טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentSnir Ohad
SubjectAcoustic Source Identification and Separation in a Riding
Vehicle by Means of Vibration Measurement at the
Sources
DepartmentDepartment of Mechanical Engineering
Supervisors Professor Emeritus Simon Braun
Professor Miriam Zacksenhouse


Abstract

The classic approach to noise cancellation is a passive acoustic approach. Passive silencing techniques such as sound absorption and isolation are inherently stable and effective for disturbances above 1000Hz. Active noise control (ANC) on the other hand uses acoustic sources to generate a secondary sound field that destructively interferes with the original undesired sound field. Due to the complexity of the acoustic transmission paths in automobile, the implementation of ANC technology is not straightforward. In the research presented here, ANC was applied to control the engine noise and the noise caused by the interaction of road and tires. When the disturbance is caused by the engine, there are many sources that radiate noise into the cabin. It is difficult to analyze these sources because the acoustic and structural paths are complex. To gain maximum reduction of the noise field, it is necessary to obtain maximum coherence between the reference signal which represents the noise and the error signal which is measured at the monitoring point where response is of particular interest. Part of know-how of the experimental engineer consists of building up a proper experimental protocol involving instrumentation sufficiently representative of sources, but it is not always possible. In fact, all measured signals are mixture of the contribution of all sources. In this research, acoustic identification and application of ANC in the driver's cabin of a Hummer is discussed. The fact that there is correlation between acoustic noise and vibration of an object has been investigated and then utilized for improving the coherence between the noise sources (engine and wheel) and the monitoring point at the driver's cabin. In order to identify the main vibro-acoustic sources, the experiment consisted of three major runs in which sources were being "turned" on and off.  An idle run in which the engine worked without moving the vehicle was used to identify the engine source. A run of 30km/h drive on asphalt was used to identify both engine and wheel sources. At the last run the engine was turned off from a 30km/h drive to obtain better identification of the noise wheel source without the engine interference. Time domain and spectral methods were utilized for source identification and coherence investigation. However the main goal of the thesis is to estimate the performance of an ANC which utilizes noise and vibration of the sources as reference signals to reduce noise field at the driver's cabin.