Decomposition of measured velocity fields within spark ignition engines using discrete wavelet transforms

 Two techniques are generally used to extract turbulence information from engine flows: the ensemble and cyclic, or scale-based average. A new method is proposed here that uses an energy-based cyclic averaging technique. The energy-based filter proves to be superior in removing the coherent events associated with the cycle-to-cycle variations in the mean flow. Received: 8 June 1998/Accepted: 20 March 2000