In this paper, a new induction motor diagnosis methodology is proposed. The approach is based on obtaining a 2-D time-frequency plot representing the time-frequency evolution of the main components in an electrical machine transient current. The identification of characteristic patterns in the time-frequency plane caused by many of the fault-related components enables a reliable machine diagnosis. Unlike other continuous-wavelet-transform-based methods, this work uses frequency B-spline (FBS) wavelets. It is shown that these wavelets enable an efficient filtering in the region neighboring the main frequency, as well as enable a high level of detail in the time-frequency maps. As a consequence, the evolution of the most important current components is precisely traced. These characteristics make it easy to identify the patterns related to the fault components. The technique is applied to the experimental no-load start-up current of motors in a healthy state and with broken bars; the FBS capabilities are revealed. One of the novelties of this paper is the fact that the diagnosis is carried out via the identification not only of the traditional lower sideband harmonic but also of the upper sideband harmonic and four additional fault-related components.

During the last years, several time-frequency decomposition tools have been applied for the diagnosis of induction motors, for those cases in which the traditional procedures, such as motor current signature analysis, cannot yield the necessary response. Among them, the Cohen distributions have been widely selected to study transient and even stationary operation due to their high-resolution and detailed information provided at all frequencies. Their main drawback, the cross-terms, has been tackled either modifying the distribution, or carrying out a pretreatment of the signal before computing its time-frequency decomposition. In this paper, a filtering process is proposed that uses advanced notch filters in order to remove constant frequency components present in the current of an induction motor, prior to the computation of its distribution, to study rotor asymmetries and mixed eccentricities. In transient operation of machines directly connected to the grid, this procedure effectively eliminates most of the artifacts that have prevented the use of these tools, allowing a wideband analysis and the definition of a precise quantification parameter able to follow the evolution of their state.

The diagnosis of induction motors through the spectral analysis of the stator current allows for the online identification of different types of faults. One of the major difficulties of this method is the strong influence of the mains component of the current, whose leakage can hide fault harmonics, especially when the machine is working at very low slip. In this paper, a new method for demodulating the stator current prior to its spectral analysis is proposed, using the Teager-Kaiser energy operator. This method is able to remove the mains component of the current with an extremely low usage of computer resources, because it operates just on three consecutive samples of the current. Besides, this operator is also capable of increasing the signal-to-noise ratio of the spectrum, sharpening the spectral peaks that reveal the presence of the faults. The proposed method has been deployed to a PC-based offline diagnosis system and tested on commercial induction motors with broken bars, mixed eccentricity, and single-point bearing faults. The diagnostic results are compared with those obtained through the conventional motor current signature analysis method.

The stray flux that is present in the vicinity of an induction motor is a very interesting information source to detect several types of failures in these machines. The analysis of this quantity can be employed, in some cases, as a supportive tool to complement the diagnosis provided by other quantities. In other cases, when no other motor quantities are available, stray flux analysis can become one of the few alternatives to evaluate the motor condition. Its non-invasive nature, low cost and easy implementation makes it a very interesting option that requires further investigation. The aim of this work is to evaluate the suitability of the stray flux analysis under the starting transient as a way to detect certain faults in induction motors (broken rotor bars and misalignments), even when these types of faults coexist in the motor. To this end, advanced signal processing tools will be applied. Several positions of the flux sensors are considered in this study. Also, for the first time, a fault indicator based on the stray flux analysis under the starting is introduced and its sensitivity is compared versus other indicators relying on other quantities. It must be emphasized that, since the capture of the transient and steady-state flux signals can be carried out in the same measurement, the application of the approach presented in this work is straightforward and its derived information may become crucial for the diagnosis of some faults.

International audienceMotor failure due to bearing currents is an issue that has drawn an increasing industrial interest over recent years. Bearing currents usually appear in motors operated by variable frequency drives (VFD); these drives may lead to common voltage modes which cause currents induced in the motor shaft that are discharged through the bearings. The presence of these currents may lead to the motor bearing failure only few months after system startup. Vibration monitoring is one of the most common ways for detecting bearing damages caused by circulating currents; the evaluation of the amplitudes of well-known characteristic components in the vibration Fourier spectrum that are associated with race, ball or cage defects enables to evaluate the bearing condition and, hence, to identify an eventual damage due to bearing currents. However, the inherent constraints of the Fourier transform may complicate the detection of the progressive bearing degradation; for instance, in some cases, other frequency components may mask or be confused with bearing defect-related while, in other cases, the analysis may not be suitable due to the eventual non-stationary nature of the captured vibration signals. Moreover, the fact that this analysis implies to lose the time-dimension limits the amount of information obtained from this technique. This work proposes the use of time-frequency (T-F) transforms to analyse vibration data in motors affected by bearing currents. The experimental results obtained in real machines show that the vibration analysis via T-F tools may provide significant advantages for the detection of bearing current damages; among other, these techniques enable to visualize the progressive degradation of the bearing while providing an effective discrimination versus other components that are not related with the fault. Moreover, their application is valid regardless of the operation regime of the machine. Both factors confirm the robustness and reliability of these tools that may be an interesting alternative for detecting this type of failure in induction motors

There exists no single quantity able to diagnose all possible failures taking place in induction motors. Currents and vibrations monitoring are rather common in the industry, but each of these quantities alone can only detect some specific failures. Moreover, even for the specific faults that a quantity is supposed to detect, many problems may rise. As a consequence, a reliable and general diagnosis system cannot rely on a single quantity. On the other hand, it would be desirable to rely on quantities that can be measured in a noninvasive way, which is a crucial requirement in many industrial applications. This paper proposes a twofold method to detect electromechanical failures in induction motors. The method relies on analysis of currents (steady state + transient) combined with analysis of infrared data captured by using appropriate cameras. Each of these noninvasive techniques may provide complementary information that may be very useful to diagnose an enough wide range of failures. In the present paper, the detection of three illustrative faults is analyzed: broken rotor bars, cooling system problems and bearing failures. The results show the potential of the methodology that may be particularly suitable for large, expensive motors, where the prevention of eventual failures justifies the costs of such system, due to the catastrophic implications that these unexpected faults may have.