Skip to main content

About this book

This book provides readers with a timely snapshot of the potential offered by and challenges posed by signal processing methods in the field of machine diagnostics and condition monitoring. It gathers contributions to the first Workshop on Signal Processing Applied to Rotating Machinery Diagnostics, held in Setif, Algeria, on April 9-10, 2017, and organized by the Applied Precision Mechanics Laboratory (LMPA) at the Institute of Precision Mechanics, University of Setif, Algeria and the Laboratory of Mechanics, Modeling and Manufacturing (LA2MP) at the National School of Engineers of Sfax. The respective chapters highlight research conducted by the two laboratories on the following main topics: noise and vibration in machines; condition monitoring in non-stationary operations; vibro-acoustic diagnosis of machinery; signal processing and pattern recognition methods; monitoring and diagnostic systems; and dynamic modeling and fault detection.

Table of Contents


Feature Selection Scheme Based on Pareto Method for Gearbox Fault Diagnosis

Fault diagnosis based on pattern recognition approach has three main steps viz. feature extraction, sensitive features selection, and classification. The vibration signals acquired from the system under study are processed for feature extraction using different signal processing methods. Followed by feature selection process, classification is performed. The challenge is to find good features that discriminate the different fault conditions of the system, and increase the classification accuracy. This paper proposes the use of Pareto method for optimal feature subset selection from the pool of features. To demonstrate the efficiency and effectiveness of the proposed fault diagnosis scheme, numerical analyses have been performed using the Westland data set. The Westland data set consists of vibration data collected from a US Navy CH-46E helicopter gearbox in healthy and faulty conditions. First, features are extracted from vibration signals in time, spectral, and time-scale domain, then ranked according to three different criterions namely: Fisher score, correlation, and Signal to Noise Ratio (SNR). Afterword, data formed by only the selected features is used as input for the classification problem. The classification task is achieved using Support Vector Machines (SVM) method. The proposed fault diagnosis scheme has shown promising results. Using only the feature subset selected by Pareto method with Fisher criterion, SVMs achieved 100% correct classification.
Ridha Ziani, Hafida Mahgoun, Semcheddine Fedala, Ahmed Felkaoui

Intelligent Gear Fault Diagnosis in Normal and Non-stationary Conditions Based on Instantaneous Angular Speed, Differential Evolution and Multi-class Support Vector Machine

The gearboxes are among the most important elements of rotating machines and consequently they require an effective condition monitoring strategy. However, many machines operate over a wide range of the rotational speed and most analysis of rotating machines are based on investigating the vibrations with a constant speed. Therefore, techniques developed for constant conditions cannot be applied directly.
The angularly sampled Instantaneous Angular Speed (IAS) carry a considerable amount of information on the health and usage status of rotating machinery. Thus, it represents a potential source of relevant information in intelligent fault detection and diagnosis systems, but also to construct Feature Vector (FV) to further get robust and effective classification methods for different running speed or load conditions.
This paper presents an intelligent gear fault diagnosis based on Instantaneous Angular Speed (IAS), Differential Evolution (DE) and multi-class Support Vector Machine (SVM) in normal and non-stationary conditions. For this purpose, features are extracted from IAS. Then, the DE selection algorithm is applied in order to select the most relevant features. The classification is performed by SVM in order to improve the detection and identification of gear defects. The methodology is applied in normal and non-stationary conditions, with six pinion fault conditions. The experimental results prove that the proposed method is able to detect the fault conditions of the gearbox effectively.
Semchedine Fedala, Didier Rémond, Ahmed Felkaoui, Houssem Selmani

Effect of Input Data on the Neural Networks Performance Applied in Bearing Fault Diagnosis

The aim of this paper is to study the effect of input parameters choice of the artificial neural network (ANN), in order to obtain the best performances of fault classification. The purpose of this network is to automate the electric motor bearing diagnosis based on vibration signal analysis. The choice of the components of ANN’s inputs (training and testing) has a big challenge for prediction of the machines faults diagnosis. The vibration signals collected from the test rig (Bearing Data Center) are preprocessed, to extract the most appropriate monitoring indicators to analyze the health of the experimental device.
To improve the performance of the neural network, we use three different dataset: the first contains only time indicators, while the second contains the frequency indicators, and the third set is a combination of these two indicators. A comparison between the effects of each feature on the ANN performances, allowed us to choose the optimal structure of input data. The obtained results show that the combined dataset give the best performances compared to the two others dataset.
Hocine Fenineche, Ahmed Felkaoui, Ali Rezig

Bearing Diagnostics Using Time-Frequency Filtering and EEMD

The ensemble empirical mode decomposition (EEMD) was largely used in the diagnosis of the rotating machines but the EEMD shows a limitation with the detection of the impulses that are influenced by the presence of noise, the mode mixing, and the end effect. To detect the shocks due to the defect at an early stage, we propose to use the Time-frequency filtering (TFF) which was recently proposed by Flandrin. This method allows us to denoise the signal and gives promising results in the detection of the defects on machine elements.
In this work first, we show by simulated bearing signal the advantage of TFF compared to the EEMD in the detection of impulses. Then, we analyze real vibration bearing signals by using the two different time-frequency methods, ensemble empirical mode decomposition (EEMD) and Time-frequency filtering (TFF), and then we compare the results given by using the two methods separately and the results by a new method when we combine the two methods. The filtered modes are analyzed by calculation of the spectrum, which gives more information about the defect and allows us to read it frequencies and detect it at an early stage.
Hafida Mahgoun, Ridha Ziani

The Time-Frequency Filtering (TFF) Method Used in Early Detection of Gear Faults in Variable Load and Dimensions Defect

In stationary condition, a local gear fault is presented by periodic impulses. However, under variable load, the vibration signal is non-stationary and the periodic impulses are masked by the noise and the part of the signal due to the load. The use directly of the time-frequency methods doesn’t allow detecting these impulses. In this study, we propose to use two different time-frequency methods, ensemble empirical mode decomposition (EEMD) and time-frequency filtering (TFF) to analyze the vibration signal. First, the EEMD method is used to decompose the vibration signals in many modes. Then each mode is filtered and denoised by using the TFF method. In this paper, we propose to compare the results given by using the two methods separately and the results when we combine the two methods.
Hafida Mahgoun, Fakher Chaari, Ahmed Felkaoui, Mohamed Haddar

Comparison Between Hidden Markov Models and Artificial Neural Networks in the Classification of Bearing Defects

In this paper a comparative study between two classification methods was presented, the first one belongs to the statistical domain in this case the Hidden Markov Models (HMM), the second is an Artificial Intelligence (AI) tool known as of Artificial Neural Networks (ANN), given their popularity in recent years and the interest shown by researchers in these methods, as to their performance and efficiency in the field of classification mainly. Indeed, the two classification tools were tested on data collected from vibratory signals on a test bench at the Bearing Data Center of Case Western Reserve University, and after being put in the appropriate form by an adequate signal processing and analysis to facilitate implementation. In this study, we have tried to identify the advantages and disadvantages of both tools in the field of classification of rotating machine defects, with the aim of accessing other work for the implementation of a classifier as effective as efficient. The results obtained are described as satisfactory and encouraging by their compatibility with those obtained by others implemented by other research but in other fields such as speech processing or image processing, which will give the character of originality to our work once completed.
Miloud Sedira, Ridha Ziani, Ahmed Felkaoui

On-line Adaptive Scaling Parameter in Active Disturbance Rejection Controller

Active Disturbance Rejection Controller (ADRC) is considered one of the most famous model free controllers in the industry. This introduced scheme of control, do not require the exact modeling of the system equations and used to reject online any types of perturbations. However, the drawback of this tool is the hard task of tuning multi-parameters and takes a long time to achieve performances requirements. In this contribution, an optimization of a scaling parameter which has an important effect in the dynamic behavior of controlled system. There has been some research concentrate in estimate the parameters uncertainties from input and output signals of the body mass in vehicle system. This kind of estimation is based on differential algebra which is known by its simplicity of implementation, fast and robust to noise marring any measured signals. Furthermore, the combination of this algebraic methodology with aforementioned control low is easy. For the purpose of improving the effectiveness of ADRC controller, this paper use to predict this unknown variation and it was incorporated in the equation of control. Using this time varying parameter instead of an empirical one, simulations results show an amelioration of the energy consumption and an increase of the ride comfort.
Maroua Haddar, S. Caglar Baslamisli, Fakher Chaari, Mohamed Haddar

Modal Analysis of the Clutch Single Spur Gear Stage System with Eccentricity Defect

Gears are an important element in a variety of industrial applications. An unexpected failure of the gear may cause significant economic losses. For that reason, fault diagnosis in gears has been the subject of intensive research. Modal analysis can be used in the fault detection of rotating machinery. It can provide natural frequencies and vibration modes which are essential information to learn about most of dynamic characteristics of the combined system. In order to investigate the dynamic behavior of a coupled clutch-gear transmission system in the presence of gear defect, a general dynamic model is developed and a numerical modal analysis technique is achieved. Several types of gear defects that can be found in the literature. In this paper, a gear eccentricity defect is introduced in the model to study their influence on the modal properties. The distributions of modal kinetic and strain energies are presented in the case without and with defect on the geared system, and a comparative study is conducted.
Ahmed Ghorbel, Moez Abdennadher, Lassâad Walha, Becem Zghal, Mohamed Haddar

Estimation of Road Disturbance for a Non Linear Half Car Model Using the Independent Component Analysis

The identification of the road profile disturbance acting on a vehicle was the objective of many recent researches. This estimation remains very interesting since it contributes to study the dynamic behavior of the vehicle in one side and to choose a control law later in other side. However most of the used techniques have many drawbacks such us those based on direct measurements of the profile which need costly profilometers or those based on neural network algorithm which are very complicated. So the purpose of this research is to use a new method named the Independent Component Analysis (ICA) to estimate the road profile. This method is based on the so-called inverse problem. So it necessitates only the knowledge of the dynamic responses of the vehicle to identify the road disturbance. Therefore the Newmark algorithm is used in this paper to extract the dynamic responses of the system under study which is a non linear half car model. Starting from these responses, the ICA algorithm is applied. The validation of the obtained results is done using some performance criteria which are the relative error and the MAC number. Finally a good agreement is found between the original profile and the estimated one.
Dorra Ben Hassen, Mariem Miladi, Mohamed Slim Abbes, S. Caglar Baslamisli, Fakher Chaari, Mohamed Haddar

Transfer Path Analysis of Planetary Gear with Mechanical Power Recirculation

Planetary gears can transmit higher power density levels because they use multiple power paths formed by each planet branches. In order to study the propagation of vibration between components of planetary gear test bench with mechanical power recirculation, an approach to the classical transfer path analysis (TPA) method is used to improve vibration control of planetary gear test bench. This approach termed Global Transmissibility Direct Transmissibility (GTDT) avoids the drawbacks of the classical TPA which are decoupling of the active part in the measurements of the Frequency Response Functions (FRFs) of source-receiver paths and the difficult measurement of the operational forces. The Global Transmissibility Direct Transmissibility (GTDT) is two steps method: the first step is the measurements of transmissibility which requires no disassembly tests and the second step is the measurement of the operational responses which is easier than measurement of the operational forces. In fact, tri-axial accelerometers are mounted in each component of the back-to-back planetary gear test bench and the transfer functions (frequency response functions, FRFs) are measured using hammer impact test in order to measure the global transmissibilities. Then, the direct transmissibilities are computed from the global transmissibilities. Finally, reconstructed operation responses are shown in the partial path contribution (PPC) plots to compare the vibration level of each component and to know its contribution in the transfer of vibration.
Ahmed Hammami, Alfonso Fernandez del Rincon, Fakher Chaari, Fernando Viadero Rueda, Mohamed Haddar

Modeling the Transmission Path Effect in a Planetary Gearbox

In such mechanical systems, as helicopters and self-propelled cranes, designers need to use gearboxes which have an important reduction ratio within compact space. Hence, planetary gearboxes are widely used. Consequently, its monitoring presents an important task for researchers and engineers either in healthy or damaged case. Many researchers are interested on the investigation of the modulation phenomenon in planetary gearbox. It is presented in a frequency representation as side-band activity near to the gear-mesh frequency component and its harmonics. In a healthy case, the origin of this phenomenon in a planetary gearbox (stationary ring) is that the transducer, which is mounted on the external housing of the ring gear, perceived signals from all components including sun-gear, ring-gear, carrier and planet-gears which can occupy different position in one carrier period rotation. Hence, when the planet comes closer to the sensor, the vibration signal increases and vice-versa. In this work, a two dimensional linear lumped parameter model is proposed to model vibration sources. A mathematical formulation of the transmission path is introduced in order to model only the amplitude modulation phenomenon due to the change of the planet-gear position since the speed of the sun is constant. A frequency representation of numerical results is presented and analyzed.
Oussama Graja, Bacem Zghal, Kajetan Dziedziech, Fakher Chaari, Adam Jablonski, Tomasz Barszcz, Mohamed Haddar

Dynamic Behavior of Spur Gearbox with Elastic Coupling in the Presence of Eccentricity Defect Under Acyclism Regime

In this paper, the effect of eccentricity defect on the dynamic behaviour of one stage spur gearbox running under acyclism regime is studied. In fact, acyclism regime is generated by a combustion engine motor which produced fluctuations of load and speed. The motor torque is periodic and it modeled in the force’s vector. The rotational speed of the Diesel engine is a harmonic function and it generates a periodic fluctuation of the gear meshing stiffness function. This driven motor is joined to the gearbox through an elastic coupling in which the model of Nelson and Crandall is adopted. The eccentricity defect is introduced in the pinion. This defect produces an additional potential energies and kinetic energy and it is modelled through additional forces. The equation of motion is obtained using Lagrange formalism and the algorithm of Newmark is used to compute the dynamic response of the studied system and the Wigner–Ville distribution shows the dynamic behaviour of the gearbox under this cyclo-stationary regime. Results show the variability of the meshing frequency and its harmonics which excites the system. Also, natural frequencies are observed in the spectrum and Wigner–Ville distribution of the dynamic signal. Nevertheless, these methods fail to detect the frequencies of eccentricity and acyclism.
Atef Hmida, Ahmed Hammami, Fakher Chaari, Mohamed Taoufik Khabou, Mohamed Haddar


Additional information

Premium Partner

image credits