Wyniki wyszukiwania - BazTech

1

Time - frequency method and artificial neural network classifier for induction motor drive system defects classification

Behim Meriem, Merabet Leila, Saad Salah

Diagnostyka

|

2024

|

Vol. 25, No. 1

art. no. 2024110

EN

In this paper, by introducing two statistical parameters, energy and L-kurtosis, a new fault diagnostic system combining Wavelet Packet Decomposition and Multilayer Perceptron Neural Network is designed to improve efficiency and precision of induction motor defects diagnosis. This method is applied to vibratory signals of asynchronous motor running at two different rotational speeds (1500 rpm and 2000 rpm) at a sampling frequency of 8 KHz to detect three main types of defects: bearing faults, load imbalance and misalignment. These speeds are considered as the usual medium running speeds of induction motor. According to the results, the high performance and accuracy of this new faults diagnostic system is proved and confirmed, thus it can be used in the detection of other machines defects.

2

Electric vehicle motor fault diagnosis using improved wavelet packet decomposition and particle swarm optimization algorithm

Zheng Wenfang, Wang Tieying

Archives of Electrical Engineering

|

2024

|

Vol. 73, nr 2

481--498

EN

This study addresses the issue of diagnosing faults in electric vehicle motors and presents a method utilizing Improved Wavelet Packet Decomposition (IWPD) combined with particle swarm optimization (PSO). Initially, the analysis focuses on common demagnetization faults, inter turn short circuit faults, and eccentricity faults of permanent magnet synchronous motors. The proposed approach involves the application of IWPD for extracting signal feature vectors, incorporating the energy spectrum scale, and extracting the feature vectors of the signal using the energy spectrum scale. Subsequently, a binary particle swarm optimization algorithm is employed to formulate strategies for updating particle velocity and position. Further optimization of the binary particle swarm algorithm using chaos theory and the simulated annealing algorithm results in the development of a motor fault diagnosis model based on the enhanced particle swarm optimization algorithm. The results demonstrate that the chaotic simulated annealing algorithm achieves the highest accuracy and recall rates, at 0.96 and 0.92, respectively. The model exhibits the highest fault accuracy rates on both the test and training sets, exceeding 98.2%, with a minimal loss function of 0.0035. Following extraction of fault signal feature vectors, the optimal fitness reaches 97.4%. In summary, the model constructed in this study demonstrates effective application in detecting faults in electric vehicle motors, holding significant implications for the advancement of the electric vehicle industry.

3

A Rattle Signal Denoising and Enhancing Method Based on Wavelet Packet Decomposition and Mathematical Morphology Filter for Vehicle

Liang Linyuan, Chen Shuming, Li Peiran

Archives of Acoustics

|

2022

|

Vol. 47, No. 1

43--55

EN

Buzz, squeak and rattle (BSR) noise has become apparent in vehicles due to the significant reductions in engine noise and road noise. The BSR often occurs in driving condition with many interference signals. Thus, the automatic BSR detection remains a challenge for vehicle engineers. In this paper, a rattle signal denoising and enhancing method is proposed to extract the rattle components from in-vehicle background noise. The proposed method combines the advantages of wavelet packet decomposition and mathematical morphology filter. The critical frequency band and the information entropy are introduced to improve the wavelet packet threshold denoising method. A rattle component enhancing method based on multi-scale compound morphological filter is proposed, and the kurtosis values are introduced to determine the best parameters of the filter. To examine the feasibility of the proposed algorithm, synthetic brake caliper rattle signals with various SNR ratios are prepared to verify the algorithm. In the validation analysis, the proposed method can well remove the disturbance background noise in the signal and extract the rattle components with well SNR ratios. It is believed that the algorithm discussed in this paper can be further applied to facilitate the detection of the vehicle rattle noise in industry.

4

A novel deep LSTM network for artifacts detection in microelectrode recordings

Hosny Mohamed, Zhu Minwei, Gao Wenpeng, Fu Yili

Biocybernetics and Biomedical Engineering

|

2020

|

Vol. 40, no. 3

1052--1063

EN

Microelectrode recording (MER) signals are world-widely used for validating the planned trajectories in the procedure of deep brain stimulation (DBS) surgery to obtain accurate position of electrodes inside the brain structure. Besides, MER signals are important source for studying extracellular neuronal activity and DBS biomarkers, such as, spike clustering and sorting. However, MER signals are prone to several artifacts derived from electrical equipment in the operating room, electrode movement and patient activities, etc., which reduce the signal-to-noise ratio of the MER signals. Therefore, in this paper, we propose a novel deep learning architecture based on long short-term memory (LSTM) network for automatic artifact detection in MER signals. Frequency and time-domain features were extracted from the raw MER signals and fed to the deep LSTM network. A manually annotated MER database obtained from 17 Parkinson's disease (PD) patients were used to validate the proposed architecture. The proposed architecture achieved promising results of 97.49% accuracy, 98.21% sensitivity and 96.87% specificity on an unseen test set. To our best knowledge, this is the first study to use LSTM network for artifacts detection in MER signals. The MER data will be available at http://homepage.hit.edu.cn/wpgao.

5

Pattern recognition for partial discharge in GIS based on pulse coupled neural networks and wavelet packet decomposition

Zhou J., Tang J., Zhang X., Tao J.

Przegląd Elektrotechniczny

|

2012

|

R. 88, nr 5b

44-47

EN

Based on the characteristics of partial discharge (PD) defects in gas insulated switchgear (GIS), four typical single defects were designed for the present paper. PD three-dimensional (3D) patterns were constructed based on the ultra high frequency detection systems. The pulse-coupled neural networks (PCNN) and wavelet packet decomposition (WPD) method were used in PD feature extraction. The recognition results show that the proposed method used in PD feature extraction can effectively improve the accuracy of pattern recognition rate.

PL

Przeanalizowano defekty wyłącznika gazowego z wyładowaniem niezupełnym. Defekty te przedstawiane są jako obrazy 3D. Do ekstrakcji cech tych obrazów wykorzystuje się transformatę falkową i impulsowo sprzężone sieci neuronowe.

6

Machined surface quality estimation based on wavelet packets parameters of the surface image

Zawada-Tomkiewicz A.

Pomiary Automatyka Kontrola

|

2010

|

R. 56, nr 6

606-609

EN

Machined surface image, destined for monitoring, was represented by the diagnostic feature vector, correlated with maladjustment of the process. Maladjustment is manifested by the increase of the signal random component in relation to deterministic one. Such a behavior of the system was the basis for diagnostic measure elaboration. From the definition of entropy, it increases with the increase of the random component. So, the entropy of normalized energy vector for optimal decomposition tree coefficients was selected as the diagnostic measure. The correlation between the entropy of energy vector of decomposition tree coefficients and machined surface parameters and tool wear was demonstrated.

PL

Środowisko zautomatyzowanego wytwarzania wymaga szybkich pomiarów chropowatości, jeszcze przed ostatecznym ukształtowaniem wyrobu. Do tej pory nie udało się opracować takiego układu, który spełniałby wymagania systemów sterowania produkcją, ze względu na wiele niekontrolowanych czynników wpływających na ostateczną jakość powierzchni. Jednym z możliwych do zastosowań przemysłowych układów inspekcji powierzchni jest system bazujący na obrazie tej powierzchni, szerzej omówiony w [1-4]. Obraz powierzchni po obróbce toczeniem, podobnie jak powierzchnia, jest strukturą kierunkową, w której dane zebrane w kierunku prostopadłym do śladów obróbki, w kolejnych chwilach czasu, reprezentują przebieg procesu skrawania. W artykule dane jasności obrazu powierzchni, zebrane w kierunku prostopadłym do kierunkowości, modelowano przy zastosowaniu pakietów falkowych. Problem analizy czasowo-częstotliwościowej obrazów powierzchni po toczeniu sprowadził się do przeprowadzenia odpowiedniego schematu dekompozycji przestrzeni czasowo-częstotliwościowej i wyznaczenia wektorów składowych [5-7]. Dobór drzewa dekompozycji pozwolił na ograniczenie liczby analizowanych wektorów składowych do ośmiu w2,2, w2,3, w3,3, w4,0, w4,1, w4,2, w4,3.. Zaproponowano miarę nieuporządkowania energii (entropia znormalizowanej energii) dla opisu poszczególnych składowych drzewa dekompozycji [8], które skutecznie charakteryzują nieregularność profilu obrazu powierzchni i rozregulowanie procesu. Potwierdzono zależność statystyczną między wartościami miar nieuporządkowania energii poszczególnych składowych a parametrami powierzchni i zużycia ostrza.

7

A novel approach of analog fault classification using a support vector machines classifier

Cui J., Wang Y.

Metrology and Measurement Systems

|

2010

|

Vol. 17, nr 4

561-581

EN

In order to make the analog fault classification more accurate, we present a method based on the Support Vector Machines Classifier (SVC) with wavelet packet decomposition (WPD) as a preprocessor. In this paper, the conventional one-against-rest SVC is resorted to perform a multi-class classification task because this classifier is simple in terms of training and testing. However, this SVC needs all decision functions to classify the query sample. In our study, this classifier is improved to make the fault classification task more fast and efficient. Also, in order to reduce the size of the feature samples, the wavelet packet analysis is employed. In our investigations, the wavelet analysis can be used as a tool of feature extractor or noise filter and this preprocessor can improve the fault classification resolution of the analog circuits. Moreover, our investigation illustrates that the SVC can be applicable to the domain of analog fault classification and this novel classifier can be viewed as an alternative for the back-propagation (BP) neural network classifier.

8

Dekompozycja falkowa profilu powierzchni obrobionej po toczeniu

Zawada-Tomkiewicz A.

Pomiary Automatyka Kontrola

|

2009

|

R. 55, nr 4

243-246

PL

Powierzchnię obrobioną po toczeniu można scharakteryzować jako zbiór różnych składowych częstotliwości. Jedną z metod opisu tego typu powierzchni jest dekompozycja przy zastosowaniu transformaty pakietów falkowych. Celem pracy była optymalizacja struktury drzewa uzyskanych metodą pakietów falkowych dla dekompozycji profilu powierzchni obrobionej. Dokonano rozdziału sygnału profilu dla 1480 przypadków różnych profili powierzchni i uogólniono statystycznie dobór najlepszego drzewa.

EN

The optimisation of wavelet packets tree structure for turned surface decomposition is the aim of the work. The machined surface after the process of turning is characterised by the set of irregularities left by tool wedge pass. Tool geometry, feed rate and tool chipping are the basic factors influencing the surface features [1, 2]. The machined surface profile can be described by a continuous spectrum of wavelength. The information about the tool chipping and other maladjustments is contained in higher frequency components [3, 4, 5]. One of the techniques of machined surface profile decomposition is a wavelet packet transform. It enables to select the appropriate wavelet function and, simultaneously, to perform deep and complete analysis [6, 7, 8]. The wavelet packet decomposition was performed with use of the Coiflet wavelet function. The previous investigations had confirmed the usefulness of this wavelet function for the turned surface description [8]. For 1480 different machined surface profiles the best tree was statistically elaborated with use of the entropy criterion [10, 11, 12]. The discrete wavelet transform tree was indicated to be optimal for 69% machined surface profile decompositions. The completion of this tree with additional detail vectors enabled the analysis to be more thorough. The statistically optimal tree contained 97% optimal subtrees (Tab. 1.) and was treated as an optimal one.

9

Improvement of glass break acoustic signal detection via application of Wavelet Packet Decompositions

Bemke I., Zielonko R.

Metrology and Measurement Systems

|

2008

|

Vol. 15, nr 4

513-526

EN

The main subject of the authors' research are non-contact methods of glass break detection based on analysis of the acoustic signal generated during the event. This problem has essential meaning for modern cost-effective alarm systems, particularly those installed into big buildings. The main diffculties of the matter are: transient, stochastic character of the signal, great number of similar sounds (false signals, mainly accidental glass hits without break) and variability of many parameters (e.g. size and thickness of the glass pane, distance to detector). During research the authors developed a detection algorithm based on Wavelet Transformation (WT) and found some measures allowing to extract distinctive features from the signal and their classification. The obtained detection effciency (>90%) is satisfactory, but immunity against false signals (near to 80%) does not reach the assumed level. Because Wavelet Packet Decomposition (WPD) provides a more detailed analysis in the frequency domain than WT and does better extraction of time-frequency interdependencies of the signals, the authors decide to use it for algorithm improvement. This paper discusses results of WPD application to improve system performance and to increase the immunity against false signals. In the paper, on the background of a description of the problem, a theoretical basis of the WPD method and results of the investigation of its effectiveness are presented.