Learning-based fault diagnosis of air brake system using wheel speed data

2021 ◽  
pp. 095440702110637
Author(s):  
Radhika Raveendran ◽  
KB Devika ◽  
Shankar C Subramanian
Keyword(s):  
Learning Algorithm ◽  
Dynamic Performance ◽  
Initial Step ◽  
Wheel Speed ◽  
Operating Conditions ◽  
Brake System ◽  
Gaussian Kernel ◽  
Support Vector ◽  
Wide Range ◽  
Air Brake System

Faults in the air brake system used in Heavy Commercial Road Vehicles (HCRVs) would adversely affect the vehicle’s dynamic performance, and hence their prompt detection is critical for vehicle safety. This paper first investigates the effect of air brake system faults through extensive hardware-in-loop experiments. These faults were observed to degrade the braking response, yaw stability, and vehicle braking distance. In many countries, an antilock brake system is mandatory in HCRVs, and wheel speed data are readily available. Inspired by this, the feasibility of using wheel speed data to detect faults is investigated in this study. As an initial step of predictive maintenance, a fault diagnostic scheme based on a supervised learning algorithm, Support Vector Machine (SVM) that uses only wheel speed data has been developed. The SVM algorithm’s efficacy was tested for 1937 test cases that encompassed a wide range of operating conditions. It was found that a Gaussian kernel SVM (G-SVM) provided a good classification accuracy of 96.54%, demonstrating its ability to predict a faulty condition accurately. The standard deviation of G-SVM’s prediction accuracy for five groups of data sets with 100 instances was found to be 1.57%, which shows that the model is more precise to predict the fault/no-fault condition of the air brake system.

Adapting SVM for data sparseness and imbalance: a case study in information extraction

2009 ◽  
Vol 15 (2) ◽  
pp. 241-271 ◽  
Author(s):  
YAOYONG LI ◽  
KALINA BONTCHEVA ◽  
HAMISH CUNNINGHAM
Keyword(s):  
Active Learning ◽  
Language Learning ◽  
Information Extraction ◽  
Language Processing ◽  
Learning Algorithm ◽  
Machine Learning Algorithms ◽  
Training Data ◽  
Support Vector ◽  
Passive Learning ◽  
Wide Range

AbstractSupport Vector Machines (SVM) have been used successfully in many Natural Language Processing (NLP) tasks. The novel contribution of this paper is in investigating two techniques for making SVM more suitable for language learning tasks. Firstly, we propose an SVM with uneven margins (SVMUM) model to deal with the problem of imbalanced training data. Secondly, SVM active learning is employed in order to alleviate the difficulty in obtaining labelled training data. The algorithms are presented and evaluated on several Information Extraction (IE) tasks, where they achieved better performance than the standard SVM and the SVM with passive learning, respectively. Moreover, by combining SVMUM with the active learning algorithm, we achieve the best reported results on the seminars and jobs corpora, which are benchmark data sets used for evaluation and comparison of machine learning algorithms for IE. In addition, we also evaluate the token based classification framework for IE with three different entity tagging schemes. In comparison to previous methods dealing with the same problems, our methods are both effective and efficient, which are valuable features for real-world applications. Due to the similarity in the formulation of the learning problem for IE and for other NLP tasks, the two techniques are likely to be beneficial in a wide range of applications1.

A Robust Pressure Controller for a Variable Speed AC Motor Pump: Application to Aircraft Hydraulic Power Packages

2018 ◽  
Author(s):  
Nils Trochelmann ◽  
Phillip Bischof Stump ◽  
Frank Thielecke ◽  
Dirk Metzler ◽  
Stefan Bassett
Keyword(s):  
Distribution Systems ◽  
Controller Design ◽  
Dynamic Performance ◽  
Operating Conditions ◽  
Gear Pump ◽  
Variable Speed ◽  
Hydraulic Power ◽  
Worst Case ◽  
Wide Range ◽  
Pressure Controller

Highly integrated electro-hydraulic power packages with electric motor-driven pumps (EMP) are a key technology for future aircraft with electric distribution systems. State of the art aircraft EMPs are robust but lack efficiency, availability, and have high noise emissions. Variable speed fixed displacement (VSFD-) EMPs, combining a permanent magnet synchronous motor and an internal gear pump, show promising properties regarding noise reduction and energy efficiency. Though, meeting the strict dynamic requirements is tough with this EMP-concept. Speed limitations and inertia impose strong restrictions on the achievable dynamic performance. Moreover, the requirements must be met under a wide range of operating conditions. For a prototype aircraft VSFD-EMP a robust pressure controller design is proposed in this paper. In a first step the operating conditions of the EMP are defined, analyzing environmental conditions and impacts of the interfacing aircraft systems. Nonlinear and linear control design models are developed and validated by measurements at an EMP test rig built for this project. A conventional cascade pressure control concept is selected. This is motivated by the demand for simple, reliable, and proven solutions in aerospace applications. A controller is designed by applying classical loop shaping techniques. Robust stability and performance of the system are investigated through a subsequent μ-analysis. Finally, the controller is tested under nominal and worst case conditions in nonlinear simulations.

scholarly journals Fast Incremental SVDD Learning Algorithm with the Gaussian Kernel

2019 ◽  
Vol 33 ◽  
pp. 3991-3998 ◽  
Author(s):  
Hansi Jiang ◽  
Haoyu Wang ◽  
Wenhao Hu ◽  
Deovrat Kakde ◽  
Arin Chaudhuri
Keyword(s):  
Outlier Detection ◽  
Learning Algorithm ◽  
Large Data ◽  
Real Data ◽  
Gaussian Kernel ◽  
Support Vector ◽  
Support Vector Data Description ◽  
Detection Accuracy ◽  
Single Class ◽  
Support Vectors

Support vector data description (SVDD) is a machine learning technique that is used for single-class classification and outlier detection. The idea of SVDD is to find a set of support vectors that defines a boundary around data. When dealing with online or large data, existing batch SVDD methods have to be rerun in each iteration. We propose an incremental learning algorithm for SVDD that uses the Gaussian kernel. This algorithm builds on the observation that all support vectors on the boundary have the same distance to the center of sphere in a higher-dimensional feature space as mapped by the Gaussian kernel function. Each iteration involves only the existing support vectors and the new data point. Moreover, the algorithm is based solely on matrix manipulations; the support vectors and their corresponding Lagrange multiplier αi’s are automatically selected and determined in each iteration. It can be seen that the complexity of our algorithm in each iteration is only O(k2), where k is the number of support vectors. Experimental results on some real data sets indicate that FISVDD demonstrates significant gains in efficiency with almost no loss in either outlier detection accuracy or objective function value.

Dynamic Force Control of Displacement Controlled Drives by Means of an Iterative Learning Algorithm

2013 ◽  
Author(s):  
Tobias Radermacher ◽  
Jürgen Weber ◽  
Dominik Dorner
Keyword(s):  
Learning Algorithm ◽  
Control Strategies ◽  
Dynamic Performance ◽  
Iterative Learning ◽  
Dynamic Force ◽  
Swash Plate ◽  
Wide Range ◽  
Drive Trains ◽  
Efficient Alternative ◽  
Valve Control

Displacement controlled drive trains are applied more and more in industry because they are an energy efficient alternative to valve control. One major burden, for the use in a wide range of application, is the lower dynamic performance, which is caused by the large inertial forces of either swash plate pumps or speed controlled constant pumps. Additional challenges are the lack of control strategies, which determine the overall system dynamic performance. Starting with a comparison of four different linear drive trains (valve control, electro-mechanic, pump control, speed control) it is shown that the potential of a speed controlled drive train can be obtained by the use of an iterative learning controller. Once the controller is parameterized it is able to improve the tracking behavior significantly without having any effect on the plants stability.

An Investigation of Friction Dynamics of Mass Spring System With Base Excitation

Tribology ◽  
2005 ◽  
Author(s):  
A. Fawzy ◽  
Y. K. Youness ◽  
A. M. A. El-Butch ◽  
I. M. Ibrahim
Keyword(s):  
Dynamic Performance ◽  
Excitation Frequency ◽  
Operating Conditions ◽  
System Dynamic ◽  
Base Excitation ◽  
Wide Range ◽  
Excited Oscillation ◽  
Spring System ◽  
Base Displacement ◽  
Mass Spring

The friction force between coupled machine members in relative motion exerts sometimes significant influence on system dynamic behavior, giving rise to undesirable self-excited oscillation. The aim of this paper is to investigate the contribution of each design parameter and operating conditions on the system dynamic performance theoretically and experimentally. The theory presents a comparative analysis between different dynamic coefficient of friction which may exist between the sliding mass which connected to two springs and a harmonic base excitation surfaces. Furthermore, results are obtained for arbitrary values of natural frequency, excitation frequency and amplitude of exciting base displacement. A test rig has been constructed to test several contacting material combinations. Results for sliding mass acceleration have been obtained for a wide range of exciting base frequencies, masses of slider and amplitudes of exciting base displacement, good agreement with theoretical results have been achieved.

scholarly journals SlurryNet: Predicting Critical Velocities and Frictional Pressure Drops in Oilfield Suspension Flows

Energies ◽  
2021 ◽  
Vol 14 (5) ◽  
pp. 1263
Author(s):  
Alireza Sarraf Shirazi ◽  
Ian Frigaard
Keyword(s):  
Machine Learning ◽  
Critical Velocity ◽  
Flow Regimes ◽  
Operating Conditions ◽  
Support Vector ◽  
Slurry Flow ◽  
Pressure Drops ◽  
Integrated Method ◽  
Physical Knowledge ◽  
Wide Range

Improving the accuracy of the slurry flow predictions in different operating flow regimes remains a major focus for multiphase flow research, and it is especially targeted at industrial applications such as oil and gas. In this paper we develop a robust integrated method consisting of an artificial neural network (ANN) and support vector regression (SVR) to estimate the critical velocity, the slurry flow regime change, and ultimately, the frictional pressure drop for a solid–liquid slurry flow in a horizontal pipe, covering wide ranges of flow and geometrical parameters. Three distinct datasets were used to develop machine learning models with totals of 100, 325, and 125 data points for critical velocity, and frictional pressure drops for heterogeneous and bed-load regimes respectively. For each dataset, 80% of the data were used for training and the rest 20% for evaluating the out of sample performance. The K-fold technique was used for cross-validation. The prediction results of the developed integrated method showed that it significantly outperforms the widely used existing correlations and models in the literature. Additionally, the proposed integrated method with the average absolute relative error (AARE) of 0.084 outperformed the model developed without regime classification with the AARE of 0.155. The proposed integrated model not only offers reliable predictions over a wide range of operating conditions and different flow regimes for the first time, but also introduces a general framework of how to utilize prior physical knowledge to achieve more reliable performances from machine learning methods.

scholarly journals Autism Spectrum Disorder Classification Using Deep Learning

2021 ◽  
Vol 17 (08) ◽  
pp. 103
Author(s):  
Abdulrazak Yahya Saleh ◽  
Lim Huey Chern
Keyword(s):  
Artificial Intelligence ◽  
Autism Spectrum Disorder ◽  
Deep Learning ◽  
Learning Algorithm ◽  
Learning Algorithms ◽  
Image Data ◽  
Autism Spectrum ◽  
Spectrum Disorder ◽  
Support Vector ◽  
Wide Range

<p class="0abstract">The goal of this paper is to evaluate the deep learning algorithm for people placed in the Autism Spectrum Disorder (ASD) classification. ASD is a developmental disability that causes the affected people to have significant communication, social, and behavioural challenges. People with autism are saddled with communication problems, difficulties in social interaction and displaying repetitive behaviours. Several methods have been used to classify the ASD from non-ASD people. However, there is a need to explore more algorithms that can yield better classification performance. Recently, deep learning methods have significantly sharpened the cutting edge of learning algorithms in a wide range of artificial intelligence tasks. These artificial intelligence tasks refer to object detection, speech recognition, and machine translation. In this research, the convolutional neural network (CNN) is employed. This algorithm is used to find processes that can classify ASD with a higher level of accuracy. The image data is pre-processed; the CNN algorithm is then applied to classify the ASD and non-ASD, and the steps of implementing the CNN algorithm are clearly stated. Finally, the effectiveness of the algorithm is evaluated based on the accuracy performance. The support vector machine (SVM) is utilised for the purpose of comparison. The CNN algorithm produces better results with an accuracy of 97.07%, compared with the SVM algorithm. In the future, different types of deep learning algorithms need to be applied, and different datasets can be tested with different hyper-parameters to produce more accurate ASD classifications.</p>

scholarly journals Fault classification in smart distribution network using support vector machine

2020 ◽  
Vol 18 (3) ◽  
pp. 1148
Author(s):  
Ong Wei Chuan ◽  
Nur Fadilah Ab Aziz ◽  
Zuhaila Mat Yasin ◽  
Nur Ashida Salim ◽  
Norfishah A. Wahab
Keyword(s):  
Machine Learning ◽  
Support Vector Machine ◽  
Distribution System ◽  
Learning Algorithm ◽  
Work Load ◽  
Fault Classification ◽  
Support Vector ◽  
Initial Finding ◽  
Wide Range ◽  
Type Classification

<span>Machine learning application have been widely used in various sector as part of reducing work load and creating an automated decision making tool. This has gain the interest of power industries and utilities to apply machine learning as part of the operation. Fault identification and classification based machine learning application in power industries have gain significant accreditation due to its great capability and performance. In this paper, a machine-learning algorithm known as Support Vector Machine (SVM) for fault type classification in distribution system has been developed. Eleven different types of faults are generated with respect to actual network. A wide range of simulation condition in terms of different fault impedance value as well as fault types are considered in training and testing data. Right setting parameters are important to learning results and generalization ability of SVM. Gaussian radial basis function (RBF) kernel function has been used for training of SVM to accomplish the most optimized classifier. Initial finding from simulation result indicates that the proposed method is quick in learning and shows good accuracy values on faults type classification in distribution system. The developed algorithm is tested on IEEE 34 bus and IEEE 123 bus test distribution system. </span>

scholarly journals LQR Tuning Using AIS for Frequency Oscillation Damping

2020 ◽  
Vol 3 (4) ◽  
pp. 112
Author(s):  
Muhammad Abdillah ◽  
Teguh Aryo Nugroho ◽  
Herlambang Setiadi
Keyword(s):  
Optimal Control ◽  
Power System ◽  
Clonal Selection ◽  
Dynamic Performance ◽  
Operating Conditions ◽  
Load Frequency Control ◽  
Primary Control ◽  
Frequency Oscillation ◽  
Lqr Control ◽  
Wide Range

Commonly, primary control, i.e. governor, in the generation unit had been employed to stabilize the change of frequency due to the change of electrical load during system operation. But, the drawback of the primary control was it could not return the frequency to its nominal value when the disturbance was occurred. Thus, the aim of the primary control was only stabilizing the frequency to reach its new value after there were load changes. Therefore, the LQR control is employed as a supplementary control called Load Frequency Control (LFC) to restore and keep the frequency on its nominal value after load changes occurred on the power system grid. However, since the LQR control parameters were commonly adjusted based on classical or Trial-Error Method (TEM), it was incapable of obtaining good dynamic performance for a wide range of operating conditions and various load change scenarios. To overcome this problem, this paper proposed an Artificial Immune System (AIS) via clonal selection to automatically adjust the weighting matrices, Q and R, of LQR related to various system operating conditions changes. The efficacy of the proposed control scheme was tested on a two-area power system network. The obtained simulation results have shown that the proposed method could reduce the settling time and the overshoot of frequency oscillation, which is better than conventional LQR optimal control and without LQR optimal control.

The Box-Cox Transformation-Based ARFNNs for Identification of Nonlinear MR Damper System with Outliers and Skewness Noises

2013 ◽  
Vol 284-287 ◽  
pp. 2120-2123
Author(s):  
Pi Yun Chen ◽  
Yu Yi Fu ◽  
Kuo Lan Su ◽  
Jin Tsong Jeng
Keyword(s):  
Neural Networks ◽  
Fuzzy Inference ◽  
Learning Algorithm ◽  
Mr Damper ◽  
Fuzzy Neural Networks ◽  
Gaussian Kernel ◽  
Support Vector ◽  
Fuzzy Neural ◽  
Inference Systems ◽  
Box Cox Transformation

In this paper, the Box–Cox transformation-based annealing robust fuzzy neural networks (ARFNNs) are proposed for identification of the nonlinear Magneto-rheological (MR) damper with outliers and skewness noises. Firstly, utilizing the Box-Cox transformation that its object is usually to make residuals more homogeneous in regression, or transform data to be normally distributed. Consequently, a support vector regression (SVR) method with Gaussian kernel function has the good performance to determine the number of rule in the simplified fuzzy inference systems and initial weights in the fuzzy neural networks. Finally, the annealing robust learning algorithm (ARLA) can be used effectively to adjust the parameters of the Box-Cox transformation-based ARFNNs. Simulation results show the superiority of the proposed method for the nonlinear MR damper systems with outliers and skewness noises.

Sign in / Sign up

Export Citation Format

Share Document