Abstract
Vibration and noise caused by rotating or alternately moving parts and the physical and chemical processes inside an internal combustion engine contain a lot of information about what happens inside the machine. We determined several positions on certain engine types for 3 axis vibration sensors and acoustic microphones to collect the most suitable signals in order to observe some pre-selected phenomena.
The art of extracting useful information from time based signals is especially difficult in the case of an internal combustion engine because rotation speed is not constant. The angular speed is changing during a single working cycle, additionally the ECU regulates the speed in the course of several turns even at nominally constant speed and eventually we want to accelerate or decelerate the engine. Using the machine’s own angular position sensors and additional sensors fixed outside we developed a method to follow the fine details of angular speed fluctuations e.g. because of the activity of single cylinders.
FFT is a commonly used method for changing to frequency domain from time domain to analyze the data measured with time. However in our case the result has limited applicability because of the spectral leakage effect as frequency components shift with speed change. To solve this problem, several research groups and great firms developed the order analysis theory and slightly different practical techniques with the help of which order spectra is calculated on the base of fixed small angular differences instead of using FFT by fixed short time interval bases.
There are ready made equipments on the market but to achieve a flexible system for developing and testing our methods we chose NI LabView software with Vibration Analysis Library and developed the related hardware device system.
In the first phase of the project we used external incremental rotary encoder signal for the order analysis function. We made on-line analysis and stored the results. In the course of the second phase we first collected and stored the analog vibration signals and several digital angle signals. We made the analysis on the stored information - which has a rather big volume. The basic result is that we can observe phenomena in details and determine relations between the engine operation and the vibration and noise signals.
Keywords: combustion engine, failure diagnosis, order analysis, vibration, noise measurement