We have demonstrated that there are several possibilities in the approach to diagnosis of objects in motion while using a stationary monitoring station, even in spite of the disturbance caused by the Dopplers effect. Numerical model discover that proper selection of the conditions and methods of measurement as well as diagnostic parameters already at the stage of designing the entire diagnostic system. The computer simulations we conducted demonstrate that the method of task-oriented dynamic signal resampling, that was presented in this paper, effectively reduces the non-stationarity of a signal. The characteristic feature of the method is lack of preliminary assumptions as regards the phenomena causing non-stationarity. The simulations we conducted demonstrate that the possibilities of this method are not restricted to eliminating the non-stationarity caused by linear change of a frequency of signal only. The method also enables elimination of other, complex and non-linear forms of non-stationarity.Effective reduction of various forms of non-stationarity is conditioned by the availability of the course of instantaneous frequency of which defines the eliminated non-stationarity. It is this very course that defines the task while leading to a situation where the method becomes oriented on eliminating the non-stationarity indicated this way. The method of determining the course of the instantaneous frequency of signal is discretionary (free), independent of the method itself. Usually, however, it is conditioned by a specific task and form of the removed non-stationarity. One of the applied methods can include determination of the course of a instantaneous frequency by means of a method relying on the Hilberts transform. This method is effective particularly in the case of signals characterized by non-complex frequency structure. If the method is used for more complex signals, then the course of instantaneous frequency should be defined in a specifically selected frequency band. This restriction is necessary since the method seeks all the forms of changes of the instantaneous frequency and not only the changes of the frequency we intend to eliminate. Thus, the outcome of applying this method of determining the frequency could be elimination of the frequency modulation of signal which was not intended by us (see the experiment with the amplitude and frequency modulated signal). Hence, if the method relying on the Hilbert transform is used, then we should select carefully the frequency band to be used for determining the course of the instantaneous frequency, so that the bandwidth contains the form of the non-stationarity of signal that we really want to eliminate. Results of the experiment confirmed the possibility of Doppler Effects removal from the real signal in the field measurements. The difference between results obtained with theoretical frequency curve and curve obtained using signal filtration were enclosed in its mean values. Another conclusion refers to Doppler and beat effect. It is possible to efficient use of the effects by choosing proper measure parameters and capable method for envelop identification. Numerical experiment shows that empirical mode decomposition method is sufficient for such a purpose. All described methods and possibilities could be used in the development of the ground pass-by vehicle diagnostic station capable of monitoring the technical state of passing vehicles.
Acoustic signal processing and analysis for the byside monitoring station
2008
16 Seiten, 18 Bilder, 9 Quellen
Conference paper
Storage medium
English
Combustion Monitoring Based on Engine Acoustic Emission Signal Processing
SAE Technical Papers | 2009
|Combustion monitoring based on engine acoustic emission signal processing
Automotive engineering | 2009
|2009-01-1024 Combustion Monitoring Based on Engine Acoustic Emission Signal Processing
British Library Conference Proceedings | 2009
|