bullet Sensors & Transducers Journal

    (ISSN: 2306-8515, e-ISSN 1726-5479)

   

   About this Journal

   Information for Authors

   Editorial Board

   Editorial Calendar

   Current Issue

   Browse Journal

S&T journal's cover

Submit Press Release

Submit White Paper

25 Top Downloaded Articles (2007-2012)

Search

Contact Us

 

 

 

 

Vol. 251, Issue 4, April 2021, pp. 1-10

 

Bullet

 

Real-time Faults Prediction by Deep Learning with Multi-sensor Measurements over IoT Networks
 

1, 2 G. Y. Luo, 3, * Y. Q. Luo and 1 H. G. Gan 1 School of Physics and Materials Science, Guangzhou University, Guangzhou 510006, China

2 SHJ Medical Gas Specialists, Asheridge Road, Chesham HP5 2QA, UK
3 Department of Mathematics, London School of Economics and Political Science, London WC2A 2AE, UK

E-mail: gaoyong.luo@yahoo.com

 

Received: 28 March 2021 /Accepted: 28 April 2021 /Published: 30 April 2021

Digital Sensors and Sensor Sysstems

 

Abstract: The advancements in sensing, data processing, and communication technology have enabled machine learning systems to make confident decisions and made it possible to optimise the operations and maintenance of the physical assets, manufacturing systems and processes of prediction and prevention. In many industrial applications, it is critical to use deep neural networks that make predictions both fast and accurate, and can be applied to coupled multiple-input multiple-output (MIMO) system of complex industrial processes. However, due to the difficulty in correctly interpreting the multi-sensor data and extracting the desired information, and the strong nonlinearity and its nearly instantaneous response to disturbances, it is still very challenging to achieve accurate faults prediction and optimised performance in such a complex MIMO system. In this paper, we propose to transform the raw multi-sensor data into the time-frequency domain by developing fast lifting wavelet transform with computational efficiency to obtain a big feature vector containing all the relevant features from all of the sensors, and giving more informative signatures for faults prediction. To raise the power and capabilities of machine learning, we propose a novel machine learning system designed by building IoT networks to remotely collect data, and developing deep wavelet neural networks (DWNN) with Gaussian (Mexican hat) wavelet derived as activation functions to improve nonlinear fitting and convergence speed, and to process the real-time data for faults prediction. Experimental results demonstrate that features extracted in time-frequency domain can reveal the presence of a fault, and its type and cause can be explained by the trained DWNN over the IoT communication networks, where the sensing capabilities and the computational power are provided by the designed controller, transmitter and cloud server to track everything that is relevant to operations, such that by deep learning with real-time data analytics we can have a knowledge base from which to predict faults, correct errors, optimize system performance and maximise efficiency.

 

Keywords: Real-time faults prediction, Machine learning, Deep wavelet neural network, Activation function, Data analytics in time-frequency domain, Wavelet analysis for feature extraction.

 

Acrobat reader logo Click <here> or title of paper to download the full pages article in pdf format

 

 

This work is licensed under a Creative Commons 4.0 International License

 

 Creative Commons License
 

 

 

 

 


1999 - 2021 Copyright , International Frequency Sensor Association (IFSA) Publishing, S.L. All Rights Reserved.


Home - News - Links - Archives - Tools - Voltage-to-Frequency Converters - Standardization - Patents - Marketplace - Projects - Wish List - e-Shop - Sensor Jobs - Membership - Videos - Publishing - Site Map - Subscribe - Search

 Members Area -Sensors Portal -Training Courses - S&T Digest - For advertisers - Bookstore - Forums - Polls - Submit Press Release - Submit White Paper - Testimonies - Twitter - Facebook - LinkedIn