Visualization of bearing noise using adaptive high-resolution near-field beamforming

Beamforming (BF) is an array processing technique that has been extensively applied in angular source localization applications such as sonar and radar, mostly from measurement in the far field. However, near-field source resolution performance of BF procedures can be significantly improved by using channel weights that are inversely proportional to the distance from the hypothesized source location to the measurement locations. Visualization of noise radiated from bearings is challenging since the measurement is usually corrupted by scattering and reflection from bearing housings that are located very close to sources even though the measurement may be taken in an anechoic environment. In this article, adaptive weightings are used in near-field BF procedure, and shown to give very good visualization of sound radiation of rolling bearings that are a complex sound source in a highly scattered and noisy environment. It is shown that for a machine made of several components, this adaptive near-field BF technique provides good imaging performance and allows identification of the primary sources of sound radiation at a broad range of frequencies.