Search

In the previous reports, a methodology for estimating noise radiation including airborne noise from a mechanical system under the operational condition by component test of a certain active subsystem using the in-situ blocked force approach, which is normally used for estimating only structure borne noise, was proposed and verified by numerical simulations of noise radiation from a cantilever rectangular plate. This paper describes experimental verification of the proposed methodology by application to rattle noise from a column type electric power steering system (EPS) for a vehicle. Therefore an EPS corresponds to the active subsystem and the other parts such as vehicle body correspond to passive subsystems. At first, a method of measuring rattle noise from an EPS under the operational conditions, i.e., running tests with a vehicle, was established. Since it is possible to apply the in-situ blocked force approach also to vehicle running tests and to estimate rattle noise, the measured and estimated rattle noises under an identical operational condition were compared and validity of the proposed methodology was confirmed. Next, a component test bench which enables to reproduce vibration behavior of an EPS under the operational conditions was developed, and rattle noise under an operational condition was estimated by means of the proposed methodology using both vibration accelerations measured with this component test bench and vibro-acoustic transfer functions measured with the vehicle under a static condition. Finally, the proposed methodology was verified by comparing the estimated and measured rattle noises under the operational condition. The proposed methodology is expected to replace time-consuming operational tests of a whole mechanical system to simple component tests of the active subsystem.

The in-situ blocked force approach is suitable for estimating noise radiation from a mechanical system by component tests of its active subsystem because blocked force is a specific characteristic of the active subsystem and independent from the passive subsystem or supporting structure of the component test bench for the active subsystem. However, it is reported that the in-situ blocked force approach is a methodology for estimating structure borne noise and cannot be applied to airborne noise. In this research it was found that the in-situ blocked force approach can be approximately applied to airborne noise in eigenfrequencies of the mechanical system. Therefore it was proposed to employ the in-situ blocked force approach to estimate airborne noise as an approximate method at eigenfrequencies of the mechanical system in the previous paper. In this paper the proposed estimation method was verified by numerical analysis of noise radiation from a cantilever rectangular plate. Eigenfrequencies and vibration responses of the cantilever rectangular plate were obtained by FEM and sound pressures at an observed point were calculated on the assumption that point sound sources are distributed at the all FE nodes on the whole plate surface. Furthermore, a means based on discarding small singular values in the inverse matrix method was proposed in order to reduce the estimation error which is increased in the peak frequencies of blocked forces.

This paper proposes a methodology for estimating airborne noise from a mechanical system under the operational conditions by component tests of a certain active subsystem which is a part of the mechanical system and includes all of vibration sources in the mechanical system. The mechanical system consists of the active and passive subsystems, and supporting structure of the active subsystem is different from that under the component test conditions. Hence airborne noise from the active system under these two conditions are different from each other. Therefore the in-situ blocked force approach is expected to estimate airborne noise under the operational conditions by the component tests, because blocked force is a specific characteristic of the active subsystem and independent from its supporting structure. However it is reported that the in-situ blocked force approach is suitable for estimating structure borne noise and cannot be applied to airborne noise, it has been found that the in-situ blocked force approach can be approximately applied to airborne noise in resonant frequency of the mechanical system. This paper describes theoretical derivation of the approximation error caused by applying the in-situ blocked force approach to airborne noise. Furthermore, the error is evaluated by means of a 3-DOF simple model.