Title:
MICROPHONE ARRAY-BASED METHOD FOR LOCATING AND IDENTIFYING FAULT SIGNAL IN INDUSTRIAL EQUIPMENT
Document Type and Number:
WIPO Patent Application WO/2022/116570
Kind Code:
A1
Abstract:
Provided is a microphone array-based method for locating and identifying a fault signal in an industrial equipment: capturing sound signals, and dividing into a training set, a validation set, and a testing set; performing feature extraction on a sound signal in the training set, and extracting a phase spectrogram and amplitude spectrogram of the frequency spectrogram; sending the output of the feature extraction module as input to a convolutional neural network (CNN), and in each layer of the CNN, using a 2D CNN to learn translation invariance in a frequency spectrogram; between CNN layers, using batch normalization to normalize the output, and using a maximum pooling layer along the frequency axis to reduce the dimensionality; sending the output from the CNN layer to a recurrent neural network (RNN) layer; using a linear activation function; inputting the output of a fully connected layer to two parallel branches of the fully connected layer, and using for fault identification and fault localization, respectively. Using a contactless method based on audio signals, there are fewer restrictions by the environment and exposure, and fault localization and fault identification are performed in parallel, improving operational efficiency.
More Like This:
Inventors:
LUAN FENG (CN)
LI XU (CN)
ZHANG ZIMING (CN)
WU YAN (CN)
HAN YUEJIAO (CN)
ZHANG DIANHUA (CN)
LI XU (CN)
ZHANG ZIMING (CN)
WU YAN (CN)
HAN YUEJIAO (CN)
ZHANG DIANHUA (CN)
Application Number:
PCT/CN2021/109200
Publication Date:
June 09, 2022
Filing Date:
July 29, 2021
Export Citation:
Assignee:
UNIV NORTHEASTERN (CN)
International Classes:
G01H17/00; G06K9/62; G06N3/04; G06N3/08
Foreign References:
| CN112560913A | 2021-03-26 | |||
| CN111160167A | 2020-05-15 | |||
| CN107154037A | 2017-09-12 | |||
| US20200371491A1 | 2020-11-26 |
Attorney, Agent or Firm:
SHENYANG DONGDA INTELLECTUAL PROPERTY AGENCY CO., LTD (CN)
Download PDF: