Research on Low-Frequency Acoustic Source Localization Based on Bayesian Methods
Tianhua Shen, Baohua Li, Yuanjun Dai, Fanglong Zhu · 2025
Early diagnosis of mechanical faults is crucial for ensuring the safe operation of equipment, and low-frequency acoustic source localization technology, as a non-invasive detection method, holds significant application value in mechanical fault diagnosis. However, traditional beamforming methods (such as CBF) suffer from insufficient resolution and poor noise resistance under low-frequency and low signal-to-noise ratio (SNR) conditions, making it difficult to accurately identify the source locations of mechanical faults. To address these challenges, this paper proposes a low-frequency acoustic source localization method based on Sparse Bayesian Learning (SBL), aiming to improve the localization accuracy and noise resistance of mechanical fault sources. First, a signal model for acoustic source localization based on a microphone array is constructed, and the distribution of mechanical fault sources is modeled using sparse priors, transforming the source localization problem into a sparse signal recovery problem. By maximizing the posterior probability, high-precision estimation of mechanical fault source locations is achieved. Second, a dynamic noise adjustment mechanism is proposed to address the complex noise interference commonly found in mechanical operating environments, further enhancing the robustness of the algorithm. Simulation and experimental results demonstrate that the proposed method exhibits high spatial resolution and source identification capability under lowf-requency and low-SNR conditions, particularly showing stronger anti-interference ability and localization accuracy in complex noise environments. Compared to traditional CBF methods, the proposed method significantly improves the performance of mechanical fault source localization, providing reliable technical support for early fault diagnosis and condition monitoring of mechanical equipment. Future research will focus on optimizing the computational efficiency of the algorithm and exploring its potential applications in more complex mechanical operating environments.