Active vibration control of flexible beam incorporating recursive least square and neural network algorithms

In recent years, active vibration control (AVC) has emerged as an important area of scient ific study especially for vibrat ion suppression of flexible structures. Flexible structures offer great advantages in contrast to the conventional structures, but necessary action must be taken for cancelling...

Full description

Saved in:

Bibliographic Details
Main Author:	Abd. Jalil, Nurhanafifi
Format:	Thesis
Language:	English
Published:	2017
Subjects:	TJ Mechanical engineering and machinery
Online Access:	http://eprints.utm.my/id/eprint/81717/1/NurhanafifiAbdJalilPFKM2017.pdf
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my-utm-ep.81717
record_format	uketd_dc
spelling	my-utm-ep.817172019-09-22T07:25:59Z Active vibration control of flexible beam incorporating recursive least square and neural network algorithms 2017 Abd. Jalil, Nurhanafifi TJ Mechanical engineering and machinery In recent years, active vibration control (AVC) has emerged as an important area of scient ific study especially for vibrat ion suppression of flexible structures. Flexible structures offer great advantages in contrast to the conventional structures, but necessary action must be taken for cancelling the unwanted vibration. In this research, a simulation algorithm represent ing flexible beam with specific condit ions was derived from Euler Bernoulli beam theory. The proposed finite difference (FD) algorithm was developed in such way that it allows the disturbance excitat ion at various points. The predicted resonance frequencies were recorded and validated with theoretical and experimental values. Subsequent ly, flexible beam test rig was developed for collecting data to be used in system ident ificat ion (SI) and controller development. The experimental rig was also utilised for implementation and validat ion of controllers. In this research, parametric and nonparametric SI approaches were used for characterising the dynamic behaviour of a lightweight flexible beam using input - output data collected experimentally. Tradit ional recursive least square (RLS) method and several artificial neural network (ANN) architectures were utilised in emulat ing this highly nonlinear dynamic system here. Once the model of the system was obtained, it was validated through a number of validation tests and compared in terms of their performance in represent ing a real beam. Next, the development of several convent ional and intelligent control schemes with collocated and non-collocated actuator sensor configurat ion for flexible beam vibrat ion attenuation was carried out. The invest igat ion involves design of convent ional proportional-integral-derivat ive (PID) based, Inverse recursive least square active vibrat ion control (RLS-AVC), Inverse neuro active vibration control (Neuro-AVC), Inverse RLS-AVC with gain and Inverse Neuro-AVC with gain controllers. All the developed controllers were tested, verified and validated experimentally. A comprehensive comparat ive performance to highlight the advantages and drawbacks of each technique was invest igated analyt ically and experimentally. Experimental results obtained revealed the superiorit y of Inverse RLS-AVC with gain controller over convent ional method in reducing the crucial modes of vibration of flexible beam structure. Vibration attenuation achieved using proportional (P), proportional-integral (PI), Inverse RLS-AVC, Inverse Neuro- AVC, Inverse RLS-AVC with gain and Inverse Neuro-AVC with gain control strategies are 9.840 dB, 6.840 dB, 9.380 dB, 8.590 dB, 17.240 dB and 5.770 dB, respectively. 2017 Thesis http://eprints.utm.my/id/eprint/81717/ http://eprints.utm.my/id/eprint/81717/1/NurhanafifiAbdJalilPFKM2017.pdf application/pdf en public http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:126145 phd doctoral Universiti Teknologi Malaysia Mechanical Engineering
institution	Universiti Teknologi Malaysia
collection	UTM Institutional Repository
language	English
topic	TJ Mechanical engineering and machinery
spellingShingle	TJ Mechanical engineering and machinery Abd. Jalil, Nurhanafifi Active vibration control of flexible beam incorporating recursive least square and neural network algorithms
description	In recent years, active vibration control (AVC) has emerged as an important area of scient ific study especially for vibrat ion suppression of flexible structures. Flexible structures offer great advantages in contrast to the conventional structures, but necessary action must be taken for cancelling the unwanted vibration. In this research, a simulation algorithm represent ing flexible beam with specific condit ions was derived from Euler Bernoulli beam theory. The proposed finite difference (FD) algorithm was developed in such way that it allows the disturbance excitat ion at various points. The predicted resonance frequencies were recorded and validated with theoretical and experimental values. Subsequent ly, flexible beam test rig was developed for collecting data to be used in system ident ificat ion (SI) and controller development. The experimental rig was also utilised for implementation and validat ion of controllers. In this research, parametric and nonparametric SI approaches were used for characterising the dynamic behaviour of a lightweight flexible beam using input - output data collected experimentally. Tradit ional recursive least square (RLS) method and several artificial neural network (ANN) architectures were utilised in emulat ing this highly nonlinear dynamic system here. Once the model of the system was obtained, it was validated through a number of validation tests and compared in terms of their performance in represent ing a real beam. Next, the development of several convent ional and intelligent control schemes with collocated and non-collocated actuator sensor configurat ion for flexible beam vibrat ion attenuation was carried out. The invest igat ion involves design of convent ional proportional-integral-derivat ive (PID) based, Inverse recursive least square active vibrat ion control (RLS-AVC), Inverse neuro active vibration control (Neuro-AVC), Inverse RLS-AVC with gain and Inverse Neuro-AVC with gain controllers. All the developed controllers were tested, verified and validated experimentally. A comprehensive comparat ive performance to highlight the advantages and drawbacks of each technique was invest igated analyt ically and experimentally. Experimental results obtained revealed the superiorit y of Inverse RLS-AVC with gain controller over convent ional method in reducing the crucial modes of vibration of flexible beam structure. Vibration attenuation achieved using proportional (P), proportional-integral (PI), Inverse RLS-AVC, Inverse Neuro- AVC, Inverse RLS-AVC with gain and Inverse Neuro-AVC with gain control strategies are 9.840 dB, 6.840 dB, 9.380 dB, 8.590 dB, 17.240 dB and 5.770 dB, respectively.
format	Thesis
qualification_name	Doctor of Philosophy (PhD.)
qualification_level	Doctorate
author	Abd. Jalil, Nurhanafifi
author_facet	Abd. Jalil, Nurhanafifi
author_sort	Abd. Jalil, Nurhanafifi
title	Active vibration control of flexible beam incorporating recursive least square and neural network algorithms
title_short	Active vibration control of flexible beam incorporating recursive least square and neural network algorithms
title_full	Active vibration control of flexible beam incorporating recursive least square and neural network algorithms
title_fullStr	Active vibration control of flexible beam incorporating recursive least square and neural network algorithms
title_full_unstemmed	Active vibration control of flexible beam incorporating recursive least square and neural network algorithms
title_sort	active vibration control of flexible beam incorporating recursive least square and neural network algorithms
granting_institution	Universiti Teknologi Malaysia
granting_department	Mechanical Engineering
publishDate	2017
url	http://eprints.utm.my/id/eprint/81717/1/NurhanafifiAbdJalilPFKM2017.pdf
_version_	1747818396820439040

Active vibration control of flexible beam incorporating recursive least square and neural network algorithms

Similar Items