Comparative Analysis of Wiener Filtering Techniques for Combustion and Mechanical Noise Breakdown in Diesel Engines

The primary objective of this thesis is to compare various Wiener filtering techniques documented in the literature for decomposing total engine noise into its combustion and mechanical components, as well as to introduce and evaluate an innovative methodology that combines aspects of the existing techniques. The Wiener filtering technique consists of a coherence method in which the in-cylinder pressure is taken as a reference for the combustion process. By analyzing the coherence between the in-cylinder pressure and the engine noise, the combustion contribution to the engine noise can be estimated. This study is conducted on an internal combustion diesel engine. A brief introduction to the combustion process and the engine noise contribution has been provided, including the use of in-cylinder pressure as a reference for assessing combustion noise. The methodologies investigated include the classical Wiener filter and the cyclostationary Wiener filter. These techniques are employed to estimate the transmissibility functions, which are then used to determine the combustion noise contribution through a convolution process with the in-cylinder pressure signals. The innovative approach proposed in this thesis applies windowing to the microphone signal, similar to the cyclostationary Wiener filter, but maintains the stationarity and ergodicity assumptions of the classical approach, relying on temporal averaging rather than cycle-by-cycle averaging. This comparative analysis is crucial in the automotive field, particularly for OEMs, as it provides a fast and reliable method to estimate combustion noise. This method allows for the assessment of sound emissions resulting from different engine calibration parameters without the need for extensive bench testing. By establishing accurate transmissibility functions, the combustion noise can be reliably estimated from in-cylinder pressure signals, aiding in compliance with increasingly stringent noise regulations.