M. CHERIF Bilal djamal eddine

It is critical to accurately detect IGBT (Insulated Gate Bipolar Transistor) switch faults in order to ensure the reliability and robustness of three-phase inverters. In this work, a new approach for the enhancement of the IGBTs open-circuit faults of a three-phase diagnosis inverter is proposed. This approach is based on an Enhanced version of the Variational Mode Decomposition algorithm (EVMD) combined with wavelet packet analysis (WPA) and scalar indicators such as the average and variance value. Firstly, the three-phase current signals are denoised using an improved version of the denoising technique based on a WPA. Secondly, the denoised current signals are decomposed by EVMD algorithm into Band Limited Intrinsic Mode Functions (BLIMFs). Thirdly, BLIMFs that have the highest correlation coefficients with current signals are selected as useful modes. Fourthly, the variance of the BLIMFs that have the highest correlation coefficient among those selected is calculated in order to detect the faulty phase and, consequently, the faulty arm. Finally, the average of the BLIMF that has the lowest correlation coefficient among those selected for the faulty phase is calculated to localize the fault position in either the upper or the lower IGBTs of the faulty arm. The proposed approach is applied to experimental signals and the results obtained show well the efficiency of the performance of the proposed approach in the diagnosis of open-circuit faults.

To ensure the reliability and robustness of the three-phase inverter, it is important to accurately detect the faults of insulated gate bipolar transistor (IGBT) switches. Signal processing is widely used for monitoring and fault diagnosis of the three-phase inverters. The current signals are usually noisy, which can lead to a loss of fault information. This article proposes a new approach for identifying and detecting the IGBTs open-circuit faults of a three-phase inverter. This approach is based on an improved version of the variational mode decomposition (IVMD) technique associated with correlation coefficients and statistical indicators (variance and mean value). First, the IVMD is applied to the three-phase current signals to obtain an elementary function called the band-limited intrinsic mode functions (BLIMFs). Second, the correlation coefficients between the original signals and theirs BLIMFs are used to select the relevant modes. The variances of the modes that have the highest correlation coefficients among those selected are calculated to detect the faulty phase and, therefore, the faulty leg. Finally, the mean value of the mode that has the lowest correlation coefficients among those selected of the faulty phase is calculated to localize the fault position in either the upper or the lower IGBTs of the faulty leg. Different experimental data are used to test the effectiveness of the proposed approach. The results obtained show the merits of the proposed approach for the diagnosis and detection of open-circuit IGBTs compared to the conventional method.

Vibration analysis for conditional preventive maintenance is an essential tool for the industry. The vibration
signals sensored, collected and analyzed can provide information about the state of an induction motor.
Appropriate processing of these vibratory signals leads to define a normal or abnormal state of the whole
rotating machinery, or in particular, one of its components. The main objective of this paper is to propose
a method for automatic monitoring of bearing components condition of an induction motor. The proposed
method is based on two approaches with one based on signal processing using the Hilbert spectral envelope
and the other approach uses machine learning based on random forests. The Hilbert spectral envelope allows
the extraction of frequency characteristics that are considered as new features entering the classifier. The
frequencies chosen as features are determined from a proportional variation of their amplitudes with the
variation of the load torque and the fault diameter. Furthermore, a random forest-based classifier can validate
the effectiveness of extracted frequency characteristics as novel features to deal with bearing fault detection
while automatically locating the faulty component with a classification rate of 99.94%. The results obtained
with the proposed method have been validated experimentally using a test rig

The principal objective of this paper is to detect and automatically monitor switch open-circuit faults in a two-level three-phase voltage source inverter fed induction motor from the processing of its current signals. The proposed diagnostic method uses both signal processing techniques and machine learning techniques in order to detect and localize the switch under an open-circuit fault. First, the Hilbert-Huang transform using the empirical ensemble mode decomposition is employed for each phase current signal, which leads to extracting the intrinsic mode functions. In order to optimally choose the function indicating the open-circuit fault harmonic, two factors, namely, the root mean square and the correlation coefficient are calculated out for each function. In this regard, two criteria are proposed that lead to choose the optimal function giving better information about the defected phase. The spectral envelope of the optimal function permits extra0cting the fault harmonic of the switch. Second, different machine learning techniques are applied to locate and classify the switch open-circuit faults with the hyper-parameters optimization for a better design of the different models. Finally, a comparative study of the different machine learning techniques is carried out for determining the best classifier for the open-circuit faults. The experimental results effectively demonstrate a very high classification rate of 98.98%.

Stator current analysis for preventive maintenance is an essential tool for industries. Its use is intended to serve three levels of analysis: supervision, diagnosis and monitoring of the state of damage to equipment. The main objective of this paper is to propose a diagnosis and monitoring method based on the analysis of the stator current for the detection and localization of a short-circuit fault occurred on the inverter (insulation fault of a phase). The proposed method uses signal processing techniques (temporal and spectral domain) combined with a machine learning technique to locate the faulty phase. The study begins with the application of the fast Fourier transform (FFT) to detect the harmonic characterizing the short-circuit fault of a phase of the inverter, and then a statistical study based on the skewness calculation is performed at the stator current spectrum for each phase. The second part of the study applies the random forest RF to locate the faulty phase. The features used to train the RF model are the amplitude of the harmonic f150 and the value of the skewness. The results obtained by RF show a good performance with a very high classification rate equal to 98.98%.

Recently, fault detection in asynchronous motors has paid attention of many researchers. The monitoring of these machines is performed through of the use of several physical quantities, among them vibration analysis has a crucial importance for early detection of rolling bearing faults in induction motor (IM), which they represent about 41% of IM’s ensemble defects. Commonly, the induction motor operates under non-stationary operating conditions (varying speed, fluctuating load …), and that leads to the birth of non-stationary vibration signals. The vibration signals produced from the bearing cannot generate any information about the state of the machine. Therefore, a proper analysis of these signals by means of different signal processing tools allows us to determine if the entire rotating machinery is in a normal or abnormal state. In the field of bearing fault detection, signal processing though of fault diagnosis methods has taken preponderant place. Among these methods, Fast Fourier Transform is most frequently used enabling the signal decomposition without losing any information, but it is limited to non-stationary signals such as it cannot provide the temporal location of the appearance of another shock after the born of a first one. To overcome this limitation the Discret Wavelet Transform is used providing both of time and frequency location. In this paper, we propose a diagnosis method for the identification of bearing faults, which serves to combine the DWT (Discret Wavelet Transform) and the envelope analysis. The DWT decomposes the signals of the outer race defect and the inner race defect of the bearing to obtain details. These details will be then subjected to a statistical analysis based on the kurtosis and energy coefficient (EC) in order to select the optimum wavelet details including significant harmonics corresponding to the fault cases. The envelope analysis is then applied to the selected details for extracting the frequencies’ characteristics of bearing faults. The calculated theoretical faults following the equations (1)-(3) mentioned in [1] are the rotation frequency 𝑓𝑟=28.83 Hz; the inner race frequency 𝑓𝑖𝑟=156.34 Hz; the outer race frequency 𝑓𝑜𝑟=103.12 Hz, this all from exploiting the data of vibration signals that are measured at a sampling frequency of 12000 Hz and a motor speed of 1730 RPM, available in CWRU [2]. The selection of the details obtained from the signals corresponding to the outer race and the inner race defects must respond to the greatest value of both the kurtosis and the energy value. As shown in table I, detail 1 matches to the greatest value of the kurtosis and the EC. For the healthy bearing case, the chosen detail has the smallest values that confirms its undamaged case. Taken into account that a value of kurtosis lower than 3 belongs to a good state of the bearing. The results obtained by this combined method are compared with those obtained theoretically that demonstrates well its effectiveness.

Stator current signals are widely used in fault detection and monitoring of voltage inverters. This paper proposes a method of automatic faultdiagnosis voltage inverters. The emphasis is on a signal processing technique and an artificial intelligence technique to detect and locate the open-circuit fault of an IGBT in an inverter feeding induction motor. The first technique is based on the fourier transform (FT) to detect the amplitude of the three harmonics (continuous component: f0, the amplitude of the fundamental harmonic: f50 and the harmonic which characterizes the open-circuit fault of an IGBT: f100) for each phase current (ias, ibs and ics). The second technique is based on the artificial neural network (ANN), to locate the faulty IGBT. The characteristics used to train the ANN model are the amplitudes of three frequencies f0, f50 and f100. The results obtained by the ANN show a good performance with a very high classification rate equal to 97.50%.

The main objective of this article is to contribute the automatic fault diagnosis of broken rotor bars in three-phase squirrel-cage induction motor using vibration analysis. In fact, two approaches are combined to do so, based on signal processing technique and artificial intelligence technique. The first technique is based on discrete wavelet transform (DWT) to detect the harmonics that characterize this fault, using the Daubechies wavelet vibration analysis according to three axes (X, Y, Z). This application permits having the approximation mode function and the details (recd). To exact choice of reconstruction details which contains the information of the broken rotor bars faults, two statistical studies based on the root mean square values (RMS) and Kurtosis shock factor calculation are carried out for each (recd). The choice of (recd) is conditioned by (RMS) and Kurtosis values as: RMSrecd1 < RMSrecd2 and Kurtosisrecd1 > Kurtosisrecd2. Experimental results showed that (recd1 and recd2) satisfied the condition set for (RMS) and Kurtosis values. At the end of first technique, a spectral envelop of recd1 is adopted to detect the broken rotor bars fault and the second technique based on artificial neural network (ANN) is used to identify the number of broken rotor bars. The characteristics of features used as input variables of ANN are the RMS of recd1 and recd2, and the Kurtosis shock factor of recd1 and recd2. The experimental results demonstrated the high efficiency of the proposed method with rotor broken bars fault classification rate of 98.66%.

Vibration signals are widely used in the detection and monitoring of broken rotor bar (BRB) faults. These signals are generally noisy by other sources, which can therefore lead to a loss of information on BRB fault. This paper proposes a denoising method in order to improve the statistical factor sensitivity (correlation coefficient: CC) and the spectral envelope for the early detection of failure of rotor bars. The proposed method is based on a DWT decomposition using the sliding window (db 27 ) associated with an optimized thresholding operation. First, the DWT is applied to the vibration signals to get the approximations and details. Second, every detail is reconstructed, in order to denoise every reconstructed detail. For the exact choice of reconstructed and denoised detail (recd), a statistical study based on the calculation of the correlation coefficient of each reed is carried out. This coefficient is compared to the threshold coefficient. This condition is met in this paper by recd 3 and recd 4 . A spectral envelope of drecd 3 and drecd 4 is then applied to detect the harmonics, which characterize BRB faults.

Quadratic Time-Frequency Distributions (TFDs) become a standard tool in many fields producing nonstationary signatures. However, these representations suffer from two drawbacks: First, bad time-frequency localization of the signal's autoterms due to the unavoidable crossterms generated by the bilinear form of these distributions. This results on bad estimation of the Instantaneous Frequency (IF) laws and decreases, in our case, the ability to precisely decide the existence of a motor fault. Secondly, the TFD's parameterization is not always straightforward. This paper deals with faults' detection in two-level inverter feeding induction motors, in particular open-circuit Insulated Gate Bipolar Transistor (IGBT) faults. For this purpose, we propose the use of a recent high-resolution TFD, referred as PCBD for Polynomial Cheriet-Belouchrani Distribution. The latter is adjusted using only a single integer that is automatically optimized using the Stankovic concentration measure, otherwise, no external windows are needed to perform the highest time-frequency resolution. The performance of the PCBD is compared to the best-known quadratic representations using a test bench. Experimental results show that the frequency components characterizing open-circuit faults are best detected using the PCBD thanks to its ability to suppress interferences while maintaining the signal's proper terms

The reliability of a motor control based on a variable speed drive is an important issue for industrial applications. Most of these machines are inverter based induction motors and are used in specific and complex industrial installations. Unlike the induction motor, the feeding part is very delicate and sensitive to faults. In order to improve system performance, it is therefore very important for a researcher to know the impact of a fault on the whole of his drive system. This paper discusses the short-circuit fault of the DC-link capacitor of an inverter fed induction motor. The simulation results of this type of faults are presented and its impact on the behavior of the rectifier, the inverter as well as the induction motor analyzed and interpreted.

Abstract—The main objective of this paper is to propose a
method that contributes to the automatic diagnosis of the IGBT
open-circuit fault of an inverter for detecting and localizing the
fault using the stator current spectral analysis technique. The
proposal focusses on the use of the combination of signal
processing and artificial intelligence techniques for the detection
and localization of the fault. The proposed diagnosis method
begins first by using the Hilbert-Huang transform (HHT) to detect
the harmonic characterizing the fault based on the complete
empirical ensemble mode decomposition (CEEMD) of the threestator
currents (ias, ibs, ics). The CEEMD provides the intrinsic
mode function (IMF) which contains information of the IGBT
open-circuit fault. For the exact choice of the IMF, a statistical
study based on the calculation of the root mean square values
(RMS) is carried out for each IMF. The IMF choice depends on
the condition that the RMS values of the inverter upper IGBTs are
always lower than the RMS values of the complementary ones.
The results obtained can be seen to respond well to the RMS
condition and the spectral envelope of the IMF1 makes it possible
to detect the harmonic characterizing the inverter IGBT opencircuit
fault. The proposed diagnosis method then moves to the
use of the artificial neural network (ANN) to localize the faulty
IGBT. The results obtained using the proposed method are
validated experimentally and demonstrate well their effectiveness
with a very high classification rate.

The purpose of this paper is to design the robust fault-tolerant control FTC open-circuit fault IGBT's, first of all. The modeling and control of the induction motor in the healthy inverter and in the faulty inverter (open-circuit fault at the IGBT switch) are proposed. Furthermore, the technique for detection and location the open-circuit fault an IGBT based on the Park vector combined with the polar coordinate. In order to ensure the service continuity of the system, two methods of tolerance are developed: the first method, the indirect vector control with H∞ controller of the induction motor fed by a three-phase inverter based on the fault compensation. The second method, the indirect vector control of the induction motor fed by a three-phase inverter with the redundant leg. Finally, comparative study between the two techniques of tolerance is carried out. The performance of each technique is confirmed by experimental results.

Nowadays, fault detection in induction motors has gained importance. Motor health monitoring is performed to diagnose their operating condition using vibration signals. These signals are processed using different signal processing methods to extract the characteristic parameters permitting localization of the fault. In this paper, we propose a diagnostic method based on Hilbert and Discrete Wavelet Transforms for the detection of bearing faults in asynchronous machines. The discrete wavelet transform (DWT) is intended to provide the detail coefficients while the Hilbert transform (HT) is used to obtain the temporal envelope then the envelope spectrum of the detail. The kurtosis value indicates the optimum decomposition wavelet level containing the significant frequencies corresponding to faults for early detection. The result obtained by HT-DWT is more suitable for the analysis of emergency signals. This technique is effective for either stationary or nonstationary signals. Healthy case is compared to faulty case in order to extract frequencies characterizing different faults. The validation of this approach is evaluated by comparing theoretical with experimental results.

Although the induction motor (IM) is known for its robustness and low cost of construction, it can happen that the it presents electrical, magnetic or mechanical faults. In this paper, a Hilbert-Huang Transform (HHT) spectral envelope of the stator current signature is proposed for the diagnosis of an IM rotor broken bar (RBB) fault. Firstly, the classical well-known spectral analysis Fast Fourier Transform (FFT) technique is applied to the stator current signal to obtain the harmonic frequency characterizing the RBB fault. Secondly, the HHT technique based on the ensemble empirical mode decomposition (EEMD) algorithm is proposed and applied to the stator current signal to obtain a function called the intrinsic mode function (IMF) containing the frequency that is related to the harmonic frequency characterizing the RBB fault. A comparative study is then carried out to illustrate the HHT technique merits and its superiority with respect to the FFT classical one. To test the effectiveness of the proposed HHT technique and validate the simulation results obtained, several experimental tests are also conducted using a test bench.

In the field of Wind Energy Conversion Systems(WECS), the prediction and early detection of wind power converter failures is one of the most promising ways to control and optimize the operational costs. This chapter presents the detection of converter open switch faults and the fault tolerant of power converters fed Doubly Fed Induction Generator (DFIG) in Wind Energy Conversion Systems (WECS). The proposed control approach is based on using the grid side converter to regulate the DC link voltage constant. The mean of the rotor side converter is to track the maximum power point for the wind turbine and to maintain unity power factor at stator terminals. The description of the proposed system is presented with the detailed dynamic modeling equations. The mean value of the rotor current diagnosis technique is adopted for fault detection. The fault tolerant topology used for service continuity is that with a redundant leg. The simulations are performed using the Matlab/Simulink environment and Sim power. The fault operation is analyzed and the detection method and fault tolerant topology used are tested with encouraging results.

In the recent years, several research works are focusing on the use of STATCOM in electrical networks because it is used to regulate the voltage, to improve the dynamic stability of the power system besides allowing better management of the power flow. All these positive tasks have guaranteed an important position of STATCOM within a family of Flexible Alternating Current Transmission System (FACTS). In this paper study, the control and operation of a three levels 48-pulse GTO based STATCOM is implemented with series connected transformers. The system may, unfortunately, be prone to GTO switch faults and therefore may affect reactive power transiting. In this paper, a new diagnostic approach is proposed based on the Single-Sided Amplitude Spectrum (SSAS) method of the three-leg converter currents for detection and localization of open-circuit faults. The integration of the STATCOM reconfigurable fault tolerant to the system is also considered to ensure service continuity. Several results are presented and discussed in this paper to illustrate the performance of the STATCOM fault-tolerant diagnostic

Currently, with the power electronics evolution, a major research axis is oriented towards the diagnosis of converters supplying induction machines. Indeed, a converter such as the inverter is susceptible to have structural failures such as faulty leg and/or open-circuit IGBT faults. In this paper, the detection of the faulty leg and the localization of the open-circuit switch of an inverter are investigated. The fault detection technique used in this work is based essentially upon the monitoring of the root mean square (RMS) value and the calculation of the mean value of the three-phase currents. In the first part of the paper work, the faulty leg is detected by monitoring the RMS value of the three-phase currents and comparing them to the nominal value of the phase current. The second part, the open-circuit IGBT fault is localized simply by knowing the polarity of the calculated mean value current of the faulty phase. The work is first accomplished using simulation work and then the obtained simulation results are validated by experimental work conducted in our LDEE laboratory to illustrate the effectiveness, simplicity and rapidity of the proposed technique.

The paper investigates the detection and location of IGBT open-circuit faults in two-level inverter fed induction motor controlled by indirect vector control strategy. The investigation proposes two new approaches entirely based on the Artificial Neural Network (ANN) for the extraction of the exact fault angle corresponding to the IGBT switch open-circuit fault. The first approach (Approach1) based on the Clark currents transform calculates the average value of the Clark currents to find the exact fault angle θ. The second approach (Approach2) based directly on the three-phase stator currents (without any transformation) calculates the average value of the three-phase currents to determine the exact fault angle between the phases (θab, θbc, θca). The paper conducts also a comparative study between the two approaches to assess the merits of each one of them. Experimental work is conducted to illustrate the effectiveness of the techniques and validate the results obtained