In structural vibration, most of the times, the analysis is done in spectral domain, by using FRF curves, Modal Analysis, ODS, etc. Based on these spectral analysis result, we will be able to review the dynamic characteristic of the structure. Furthermore, we should be able to understand how to improve the structural design based on these analysis result.
But when we move to engine sound and vibration evaluation, we will have more methods and approaches that can be used to troubleshoot or improve the sound characteristics.
In many cases, engine is considered as source of the vibration and sound. The fact that sound can be transferred through structure borne, air borne, or liquid borne makes acoustic/sound analysis becomes more challenging.
Traditionally, engine sound evaluation is done by an experienced engineer by directly listening to the engine sound.
In some cases, a stethoscope is used to check if there is any abnormal sound from a particular engine component, before deciding to go to even further detail check.
The disadvantage of this traditional method lies in the fact that it is really depending on the engineer experience. The longer the engineer has been exposed to various kind of engine defects, the more accurate and faster the analysis can be performed.
This skill that is gained through long experience, makes it not easy to be transferred to other engineer.
The fact that this traditional method does not include any parameter quantification, makes the repeatability of this approach not measurable. This makes the time needed to perform engine sound evaluation pretty much depending on the availability of the experienced engineer.
In order to increase the accuracy of engine sound evaluation, parameter quantification should be performed. By having these, a valid database can be built. The knowledge and experience can be transferred easier, the repeatability of the method will be increased.
Physically, sound is a vibration that propagates as an acoustic wave, through a transmission medium such as a gas, liquid or solid. For this reason, vibration and sound pressure are most common parameters to be measured for further analysis.
By using microphone and vibration sensor (accelerometers), we can record the data, then by comparing the data with normal engine signature, we may be able to localise and narrow down the probable root cause of the abnormal sound found during engine sound evaluation.
We need to record the data from the event that is producing the abnormal sound. This data recording can be done at stationary RPM or also by sweeping the RPM from idle to the highest one.
We may use spectrum analysis or order analysis to see the spectrum content from the data. If we have done sweeping RPM, order analysis can be used to analyse the spectrum at particular RPM or at particular order.
We can try to identify some critical frequencies that may cause the abnormal tone by comparing the spectrum content with the data from normal engine signature.
By replaying the audio and filtering the critical frequencies out from the data, we may validate the frequency that is causing abnormal tone.
By comparing the data to the spectrum coming from the vibration sensor (accelerometer), we may be able to localise and narrow down the probable root causes even further.
By using Simcenter Testlab Audio Replay and Filtering, below filtering features can be utilized to perform engine sound evaluation more accurate and even faster.
Virtually unlimited number of real-time filters can be applied simultaneous
Filter types include:
Low pass
High pass
Band pass
Band stop
Fixed Notch
Tracked order Filter (pass)
Tracked order Filter (stop)
Online filter tuning: during replay any filter can be turned on/off and any characteristic (center frequency, cut-off, width, etc.) of any filter can be modified. Sets of filters can be saved and reload)
Do you have any issues with engine sound? We will be happy to assist you.
