Noise controlled diesel engine
Generally speaking, cylinder pressure development in the combustion engine is regarded as a key variable for describing processes taking place inside the engine. Integrating this into cylinder-pressure-based engine management makes it possible to control combustion on a cylinder-selective basis. The use of cylinder pressure sensors, however, increases the cost of production. In diesel engines, combustion noise is also particularly dominant in the low engine speed and load range. For this reason, the following thesis examines how to obtain the combustion variables required for a cylinder-pressure-based engine management system from structure-borne sound signals and integrate diesel knocking noise behavior into the control structure. Based on structure-borne sound, an engine management system is developed that uses virtual sensors to deliver the combustion variables of cylinder pressure and provide an objective characteristic value of diesel knocking noise.
The first step in developing the virtual cylinder
pressure sensor involves determining suitable structure-borne sound
positions by means of coherence analysis and then examining them in
the time and frequency range on the basis of
Smoothed-Pseudo-Wigner-Ville distribution. This reveals signal
components of combustion and injection in the structure-borne sound
signal. In the next step, these are extracted and form the basis for
the virtual sensor. In conjunction with control unit variables,
models estimate the required combustion variables. The virtual
cylinder pressure sensor is validated on the basis of a steady-state
operating point and load increase.
Engine noise is assessed using objective characteristic values based on subjective evaluations. Diesel knocking noise is assessed here in the part-load range with the diesel noise rating. The diesel noise rating is calculated on the basis of the loudness and modulation spectrum from the airborne-sound signal. Sensing airborne sound directly in the vehicle is not practicable. Instead, the approach taken calculates loudness and modulation directly from the structure-borne sound signal. The variables calculated from structure-borne sound provide the basis for the virtual noise sensor. In combination with control unit variables, models are generated that are capable of calculating the diesel noise rating of the airborne-sound signal with good results.
Finally, the engine management system based on structure-borne sound is put into operation at an operating point selected with the aid of center of heat release control and simultaneously determined diesel noise rating. Comparing cylinder-pressure-controlled engine operation with engine operation controlled on the basis of structure-borne sound reveals no significant differences.