Novitasari, Rizka Wahyu (2014) Evaluasi Unjuk Kerja Microphone Array Menggunakan Metode Time-Frequency Independent Component Analysis. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
2412105025-Undergraduate-Thesis.pdf - Published Version Download (2MB) | Preview |
Abstract
Penggunaan sensor kontak dalam melakukan
Monitoring vibrasi telah dikenal luas untuk mengetahui
kondisi suatu mesin berputar, sehingga dapat diambil langkah
korektif dan antisipatif yang tepat. Dibalik hasil reliabel dari
penggunaan sensor kontak dalam merepresentasikan kondisi
mesin berputar, tingkat keamaan bagi pengguna merupakan
suatu aspek yang banyak disorot. Salah satu jawaban dari
fenomena ini adalah dengan menggunakan sensor non-kontak
dengan memanfaatkan fitur emisi suara yang dibangkitkan
oleh getaran dari mesin berputar tersebut. Penelitian ini
membahas sensor akustik yang digunakan untuk mendeteksi
kerusakan suatu mesin dengan dua jenis microphone dengan
karakter respon yang berbeda serta sensor vibrasi. Hasil data
baseline, dengan menggunakan tiga variasi sensor diperoleh
ciri yang sama dengan teori. Pemisahan sinyal suara pompa
yang memiliki multi kerusakan dilakukan dengan metode ICA
(TDICA, FDICA dan MSICA) dengan variasi jarak
microphone array 10 cm; 20 cm; 30 cm dan variasi jarak
sensor – sumber 15 cm; 30 cm; 45 cm. Hasil dari ketiga teknik
ICA yang memiliki nilai MSE yang paling rendah adalah
0.043 dengan jarak microphone array 10 cm dan jarak sensor
– sumber 30 cm. Evaluasi nilai MSE menunjukkan tidak
terdapat perbedaan yang signifikan antara performa ketiga
metode yang digunakan. Ditinjau dari perbedaan spektrum
sinyal estimasi dan sinyal baseline terdapat peningkatan amplitudo, sehingga unjuk kerja microphone array terhadap
data baseline dapat dikatakan baik.
========================================================================================================================
The use of contact sensors in performing vibration
monitoring has been widely known to determine the condition
of an engine rotating, so that corrective measures can be taken
and appropriate anticipatory. Behind reliable results from the
use of contact sensors to represent the condition of rotating
machines, the user's security level is an aspect that many
highlighted. One answer to this phenomenon is to use a noncontact
sensors by utilizing the features emisis sound
generated by the vibration of the spinning machine. This study
discusses the acoustic sensors are used to detect damage to a
machine with two types of microphones with different response
characters and vibration sensors. The results of the baseline
data, using three variations of the same sensor characteristics
obtained with the theory. The separation of the speech signal
which has a multi-pump damage done by the method of ICA
(TDICA, FDICA and MSICA) with variation of microphone
array distance of 10 cm; 20 cm; 30 cm and the variation of the
distance sensor - the source of 15 cm; 30 cm; 45 cm. The
results of these three ICA technique that has the lowest MSE
value is 0.043 with an array microphone distance of 10 cm
and a distance sensor - the source of 30 cm. Evaluation of the
MSE showed no significant difference between the
performance of the three methods used. Judging from the
difference spectrum of the estimated signal and the baseline
signal there is an increase in the amplitude, so that the performance of the microphone array to the baseline data can
be said to be good.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSF 621.382 84 Nov e |
| Uncontrolled Keywords: | Sensor vibrasi, sensor akustik, unjuk kerja, evaluasi, microphone array |
| Subjects: | T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Users 13 not found. |
| Date Deposited: | 21 Aug 2017 07:43 |
| Last Modified: | 21 Aug 2017 07:43 |
| URI: | http://repository.its.ac.id/id/eprint/48694 |
Actions (login required)
 |
View Item |