Papers by Keyword: Intrinsic Mode Function (IMF)

Abstract: Heavy mechanical devices of complex industrial process produce soundly mechanical vibration and acoustical signals. Some difficult-to-measure key process parameters have direct relationship with these signals. A newly ensemble empirical mode decomposition (EEMD), Fast Fourier Transform (FFT), Mutual information (MI), and Kernel partial least squares (KPLS) based modeling approach is proposed to measure these process parameters. At first, different scale intrinsic mode functions (IMFs) of mechanical vibration and acoustical signals are obtained using EEMD technology. Then, FFT transforms these multi-scale IMFs into frequency domain, and MI based feature selection method selects interesting frequency spectral features. Finally, KPLS constructs the final soft sensor models using the selected features. Experimental results based on vibration and acoustical signals of ball mill demonstrate this approach is more effective than other exist multi-scale decomposition based methods.

Abstract: Phase is an important feature of the local field potential (LFP) and plays a significant role in transmission and processing information in visual system. In this paper, the LFP of Long Evans rats primary visual cortex is recorded by the microelectrode array through the visual stimuli of the checkerboard and different orientation gratings. Then, a multi-mode phase extraction model based on the firing spikes was built. We found that neurons selective orientation information using the third intrinsic mode functions of local field potential during firing spikes.

Abstract: Low frequency oscillations due to the lack of damping may occur in power systems under normal operation and will cause system instability. These oscillations are essentially nonlinear power responses which are difficult to extract the inherent characteristics by the time domain method. This paper aims to analyze nonlinear power responses by using the Hilbert-Huang transform (HHT) which is a time-frequency signal processing method which comprises steps of the empirical mode decomposition and the Hilbert transform. Dynamic power system responses, including generator output power and line power are to be processed by the HHT and a set of intrinsic mode functions and the associated Hilbert spectrum are obtained. The generator with most effects on the system will be accordingly found out through the time-frequency analysis and the power system stabilizer will be placed at the generator. Numerical results from a sample power system are demonstrated to show the validity of the time-frequency approach in the study of power system low frequency oscillations.

Abstract: Speech endpoint detection plays an important role in speech signal processing. In this paper, a method of speech endpoint detection based on empirical mode decomposition is introduced for accurately detecting the speech endpoint. This method used in speech signal decomposition gets a set of intrinsic mode functions (IMF). An IMF which contained a lot of noise must be filtered, and the rest of IMFs can be reconstructed to a new speech signal. The speech endpoint is detected by average magnitude difference function precisely. Simulation experiments show that the method proposed in this paper can eliminate the impact of noise effectively and detect the speech signal endpoint accurately.

Abstract: A new pitch detection method is designed by the recurrence analysis in this paper, which is combined of Empirical Mode Decomposition (EMD) and Elliptic Filter (EF). The Empirical Mode Decomposition (EMD) of Hilbert-Huang Transform (HHT) are utilized tosolve the problem, and a noisy voice is first filtered on the elliptic band filter. The two Intrinsic Mode Functions (IMF) are synthesized by EMD with maximum correlation of voice, and then the pitch be easily divided. The results show that the new method performance is better than the conventional autocorrelation algorithm and cepstrum method, especially in the part that the surd and the sonant are not evident, and get a high robustness in noisy environment.

Abstract: The spatio-temporal variations of extreme temperature in North China during 1954-2008 are analyzed based on Empirical Mode Decomposition (EMD) method. The results show that the interannual-to-interdecadal variabilities of extreme temperature in North China. 3-4 years and 7-8 years interannual variabilities as well as a decreasing trend are dominant while 15 years oscillation takes second place in the extreme low temperature variation. Meanwhile, 3-4 year interannual variabilities are dominant while 6-7 years, 14-15 years 40 years oscillation as well as an increasing trend takes second place in the extreme high temperature variation. 3-4 years interannual variabilities of both extreme low and high temperature are closely related to El Nino-Southern Oscillation (ENSO), while interdecadal variabilities and trend of extreme temperature are obviously associated with the Pacific Decadal Oscillation (PDO). Besides, PDO plays an important role on interdecadal modulation of interannual to decadal variabilities of extreme temperature in North China.

Abstract: A kind of soft sensing is proposed by combining empirical mode decomposition(EMD) with support vector machine optimized by improved particle swarm optimization (IPSO-SVM). EMD is a highly adaptive decomposition and can decompose any complicated signal into so called Intrinsic Mode Functions (IMF), which not only has excellent performance of feature extraction but also can reduce the dimension of the model input data space. we can extracts IMF energy feature as the input feature vectors of IPSO-SVM. Support vector machine (SVM) has been successfully employed to solve regression problem but it is difficult to select appropriate SVM parameters. A new SVM model based on adaptive particle swarm optimization (APSO) for parameter optimization is proposed which not only has strong global search capability, but also is very easy to implement. The proposed method is used to build soft sensing of diesel oil solidifying point. Compared with other two models, the result shows that IPSO-SVM approach has a better prediction and generalization.

Abstract: With the continuous development of power market, the precision requirement for short-term power load forecasting is constantly being improved. In order to obtain higher prediction accuracy, this paper put forward a method of combining empirical mode decomposition (EMD) with echo state network (ESN) for short-term power load forecasting. First, original data had been decomposed into several independent components, whose features were obvious. A corresponding echo state network was built for each component. Then, each component should be trained and predicted by its corresponding echo state network. The experimental results showed that this method has a better prediction accuracy compared with traditional neural network method.

Abstract: Obstructive sleep apnea (OSA) is one of the most important sleep disorders. The gold standard diagnosis of OSA is overnight PSG examination that is time-consuming and labor intensive. Overnight ECG signal was developed to examine OSA, with easy implementation and portable equipment. There were various ECG derived features used for OSA identification, in this study, intrinsic mode function (IMF) was developed. IMF is a byproduct of Hilbert-Huang transform. IMF decompose original signal into various sub components, due to its complexity. In this study, some novel IMF derived features were used to examine the OSA duration measured from ECG signal, compared with traditional HRV features.

Abstract: Hilbert-Huang transform (HHT) is a signal processing technique is relatively strong and effective today, to overcome the limitations of the used techniques have been widely used as traditional Fourier transform or Wavelet transform. Hilbert-Huang transform can handle effectively nonstationary and nonlinear signal. In this paper present how to use the HHT method to detect damage of the beams under the effect of moving load. Location damage is determined by observing changes in the first instantaneous frequency curve (IF1), damage location is the highest peak in the IF curve. This paper briefly described the theoretical basis and then applies through numerical simulation. These numerical simulations are simply supported beams damaged at different locations and with different levels of damage under the effect of the moving velocity case different. The results of numerical simulation shows relative accuracy compared with the theory and practice is assumed.