Empirical mode decomposition based probabilistic neural network for faults classification

Signal is a non-linear and non- stationary signal, the empirical mode decomposition (EMD) can decompose the signal into several intrinsic mode functions(IMF), the IMF can reflect the characteristic in frequency changing of logging data along depth, IMF curves of different steps might reveal cycle characteristics of different scale, which are in correspondence with geological cycles o different scales, so that different scale sequences might be detected clearly from IMF curves. Empirical mode decomposition method is used to demarcat stratigraphic sequence and stratigraphic correlation in this paper ,the result show that EMD method can better show the internal vibration of logging signal and reflect the characteristics of the stratigraphic sequence ,and EMD method can provide technical support for stratigraphic correlation.

The fault location of VSC-HVDC section for grid integrated offshore wind farm is found by Empirical Mode Decomposition (EMD) using Hilbert transform (HT). The intrinsic mode function (IMF) of the fault generated in traveling wave signal retrieved at rectifier and inverter end is extracted by EMD. The HT is used to detect the magnitude and phase of each IMF. The HT of first IMF (IMF1-H) is evaluated for fault location objective and is further compared with wavelet transform (WT). The efficacy of the proposed fault location scheme is reflected in the simulation result.

Detection of Subtle Damage in Structures through Smart Signal Reconstruction

Sleep disorders has a vital effect on mental depression and many other diseases of human body. Diagnosing the sleep disorder in an early curable stage may help to provide better treatment and save the life. The EEG (Electroencephalogram) signal is one of the most uses bio signal for capturing brain activities to detect and diagnosis the sleep disorders. Empirical mode decomposition (EMD) is an efficient time-frequency data analysis technique for diagnosing disease by analyzing EEG signal. However, it is a challenging issue to select the optimal intrinsic mode functions (IMFs) of Empirical mode decomposition (EMD) for extracting discriminant properties of EEG signals to diagnosis the sleep disorder. From this point of view, this paper presents a model to select optimal IMF of EMD for diagnosing the sleep disorder using EEG brain signal. In this proposed model, EMD is applied to decompose and analyze EEG signal for extracting biomarker/feature of sleep disorders. During the EMD decomposition process, different levels of IMF are extracted and features, i.e., Shannon Entropy, Spectral Entropy, Standard deviation, Skewness and Kurtosis are calculated from those IMFs for detecting the sleep disorders. In identification process, the multiclass support vector machine (MC-SVM) classification algorithm is used and sleep disorders are classified based on trained knowledge. Finally, the performance of proposed model is evaluated for different IMFs of EMD and find the optimal IMF for sleep disorder diagnosis. For evaluating the proposed model, a benchmark dataset including 4 types of data such as Apnea, REM, PLM and healthy subjects are used in experiment. According to the experimental result, the proposed model achieves the optimal classification performance for IMF 8, i.e., 93.24% average classification accuracy.

Using filter bank property to simplify the calculations of Empirical Mode Decomposition

A relationship between Intrinsic Mode Functions (IMFs), derived from the Empirical Mode Decomposition (EMD), and the slow-flow model of a no nlinear dynamical system has been exploited in the development of the Slow Flow Model Identification (SFMI) method for strongly nonlinear systems, in which the physical parameters of such systems are identified from experimental data. Both the slow flows and IMFs provide the means to expand a general multicomponent signal in terms of a series of simpler, dominant, monocomponent signals. The slow flows are obtained analytically, for example through application of the method of complexification and averaging (CxA), which transforms the equations of motion into a set of approximate equations in amplitude and phase for each modeled frequency component. In contrast, the EMD characterizes a signal through the envelope and phase of its elemental components, the IMFs. Thus, between nonlinear transitions, the equations derived using the CxA method govern the amplitude and phase of the modeled IMFs. Application of SFMI has, until now, been limited to low-dimensional systems subjected to impulsive excitation. Herein, the method is extended to identificatio n of a planar rigid airfoil

The Fourier transform (FT) has been the most popular method for analyzing large rotating machine shaft vibration problems, but it assumes that these vibration signals are linear and stationary. However, in reality this is not always true. Nonlinear and non-stationary shaft vibration signals are often encountered during the start-up and shut-down processes of the machines. Additionally, mechanical faults, for example rotor-to-stator rubbing, fluid excitation, part-loosening, and shaft cracking, are nonlinear. Owing to these reasons, an accurate analysis of shaft vibration cannot always be achieved by using the FT. An alternative tool, the wavelet transform (WT), is now being used to improve the situation. But the efficiency is a problem especially when applying the WT to the accurate analysis of a large-scale, lengthy data. In view of the powerful capability of empirical mode decomposition (EMD) to process nonlinear/non-stationary signals, its algorithm efficiency and its satisfactory performance in minimizing energy leakage, the EMD is used in this paper to analyze the problem, the signals investigated are adaptively decomposed into a finite number of intrinsic mode functions (IMFs). The principal IMFs, identified using an energy-distribution threshold, dominate the signals’ oscillation. So, ‘purified’ shaft vibration signals can be reconstructed from these principal IMFs. To remove interference present in principal IMFs, an adaptive band-pass filter is designed, whose central frequency is automatically set to the frequency dominating the IMF being investigated. To facilitate the observation of transient shaft vibration, a transient shaft orbit (TSO) is constructed by introducing timescale into the orbit drawing process. Nine mathematical criteria are also proposed to evaluate the shaft vibrations exhibited in the IMFs and TSOs. The novelty of this approach is that the EMD provides an adaptive, effective, and efficient way to obtain ‘purified’ shaft vibration signals, which describe the transient shaft vibration more vividly and precisely, reducing misinterpretation of the machine running condition. The approach is validated by two practical examples: • Part-loosening incident on a large centrifugal compressor. • Fluid excitation incident on a centrifugal compressor. Calculations of the mathematical criteria for seven different running conditions of these large rotating machines are also presented. It is demonstrated that the proposed technique provides a feasible and reliable way to interpret shaft vibration, by which means the machine condition can be diagnosed correctly.

Primal sketch performs an important role in early vision. In this paper, we propose a novel method to obtain the primal sketch of natural images by integrating empirical mode decomposition (EMD) techniques and image gradient. 2D EMD approach can decompose the image into a finite number of intrinsic mode functions (IMF), and each one represents the original image in a different scale, with the 1st IMF representing the finest scale. To enhance the information represented by the IMF, we multiply the 1st IMF by the image gradient. This enhanced IMF highlights intensity changes in the image. By linking all the maximal points in the enhanced IMF, we obtain a primal sketch of the original image. Compared with the existed primal sketch extraction methods, our method is fully driven by the image data, and it needs neither to choose filters nor to learn the image bases. The experiment results show that our method is fast and effective.

This paper presents a technique for the feature extraction of voltage signal for classification of open-circuited (O-C) faults in power electronic converters (PECs) connected to grid tied wind power system. The Empirical Mode Decomposition (EMD) was applied in the PEC output for the detection of the fault, as well as identification of the abnormal power switching device in the PEC. This method is applied on 12 types of signals including normal and faulted output voltage signals. Six features are computed from each Intrinsic Mode Functions (IMFs) obtained from EMD which are: Root Mean Square (RMS), Mean, Median, Variation, Standard Deviation and Entropy. The classification of the fault is achieved using Neural Networks (NN) and a comparison of classification precision between Probabilistic Neural Network (PNN) and Back-Propagation Multi-Layer perceptron (BPMLP) is also done. The overall classification accuracy concludes that EMD allows receiving better performance in the fault classification than other traditional methods like support vector machines, Wavelet.

In this paper, a statistical analysis of electroencephalogram (EEG) signals is carried out in the empirical mode decomposition (EMD) domain using a publicly available benchmark EEG database. First, the intrinsic mode functions (IMF) are extracted from the EEG signals in the EMD domain. Next, the investigation was carried whether the Bessel k-form (BKF) probability density function (pdf) can appropriately model the IMFs extracted in EMD domain of the EEG signals. It is shown that on an average, the BKF pdf is a suitable prior to model the first five IMFs extracted from various types of EEG recordings. After that, it is shown that the BKF parameters can distinguish among the EEG signals at those five IMF levels quite well. The analysis is further confirmed through the p-values obtained by one way analysis of variance (ANOVA). Thus, the BKF parameters in the EMD domain may be used to characterize EEG signals and help the electroencephalographers in developing fast and effective classifiers for seizure seizure detection.

Fault classification in power systems using EMD and SVM

The prediction of mold level is a basic and key problem of continuous casting production control. Many current techniques fail to predict the mold level because of mold level is non-linear, non-stationary and does not have a normal distribution. A hybrid model, based on empirical mode decomposition (EMD) and support vector regression (SVR), is proposed to solve the mold level in this paper. Firstly, the EMD algorithm, with adaptive decomposition, is used to decompose the original mold level signal to many intrinsic mode functions (IMFs). Then, the SVR model optimized by genetic algorithm (GA) is used to predict the IMFs and residual sequences. Finally, the equalization of the predict results is reconstructed to obtain the predict result. Several hybrid predicting methods such as EMD and autoregressive moving average model (ARMA), EMD and SVR, wavelet transform (WT) and ARMA, WT and SVR are discussed and compared in this paper. These methods are applied to mold level prediction, the experimental results show that the proposed hybrid method based on EMD and SVR is a powerful tool for solving complex time series prediction. In view of the excellent generalization ability of the EMD, it is believed that the hybrid algorithm of EMD and SVR is the best model for mold level predict among the six methods, providing a new idea for guiding continuous casting process improvement.

The Empirical Mode Decomposition (EMD) is a signal analysis method that separates multi-component signals into single oscillatory modes called intrinsic mode functions (IMFs). These IMFs can generally be associated to a physical meaning of the process from which the signal is obtained. When the phenomena of mode mixing occurs, as a result of the EMD sifting process, the IMFs may lose their physical meaning hindering the interpretation of the results of the analysis. Previous research presents a rigorous mathematical analysis that shows how EMD behaves in the case of a composite two-components signal, explaining the roots of the mode mixing problem. Also, the frequency-amplitude region within which a good separation is achieved with EMD is well identified and discussed. However, a solution that offers good IMF separation when components reside within the same octave is not yet available. In this paper, a method to separate spectral components that reside within the same octave, is presented. This method is based on reversing the conditions by which mode mixing occurs presented in the paper One or Two frequencies? The Empirical Mode Decomposition Answers, in [3]. Numerical experiments with signals containing spectral components within the same octave shows the effective separation of modes that EMD can perform after this principle is applied. This separation technique has potential application for identifying the cause of different oscillatory modes with spectral proximity present in the smart grid.

Automated classification of lung sound signals based on empirical mode decomposition

This study proposes a novel, composite time-frequency distribution (TFD) constructed using a multiple-view approach. This composite TFD utilises the intrinsic mode functions (IMFs) of the empirical mode decomposition (EMD) to generate each view that are then combined using the arithmetic mean. This process has the potential to eliminate the inter-component interference generated by a quadratic TFD (QTFD), as the IMFs of the EMD are, in general, monocomponent signals. The formulation of the multiple-view TFD in the ambiguity domain results in faster computation, compared to a convolutive implementation in the time-frequency domain, and a more robust TFD in the presence of noise. The composite TFD, referred to as the EMD-TFD, was shown to generate a heuristically more accurate representation of the distribution of time-frequency energy in a signal. It was also shown to have performance comparable to the Wigner-Ville distribution when estimating the instantaneous frequency of multiple signal components in the presence of noise.