Purpose Vibration-based condition monitoring techniques are widely used for diagnosing faults in rotating machines. These techniques are implemented in the time domain, the frequency domain, or both. However, the composite and noisy nature of the raw data collected requires a preprocessing stage such as filtering and decomposition using in-depth processing techniques. Moreover, these methods require good frequency resolution and involve examining a broad frequency range to discern both healthy and faulty cases. In this work, we introduce a simple and fast diagnostic scheme for wind turbine gear teeth wear based on time domain analysis. Methods The proposed method is based on the local minima interpolation of a filtered version of the vibration signal following time synchronous averaging (TSA) technique. Given tachometer signal, the TSA of the vibration data is performed using MTALAB software. Then, local minima of the filtered signal are interpolated using the Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) function. The variance of the interpolated curve built a gear fault index. Results The derived fault index resulting of the proposed technique allows a substantial distinction between the healthy and faulty cases. Its efficiency is validated using 10 real-world datasets of vibration stemmed from a wind turbine planetary gearbox. Conclusion The proposed method boasts a low computation time and ease of interpretation, specifically beneficial for gearbox fault diagnosis purposes.

Induction generator is the subject of various electrical and mechanical fault that involve the wind turbine reliability. As a few papers studied the electrical imbalanced rotor fault in the wound rotor induction generator, we present a MCSA technique result for rotor electrical imbalance fault using Blackman windows with show its improvement for spectral accuracy. As well as the application of signal processing methods such as FFT, EMD and RLMD, in the context of fault diagnosis and highlighting the effect of rotor imbalance on the operation of dfig.

Condition monitoring of gearboxes is mainly based on vibration analysis. Traditionally, spectral lines related to shaft speeds and gear meshing frequency (GMF) are the key components that serve as fault indicators. However, due to the noise level and the dependence of signal energy on operating conditions, this technique can fail. In this article, we present an alternative technique based on the calculation of the area covered by a spectral band around the GMF. This band is observed in the envelope spectrum of the vibration data and covers the modulating components characterizing the appearance of the defects. The experimental results obtained from a gearbox test bench operating under different speeds and load levels show the effectiveness of the proposed fault indicator in the detection of combined faults. Compared to certain other known methods for diagnosing gear faults, the proposed technique shows its superiority in terms of processing time and simplicity of implementation. (c) 2023 Elsevier Ltd. All rights reserved.

Monitoring is a crucial part of an overall process of controlling and supervising systems, as it aims to detect, locate and diagnose faults that may affect their operational safety, performance, and reliability. One of our primary objectives is to develop an effective monitoring tool for transmission systems that use electrical measurements, particularly stator currents, which are already available in machine control devices. In this paper, we propose a dynamic model for a V-belt transmission system to demonstrate the impact of faults on the driving induction motor. Subsequently, we discuss two diagnostic techniques based on spectral analysis of the stator currents: Motor Current Signature Analysis (MCSA), and the Extended Park Vector Approach (EPVA). To evaluate the effectiveness of these techniques, we introduce a side-cut-off fault by 5 cm on the belt. The results of both theoretical and experimental studies revealed that the EPVA exhibits a greater richness in characteristic frequencies associated with belt faults compared to the MCSA method. In addition, EPVA demonstrates higher sensitivity to V-belt defects as evidenced by a large difference in frequency amplitude between the normal and the defect states.

As the wind turbine operates in harsh conditions, numerous of its components are critical and present an important downtime for maintenance. In this paper, we propose a fault diagnosis algorithm to detect and locate the defects affecting the generator rotor and the pinion of the gearbox lay shaft in a real 25 kW wind turbine drivetrain. The induction generator was used as a fault sensor for gear teeth damage. Through the use of the wavelet packet transform, and the local mean decomposition combined with the Fast Fourier Transform, the detection of gear meshing frequency in the stator current reflects teeth faults. Hence, the principal component analysis of the stator current gives a suitable classification for the gearbox states under different working stages. The obtained results have been significant, despite the use of a short duration and a low sampling frequency of the experimental data.

Asynchronous drives are widely used in many industrial applications because of their low cost, high performance and robustness. However, faulty operations may appear during the lifetime of the system. This paper deals with the use of motor current signature analysis (MCSA) as a diagnostic technique for the influence of V-belt drive faults on three-phase induction motors. For this purpose, we create faults on the belt such as axial misalignment and side-cut-off fault.

In this work a fault signature based on the Park Lissajou’s curve is suggested. It was applied to an induction machine driving a centrifugal fan, that operating in various industrial systems. We propose a new index extracted from Park Lissajou’s curve for the detection and localization of unbalanced voltages. In fact, the range ratio is determined to distinguish the two states; normal (balanced voltage) and faulty (unbalanced voltage). The ellipse orientation angle α is calculated to differentiate between the unbalance in phases B and C. This approach is simple, it shows its effectiveness and its robustness in detecting and localizing of studied fault.

Wind turbine is among the renewable electrical energy sources, this source of electricity raises its importance from its continuity and its cleanliness. This justifies the importance of its reliability. In this way, gearbox of wind turbine must be monitored. In this paper, simulated wind system is implemented in laboratory, vibration and generator stator current are sampled from sensors and analyzed through FFT techniques. Resulting spectrums are investigated in order to characterize a healthy gearbox before introducing a tooth fault. Differences between theoretical and practical characteristic frequencies of the gearbox can affect the spectrums interpretation and consequently the diagnosis decision cannot be efficient. This difference is caused by losses in practical frequencies. The present work gives a formulation of the gearbox transmission efficiency to determine the origin of the losses in gearbox practical frequencies.

•A low complexity fault detector based on the EEMD.•A fault detector based on dominant IMF extraction.•Use of Pearson correlation for the closest IMF cancellation.

Even under normal operating conditions, Gear-based systems naturally generate four particular frequencies: the input and output mechanical speeds as well as the gear meshing and the hunting tooth frequencies. Thereby, through amplitude monitoring of these components and their harmonics, the gear state can be easily monitored and successfully assessed. Based on this fact, this paper discusses the fitness of the vibration data and the load torque (mechanical signature) to detect these frequencies. Furthermore, when the system is driven by an induction machine these frequencies will affect the stator currents. Thus, the appropriateness of the spectral analyzing techniques based such amounts as alternate for monitoring gear-based systems will also be discussed. Moreover, for improving the sensitivity detection of these particular frequencies an original preprocessing technique is proposed and its effectiveness is evaluated for spectral analysis of mechanical as well as electrical experimental data.

In this work we discuss the efficiency of various stator current signatures for monitoring shaft misalignment that could affect mechanical systems when driven by an induction machine speed-controlled. In that respect, we have considered the Motor Current Signature Analysis, the Current Space Vector, and the Current Park Loci, under no-load as well as under full load conditions. On the other hand, the effect of the machine's input-frequency on the detection efficiency is also discussed. Thus, we have assessed these three techniques when the motor is operating under full speed's condition and under half speed's condition. The presented results are experimentally validated on a laboratory's test-rig simulating an angular shaft misalignment.

The machine's air-gap cannot be perfectly smooth. Since a static or a dynamic eccentricity occurs, a mixed eccentricity behavior is observed and manifests by amplitude modulation of the stator current. Many other failures related to the motor condition or its driven load produce this same effect. Hence, by loading an induction machine, eccentricities incidence may be masked. Furthermore, because of power supply imperfections, the detection of characteristic frequencies around principal slot harmonics is not always practicable. In this paper, from theoretical development of eccentricities effects on analytical expressions of rotor and stator currents, a machine classification is proposed and related defect frequencies are predicted. The weakness of the motor current signature analysis is experimentally verified, and an alternative based on spectral analysis of the stray flux is proposed. Experimental results have shown an excellent capability of the suggested frequency signature for detecting and distinguishing the factual eccentricity.

Vibration analysis is the most used technique for gearbox fault diagnosis based on gear meshing frequency (GMF) magnitude tracking. The originality of this paper is illustrated in two steps. The meshing between two damaged teeth repeating to the Hunting tooth frequency (HTF) is clearly detected by performing spectral analysis to vibration envelope signal and the motor stator current, unlike the confirmation of [1][2] that this frequency is very low and cannot be measurable. In second step, gearbox input shaft break was predicted by the GMF harmonics magnitude comparison.

This paper deals with a fault detection method based on an empirically data-driven approach combined to a statistical tool. This approach is an enhanced version of the empirical mode decomposition. The proposed fault detector application to bearing defects in wind turbine based on induction generator clearly shows that it is well suited for stationary and non-stationary behavior regardless the rank of the intrinsic mode function introduced by the fault.

This paper deals with the use of the stator current signature analysis as a technique for the diagnostic of static, dynamic and mixed air-gap-eccentricity conditions in working three-phase induction motors. Associated spectra are analyzed and attribute fault harmonics are determined for an induction motor, under several types of air-gap eccentricity conditions. The effect of load levels in air-gap eccentricity fault detection is presented. The experimental results have revealed the potential of the spectral analysis of the stator current for the air-gap eccentricity fault detection.

In this study, we address the topic of monitoring the drive speed of synchronous generators by time-frequency analysis of electromagnetic quantities available at the machine's outputs. In this regard, after recalling the commonly used techniques to deal with signals of scalable spectral content, we present and discuss the detection of an abrupt variation of the drive speed by suitable processing of load currents. Afterward, we propose an alternative technique based on the instantaneous frequency estimation of the stray flux. Through experimental trials conducted on a laboratory test-rig, the effectiveness of the suggested approach has been assessed under various load conditions, and its performance has been judged against those stemming from load currents processing. Promising results have been concluded in terms of generator speed monitoring by the use of a low-cost noninvasive sensor.

Online induction machine faults diagnosis is a concern to guarantee the overall production process efficiency. Nowadays, the industry demands the integration of smart wireless sensors networks (WSN) to improve the fault detection in order to reduce cost, maintenance and power consumption. Induction motors can develop one or more faults at the same time that can produce sever damages. The origin of most recurrent faults in rotary machines is in the components: stator, rotor, bearing and others. This work presents a novel methodology for the online faults diagnosis in induction motors. This technique uses the smart WSN to obtain the machine condition based on the motor stator current analysis. The implementation of the proposed smart sensor methodology allows the system to perform online fault detection in a fully automated way. Simulation results presented show the efficiency of the proposed method to detect simple and multiple faults in induction machine. It provides detailed analysis to address challenges in designing and deploying WSNs in industrial environments, and its reliability.

For economic and environmental reasons, wind turbines are becoming a potential renewable power source that could replace conventional fossil-fuelled plants. In remote areas where the power grid is unavailable, these wind plants may be equipped with self-excited induction generators. Order to maximize their productivity, the generators condition has to be continually monitored. For this purpose, many processing techniques have been interested to the analysis of fluently known signals such that vibration, ultrasound, acoustic emission, temperature, electrical amounts, etc. In this work, we present an innovative approach for monitoring the drive speed of such generator. The proposed technique is based on estimation of the instantaneous frequency related to the signal stemming from a stray flux sensor. Experimental investigations conducted on a laboratory test-rig have shown promising results in terms of speed monitoring by the employ of a low-cost sensor.

Elderly loads cannot impose to the driving rotation process a strict unvarying torque. This truth can occur even with new loads because of inherent manufacturing imperfections. In the majority of circumstances, the load torque varies according to the machine rotor position and these variations, which are not related to the motor health condition, involve an amplitude modulation of the machine stator current at rotational frequency. Thus, harmonics caused by an eventual mechanical unbalance can be overlapped and the user can not distinguish a normal operating condition from a failure mode. Since an induction machine becomes sufficiently loaded, classical diagnosis technique based on the spectral analysis of the stator phase current becomes insufficient to monitor such fault incidence; in this paper we suggest an alternative diagnosis practice based on the analysis of the complex cepstrum. Simulation and experimental results show the effectiveness of the proposed method.

In brushless excitation systems, the rotating diodes can experience open- or short-circuits. For a three-phase synchronous generator under no-load, we present theoretical development of effects of diode failures on machine output voltage. Thereby, we expect the spectral response faced with each fault condition, and we propose an original algorithm for state monitoring of rotating diodes. Moreover, given experimental observations of the spectral behavior of stray flux, we propose an alternative technique. Laboratory tests have proven the effectiveness of the proposed methods for detection of fault diodes, even when the generator has been fully loaded. However, their ability to distinguish between cases of diodes interrupted and short-circuited, has been limited to the no-load condition, and certain loads of specific natures.