Change search
ReferencesLink to record
Permanent link

Direct link
Using the Singular Value Decomposition and Multiple Linear Regression on sound and vibration signals from a truck engine
2001 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

The sound- and accelerometer signals from running engines are complex. In the engine, an amount of events take place, and the sound- and accelerometer signals are a mix of the sounds and vibrations caused by these events. In this report, some ideas of how to find and identify the sounds and vibrations from these events are tested on an existing set of data. Two different methods are tested in this report. The first is a periodic signal component extraction method, based on the singular value decomposition. The idea is to divide the engine sound signal into periodic components with the singular value decomposition, and then see if they correspond to some known engine events. One component that corresponds to the engine cycle is extractable. Since the extractable component corresponds to the engine cycle, it is possible to gain information of the engine speed with this method. The idea of the other method is to find out how some engine parameters influence the accelerometer signal with the method of multiple linear regression. A model of the accelerometer signal is fitted to a selection of engine parameter values, and from the model coefficients calculated in the fitting process, it is possible to gain information of how the different parameters influence the accelerometer signal. From the investigated set of data, it is possible to see what influence the engine speed and load has on the accelerometer signal. It is hard to determine the influence of the other investigated engine parameters, since the correlation between some engine parameter values in the examined data set is substantial. To see if this method can be used to gain information of how these engine parameters influence the accelerometer signal, some more measurements must be made.

Place, publisher, year, edition, pages
Keyword [en]
Keyword [sv]
URN: urn:nbn:se:ltu:diva-50657ISRN: LTU-EX--01/001--SELocal ID: 7e5fa4c2-b228-47a9-8ce9-5311af7d5d0aOAI: diva2:1024019
Subject / course
Student thesis, at least 30 credits
Educational program
Engineering Physics, master's level
Validerat; 20101217 (root)Available from: 2016-10-04 Created: 2016-10-04Bibliographically approved

Open Access in DiVA

No full text

Search outside of DiVA

GoogleGoogle Scholar

Total: 7 hits
ReferencesLink to record
Permanent link

Direct link