Search

In rotating machine diagnosis different spectral tools are used to analyse vibration signals. Despite the good diagnostic performance such tools are usually refined, computationally complex to implement and require oversight of an expert user. This paper introduces an intuitive and easy to implement method for vibration analysis: amplitude cyclic frequency decomposition. This method firstly separates vibration signals accordingly to their spectral amplitudes and secondly uses the squared envelope spectrum to reveal the presence of cyclostationarity in each amplitude level. The intuitive idea is that in a rotating machine different components contribute vibrations at different amplitudes, for instance defective bearings contribute a very weak signal in contrast to gears. This paper also introduces a new quantity, the decomposition squared envelope spectrum, which enables separation between the components of a rotating machine. The amplitude cyclic frequency decomposition and the decomposition squared envelope spectrum are tested on real word signals, both at stationary and varying speeds, using data from a wind turbine gearbox and an aircraft engine. In addition a benchmark comparison to the spectral correlation method is presented.

Nominated for the price of best paper (Condition Monitoring Non-Stationary Operations) -; International audience; Gear wear is hard to monitor in lifting cranes due to the difficulties to provide appropriate models of such complex systems with varying functioning modes. Statistical machine learning offers an elegant framework to circumvent these difficulties. This work explores recent advances in statistical machine learning to provide a data-driven model-free approach to monitor lifting cranes, by investigating a large number of indicators extracted from vibration signals. The principal contributions of this paper are twofold. Firstly, it explores the recently introduced Neighbor Retrieval Visualizer (NeRV) method for nonlinear information retrieval. The extracted information allows to construct a low-dimensional representation space that faithfully depicts the evolution of the system. Secondly, it proposes a simple and efficient detection method to detect abnormal evolution and abrupt changes of the system at hand, using the distance measure with neighborhood retrieval in the same spirit as NeRV. The relevance of the proposed methods, for visualizing the evolution and detecting abnormality, is demonstrated with experiments conducted on real data acquired on a lifting crane benchmark operating for almost two years with more than fifty indicators extracted from vibration signals.

International audience; Nowadays, the vibration analysis of rotating machine signals is a well-established methodology, rooted on powerful tools offered, in particular, by the theory of cyclostationary (CS) processes. Among them, the squared envelope spectrum (SES) is probably the most popular to detect random CS components which are typical symptoms, for instance, of rolling element bearing faults. Recent researches are shifted towards the extension of existing CS tools – originally devised in constant speed conditions – to the case of variable speed conditions. Many of these works combine the SES with computed order tracking after some preprocessing steps. The principal object of this paper is to organize these dispersed researches into a structured comprehensive framework. Three original features are furnished. First, a model of rotating machine signals is introduced which sheds light on the various components to be expected in the SES. Second, a critical comparison is made of three sophisticated methods, namely, the improved synchronous average, the cepstrum prewhitening, and the generalized synchronous average, used for suppressing the deterministic part. Also, a general envelope enhancement methodology which combines the latter two techniques with a time-domain filtering operation is revisited. All theoretical findings are experimentally validated on simulated and real-world vibration signals.

International audience; Condition monitoring of rotating machines in speed-varying conditions remains a challenging task and an active field of research. Specifically, the produced vibrations belong to a particular class of non-stationary signals called cyclo-non-stationary: although highly non-stationary, they contain hidden periodicities related to the shaft angle; the phenomenon of long term modulations is what makes them different from cyclostationary signals which are encountered under constant speed regimes. In this paper, it is shown that the optimal way of describing cyclo-non-stationary signals is jointly in the time and the angular domains. While the first domain describes the waveform characteristics related to the system dynamics, the second one reveals existing periodicities linked to the system kinematics. Therefore, a specific class of signals – coined angle-time cyclostationary is considered, expressing the angle-time interaction. Accordingly, the related spectral representations, the order-frequency spectral correlation and coherence functions are proposed and their efficiency is demonstrated on two industrial cases.

In the field of vibration based condition monitoring a trusted symptom of a defective bearing is the observation of peaks, at characteristic frequencies, in the squared envelope spectrum (SES). If a machine is operating in a varying speed regime the SES is computed on the order tracked signal, i.e. the signal resampled at constant angular increments, and the SES can still be used for diagnostic. Despite its versatility a common problem with the SES is that peaks from other sources of vibrations, as for instance gears, can prevent the diagnosis of a defective bearing. Therefore pre-processing techniques are applied to the vibrational signal before the computation of the SES to enhance the signal from the bearings. Among these techniques cepstral pre-whitening (CPW) has gained much attention offering a remarkable capability of eliminating, in a blind way, both harmonics and modulation side-bands of the unwanted components. In the case of a varying speed regime the usual procedure consists of three steps: order track the signal, calculate the CPW, evaluate the SES. In this paper on the contrary the CPW is applied before the step of order tracking; therefore the proposed approach is: CPW the raw time signal, order tracking, evaluation of the SES. The remarkable observation is that for this approach the cepstrum does not present peaks at characteristic quefrencies, being the raw signal acquired in a varying speed regime. However this paper shows by means of numerical simulations and analysis of experimental data, that with the proposed methodology the masking components coming from the gears are suppressed and the signal from the defective bearing is enhanced.

International audience; In rotating machinery vibration analysis, the synchronous average is perhaps the most widely used technique for extracting periodic components. Periodic components are typically related to gear vibrations, misalignments, unbalances, blade rotations, reciprocating forces, etc. Their separation from other random components is essential in vibration-based diagnosis in order to discriminate useful information from masking noise. However, synchronous averaging theoretically requires the machine to operate under stationary regime (i.e. the related vibration signals are cyclostationary) and is otherwise jeopardized by the presence of amplitude and phase modulations. A first object of this paper is to investigate the nature of the nonstationarity induced by the response of a linear time-invariant system subjected to speed varying excitation. For this purpose, the concept of a cyclo-non-stationary signal is introduced, which extends the class of cyclostationary signals to speed-varying regimes. Next, a “generalized synchronous average’’ is designed to extract the deterministic part of a cyclo-non-stationary vibration signal—i.e. the analog of the periodic part of a cyclostationary signal. Two estimators of the GSA have been proposed. The first one returns the synchronous average of the signal at predefined discrete operating speeds. A brief statistical study of it is performed, aiming to provide the user with confidence intervals that reflect the "quality" of the estimator according to the SNR and the estimated speed. The second estimator returns a smoothed version of the former by enforcing continuity over the speed axis. It helps to reconstruct the deterministic component by tracking a specific trajectory dictated by the speed profile (assumed to be known a priori).The proposed method is validated first on synthetic signals and then on actual industrial signals. The usefulness of the approach is demonstrated on envelope-based diagnosis of bearings in variable-speed operation.

In speed-varying conditions, the assumption of cyclostationarity of rolling-element bearing vibrations is jeopardized. The emitted signal comprises an interaction between (i) time-dependent components related to the system dynamics (e.g. transfer function) and (ii) angle-dependent mechanisms related to the system kinematics (e.g. impact, load modulations…). This necessarily implies the inadequacy of classical cyclostationary tools no matter a temporal or angular vision is adopted. This consequently calls for an angle ⧹ time approach which preserves—via the angle variable—the cyclic evolution of the signal while maintaining—via the time variable—a temporal description of the system dynamics. The first object of this paper is to analytically characterize bearing fault vibrations and explore its angle ⧹ time cyclostationary property. The second object is to experimentally validate these results on real-world vibration signals and demonstrate the optimality of the angle–time approach over classical approaches for rolling element bearing diagnosis.

A new technique of diagnosing data for broken rotor bars in induction motors derived from two of the three stator currents, the Beirut diagnostic procedure (BDP) is presented in this paper. The theoretical principles directly related to the application of this diagnostic technique are described, emphasizing the use of a severity factor in order to evaluate the extension of the fault. Defining the severity factor as the normalized amplitude of the fault characteristic frequency enables us to draw up a table of comparison of several usual electric diagnostic methods. Besides the traditional one-phase current spectrum analysis, values of the severity factor related to electrical signals like the instantaneous powers, the current space vector modulus and finally related to the new Beirut diagnostic method are analyzed with respect to the variation of the power factor angle and of the sum of the two current side-band angular displacement. The BDP offers several advantages over the usual motor current signature analyses (MCSA) methods: it is shown how the proposed severity factor applied to the new diagnostic technique is not dependent on motor parameters such as the power factor angle and the fault type which is not the case of the instantaneous powers. In addition, the BDP has the advantage of detecting easily fault characteristic frequencies, which is not possible via diagnostic methods that use the detection of two side-band components as in the simple current spectrum. By theoretical analysis, computer simulations, and laboratory experiments, it is shown that the new method enhances the reliability of diagnostics of broken rotor bars in induction motor.

This paper reports quantitative analyses of spectral fault components in five noninvasive diagnostic procedures that use input electric signals to detect different types of abnormalities in induction motors. Besides the traditional one phase current spectrum analysis “SC”, the diagnostic procedures based on spectrum analysis of the instantaneous partial powers “Pab”, “Pcb”, total power “Pabc”, and the current space vector modulus “csvm” are considered. The aim of this comparison study is to improve the diagnosis tools for detection of electromechanical faults in electrical machines by using the best suitable diagnostic procedure knowing some motor and fault characteristics. Defining a severity factor as the increase in amplitude of the fault characteristic frequency, with respect to the healthy condition, enables us to study the sensitivity of the electrical diagnostic tools. As a result, it is shown that the relationship between the angular displacement of the current side-bands components at frequencies (f±fosc) is directly related to the type of induction motor faults. It is also proved that the total instantaneous power diagnostic procedure was observed to exhibit the highest values of the detection criterion in case of mechanical faults while in case of electrical ones the most reliable diagnostic procedure is tightly related to the value of the motor power factor angle and the group motor-load inertia. Finally, simulation and experimental results show good agreement with the fault modeling theoretical results.

This article presents a comparative investigation of various media for non-invasive diagnosis of mechanical abnormalities in induction motors. Stator voltages and stator currents as well as noises on these signals are simulated first for a fault-free motor then for a motor with mechanical abnormalities. These signals are subsequently employed to compute the instantaneous powers P ab , P cb , P abc and last the current Park vector modulus known as the extended Park vector approach (EPVA) method. Waveforms of these simulated signals were analyzed using the power spectral density transformation. Commercially available diagnostics systems use the fact that the amplitude of some components known as fault characteristic frequencies in these electrical signals increase when mechanical abnormalities occur. The amplitude increase of these characteristic frequencies is employed as a criterion in order to investigate noise immunity of the five diagnosis methods, thus making possible their classification. Si...

Rotor asymmetries in induction machines perturb many components such as, flux patterns and electromagnetic torque. The supervision of these signals enables early detection of such faults and help to machine diagnostic. This paper studies the detection of rotor imperfection by spectral analysis of the electromagnetic torque computed by two rotor flux estimators. In the first approach, the Kalman filter is used assuming to be known the mechanical velocity. The second approach uses Extended Kalman Filter (EKF) for speed estimation. Experimental results show the great capability of these methods to detect this type of faults.

International audience; This paper proposes a model-based technique for detecting wear in a multistage planetary gearbox used by lifting cranes. The proposed method establishes a vibration signal model which deals with cyclostationary and autoregressive models. First-order cyclostationarity is addressed by the analysis of the time synchronous average (TSA) of the angular resampled vibration signal. Then an autoregressive model (AR) is applied to the TSA part in order to extract a residual signal containing pertinent fault signatures. The paper also explores a number of methods commonly used in vibration monitoring of planetary gearboxes, in order to make comparisons. In the experimental part of this study, these techniques are applied to accelerated lifetime test bench data for the lifting winch. After processing raw signals recorded with an accelerometer mounted on the outside of the gearbox, a number of condition indicators (CIs) are derived from the TSA signal, the residual autoregressive signal and other signals derived using standard signal processing methods. The goal is to check the evolution of the CIs during the accelerated lifetime test (ALT). Clarity and fluctuation level of the historical trends are finally considered as a criteria for comparing between the extracted CIs. This study reveals the most relevant features to be used for damage detection and condition monitoring of the gear system. It is also shown that the proposed procedure using a combination of cyclostationarity and autoregressive modeling enhance the ability to detect and diagnose mechanical wear in multi-staged planetary gears.; This paper proposes a model-based technique for detecting wear in a multistage planetary gearbox used by lifting cranes. The proposed method establishes a vibration signal model which deals with cyclostationary and autoregressive models. First-order cyclostationarity is addressed by the analysis of the time synchronous average (TSA) of the angular resampled vibration signal. Then an autoregressive model (AR) is applied to the TSA part in order to extract a residual signal containing pertinent fault signatures. The paper also explores a number of methods commonly used in vibration monitoring of planetary gearboxes, in order to make comparisons. In the experimental part of this study, these techniques are applied to accelerated lifetime test bench data for the lifting winch. After processing raw signals recorded with an accelerometer mounted on the outside of the gearbox, a number of condition indicators (CIs) are derived from the TSA signal, the residual autoregressive signal and other signals derived using standard signal processing methods. The goal is to check the evolution of the CIs during the accelerated lifetime test (ALT). Clarity and fluctuation level of the historical trends are finally considered as a criteria for comparing between the extracted CIs.This study reveals the most relevant features to be used for damage detection and condition monitoring of the gear system. It is also shown that the proposed procedure using a combination of cyclostationarity and autoregressive modeling enhance the ability to detect and diagnose mechanical wear in multi-staged planetary gears.

This paper proposes a model-based technique using a combination of cyclostationary and autoregressive signal modelling in order to detect wear in a multistage planetary gear of lifting cranes. The first-order cyclostationarity is exploited by the analysis of the Time Synchronous Average part (TSA) of the angular resampled vibration signal. Then an autoregressive model (AR) is applied to the TSA part in order to extract a residual signal containing pertinent fault signatures. The paper also explores the efficiency of a number of methods commonly used in vibration monitoring. Condition monitoring indicators are then extracted from different treated signals. In the experimental part, all these techniques are applied to a test bench data of a lifting winch. The goal is to trend the evolution of the extracted features during the test. This study reveals that the proposed procedure using this combination enhances the ability to detect and diagnose mechanical wear of winch planetary gears.

International audience; A keep-it-simple (KIS) solution is introduced that interpolates a signal up to an arbitrary accuracy. It consists of interpolating the complex envelopes at the output of a perfect-reconstruction filter-bank: if K subbands are used, the proposed interpolation scheme is shown to have a similar figure of merit as if the signal was initially oversampled by factor K at the acquisition stage, yet without the storage burden implied by the latter method. (c) 2012 Elsevier Ltd. All rights reserved.

978-3-642-32614-1; Robotic cooperative tasks impose, in many cases, a grasping action. Grasping by coiling it is one of the most versatile action. The present article propose a frequency stability criterion based on the Kahman - Yakubovich - Popov Lemma for the hyper-redundant arms with continuum element that performs the grasping function by coiling. Dynamics of the biomimetical robot during non-contact and contact operations, for the position control, is studied. An extension of the Popov criterion is developed. The P control algorithms based on SMA snake-type robot actuators are introduced. Numerical simulations and experimental results of the snake type robot motion toward an imposed target are presented.

A new method of obtaining diagnostic data from induction motors, derived from the three supply currents using principal components analysis, is presented in this paper. The main advantage of this new diagnostic tool is its ability to extract automatically the characteristic frequencies relative to the different machine operating modes. This is accomplished using the proportion of the power attributed to the first principal component and/or using the sensor contribution to the power at specific frequencies. Thus, the new diagnostic method gives a good basis for an automatic non intrusive condition monitoringfor rotating machinery.

A new method of obtaining diagnostic data from induction motors, derived from the three supply currents using principal components analysis, is presented in this paper. The techniques presented here focus on extracting relevant information from spectral matrices. These techniques are qualified as parsimonious tools for exploring the behaviour of current vector valued signals in the frequency domain with minimal loss of information. In fact, the new diagnostic method obtains data from the three stator currents by exploring special fault characteristic frequencies in the power spectral density of the first principal component. The main advantage of this new diagnostic tool is its ability to extract automatically the characteristic frequencies relative to the different machine operating modes. This is accomplished using the proportion of the power attributed to the first principal component and/or using the sensor contribution to the power at specific frequencies. Thus, the new diagnostic method gives a good basis for an automatic non intrusive condition monitoring for rotating machinery.

This paper presents a comparative investigation of different methods for non-invasive diagnosis of mechanical abnormalities in induction motors. Stator voltages and stator currents, as well as noise on these signals, are simulated first for a fault-free motor, and then for a motor with mechanical abnormalities. These signals are subsequently used to compute the instantaneous powers "P/sub ab/", *P/sub cb/", "P/sub abc/". The last method to be examined is the current Park vector modulus, known as the extended Park vector approach "EPVA" method. Waveforms of these simulated signals were analyzed using the power spectral density transformation. The amplitude increase of these characteristic frequencies is used as a criterion in order to investigate noise-immunity of the five diagnosis methods, thus making possible their classification. As a result, lhe extended Park vector approach "EPVA" was observed to exhibit the highest noise-immunity. The "EPVA" therefore improves the reliability of diagnostics of induction motor drives.

Rotor asymmetries lead to perturbations of air-gap flux patterns in induction machines. These perturbations in flux components affect a number of components including currents and electromagnetic torque. The supervision of these signals enables early detection of such faults and assists in fault diagnosis. This paper studies the detection of rotor imperfections by spectral analysis of the electromagnetic torque, computed by two stator flux estimators using only noninvasive sensors such as current and voltage sensors. In a first approach, the Extended Luenberger Observer (ELO) is used to estimate stator flux components as well mechanical velocity. A second approach uses a nonlinear High Gain Observer (HGO) for the same purposes. Experimental results and comparison show the significant potential of these methods in detecting these types of faults.

Rotor asymmetries lead to perturbations of airgap flux patterns in induction machines. These perturbations in flux components affect the electromagnetic torque, as well as stator currents and voltages. The supervision of these signals enables the detection of underlying mechanical asymmetries. The effect of mechanical imperfections on stator signals can be negligible in certain cases. Thus, this makes the detection of minor rotor imperfections a difficult task when the load is weak. In this paper we study the detection of rotor imperfections by spectral analysis of the electromagnetic torque. This torque is computed by rotor flux observers based on two induction motor models : fourth-order and second-order model. A comparison of spectral analysis methods applied to the torque shows the significance of the method we describe, which makes use of mechanical velocity, in order to detect low levels of asymmetry in the rotor cage at low level load. Finally, it is to be noted that this method is applied to real signals with and without imperfections.

International audience; This paper presents a condition monitoring methodology applied to the suction roll and the Press roll of a paper machine. Experimental results obtained for the detection and identification of many defects that may occur to different mechanical components are presented. To this end, many fault indicators are calculated using a set of signal processing methods. We endeavor to propose robust fault indicators with respect to the variations of the operation parameters as the speed variation. Cyclostationary and cepstral approaches are used in order to make vibration source separation and to extract pertinent indicators closely related to the health of the paper machine. AStrion strategy, a stand-alone, data-driven and automatic tracking analyzer, is applied in order to characterize a sensor failure on the suction roll and a fault on the motor that drives the press roll. The trends of these parameters have shown the effectiveness of these methods to detect and identify the failure modes of the equipment thus allowing the reduction of the overall maintenance cost. This work has been done within the SUPREME project, funded by the European Commission, under the FP7 program.