Lecture Notes in Networks and Systems 63

Hamido Fujita
Duy Cuong Nguyen
Ngoc Pi Vu
Tien Long Banh
Hermann Horst Puta Editors

Advances in
Engineering
Research and
Application
Proceedings of the International
Conference, ICERA 2018
Lecture Notes in Networks and Systems

Volume 63

Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
The series “Lecture Notes in Networks and Systems” publishes the latest
developments in Networks and Systems—quickly, informally and with high quality.
Original research reported in proceedings and post-proceedings represents the core
of LNNS.
Volumes published in LNNS embrace all aspects and subfields of, as well as
new challenges in, Networks and Systems.
The series contains proceedings and edited volumes in systems and networks,
spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor
Networks, Control Systems, Energy Systems, Automotive Systems, Biological
Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems,
Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems,
Robotics, Social Systems, Economic Systems and other. Of particular value to both
the contributors and the readership are the short publication timeframe and the
world-wide distribution and exposure which enable both a wide and rapid
dissemination of research output.
The series covers the theory, applications, and perspectives on the state of the art
and future developments relevant to systems and networks, decision making, control,
complex processes and related areas, as embedded in the fields of interdisciplinary
and applied sciences, engineering, computer science, physics, economics, social, and
life sciences, as well as the paradigms and methodologies behind them.

Advisory Board
Fernando Gomide, Department of Computer Engineering and Automation—DCA, School of
Electrical and Computer Engineering—FEEC, University of Campinas—UNICAMP,
São Paulo, Brazil
e-mail: gomide@dca.fee.unicamp.br
Okyay Kaynak, Department of Electrical and Electronic Engineering, Bogazici University,
Istanbul, Turkey
e-mail: okyay.kaynak@boun.edu.tr
Derong Liu, Department of Electrical and Computer Engineering, University of Illinois
at Chicago, Chicago, USA and Institute of Automation, Chinese Academy of Sciences,
Beijing, China
e-mail: derong@uic.edu
Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta,
Alberta, Canada and Systems Research Institute, Polish Academy of Sciences, Warsaw,
Poland
e-mail: wpedrycz@ualberta.ca
Marios M. Polycarpou, KIOS Research Center for Intelligent Systems and Networks,
Department of Electrical and Computer Engineering, University of Cyprus, Nicosia, Cyprus
e-mail: mpolycar@ucy.ac.cy
Imre J. Rudas, Óbuda University, Budapest Hungary
e-mail: rudas@uni-obuda.hu
Jun Wang, Department of Computer Science, City University of Hong Kong
Kowloon, Hong Kong
e-mail: jwang.cs@cityu.edu.hk

More information about this series at http://www.springer.com/series/15179

Hamido Fujita Duy Cuong Nguyen
•

Ngoc Pi Vu Tien Long Banh

•

Hermann Horst Puta

Editors

Advances in Engineering
Research and Application
Proceedings of the International Conference,
ICERA 2018

123
Editors
Hamido Fujita Tien Long Banh
Faculty of Software and Information Science Hanoi University of Science and Technology
Iwate Prefectural University Hanoi, Vietnam
Takizawa, Iwate, Japan
Hermann Horst Puta
Duy Cuong Nguyen Institute for Automation and Systems
Faculty of Electronic Engineering Engineering
Thai Nguyen University of Technology Ilmenau University of Technology
Thai Nguyen, Vietnam (TU Ilmenau)
Ilmenau, Germany
Ngoc Pi Vu
Faculty of Mechanical Engineering
Thai Nguyen University of Technology
Thai Nguyen, Vietnam

ISSN 2367-3370 ISSN 2367-3389 (electronic)

Lecture Notes in Networks and Systems
ISBN 978-3-030-04791-7 ISBN 978-3-030-04792-4 (eBook)
https://doi.org/10.1007/978-3-030-04792-4

Library of Congress Control Number: 2018962147

© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents

Keynote Addresses
Data Analytics in Machine Learning: New Directions
and Challenges in Knowledge-Based Systems . . . . . . . . . . . . . . . . . . . . . 3
Hamido Fujita
Big Data in Engineering: Opportunities, Challenges,
and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Kai-Uwe Sattler
Effect of Ionic Liquid and Nanoparticles on PVA-co-PE-Based
GPEs for the Applications in DSSCs . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
S. Ramesh, C. Y. Tan, and K. Ramesh
Intelligent Cervical Squamous Epithelial Cells Classification System . . . 6
Nor Ashidi Mat Isa

ICERA 2018 Main Track

2S-Norm: A New Score Normalization for a GMM Based
Text-Independent Speaker Identification System . . . . . . . . . . . . . . . . . . 9
Van Huy Nguyen
A Comparison of Various Approaches to Reinforcement Learning
Algorithms for Multi-robot Box Pushing . . . . . . . . . . . . . . . . . . . . . . . . 16
Mehdi Rahimi, Spencer Gibb, Yantao Shen, and Hung Manh La
A Group Decision-Making Model with Comparative Linguistic
Expression Based Hedge Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Nguyen Tien Duy and Tran Thai Son
A Multi-objective Ions Motion Optimization for Robot Path Planning . . . 46
Jeng-Shyang Pan, Trong-The Nguyen, Shu-Chuan Chu, Thi-Kien Dao,
and Truong-Giang Ngo

v
vi Contents

A New Approach of a Tube Based Output Feedback Model

Predictive Control: Control Design for 2D Overhead Crane . . . . . . . . . 55
Dao Phuong Nam, Nguyen Hong Quang, and Nguyen Van Huong
A New Modeling of Photovoltaic Power Generation . . . . . . . . . . . . . . . . 63
Le Tien Phong and Ngo Duc Minh
A New Solution Method for Solving Transit Assignment Problems . . . . 70
Le Luong Vuong, Tran Duc Quynh, and Nguyen Quang Thuan
A Novel Despeckling Approach for Ultrasound Images
Using Adaptive OBNLM Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Bo Quoc Bao, Tong Van Luyen, Nguyen Hai Duong, and Ta Chi Hieu
A Study on Calculating Grinding Temperature . . . . . . . . . . . . . . . . . . . 84
Do Duc Trung, Nguyen Van Thien, Hoang Tien Dung,
and Tran Quoc Hung
A Study on Determination of Optimum Partial Transmission Ratios
of Mechanical Driven Systems Using a Chain Drive
and a Three-Step Helical Reducer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Nguyen Thi Hong Cam, Vu Ngoc Pi, Nguyen Khac Tuan,
Le Xuan Hung, and Tran Thi Phuong Thao
A Study on Optimization of Surface Roughness in Surface
Grinding 9CrSi Tool Steel by Using Taguchi Method . . . . . . . . . . . . . . 100
Luu Anh Tung, Vu Ngoc Pi, Do Thi Thu Ha, Le Xuan Hung,
and Tien Long Banh
A Vision-Based Wheel Disc Inspection System . . . . . . . . . . . . . . . . . . . . 109
Huu-Cuong Nguyen, Phuoc-Loc Nguyen, and Byung-Ryong Lee
A Website Defacement Detection Method Based on Machine Learning . . . 116
Xuan Dau Hoang
Acceleration Estimation for Macpherson Active Suspension
Using Dynamic Neural Networks-Based Observers . . . . . . . . . . . . . . . . 125
Huyen T. Dinh
Adaptive Control to Load Disturbance for Brushless DC Motor
Operates at Low Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Nguyen Hong Quang, Duong Quoc Tuan, Nguyen Nhu Hien,
and Tran Ngoc Anh
Adomian Decomposition Method for Thermal Analysis of a Furnace . . . 141
Kavita Goyal and Meenal Singhal
An Experiment for Nonlinear an Active Magnetic Bearing System
Using Fuzzy Logic Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Van-Sum Nguyen, Lai Khac Lai, and Thanh Hoa Lai
Contents vii

An Experimental Investigation of Dynamic Cutting Forces

in the Stable Milling Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Nhu-Tung Nguyen, Yung-Chou Kao, Gia Thinh Bui, Quan Nguyen,
Quoc-Manh Nguyen, and The-Vinh Do
An Improvement of Model Analysis for Spindle Based on Finite
Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Thi-Thao Ngo and Van-The Than
Analysis of Static Loading in Automotive Power Train System . . . . . . . 174
Quang Thanh Nguyen
Applying the Adaptive Nonlinear Backstepping Position Controller
in Order to Control Angular Position of Bipolar Permanent
Magnet Stepper Motors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
C. X. Tuyen and N. T. Huong
Chemical Metallization of Insulating Polymeric Surfaces
Through Simple Diazonium-Based Covalent Amination . . . . . . . . . . . . . 188
Viet Anh Ly, Ngoc Pi Vu, Duy Cuong Nguyen, Xuan Truong Duong,
Gul Zeb, and Xuan Tuan Le
Comparison Between DC and HiPIMS Discharges. Application
to Nickel Thin Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
J. Keraudy, D. T. Nguyen, A. Ferrec, and P-Y. Jouan
Defect Investigation in Perovskite Solar Cells by the Charge
Based Deep Level Transient Spectroscopy (Q-DLTS) . . . . . . . . . . . . . . . 204
K. Xue, C. Renaud, P. Y. Chen, S. H. Yang, and T. P. Nguyen
Defects Morphology in the Dissimilar Friction Stir Welded
T-lap Joints of AA7075 and AA5083 . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Duong Dinh Hao, Masakazu Okazaki, Tran Hung Tra,
and Quach Hoai Nam
Design of a Fuzzy Logic Controller Based on Genetic Algorithm
for Controlling Dissolved Oxygen in Wasted-Water Treatment
System Using Activated Sludge Method . . . . . . . . . . . . . . . . . . . . . . . . . 217
Do Trung Hai and Bach Van Nam
Design of a Small Scale Test Rig for Rotating
Machinery Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Pham Nguyen Quoc Huy, Dang Phuoc Vinh, Ngo Thanh Nghi,
and Tran Phuoc Thanh
Determination of the Plastic Zone Size by Using Nanotest
for Aluminum Alloy 2024T351 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Tien-Dung Do, Florent Chalon, and Pham-Tuong-Minh Duong
viii Contents

Determining Cutting Force After Surface Roughness

Measurement in Grinding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Do Duc Trung, Nguyen Dinh Man, and Phung Xuan Son
Determining RAM Speed and Billet Temperature to Ensure
Two Indicators of Surface Roughness and Extrusion Pressure
When Extruding Aluminum Alloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Nguyen Trong Mai, Do Duc Trung, Tran Duc Quy,
and Pham Van Nghe
Determining Optimal Partial Transmission Ratios of Mechanical
Driven Systems Using a V-Belt Drive and a Helical Reducer
with Second-Step Double Gear-Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Nguyen Thi Hong Cam, Vu Ngoc Pi, Nguyen Khac Tuan, Le Xuan Hung,
and Tran Thi Phuong Thao
Determining the Parameter Area at the Request of a Physical
Field Based on Shape Function Technique . . . . . . . . . . . . . . . . . . . . . . . 270
Long Pham Thanh, Thuy Le Thi Thu, and Thang Nguyen Huu
Diffusion of Circular Source in the Channels of Ventilation Systems . . . 278
Volodymyr Sokolov
Discrete Element Modeling of Steel Slag Concrete . . . . . . . . . . . . . . . . . 284
Nguyen Thi Thuy Hang, Nguyen Xuan Khanh, and Tran Van Tieng
Dynamic Responses of the One-Story Building Frame When
Changing the Bending Stiffness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Duong The Hung and Tran Viet Thang
The Effect of Damper Configurations on the Vibration
of Horizontal Washing Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Ngo Nhu Khoa, Nguyen Thi Hoa, and Nguyen Thi Bich Ngoc
Evaluating the Effect of HVOF Sprayed WC-10Co-4Cr and Hard
Chromium Electroplated Coatings on Fatigue Strength
of Axle-Shaped Machine Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
Phoi Nguyen Vinh, Ngon Dang Thien, and Cuong Le Chi
Experimental Characterization of an Ionically Conductive Fluid
Based High Flexibility Strain Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
Chi Tran Nhu, Ha Tran Thi Thuy, An Tran Hoai, Nguyen Ta Hoang,
Hoai Nguyen Thi, An Nguyen Ngoc, Trinh Chu Duc, Van Thanh Dau,
and Tung Thanh Bui
Exploiting Two-Layer Support Vector Machine to Predict Protein
SUMOylation Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
Van-Nui Nguyen, Huy-Khoi Do, Thi-Xuan Tran,
Nguyen-Quoc-Khanh Le, Anh-Tu Le, and Tzong-Yi Lee
Contents ix

Face and Hand Gesture Recognition for Secure Control of Equipment . . . 333
Dao Thi Thanh, Vu Duc Thai, and Pao-Ann Hsiung
Homogeneous Transformation and Kinematics of a Steering Tyre . . . . . 340
Dai Quoc Vo, Nam Ky Le, Hormoz Marzbani, and Reza N. Jazar
Improving Vehicle Roll Stability by LQR Active Anti-roll Bar Control . . . 350
Van Tan Vu, Olivier Sename, and Duc Tien Bui
Incorporating Unsupervised and Semi-supervised Learning
in Min-Max Neuron Network for Clustering Data . . . . . . . . . . . . . . . . . 357
Le Anh Tu, Vu Duc Thai, and Vu Dinh Minh
Influence of Engine Torque on Vehicle Ride Comfort . . . . . . . . . . . . . . 364
Nguyen Khac Tuan, Vu Van Hai, and Hoang Anh Thai
Influence of Heavy Truck Operating Condition on Dynamic
Load Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
Bui Van Cuong, Le Van Quynh, and Le Xuan Long
Influence of the Driving Frequency on Electrostatic Linear
Comb Actuator Displacement Amplitude . . . . . . . . . . . . . . . . . . . . . . . . 380
Kien Trung Hoang, Ham Cong Vu, Viet Hoang Nguyen,
and Phuc Hong Pham
Initial Tool Wear Mechanism in Dry and Lubricated Turning
of Inconel 718 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
Morvarid Memarianpour, Seyed Ali Niknam, Sylvain Turenne,
and Marek Balazinski
Near-Infrared Emitting Type-II CdTe/CdSe Core/Shell Nanocrystals:
Synthesis and Optical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
N. X. Ca, N. T. Hien, N. T. Luyen, and P. M. Tan
Numerical Investigations of Ellipsoid Shaped Filler on Heat
Transport Behavior of Reinforced Polymer Composites . . . . . . . . . . . . . 408
Satnam Singh, Rohit Kumar Singla, and Deepak Jain
Numerical Modeling and Experimental Study on Vibration
of a Horizontal Washing Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
Ngo Nhu Khoa, Nguyen Thi Hoa, and Nguyen Thi Bich Ngoc
On a Diagnostic Procedure to Automatically Classify Gear
Faults Using the Vibration Signal Decomposition and Support
Vector Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
Nguyen Phong Dien and Nguyen Trong Du
On-machine and In-laboratory Investigation of Errors of Probes
for CNC Machine Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
Michał Jankowski, Adam Wozniak, and J. R. R. Mayer
x Contents

Optimal Planning of Renewable Sources in Micro-grids Based

on Life Cycle Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
V. V. Thang and N. H. Trung
Optimization of Surface Roughness and Cutting Force in MQL
Hard-Milling of AISI H13 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
The-Vinh Do and Nguyen-Anh-Vu Le
Optimizing Number of Clusters for Energy Saving Purpose
in Wireless Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
Minh T. Nguyen and Keith A. Teague
Optimization of Matching Phase Between Two Driving Oscillations
of a TFG Using Diamond-Shaped Frame . . . . . . . . . . . . . . . . . . . . . . . . 462
Vu Van The, Tran Quang Dung, and Chu Duc Trinh
Optimum Calculation of Partial Transmission Ratios
of Mechanical Driven Systems Using a V-Belt and a Three-Step
Bevel Helical Gearbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469
Nguyen Khac Tuan, Tran Thi Phuong Thao, Nguyen Thi Hong Cam,
Le Xuan Hung, and Vu Ngoc Pi
Outage Analysis of Downlink NOMA Full-Duplex Relay Networks
with RF Energy Harvesting over Nakagami-m Fading Channel . . . . . . 477
Nguyen Trung Tan, Tran Manh Hoang, Ba Cao Nguyen,
and Le The Dung
Phase Composition, Microstructure and Properties of Porous
Ti-3.5Nb-3.5Zr Alloy Fabricated by Using a Combined Process
of Decomposition and Vacuum Sintering . . . . . . . . . . . . . . . . . . . . . . . . 488
Nguyen Van Toan, Ho Ky Thanh, and Tran Van Dung
PID Adaptive Tuning with the Principle of Receding Horizon . . . . . . . . 497
Ha Thi Thu Le, Nam Hoai Nguyen, and Phuoc Doan Nguyen
Propose a New Distance Degradation Parameter to Estimate Level
of Out-Of-Band Emission Caused by Non-linear High Power
Amplifiers for DVB-S2 System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
Doan Thanh Hai, Nguyen Van Vinh, Nguyen Tat Nam,
Nguyen Quoc Binh, and Dao Huy Du
Real-Time Optimal Control of TRMS with State Dependent
Riccati Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513
Dinh Van Nghiep, Nguyen Nhu Hien, and Nguyen Thi Phuong Chi
Real-Time Vehicle Inertial Parameters Estimation Based
on a Simplified Half-Car Vertical Vibration Model . . . . . . . . . . . . . . . . 521
Nguyen Tuan-Anh and Nguyen Cong Tuan
Contents xi

Research of Efficiency of the System of Photovoltaic Converters

for Electric Vehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532
Kirill Karpukhin and Alexey Kolbasov
Situational Nonlinear Model Predictive Control
for Autonomous Driving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
Jonas Spindler, Siegbert Hopfgarten, Evgeny Lazutkin, and Pu Li
Study of Emissions and Fuel Economy for Series - Parallel
HEVson FTP-75, US-Highway-Cycle Driving Cycles . . . . . . . . . . . . . . . 545
Kien Nguyen Trung and Tan Vu Tat
Study on Flow-Focusing Microfluidic Device with External
Electric Field for Droplet Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
Cuong Nguyen Nhu, Hang Nguyen Thu, Luan Le Van,
Trinh Chu Duc, Van Thanh Dau, and Tung Thanh Bui
The Effects of Machine Tool Pallet Change on Machine Tool
Geometric Measurement Using the Scale and Master Ball
Artefact Method (SAMBA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
Kanglin Xing, Sofiane Achiche, and J. R. R. Mayer
The Efficiency Reaches a Plateau in Inverted Schottky Quantum
Dot Solar Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566
Van Tuan Mai, Ngoc Huyen Duong, and Xuan-Dung Mai
The Hedge-Algebras-Based Controller for Robotic Arm . . . . . . . . . . . . 572
K. L. Lai, Nguyen Hai Vu, and Thanh Hoa Lai
Theoretical and Experimental Integration for Working Process
Simulation on Marine Diesel Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . 580
Do Duc Luu and Nguyen Quang Vinh
Vibroacoustic Response of a Finite Clamped Laminated
Composite Plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
Tran Ich Thinh and Pham Ngoc Thanh
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
 On a Diagnostic Procedure to Automatically
Classify Gear Faults Using the Vibration
Signal Decomposition and Support
Vector Machine

Nguyen Phong Dien1 and Nguyen Trong Du2(&)

1
School of Mechanical Engineering, Hanoi University of Science
and Technology, Hanoi, Vietnam
2
Mechanical and Power Engineering Faculty, Electric Power University,
Hanoi, Vietnam
dunt@epu.edu.vn

Abstract. This study attempts to demonstrate the usefulness of an automated

diagnostic procedure based on the Ensemble Empirical Mode Decomposition
(EEMD) method and the Support Vector Machine (SVM) for gear fault detec-
tion and classification in a two-stage helical gearbox. First, the vibration signals
measured on the gearbox casing corresponding to three conditions: normal gear,
chipped gear and broken tooth gear are decomposed into different intrinsic
modes by EEMD method. The standard SVM is then applied to solve a multi-
class problem of gear fault classification. It can be seen from the results obtained
at a gearbox test rig that the gear faults can be clearly detected and identified by
this approach.

Keywords: Signal decomposition
Gear fault
Gearbox

Support Vector Machine

1 Introduction

Fault diagnosis of gearboxes has attracted increasing interest in the last decades due to
their ubiquity and importance in industry. Some advances in signal processing methods
and measuring techniques have enabled the vibration analysis to early detect and
identify faults of critical components such as gears and bearings within gearboxes.
Recent research trends focus on developing a diagnostic system that works in an
automatic manner with the application of artificial intelligence [1–4]. According to this,
the diagnostic procedure usually involves the following steps: diagnostic feature
selection and extraction from vibration signals, calculating the instant values of
selected diagnostic parameters as input data set for classification of faults or technical
conditions, implementation of an algorithm for pattern recognition and classification, or
a classifier with supervised learning. The first step deals with the application of signal
processing methods. In fact, the signal is decomposed into different parts using various
decomposition methods that allow for the extraction of the diagnostic features from the
raw vibration signals. The results of signal decomposition allow a deeper insight into

© Springer Nature Switzerland AG 2019

H. Fujita et al. (Eds.): ICERA 2018, LNNS 63, pp. 425–432, 2019.
https://doi.org/10.1007/978-3-030-04792-4_55
426 N. P. Dien and N. T. Du

the signal components to realize diagnostic symptoms when failures occur in the
gearbox, enhance damage detection and select features such as sidebands effect [5] or
transient vibrations that are closely related to different types of gear damages. The next
step is the quantification of the selected diagnostic features by one or several sensitive
characteristic parameters as RMS, Crest factor, Kurtosis [5], signal energy in a narrow
frequency range, wavelet coefficients, etc. These parameters constitute a set of input
data for the last step which is the training and implementation of a classification
algorithm such as genetic algorithms, artificial neural networks (ANNs) [2, 4] or
support vector machines (SVMs) [1, 3] to automatically detect and classify gear faults.
This study attempts to demonstrate the usefulness of an automated diagnostic proce-
dure based on the Ensemble Empirical Mode Decomposition (EEMD) method and the
conventional SVM for gear fault detection and classification in a two-stage helical
gearbox. First, the vibration signals measured on the gearbox casing corresponding to
three conditions: normal gear, chipped gear and broken tooth gear are decomposed into
different intrinsic modes by EEMD method. The standard SVM is then applied to solve
a multi-class problem of gear fault classification. It can be seen from the results
obtained at a gearbox test rig that the gear faults can be clearly detected and identified
by this approach.

2 Signal Decomposition Using EEMD Method

As mentioned above, the extraction of diagnostic features from vibration signals is

almost based on the signal decomposition. A simple, but effective decomposition
method is the time-synchronous averaging (TSA) proposed by McFadden [5]. TSA has
become a common technique to detect gear faults in time domain for many years. By
using TSA, a vibration signal x(t) measured at a rotating machine can be divided into a
periodic part xp and a non-periodic part xnp , including the random noise components.

xðtÞ ¼ xp ðtÞ þ xnp ðtÞ ð1Þ

The other methods may have different underlying mathematical or degree of

complexity. Among them, the signal can be decomposed into harmonic functions by
Fourier transform that uses fixed harmonic basis functions. Since the amplitude and
frequency values of the resulting harmonic components are constant, this decompo-
sition method is suitable only for analysis of stationary processes. A wavelet transform
performs also signal decomposition in a fixed basis of functions called wavelets [6, 7],
but the basic functions can be preset by selecting the mother wavelet such as Coiflet,
Daubechies or Symlet wavelet. The wavelet transform result can be shown as a
decomposition of the original signal into components having low and high frequencies
according to the number of levels of decomposition. In addition, every signal com-
ponent decomposed by the wavelet transform contains parameters that determine its
scale over time. This feature makes the analysis of non-stationary processes possible.
A number of methods is developed to deal with non-stationary signals using an
adaptive basis such as Wavelet packet transform [7]. A method referred to as Empirical
Mode Decomposition (EMD) is proposed by Huang et al. [8] for adaptively
 On a Diagnostic Procedure to Automatically Classify Gear Faults 427

representing non-stationary signals as sums of the amplitude modulation and frequency

modulation components, called intrinsic mode functions (IMFs).
Xn
xðtÞ ¼ k¼1
imfk ðtÞ þ rn ðtÞ ð2Þ

where the final residue rn is a constant or a monotonic function.

The basic functions are in this case derived adaptively directly from input signal.
Flandrin et al. [9] developed an effective algorithm for EMD as the fundamental part of
the Hilbert–Huang transform [8], called EMD dyadic filtering, and promoted this
method to be a promising new addition to existing method for non-stationary signal
processing. The iterative algorithm of EMD can be summarized as follows.
Step 1. Initialize r0 = the original signal x, let k = 1; initialize hk,i−1 = rk−1, let
i = 1.
Step 2. Construct upper envelop emaxk and lower envelop emink interpolating
maximum and minimum, respectively of hk,i−1.
Step 3. Approximate the local average called the envelop mean emk by taking
average of two envelops emaxk and emink. That is, emk = (emaxk + emink)/2.
Step 4. Compute the candidate implicit mode hki = hk,i−1 − emk.
Step 5. Checking: If hki is an IMF, decompose signal rk−1 as an IMF imfk = hki and
the residue signal rk = rk−1 − imfk. Otherwise, repeat step 2 through step 5 with
i = i + 1. If rk has an implicit oscillation mode, repeat 5 steps above with k = k + 1.
The decomposition results in a family of frequency ordered IMF components. Each
successive IMF contains lower frequency oscillations than the preceding one. How-
ever, a problem of mode mixing arises in the EMD sifting process due to the presence
of closely spaced spectral components. This effect decreases the physical meaning of
IMFs and the interpretation of the results of the analysis.
The Ensemble Empirical Mode Decomposition (EEMD) method is developed as an
improved version of the EMD to overcome the mode mixing problem. In fact, the low
signal-noise ratio does not affect the decomposition process but actually enhances it to
avoid the mode mixing. Based on this, the EEMD procedure is proposed as follows
[11]:
Step 1. Add the white noise series to the original signal.
Step 2. Perform EMD and obtain IMFs.
Step 3. Repeat step 1 and step 2 with different white noise series.
Step 4. Calculate the ensemble means of corresponding IMFs as the final result.
In the implementation process of EEMD, the added white noise series cancel each
other, and the mean IMFs stays within the natural dyadic filter windows, significantly
reducing the occurrence of mode mixing.
428 N. P. Dien and N. T. Du

3 Experimental Demonstration

The experiments were carried out at a gearbox test rig. The test is conducted to
investigate the influence of different gear faults on the measured acceleration of the
gearbox casing and the feasibility of the diagnostic procedure for gear fault detection
and recognition. The gearbox is constituted by two pairs of gear pairings with gear
crack fault as shown in Fig. 1. The meshing frequencies of the input stage and the
output stage are 552 Hz and 275.2 Hz, respectively. The rotating frequencies of the
input shaft, the idle shaft and the output shaft are 34.5 Hz, 11.5 Hz and 6.88 Hz
respectively. The vibration measurement is taken with two accelerometers. An optical
tachometer mounted in proximity to the input shaft is used to generate one-pulse-per-
revolution signals.

Fig. 1. Gearbox test rig, seeded tooth gear faults and data acquisition scheme

The vibration signals were sampled at 60 kHz using a data acquisition system with
DAQ card and software LabVIEW. The following conditions were under considera-
tion: (i) normal gear, (ii) chipped gear and (iii) broken tooth gear. The amount of data
collected is large enough for further analysis.
It is well known that the main source of vibration in a geared transmission system is
usually the meshing action of the gears. Gear faults can generate modulation sidebands
in vibration spectra, called sideband effect. From experimental works, e.g. [5], the most
important components in gear vibration spectra are the tooth-meshing frequencies and
its harmonics, together with sideband structures due to the modulation effect. The
increment in the number and amplitude of sidebands may indicate a gear fault con-
dition, and the spacing of the sidebands is related to their source. In addition, the other
machine parts that rotate with the same rotating frequency as the considered gear (such
as shaft, coupling and inner bearing mounted in the same gear shaft) have different fault
frequencies that are much lower than the tooth-meshing frequency of the gear-pair.
It can be seen from Fig. 2b that the application of conventional methods in fre-
quency domain such as the sideband analysis is ineffective due to the influence of
strong noise components. The meshing frequency of the second gear stage is
 On a Diagnostic Procedure to Automatically Classify Gear Faults 429

recognized at about 275.2 Hz, but no sidebands around this meshing frequency can be
detected. The vibration signal shown in Fig. 2a is now decomposed using EEMD
method. Figure 3 shows the time plots of the 10 IMFs obtained by implementing the
procedure described in Sect. 2.

Fig. 2. (a) The raw vibration signal corresponding to the broken tooth gear condition, (b) its
frequency spectrum by FFT

Fig. 3. IMFs of the raw signal shown in Fig. 2a

430 N. P. Dien and N. T. Du

Based on the decomposition result, the diagnostic features can be now selected
from the obtained IMFs. In the case of the broken gear condition, the marginal spec-
trum of 7th IMF depicted in Fig. 4 reflects most clearly symptoms of gear faults. It
could be clearly seen that the sideband structure around the meshing frequency of
275.2 Hz is spacing at the same value as the rotating frequency of the idle shaft
(11.5 Hz), which indicates that a fault occurred on gear 3.

Fig. 4. Marginal spectrum of IMF7

To recognize and classify the fault conditions considered in this example, an

algorithm for pattern recognition and classification is applied, namely, the standard
SVM. Unlike the traditional ANN using empirical risk minimization, the SVM is based
on structural risk minimization that can obtain better results for classification tasks [1].
Recent advances in research have made SVM popular in many applications, and the
theoretical background and algorithm of the standard SVM will therefore be not
mentioned here.
 On a Diagnostic Procedure to Automatically Classify Gear Faults 431

Fig. 5. Classification results by SVM

To deal with the multi-class problem using the SVM, namely three classes corre-
sponding to normal gear, chipped gear and broken gear condition, the vibration signals
in a class are decomposed into IMFs. One IMF that is sensitive to a gear fault condition
can be selected to calculate the characteristic parameters, which are kurtosis and signal
energy of the IMF [12]. Finally, a two-dimensional characteristic vector is obtained to
be used as a training data set for the SVM. The results depicted in Fig. 5 show that
different fault conditions of the gearbox test rig are exactly detected and clearly clas-
sified by the SVM with an appropriate parameter set.

4 Conclusions

The present study deals with the application of a diagnostic procedure for automatically
detecting and classifying gear faults, based on the methods of EEMD and SVM. The
following concluding remarks are reached.
– A gearbox mainly produces non-stationary and transient mechanical vibrations due
to the gear meshing impact and the presence of gear faults. Therefore, the EEMD
can be seen as a suitable signal processing tool to extract diagnostic features for
gearboxes.
– The proposed procedure has been experimentally verified by a gearbox test rig. As
can be seen from the obtained results, the usefulness and effectiveness of the EEMD
and SVM for gear fault detection and classification are obvious. A continued
investigation on this procedure is therefore highly recommended.
– We note that the accuracy of classification result depends mainly on several
parameters selected for the SVM such as kernel function, and a large enough data
set is required for training SVM.
432 N. P. Dien and N. T. Du

References
1. Jedlinski, L., Jonak, J.: Early fault detection in gearboxes based on support vector machines
and multilayer perceptron with a continuous wavelet transform. Appl. Soft Comput. 30,
636–641 (2015)
2. Rafiee, J., Arvani, F., Harifi, A., Sadeghi, M.H.: Intelligent condition monitoring of a
gearbox using artificial neural network. Mech. Syst. Signal Process. 21, 1746–1754 (2007)
3. Widodo, A., Yang, B.-S.: Support vector machine in machine condition monitoring and fault
diagnosis. Mech. Syst. Signal Process. 21, 2560–2574 (2007)
4. Samanta, B.: Gear fault detection using artificial neural networks and support vector
machines with genetic algorithms. Mech. Syst. Signal Process. 18, 625–644 (2004)
5. Dalpiaz, G., Rivola, A., Rubini, R.: Effectiveness and sensitivity of vibration processing
techniques for local fault detection in gears. Mech. Syst. Signal Process. 14(3), 387–412
(2000)
6. Mallat, S.: A Wavelet Tour of Signal Processing. Academic Press, New York (1999)
7. Lin, J., Qu, L.: Feature extraction based on morlet wavelet and its application for mechanical
fault diagnosis. J. Sound Vibr. 234, 135–148 (2000)
8. Huang, N.E., et al.: The empirical mode decomposition and the Hilbert spectrum for
nonlinear and non-stationary time series analysis. Proc. R. Soc. A Math. Phys. Eng. Sci. 454,
903–995 (1998)
9. Flandrin, P., Rilling, G., Gonçalvés, P.: Empirical mode decomposition as a filter bank. IEEE
Signal Process. Lett. 11(2), 112–114 (2004)
10. McFadden, P.D.: Examination of a technique for the early detection of failure in gears by
signal processing of the time domain average of the messing vibration. Mech. Syst. Signal
Process. 1, 173–183 (1987)
11. Wu, Z., Huang, N.E.: Ensemble empirical mode decomposition: a noise assisted data
analysis method. Adv. Adapt. Data Anal. 1(1), 1–51 (2005)
12. Yang, Y., Wu, T.Y.: Diagnostics of gear deterioration using EEMD approach and PCA
process. Measurement 61, 75–87 (2015)