Paper No. 88
The digital signal processing techniques for identification of noise or vibration energy sources; ordinary, multiple and partial coherence techniques, either require incoherence of the measured input data or a degree of apriori knowledge of the system. This paper discusses a transformation technique for conditioning the measured spectra to determine how many real incoherent sources exist and to create a virtual image of those sources at the measurement plane. In addition, the conditioned spectra can be used to generate a “virtual” coherence function between the measured output and each of the incoherent inputs.