of this frequency component magnitude may indicate the incipience of a rotor fault but may not discriminate the nature of the fault (electrical or mechanical). Thus, it is necessary to distinguish the effects of electrical and mechanical faults in these components in order to get a reliable diagnosis of the machine. No works have been found in the literature approaching this specific issue. This paper proposes a simple and innovative methodology to distinguish the effect of electrical and mechanical faults in the mentioned ESA failure patterns based on the method of symmetrical components. Its effectiveness is validated by experiments performed on a synchronous generator test rig. The proposed condition monitoring system is simple, low cost, and low intrusive, as it only relies on stator electrical quantities. Moreover, it is in operation in a Brazilian 404-MW thermal power station.

Experimental Bearing Fault Detection, Identification, and Prognosis through Spectral Kurtosis and Envelope Spectral Analysis

Abstract Rolling element bearings are machinery elements that are widely used in aerospace and industrial applications to support rotating shafts, thereby reducing mechanical friction and heating. However, these components are subject to several kinds of faults which, in general, can be divided into single-point or localized and distributed faults. A common way to detect ball bearing localized or raceway faults is by identifying and analyzing the so-called characteristic “bearing frequencies.” This article presents an approach to bearing condition-based maintenance which comprises fault detection and identification and condition prognosis, based on spectral kurtosis and squared envelope spectral analysis of stator current. In this work, envelope spectrum analysis, based on spectral kurtosis, is applied to identify these frequencies and to evaluate its applicability as a severity estimation index. To evaluate the performance of this index, experimental tests were carried out considering bearings with outer raceway defects at different stages, simulating different fault severities. The numerical results confirm the applicability of the described methodology for detecting and identifying faults, as well as estimating the generated fault severity levels, suggesting that the proposed fault severity index can be applied for bearing monitoring in real world industrial applications.

Development of a reduced-model laboratory for testing predictive fault system in internal combustion engines

Early diagnosis of a fault is relevant in the industrial environment to increase the availability of critical equipment. In a scenario where the thermal plants have fully functioning for long periods of time, to monitor the condition of generators and their primary machines becomes increasingly imperative. This paper presents the development of a reduced-model laboratory to test a predictive system for diagnosing thermos-mechanical faults in internal combustion engines through electrical signature analysis of voltage and current signals of the generator. This laboratory serves for the identification of patterns of failure in combustion engines and the verification of the system developed for installation in SUAPE II thermal power plant, in Brazil.