Improved training of neural networks for the nonlinear active control of sound and vibration

Bouchard, M.; Paillard, Bruno; Dinh, Chon Tan Le

doi:10.1109/72.750568

Cited by 132 publications

(68 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Bentuk yang paling banyak dikenal dari kombinasi arsitektur/algoritma adaptif berdasarkan pendekatan linier adalah pengendali berbasis filter tranversal, seperti Filter FIR (Finite Impulse Response) atau Filter IIR (Infinite Im pulse Response) yang diadaptasi dengan memakai algoritma Filtered-X Least-MeanSquares (FX-LMS) [2].…”

Section: Pendahuluan Latar Belakangunclassified

“…Suatu aktuator secara khas mempunyai tanggapan nonlinear ketika beroperasi dengan suatu sinyal masukan yang mempunyai amplitudo mendekati atau di atas nilai sinyal masukan nominal, atau ketika beroperasi pada suatu frekuensi di luar rentang frekuensi nominal. Perilaku nonlinier dapat juga terjadi ketika dinamika sistem yang akan dikendalikan nonlinear [2].…”

Section: Pendahuluan Latar Belakangunclassified

See 1 more Smart Citation

Pengembangan Algoritma Pembelajaran untuk Jaringan Syaraf Tiruan Diagonal Recurrent dalam Sistem Kendali Derau Akustik Secara Aktif

Sanjaya¹,

Yacoub²

2007

jte

View full text Add to dashboard Cite

ABSTRAKSistem kendali derau aktif lebih menitikberatkan pada penggunaan algoritma adaptasi dan struktur kendali adaptif. Dalam penelitian ini ditampilkan hasil percobaan sistem kendali derau aktif pada ruang terbuka menggunakan struktur kendali Jaringan Syaraf Tiruan Diagonal Recurrent (JSTDR). Pengembangan algoritma Extended Kalman Filter (EKF) untuk struktur kendali JSTDR, yang kemudian dinamakan sebagai algoritma Extended Kalman Filter Diagonal Recurrent (EKFDR) merupakan tujuan dari penelitian ini. Masalah nonlinieritas jalur sekunder dalam sistem kendali derau aktif, dapat mempersulit proses pengendalian. Dari hasil percobaan terlihat bahwa jumlah neuron dalam JST dapat dikurangi dengan menggunakan JSTDR, tanpa mengurangi kinerja sistem kendali. JSTDR dengan algoritma EKFDR juga menunjukkan kinerja yang baik dalam mengkompensasi nonlinieritas jalur sekunder jika dibandingkan dengan pengendali linier.Kata kunci: derau aktif, jaringan syaraf tiruan, diagonal recurrent, extended kalman f ilter diagonal recurrent, nonlinieritas jalur sekunder. ABSTRACT Active Noise Control (ANC) system emphasizes at usage of adaptation algorithm and adaptive control structure. In this research is presented experimental result of active noise control system at free space using Diagonal Recurrent Artificial Neural Network control structure. The objective of this research is to develop an Extended Kalman Filter (EKF) algorithm for Diagonal Recurrent Artificial Neural Network control structure, named as Extended Kalman Filter Diagonal Recurrent (EKFDR) algorithm. Experimental result shows that amount of neuron in artificial neural

show abstract

Section: Pendahuluan Latar Belakangunclassified

Pengembangan Algoritma Pembelajaran untuk Jaringan Syaraf Tiruan Diagonal Recurrent dalam Sistem Kendali Derau Akustik Secara Aktif

Sanjaya¹,

Yacoub²

2007

jte

View full text Add to dashboard Cite

show abstract

“…The performance of linear ANC system would be degraded under this situation, so that nonlinear active controllers might be required in cases where the actuators used in ANC systems exhibit nonlinear characteristics (Strauch and Mulgrew, 1998, Bouchard et al, 1999, Efe and Kaynak, 2000, Tan and Jiang, 2001). …”

Section: Air Ductsmentioning

confidence: 99%

“…However, based on this combined structure, problems appeared for filter convergence speed and system computation load. In order to avoid these problems, an improved training of the NN was proposed (Bouchard et al, 1999) by introducing a novel heuristical algorithm for the same NN control structure. The simulation results…”

Section: Nonlinearitiesmentioning

confidence: 99%

Active control of pressure pulsation in a switched inertance hydraulic system

Pan

Johnston

Hillis

2013

Proceedings of the Institution of Mechanical Engineers, Part I:

View full text Add to dashboard Cite

The nature of digital hydraulic systems may cause pressure pulsation problems. For example, switched inertance hydraulic systems (SIHS), which are applied to adjust or control flow and pressure by a means that does not rely on dissipation of power, have noise problems due to the pulsed nature of the flow. An effective method to reduce the pulsation is important to improve system performance and increase efficiency. Although passive systems to reduce the noise have been shown to be effective in many situations, their attenuation frequency range is limited and they may be bulky. Furthermore, attenuation devices based on expansion chambers, accumulators or hoses are likely to be unsuitable for SIHS as they add compliance to the system and would impair the dynamic response. This thesis is concerned with issues relating to the development of an active noise canceller for attenuating the pressure pulsation which is caused primarily by pulsed flow from high-speed valves in SIHS.Active control methods are widely and successfully applied in the area of structureborne noise (SBN) and air-borne noise (ABN) cancellation. The idea is using the intentional superposition of waves to create a destructive interference pattern such that a reduction of the unwanted noise occurs. However, applications for fluid-borne noise (FBN) attenuation based on the 'Active noise control (ANC) principle' are rare due to the restriction of the hardware and experimental apparatus in previous researches.In this thesis, an adaptive controller has been developed for active control of pressure pulsation in hydraulic system. The principle of the adaptive LMS filter and details of the controller design are described and the implementation was carried out through simulation. The designed controller was applied on a vibration test rig initially prior to the hydraulic testing in order to investigate its advantages and limitations in practice.Extensive testing on a switched inertance hydraulic rig proved that the controller, which used a piezoelectric valve with fast response and good bandwidth, is effective and that it has several advantages over previous methods, being effective for low frequency cancellation, with a quick response, and is robust and versatile. II A novel method for the accurate measurement of unsteady flowrate in a pipe was proposed. This was applied and validated on a pipe, and was shown to give good results. This method solves the difficulty for measuring the unsteady flowrate currently by using easy-measured signals, such as pressures. It can be used widely for predicting the unsteady flowrate along the pipe. III ACKNOWLEDGEMENTS

show abstract

“…If successful, the neural net will provide an extremely fast but still accurate determination of transmission loss. Although neural nets have been used previously in acoustics [3], [4], [5], the main application has been automatic target recognition (ATR). Here, we are attempting something completely different, namely, the prediction of propagation effects.…”

Section: Introductionmentioning

confidence: 99%

A Nondimensional Parameterization for Sound Propagation in the Atmosphere

Mungiole¹,

Wilson²

2003

View full text Add to dashboard Cite

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing insb-uctions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing the burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for tailing to comply with a colIet:tion of information if it does not display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS. REPORT DATE (DD-MM-YYYY) March 2003 REPORT TYPE DISTRIBUTION/AVAILABILITY STATEMENTApproved for public release; distribution unlimited. SUPPLEMENTARY NOTESAMSCode:622784H7111 DA Project: AH71 ABSTRACTParabolic equation (PE) techniques have been successfully used to obtain numerical solutions of sound pressure attenuation in which sound propagation is affected by turbulence and vertical gradients in wind and temperature. The PE models generally produce accurate attenuation values, but the execution time is excessive for applications when near real-time results are required. To obtain sound level attenuation predictions at selected locations more quickly, we are developing an artificial neural network. As a first step in tliis effort, tlie PE and boundary conditions were modified to obtain a nondimensional version, written in the MATLAB code. Tliis nondimensional version was developed to be used to train the artificial neural network because a fewer number of parameters (seven) would be required to be specified, resulting in a reduced number of model runs to develop the training algorithm. Tliis report documents the derivation of tlie appropriate equations that are used in tlie modified (nondimensional) version of tlie acoustic propagation model. In addition, graphical data are provided tliat identify the sensitivity of sound pressure attenuation to each of tlie seven nondimensional parameters. SUBJECT TERMS

show abstract

Improved training of neural networks for the nonlinear active control of sound and vibration

Cited by 132 publications

References 13 publications

Pengembangan Algoritma Pembelajaran untuk Jaringan Syaraf Tiruan Diagonal Recurrent dalam Sistem Kendali Derau Akustik Secara Aktif

Pengembangan Algoritma Pembelajaran untuk Jaringan Syaraf Tiruan Diagonal Recurrent dalam Sistem Kendali Derau Akustik Secara Aktif

Active control of pressure pulsation in a switched inertance hydraulic system

A Nondimensional Parameterization for Sound Propagation in the Atmosphere

Contact Info

Product

Resources

About