Optimization of actuator configuration for the reduction of structure-borne noise in automobiles

Abstract

In this thesis, a strategy is proposed for the optimization of actuator configuration in the implementation of Active Structural Acoustic Control (ASAC) of road noise in an automobile suspension. First, a laboratory test bench consisting of a quarter-car suspension consisting of a wheel/suspension/lower A-arm assembly is modeled. A 12 degrees-of-freedom discrete element model of the rigid parts of the suspension is first used to produce global suspension resonances. Equivalent rigidity models of flexible components are then measured experimentally or identified with the help of genetic algorithms. The discrete element model and the equivalent rigidity models are combined to reproduce test bench Frequency Response Functions (FRFs) measured on the test bench. Second, the impedance of the test bench tire table is corrected using experimental measurements and analytical road profiles to reproduce a realistic road excitation. Afterward, a Chevrolet EPICA LS automobile is instrumented and operated on a concrete test track to observe the relative importance of road-induced vibrations compared to other noise sources in a moving car at 50 km/h. FRFs between the road excitation ant the car interior pressure level at the driver's head are measured and compared to a complete car transmission path tool (composed of the quarter-car test bench and a car frame finite element model). The FRF analysis between reference sensors and error microphones reveals the difficulty of obtaining sufficient experimental coherence to realise the active control of a suspension. Finally, an algorithm is implemented to find optimal actuator locations and orientations in the ASAC, using the suspension model and the filtered road excitation. Genetic algorithm tools are used with the suspension model for actuator positioning: the tire, the coil spring & the car panels control volumes are included into the model to constrain the evolution of the algorithm. For a given actuator configuration, the optimal control command is obtained by quadratic minimization of specific cost function (displacement at suspension links, force transmissibility & sound pressure level). Optimal actuator configurations are then suggested for future studies.