Classification-Based Fuel Injection Fault Detection of a Trainset Diesel Engine Using Vibration Signature Analysis

2020 ◽  
Vol 142 (5) ◽  
Author(s):  
Moosa Ayati ◽  
Farzad A. Shirazi ◽  
Saeed Ansari-Rad ◽  
Alireza Zabihihesari
Keyword(s):  
Diesel Engine ◽  
Fault Detection ◽  
Diesel Engines ◽  
Nearest Neighbor ◽  
Fuel Injection ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Vibration Signals ◽  
Feature Extraction Method ◽  
Time Frequency

Abstract Diesel engines are crucial components of trainsets. Automated fault detection of diesel engines can play an important role for increasing reliability of passenger trains. In this research, vibration-based fuel injection fault detection of a high-power 12-cylinder trainset diesel engine is studied. Vibration signals are analyzed in frequency and time-frequency domains to obtain possible patterns of faults. Fast Fourier transform (FFT) and wavelet packet transform (WPT) of vibration signals are used to extract several uncorrelated features. These features are chosen to increase the ability of classifiers to separate healthy and faulty engine sides, automatically. Different classification methods including multilayer perception (MLP), support vector machines (SVM), K-nearest neighbor (KNN), and local linear model tree (LOLIMOT) are used to process captured features; these methods are utilized in both “Single-sensor condition monitoring” and “Classification and fault detection” sections. It is shown that KNN networks are practical tools in the proposed fault detection procedure. The main novelty of this work comes from introducing a rich feature-extraction method based on a combination of FFT and db4 features. In addition, the complexity of computations and average running-time decrease while classification accuracy in the fuel injection fault detection procedure increases.

scholarly journals A Novel Improved Local Binary Pattern and Its Application to the Fault Diagnosis of Diesel Engine

2020 ◽  
Vol 2020 ◽  
pp. 1-15 ◽  
Author(s):  
Yanping Cai ◽  
Guanghua Xu ◽  
Aihua Li ◽  
Xu Wang
Keyword(s):  
Fault Diagnosis ◽  
Diesel Engine ◽  
Local Binary Pattern ◽  
Nearest Neighbor ◽  
Spatial Information ◽  
Wavelet Packet ◽  
Texture Feature ◽  
Support Vector ◽  
Vibration Signals ◽  
Time Frequency

Aiming at the feature extraction difficulty of vibration signals, an improved local binary pattern- (ILBP-) based diesel engine fault diagnosis approach is proposed. To effectively make use of the component spatial information in time-frequency images, local binary pattern (LBP) algorithm is applied. Also, in view of the problems that traditional LBP coding is easily interfered by singular pixel points and the relative spatial information is not prominent, an improved coding rule of the LBP operator is put forward in this paper. Compared with some typical LBP algorithms, computational complexity of the proposed ILBP algorithm is greatly reduced, and the coding sparsity is greatly improved. The ILBP operator is applied to fault diagnosis of BF4L1011F diesel engine with eight different valve conditions. For comparison, six kinds of time-frequency distribution are used to convert raw vibration signals into time-frequency images, and then circular LBP, rotation-invariant LBP, uniform LBP, and ILBP operator are applied for texture coding. Finally, nearest neighbor classifier (NNC) and support vector machine (SVM) are used for fault identification. The classification results show that the ILBP operator proposed in this paper can better describe the texture feature information in vibration time-frequency images of the diesel engine, and a good diagnostic effect can be achieved by combining wavelet packet (WP) distribution and ILBP.

scholarly journals Fault Detection of a Spherical Tank Using a Genetic Algorithm-Based Hybrid Feature Pool and k-Nearest Neighbor Algorithm

Energies ◽  
2019 ◽  
Vol 12 (6) ◽  
pp. 991 ◽  
Author(s):  
Md Hasan ◽  
Jong-Myon Kim
Keyword(s):  
Fault Detection ◽  
Nearest Neighbor ◽  
K Nearest Neighbor ◽  
Time Frequency ◽  
Reduction Techniques ◽  
Spherical Tank ◽  
Frequency Domains ◽  
Dimensionality Reduction Techniques ◽  
Linear Dimensionality Reduction ◽  
Ae Signals

Fault detection in metallic structures requires a detailed and discriminative feature pool creation mechanism to develop an effective condition monitoring system. Traditional fault detection methods incorporate handcrafted features either from the time, frequency or time-frequency domains. To explore the salient information provided by the acoustic emission (AE) signals, a hybrid of feature pool creation and an optimal features subset selection mechanism is proposed for crack detection in a spherical tank. The optimal hybrid feature pool creation process is composed of two major parts: (1) extraction of statistical features from time and frequency domains, as well as extraction of traditional features associated with the AE signals; and (2) genetic algorithm (GA)-based optimal features subset selection. The optimal features subset is then provided to the k-nearest neighbor (k-NN) classifier to distinguish between normal (NC) and crack conditions (CC). Experimental results show that the proposed approach yields an average 99.8% accuracy for heath state classification. To validate the effectiveness of the proposed approach, it is compared to conventional non-linear dimensionality reduction techniques, as well as those without feature selection schemes. Experimental results show that the proposed approach outperforms conventional non-linear dimensionality reduction techniques, achieving at least 2.55% higher classification accuracy.

scholarly journals An Investigation of Insider Threat Mitigation Based on EEG Signal Classification

Sensors ◽  
2020 ◽  
Vol 20 (21) ◽  
pp. 6365
Author(s):  
Jung Hwan Kim ◽  
Chul Min Kim ◽  
Man-Sung Yim
Keyword(s):  
Nearest Neighbor ◽  
Nuclear Industry ◽  
Classification Model ◽  
Support Vector ◽  
Insider Threat ◽  
K Nearest Neighbor ◽  
Eeg Signals ◽  
Nuclear Facilities ◽  
Time Frequency ◽  
Threat Mitigation

This study proposes a scheme to identify insider threats in nuclear facilities through the detection of malicious intentions of potential insiders using subject-wise classification. Based on electroencephalography (EEG) signals, a classification model was developed to identify whether a subject has a malicious intention under scenarios of being forced to become an insider threat. The model also distinguishes insider threat scenarios from everyday conflict scenarios. To support model development, 21-channel EEG signals were measured on 25 healthy subjects, and sets of features were extracted from the time, time–frequency, frequency and nonlinear domains. To select the best use of the available features, automatic selection was performed by random-forest-based algorithms. The k-nearest neighbor, support vector machine with radial kernel, naïve Bayes, and multilayer perceptron algorithms were applied for the classification. By using EEG signals obtained while contemplating becoming an insider threat, the subject-wise model identified malicious intentions with 78.57% accuracy. The model also distinguished insider threat scenarios from everyday conflict scenarios with 93.47% accuracy. These findings could be utilized to support the development of insider threat mitigation systems along with existing trustworthiness assessments in the nuclear industry.

scholarly journals Analysis of classification methods suitable for band limited spatially filtered EEG signal applicable to non-invasive BCI

2019 ◽  
Vol 8 (9) ◽  
pp. 1689-1694
Keyword(s):  
Feature Extraction ◽  
Nearest Neighbor ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Common Spatial Pattern ◽  
Linear Discriminant ◽  
Feature Extraction Method ◽  
Non Invasive ◽  
Band Limited ◽  
Ca Sensitivity

Electroencephalographic (EEG) signals are the preferred input for non-invasive Brain-Computer Interface (BCI). Efficient signal processing strategies, including feature extraction and classification, are required to distinguish the underlying task of BCI. This work proposes the optimized common spatial pattern(CSP) filtering technique as the feature extraction method for collecting the spatially spread variation of the signal. The bandpass filter (BPF) designed for this work assures the availability of event-related synchronized (ERS) and event-related desynchronized (ERD) signal as input to the spatial filter. This work takes consideration of the area-specific electrodes for feature formation. This work further proposes a comparative analysis of classifier algorithms for classification accuracy(CA), sensitivity and specificity and the considered algorithms are Support Vector Machine(SVM), Linear Discriminant Analysis(LDA), and K-Nearest Neighbor(KNN). Performance parameters considered are CA, sensitivity, and selectivity, which can judge the method not only for high CA but also inclining towards the particular class. Thus it will direct in the selection of appropriate classifier as well as tuning the classifier to get the balanced results. In this work, CA, the prior performance parameter is obtained to be 88.2% sensitivity of 94.2% and selectivity 82.2% for the cosine KNN classifier. SVM with linear kernel function also gives the comparable results, thus concluding that the robust classifiers perform well for all parameters in case of CSP for feature extraction.

scholarly journals Biometric authenticator algorithm based on multiresolution analysis

2020 ◽  
Vol 20 (3) ◽  
pp. 1332
Author(s):  
Soumia Kerrache ◽  
Beladgham Mohammed ◽  
Hamza Aymen ◽  
Kadri Ibrahim
Keyword(s):  
Feature Extraction ◽  
Multiresolution Analysis ◽  
Nearest Neighbor ◽  
Curvelet Transform ◽  
Principal Component ◽  
Image Features ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Feature Extraction Method ◽  
Fusion Approach

Features extraction is an essential process in identifying person biometrics because the effectiveness of the system depends on it. Multiresolution Analysis success can be used in the system of a person’s identification and pattern recognition. In this paper, we present a feature extraction method for two-dimensional face and iris authentication.  Our approach is a combination of principal component analysis (PCA) and curvelet transform as an improved fusion approach for feature extraction. The proposed fusion approach involves image denoising using 2D-Curvelet transform to achieve compact representations of curves singularities. This is followed by the application of PCA as a fusion rule to improve upon the spatial resolution. The limitations of the only PCA algorithm are a poor recognition speed and complex mathematical calculating load, to reduce these limitations, we are applying the curvelet transform. <br /> To assess the performance of the presented method, we have employed three classification techniques: Neural networks (NN), K-Nearest Neighbor (KNN) and Support Vector machines (SVM).<br />The results reveal that the extraction of image features is more efficient using Curvelet/PCA.

scholarly journals Enhanced Performance of PCG Signal using Effective Feature Extraction Method

2020 ◽  
Vol 9 (2) ◽  
pp. 809-813
Keyword(s):  
Nearest Neighbor ◽  
Heart Sound ◽  
Biomedical Signal Processing ◽  
Support Vector ◽  
Discrete Wavelet ◽  
K Nearest Neighbor ◽  
Feature Extraction Method ◽  
Signal Features ◽  
Heart Auscultation ◽  
Noisy Background

Phonocardiography (PCG) is the realistic portrayal of sounds created in the heart auscultation. PCG is an improvement for ECG. Particularly in observing of patient and biomedical research, these signals need to do the diagnosis. This paper deals with the processing of heart sound signals i.e., Phonocardiography (PCG) Signals. The primary goal of analyzing these heart sound signals is to separate the signals from the noisy background and to extract some parameters which are used for patient monitoring and for other researches. Various momentum explore ventures are going on biomedical signal processing and its applications. The performance of the PCG signal will comprise of sectioning the signal into S1 and S2 and then compare, whether the PCG is normal or abnormal. In the previous framework the different change approaches are utilized to break down the PCG signal.In the primary stage, for include extraction; acquired heart sound signals were isolated to its subgroups utilizing discrete wavelet change with Level-1 to Level-10. This upgraded strategy proposes a best component for Heart Signal Features, which are removed and changed in to other area to arrange signals. This enhanced method proposes a best feature for Heart Signal Features, which are extracted and transformed in to other domain to classify signals. In the proposed strategy the Wavelet is utilized for highlight extraction and different Statistical strategies are utilized. InformationGain (IG), Mutual Information (MI) and so on. Feature selection techniques are compared using classifiers like kNN(k-Nearest Neighbor), Naïve Bayes, C4.5 and Support Vector Machines (SVMs). MATLAB & WEKA Soft wares are used for analysis Purpose. In this paper, coiffelet technique is utilized to analyze the synthetic PCG and the classifier parameters are compared with one another.

scholarly journals A Comparative Analysis of Artificial Intelligence-Based Methods for Fault Diagnosis of Mechanical Systems

2019 ◽  
Vol 23 (1) ◽  
pp. 113-124
Author(s):  
Reihaneh Kardehi Moghaddam ◽  
Navid Moshtaghi Yazdan
Keyword(s):  
Artificial Intelligence ◽  
Fault Diagnosis ◽  
Nearest Neighbor ◽  
Composite Plates ◽  
Classification Systems ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Inference System ◽  
Time Frequency ◽  
Number Of Iterations

Abstract The present research studied fault diagnosis of composite sheets using vibration signal processing and artificial intelligence (AI)-based methods. To this end, vibration signals were collected from sound and faulty composite plates. Using different time-frequency signal analysis and processing methods, a number of features were extracted from these signals and the most effective features containing further information on these composite plates were provided as input to different classification systems. The output of these classification systems reveals the faults in composite plates. The different types of classification systems used in this research were the support vector machine (SVM), adaptive neuro-fuzzy inference system (ANFIS), k-nearest neighbor (k-NN), artificial neural networks (ANNs), Extended Classifier System (XCS) algorithm, and the proposed improved XCS algorithm. The research results were reflective of the superiority of ANFIS in terms of precision, while this method had the highest process duration with an equal number of iterations. The precision of the proposed improved XCS method was lower than that of ANFIS, but the duration of the process was shorter than the ANFIS method with an equal number of iterations.

scholarly journals EEG-Based Prediction of the Recovery of Carotid Blood Flow during Cardiopulmonary Resuscitation in a Swine Model

Sensors ◽  
2021 ◽  
Vol 21 (11) ◽  
pp. 3650
Author(s):  
Heejin Kim ◽  
Ki Hong Kim ◽  
Ki Jeong Hong ◽  
Yunseo Ku ◽  
Sang Do Shin ◽  
...  
Keyword(s):  
Blood Flow ◽  
Cardiopulmonary Resuscitation ◽  
Nearest Neighbor ◽  
Polynomial Kernel ◽  
Support Vector ◽  
Swine Model ◽  
K Nearest Neighbor ◽  
Carotid Blood Flow ◽  
Time Frequency ◽  
Svm Model

The recovery of cerebral circulation during cardiopulmonary resuscitation (CPR) is important to improve the neurologic outcomes of cardiac arrest patients. To evaluate the feasibility of an electroencephalogram (EEG)-based prediction model as a CPR feedback indicator of high- or low-CBF carotid blood flow (CBF), the frontal EEG and hemodynamic data including CBF were measured during animal experiments with a ventricular fibrillation (VF) swine model. The most significant 10 EEG parameters in the time, frequency and entropy domains were determined by neighborhood component analysis and Student’s t-test for discriminating high- or low-CBF recovery with a division criterion of 30%. As a binary CBF classifier, the performances of logistic regression, support vector machine (SVM), k-nearest neighbor, random forest and multilayer perceptron algorithms were compared with eight-fold cross-validation. The three-order polynomial kernel-based SVM model showed the best accuracy of 0.853. The sensitivity, specificity, F1 score and area under the curve of the SVM model were 0.807, 0.906, 0.853 and 0.909, respectively. An automated CBF classifier derived from non-invasive EEG is feasible as a potential indicator of the CBF recovery during CPR in a VF swine model.

scholarly journals Diagnosis of amyotrophic lateral sclerosis (ALS) disorders based on electromyogram (EMG) signal analysis and feature selection

2020 ◽  
Vol 26 (3) ◽  
pp. 155-160
Author(s):  
Aicha Mokdad ◽  
Sidi Mohammed El Amine Debbal ◽  
Fadia Meziani
Keyword(s):  
Nearest Neighbor ◽  
Biceps Brachii ◽  
Neuromuscular Disorders ◽  
Machine Learning Algorithms ◽  
Support Vector ◽  
Svm Classifier ◽  
K Nearest Neighbor ◽  
Time Frequency ◽  
Emg Signal ◽  
Frequency Features

AbstractElectromyogram signal (EMG) provides an important source of information for the diagnosis of neuromuscular disorders. In this study, we proposed two methods of analysis which concern the bispectrum and continuous wavelet transform (CWT) of the EMG signal then a comparison is made to select which one is the most suitable to identify an abnormality in biceps brachii muscle in the main purpose is to assess the pathological severity in bifrequency and time-frequency analysis applying respectively bispectrum and CWT. Then four time and frequency features are extracted and three popular machine learning algorithms are implemented to differentiate neuropathy and healthy conditions of the selected muscle. The performance of these time and frequency features are compared using support vector machine (SVM), linear discriminate analysis (LDA) and K-Nearest Neighbor (KNN) classifier performance. The results obtained showed that the SVM classifier yielded the best performance with an accuracy of 95.8%, precision of 92.59% and specificity of 92%. followed by respectively KNN and LDA classifier that achieved respectively an accuracy of 92% and 91.5%, precision of 92% and 85.4%, and specificity of 92% and 83%.

Integrating synthetic minority oversampling and gradient boosting decision tree for bogie fault diagnosis in rail vehicles

2018 ◽  
Vol 233 (3) ◽  
pp. 312-325 ◽  
Author(s):  
Linlin Kou ◽  
Yong Qin ◽  
Xunjun Zhao ◽  
Yong Fu
Keyword(s):  
Support Vector Machine ◽  
Fault Diagnosis ◽  
Fault Detection ◽  
Decision Tree ◽  
Nearest Neighbor ◽  
Naive Bayes ◽  
Naïve Bayes ◽  
Gradient Boosting ◽  
Support Vector ◽  
K Nearest Neighbor

Bogies are critical components of a rail vehicle, which are important for the safe operation of rail transit. In this study, the authors analyzed the real vibration data of the bogies of a railway vehicle obtained from a Chinese subway company under four different operating conditions. The authors selected 15 feature indexes – that ranged from time-domain, energy, and entropy – as well as their correlations. The adaptive synthetic sampling approach–gradient boosting decision tree (ADASYN–GBDT) method is proposed for the bogie fault diagnosis. A comparison between ADASYN–GBDT and the three commonly used classifiers (K-nearest neighbor, support vector machine, and Gaussian naïve Bayes), combined with random forest as the feature selection, was done under different test data sizes. A confusion matrix was used to evaluate those classifiers. In K-nearest neighbor, support vector machine, and Gaussian naïve Bayes, the optimal features should be selected first, while the proposed method of this study does not need to select the optimal features. K-nearest neighbor, support vector machine, and Gaussian naïve Bayes produced inaccurate results in multi-class identification. It can be seen that the lowest false detection rates of the proposed ADASYN–GBDT model are 92.95% and 87.81% when proportion of the test dataset is 0.4 and 0.9, respectively. In addition, the ADASYN–GBDT model has the ability to correctly identify a fault, which makes it more practical and suitable for use in railway operations. The entire process (training and testing) was finished in 2.4231 s and the detection procedure took 0.0027 s on average. The results show that the proposed ADASYN–GBDT method satisfied the requirements of real-time performance and accuracy for online fault detection. It might therefore aid in the fault detection of bogies.

Sign in / Sign up

Export Citation Format

Share Document