Gearbox Fault Detection Using Neural Networks Technology

1999 ◽  
Author(s):  
Qiao Sun ◽  
Xiaolei Li ◽  
Baoyun Xu
Keyword(s):  
Genetic Algorithm ◽  
Neural Networks ◽  
Feature Extraction ◽  
Fault Diagnosis ◽  
Learning Algorithm ◽  
Extraction Methods ◽  
Bp Network ◽  
Speed Increase ◽  
Learning Speed ◽  
Hidden Layer

Abstract This paper describes the application of neural networks to gearbox fault diagnosis. In order to increase learning speed of BP network, a modified learning algorithm was presented. Considering of the difficulty of choosing neural networks’ architecture, genetic algorithm was employed. The discussion of the effect of hidden layer nodes shows that with the increase of the number of nodes, the learning speed increase also yet result in poor generalization ability. The test of fault tolerance ability tells that neural networks have ‘bench type’ tolerance ability. This ensures that when signals were contaminated by noise or feature extraction methods were not effective, the result can still be acceptable. To test the performance of the application of neural networks on gearbox fault diagnosis, experiments of single fault and multi-faults were both implemented and diagnosed by neural networks. The results were satisfied.

Failure Modes Model of MEMS Accelerometers Based on Neural Networks

2008 ◽  
Author(s):  
Yanping Bai ◽  
Ping An ◽  
Yilong Hao
Keyword(s):  
Neural Networks ◽  
Failure Modes ◽  
Late Phase ◽  
Primary Sources ◽  
Bp Network ◽  
Bp Networks ◽  
Mems Accelerometers ◽  
Perception Model ◽  
Hidden Layer ◽  
Training Success

Fabrication of a MEMS system involves design, testing, packaging and reliability related issues. However, reliability issues that are discovered at a late phase may cause major delays in the product development going together with high costs. In this paper we study the failure modes and Mechanisms of MEMS accelerometers products and present the classification modeling of failure modes based on neural networks. In ours MEMS accelerometers, there are six failure mechanisms that have been found to be the primary sources of failure nodes. We introduce nonlinear BP network with a hidden layer and linear perception to classify for MEMS accelerometers products. Classification results show that nonlinear BP network seem to be most appropriate to approach the problem of failure modes classification than linear perception. BP neural network is capable of learning the intrinsic relations of the patterns with which they were trained. For all experiments results, the training success of rate is 100% for both methods. BP networks obtained a high forecast success of rate of over 99.5%. The linear perception model obtained a success of rate of over 95.5%. We also analyze the technology stability of MEMS products.

Transformer Fault Diagnosis Based on Improved Quantum Genetic Algorithm and BP Network

2010 ◽  
Vol 29-32 ◽  
pp. 1543-1549 ◽  
Author(s):  
Jie Wei ◽  
Hong Yu ◽  
Jin Li
Keyword(s):  
Neural Network ◽  
Genetic Algorithm ◽  
Fault Diagnosis ◽  
Bp Neural Network ◽  
Fuzzy Theory ◽  
Back Propagation ◽  
Gas Analysis ◽  
Bp Network ◽  
Quantum Genetic Algorithm ◽  
Transformer Fault

Three-ratio of the IEC is a convenient and effective approach for transformer fault diagnosis in the dissolved gas analysis (DGA). Fuzzy theory is used to preprocess the three-ratio for its boundary that is too absolute. As the same time, an improved quantum genetic algorithm IQGA (QGASAC) is used to optimize the weight and threshold of the back propagation (BP). The local and global searching ability of the QGASAC approach is utilized to find the BP optimization solution. It can overcome the slower convergence velocity and hardly getting the optimization of the BP neural network. So, aiming at the shortcoming of BP neural network and three-ratio, blurring the boundary of the gas ratio and the QGASAC algorithm is introduced to optimize the BP network. Then the QGASAC-IECBP method is proposed in this paper. Experimental results indicate that the proposed algorithm in this paper that both convergence velocity and veracity are all improved to some extent. And in this paper, the proposed algorithm is robust and practical.

scholarly journals A REVIEW OF FEATURE EXTRACTION METHODS ON MACHINE LEARNING

2021 ◽  
Vol 6 (22) ◽  
pp. 51-59
Author(s):  
Mustazzihim Suhaidi ◽  
Rabiah Abdul Kadir ◽  
Sabrina Tiun
Keyword(s):  
Machine Learning ◽  
Feature Extraction ◽  
Feature Selection ◽  
Input Data ◽  
Feature Vector ◽  
Learning Algorithm ◽  
Extraction Methods ◽  
Machine Learning Algorithm ◽  
Learning Tasks ◽  
Low Dimensional

Extracting features from input data is vital for successful classification and machine learning tasks. Classification is the process of declaring an object into one of the predefined categories. Many different feature selection and feature extraction methods exist, and they are being widely used. Feature extraction, obviously, is a transformation of large input data into a low dimensional feature vector, which is an input to classification or a machine learning algorithm. The task of feature extraction has major challenges, which will be discussed in this paper. The challenge is to learn and extract knowledge from text datasets to make correct decisions. The objective of this paper is to give an overview of methods used in feature extraction for various applications, with a dataset containing a collection of texts taken from social media.

Design and Algorithm of Intelligent Manufacturing Evaluation Model Based on Machine Learning

2021 ◽  
Author(s):  
Xianwang Li ◽  
Zhongxiang Huang ◽  
Wenhui Ning
Keyword(s):  
Machine Learning ◽  
Genetic Algorithm ◽  
Manufacturing System ◽  
Learning Algorithm ◽  
Intelligent Manufacturing ◽  
System Model ◽  
Machine Learning Algorithm ◽  
Model Design ◽  
Bp Network ◽  
Intelligent Manufacturing System

Abstract Machine learning is gradually developed and applied to more and more fields. Intelligent manufacturing system is also an important system model that many companies and enterprises are designing and implementing. The purpose of this study is to evaluate and analyze the model design of Intelligent Manufacturing System Based on machine learning algorithm. The method of this study is to first obtain all the relevant attributes of the intelligent manufacturing system model, and then use machine learning algorithm to delete irrelevant attributes to prevent redundancy and deviation of neural network fitting, make the original probability distribution as close as possible to the distribution when using the selected attributes, and use the ratio of industry average to quantitative expression for measurable and obvious data indicators. As a result, the average running time of the intelligent manufacturing system is 17.35 seconds, and the genetic algorithm occupies 15.63 seconds. The machine learning network takes up 1.72 seconds. Under the machine learning algorithm, the training speed is very high, obviously higher than that of the genetic algorithm, and the BP network is 2.1% higher than the Elman algorithm. The evaluation running speed of the system model design is fast and the accuracy is high. This study provides a certain value for the model design evaluation and algorithm of various systems in the intelligent era.

scholarly journals Intelligent Fault Diagnosis Method Using Acoustic Emission Signals for Bearings under Complex Working Conditions

2020 ◽  
Vol 10 (20) ◽  
pp. 7068
Author(s):  
Minh Tuan Pham ◽  
Jong-Myon Kim ◽  
Cheol Hong Kim
Keyword(s):  
Acoustic Emission ◽  
Feature Extraction ◽  
Deep Learning ◽  
Fault Diagnosis ◽  
Extraction Methods ◽  
High Accuracy ◽  
Time Frequency ◽  
Bearing Fault ◽  
Bearing Fault Diagnosis ◽  
Ae Signals

Recent convolutional neural network (CNN) models in image processing can be used as feature-extraction methods to achieve high accuracy as well as automatic processing in bearing fault diagnosis. The combination of deep learning methods with appropriate signal representation techniques has proven its efficiency compared with traditional algorithms. Vital electrical machines require a strict monitoring system, and the accuracy of these machines’ monitoring systems takes precedence over any other factors. In this paper, we propose a new method for diagnosing bearing faults under variable shaft speeds using acoustic emission (AE) signals. Our proposed method predicts not only bearing fault types but also the degradation level of bearings. In the proposed technique, AE signals acquired from bearings are represented by spectrograms to obtain as much information as possible in the time–frequency domain. Feature extraction and classification processes are performed by deep learning using EfficientNet and a stochastic line-search optimizer. According to our various experiments, the proposed method can provide high accuracy and robustness under noisy environments compared with existing AE-based bearing fault diagnosis methods.

FUSION OF EXTREME LEARNING MACHINE WITH FUZZY INTEGRAL

2013 ◽  
Vol 21 (supp02) ◽  
pp. 23-34 ◽  
Author(s):  
JUNHAI ZHAI ◽  
HONGYU XU ◽  
YAN LI
Keyword(s):  
Extreme Learning Machine ◽  
Learning Algorithm ◽  
Fuzzy Integral ◽  
Fast Learning ◽  
Original Dataset ◽  
Learning Speed ◽  
Feed Forward Neural Networks ◽  
Three Stages ◽  
Learning Machine ◽  
Hidden Layer

Extreme learning machine (ELM) is an efficient and practical learning algorithm used for training single hidden layer feed-forward neural networks (SLFNs). ELM can provide good generalization performance at extremely fast learning speed. However, ELM suffers from instability and over-fitting, especially on relatively large datasets. Based on probabilistic SLFNs, an approach of fusion of extreme learning machine (F-ELM) with fuzzy integral is proposed in this paper. The proposed algorithm consists of three stages. Firstly, the bootstrap technique is employed to generate several subsets of original dataset. Secondly, probabilistic SLFNs are trained with ELM algorithm on each subset. Finally, the trained probabilistic SLFNs are fused with fuzzy integral. The experimental results show that the proposed approach can alleviate to some extent the problems mentioned above, and can increase the prediction accuracy.

Application Overview of Neural Network in Fault Diagnosis

2014 ◽  
Vol 697 ◽  
pp. 419-424
Author(s):  
Ze Fan Cai ◽  
Dao Ping Huang
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Fault Diagnosis ◽  
Development Trend ◽  
System Structure ◽  
Bp Network ◽  
Rbf Network ◽  
The Third ◽  
The Neural Network ◽  
The Development Trend

This paper introduces the system structure of neural network in fault diagnosis, and summarizes some applications of neural network in fault diagnosis. The most commonly used neural network in fault diagnosis is BP network. The second is RBF network and the third is ART. For each neural network, the paper will discuss the neural network, and the introduce some applications. It also introduces the combination of neural networks and other techniques. In the last part, this paper points out the development trend of the neural network in fault diagnosis.

Wavelet Neural Network Based on Chaos Genetic Algorithm

2013 ◽  
Vol 339 ◽  
pp. 307-312 ◽  
Author(s):  
Hu Cheng Zhao
Keyword(s):  
Neural Network ◽  
Genetic Algorithm ◽  
Learning Algorithm ◽  
Wavelet Neural Network ◽  
Optimal Parameters ◽  
Chaos Optimization ◽  
Chaos Optimization Algorithm ◽  
Learning Speed ◽  
Chaos Genetic Algorithm ◽  
Remarkable Progress

To improve the performance of Wavelet Neural Network (WNN), a hybrid WNN learning algorithm, which is combination of Genetic Algorithm (GA) and Chaos Optimization Algorithm (COA) in a mutual complementarity manner, is proposed. In the algorithm, GA is first used to roughly search the optimal parameters of WNN as a whole, and then COA is adopted to perform the refined search on the basis of the result obtained by GA, which can make remarkable progress in modeling accuracy, learning speed, and overcoming local convergence or precocity. Simulation show its effectiveness.

Sign in / Sign up

Export Citation Format

Share Document