Abstract
A diagnostic method, which uses the two-sided directional power spectra of complex-valued engine vibration signals, is presented and tested with four-cylinder compression and spark ignition engines for the diagnosis of cylinder power faults. As spectral estimators, the maximum likelihood and FFT methods are compared, and the multi-layer neural network is employed for pattern recognition. Experimental results show that the success rate for identifying the misfired cylinder is much higher with the use of two-sided directional power spectra than conventional one-sided power spectra.