Acoustic Machine Inspection Using Beamforming and Dictionary Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial machine inspection faces challenges in early detection of machine faults due to corrupted acoustic signals by interference and background noise, especially in scenarios with multiple machines operating simultaneously, where existing methods struggle to effectively separate and analyze multi-channel acoustic signals.
Innovation Solution
The method employs Delay-and-Sum beamforming (DAS-BF) and dictionary learning (DL) to separate acoustic signals from multiple sources, using a two-stage approach: initial coarser separation with DAS-BF followed by refined source separation using pre-trained dictionaries, and anomaly detection by comparing separated signals with normal machine sound templates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If acoustic signals are captured in an industrial plant with multiple machines operating simultaneously, then machine fault detection capability is improved, but signal quality deteriorates due to interference and background noise
Solution Approach 1:
The patent segments the mixed acoustic signal into individual source signals using beamforming technology. The microphone array captures the mixed signal, and beamforming algorithms spatially separate the signals from different machines, allowing individual machine fault detection despite the presence of multiple operating machines and background noise.
Solution Approach 2:
The patent introduces an intermediary processing system consisting of beamforming algorithms and signal processing techniques that act as a mediator between the noisy industrial environment and the fault detection analysis. This intermediary system filters and separates the acoustic signals, removing the harmful effects of background noise and interference before analysis.
2Device complexity
If traditional acoustic analysis methods are used on mixed signals from multiple machines, then system complexity is reduced, but measurement precision deteriorates due to inability to separate source signals
Solution Approach 1:
The patent transitions from single-channel to multi-channel acoustic analysis by deploying a microphone array. This adds the spatial dimension to the analysis, allowing signals from different machines to be separated based on their spatial origins. The beamforming process exploits this additional dimensional information to achieve precise source separation.
3Measurement precision
If beamforming and dictionary learning are applied to separate acoustic sources, then source separation accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies preliminary beamforming processing to the multi-channel acoustic signals before performing dictionary learning. This preliminary action of spatial filtering and source separation preparation reduces the dimensionality and complexity of the subsequent dictionary learning process, making the overall system computationally feasible while maintaining high source separation accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively reduces interference and background noise, enabling robust anomaly detection in complex industrial settings, outperforming state-of-the-art methods in accuracy and F1 score, particularly in multi-source scenarios.
Implementation Method 1
obtaining, a plurality of beamformed source signals, by feeding the multi-channel acoustic mixed signal to a DAS-BF
Implementation Method 2
computing, a plurality of Mel-spectrograms corresponding to each of the plurality of beamformed source signals
Data Source
AI summary
This disclosure relates generally to a field of industrial machine inspection, and, more particularly, to method and system for acoustic based industrial machine inspection using Delay-and-Sum beamforming (DAS-BF) and dictionary learning (DL). The disclosed method presents a two-stage approach for anomaly detection using a multi-channel acoustic mixed signal. In the first stage, separation of a plurality of acoustic signals corresponding to the spatially distributed acoustic sources is performed at a coarser level by using the DAS-BF. Subsequently, dictionaries pre-trained using the plurality of acoustic signals of the individual source machines are utilized for generating a plurality of separated acoustic source signals. The generated plurality of separated acoustic source signals are analyzed for the anomaly detection by comparing them with a corresponding normal machine sound template.