On measurement, assessment and control of diesel engine noise

Sammanfattning: The thesis comprises six papers with the overall theme of measurement, assessment and control of diesel engine noise, with an emphasize on control. The radiation of noise is shown to be dominated by the low end of the engine sides and by the engine front. The mode shapes of the crankcase, the oil sump and the timing transmission cover are investigated. The vibrations are measured and analysed through running mode, modal analyses and SEA. The noise radiation is explored in detail with near-field measurements of sound intensity. In the engine front, the noise radiation is shown to have contributions from the timing casing, the oil-sump front and the crankshaft torsion vibration damper. The radiation from the torsion damper is analysed by a regression analysis of measured intensity data and the vibrations of the damper are investigated with a laser-doppler vibrometer. The results show that the damper vibrates in the axial directions with vibration modes that both radiate noise and interfere with the radiation from the engine structure. It is shown that there in the engine front are strong coupled vibration modes between the timing cover and the oil sump front m the frequency range 500 Hz - 1 kHz. The importance of the strong vibration modes in the crankcase and the oil-sump sides is shown. The propagation paths of noise and vibration to the engine front have been examined with an SEA powerflow analysis and by opening the front cover to measure the sound power from the timing gears. The main excitations of the front cover are found to be by engine block vibrations below 1.25 kHz and by tuning gears noise above 2 kHz. Various constructions to control the noise are tested. Two stiffeners are designed and tested to reduce vibrations m the engine low end, one ladder frame introducing stiffness at the crankcase flange and one bearing beam introducing stiffness at the main bearing caps. The stiffeners have hem evaluated by sound intensity measurements and mobility measurements. The ladder frame gave good noise reductions but the bearing beam merely caused frequency shifts of the bearing modes. A decoupling of the oil sump resulted in significant noise and vibration reductions. The timing cover is modified by increasing the damping and by decrease the radiation efficiency. Various interior panels are tested, like plexiglass and aluminium panels of different thicknesses, a rubber damping layer and a combined rubber/steel-sheet damping layer. A thin plastic sheet has low radiation efficiency and may thereby lower the noise emission. A slightly improved model for calculation of radiation efficiency of small irregularly shaped plates is suggested. The traditional SEA prediction model is shown to be ill-conditioned for engine applications. An improved model using geometric averaging is suggested and evaluated. The results show that the new SEA model is working well for frequencies down to 800 Hz for predictions of damping treatments, decoupling of the oil sump and for power-flow determinations. The concept of equivalent mass is found valuable and validations are made according to the consistency and reciprocity theories. A hemi-anechoic engine laboratory is constructed and evaluated. New efficient low-cost diffusing absorbers have been designed, the performance is evaluated with standard deviation analysis of sound pressure measurements. The performance of the absorbers is found to be compeatable with much more expensive commercial designs. Measured sound intensity in three-dimensional vectors is a powerful tool to identify and illustrate sound fields. When used in near-fields to identify complex noise sources large errors may occur. It is shown that the low relative levels of vector components and the reactivity of the sound field give large errors m repeated measurements. Analyses and comparisons are performed on a simple source and on an engine. A two- and a six-microphone probe were used that were hand-held and robot-controlled. It is concluded dig in point-intensity measurements on engineering noise sources, a 4-6 microphone probe and a precision positioner should be used to get reliable measurements.