CN-121997199-A - Fault diagnosis method and device for vibrating screen spring

CN121997199ACN 121997199 ACN121997199 ACN 121997199ACN-121997199-A

Abstract

The application provides a vibrating screen spring fault diagnosis method and device, wherein the method comprises the steps of collecting an original vibration signal of a vibrating screen spring through a vibration sensor; extracting time-frequency characteristic signals from the original vibration signals, extracting spatial characteristics and time characteristics from the time-frequency characteristic signals respectively by utilizing a pre-trained characteristic fusion neural network model, fusing the spatial characteristics and the time characteristics to obtain fusion characteristics, inputting the fusion characteristics into a pre-trained fault classifier, and outputting a fault diagnosis result. The time characteristics and the space characteristics of the signals are respectively extracted by utilizing the characteristic fusion neural network model, and finally, the spring fault diagnosis is realized by the fusion characteristics, so that the vibration screen spring fault can be timely and accurately diagnosed, and the method has important significance for guaranteeing the safe and stable operation of the coal preparation production system.

Inventors

LIAN KAI
WANG FENG
GUO YONGFENG
LU YONGHENG
NIU BEN
XIAO WANGQIANG
CAI ZHIQIN
LIN JUNZE

Assignees

神华准格尔能源有限责任公司

Dates

Publication Date: 20260508
Application Date: 20251216

Claims (10)

1. A method for diagnosing a vibrating screen spring fault, comprising: Collecting an original vibration signal of a vibrating screen spring through a vibration sensor; Extracting a time-frequency characteristic signal from the original vibration signal; Respectively extracting spatial features and time features from the time-frequency feature signals by utilizing a pre-trained feature fusion neural network model; Fusing the spatial features and the temporal features to obtain fusion features; and inputting the fusion characteristics into a pre-trained fault classifier, and outputting a fault diagnosis result.
2. The method of claim 1, wherein extracting a time-frequency characteristic signal from the raw vibration signal comprises: performing continuous wavelet transformation on the original vibration signal, and extracting corresponding first time-frequency characteristics; performing characteristic modal decomposition on the original vibration signal to obtain a series of eigen-modal functions as second time-frequency characteristics; And performing short-time Fourier transform on the original vibration signal to obtain a third time-frequency characteristic, wherein the time-frequency characteristic signal comprises at least one of the first time-frequency characteristic, the second time-frequency characteristic and the third time-frequency characteristic.
3. The method for diagnosing a spring fault in a vibrating screen according to claim 1, wherein the feature fusion neural network model comprises a dynamic graph convolution network and a long-short-term memory network, and the method for extracting spatial features and temporal features from the time-frequency feature signals by using the pre-trained feature fusion neural network model comprises the following steps: Patterning the time-frequency characteristic signal into space-time diagram data; Extracting the spatial features from the space-time diagram data by using the dynamic diagram convolution network; and extracting the time characteristic from the time-frequency characteristic signal by using the long-period and short-period memory network.
4. A method of diagnosing a vibrating screen spring fault as in claim 3, wherein extracting the spatial feature from the space-time diagram data using the dynamic diagram convolution network comprises: processing the space-time diagram data through the dynamic diagram convolution network, performing EdgeConv operations to dynamically update the diagram structure and capture local geometric features; introducing a channel attention mechanism to adaptively adjust weights of different channel features in the local geometric features; and reducing the dimension of the weighted local geometric features through a maximum pooling layer and a global average pooling layer, and integrating global space information to obtain the space features.
5. A method of diagnosing a spring failure in a vibrating screen as in claim 3, wherein extracting the time signature from the time-frequency signature signal using the long-short term memory network comprises: Performing one-dimensional convolution operation and batch standardization processing on the time sequence; Inputting the processed characteristics into the long-period and short-period memory network, and capturing long-period dependency relationship through a gating mechanism of the long-period and short-period memory network; the temporal feature is formed based on the long-term dependency.
6. The method for diagnosing a spring fault in a vibrating screen according to claim 3, wherein the feature fusion neural network model further comprises a feature fusion sub-network, and wherein the step of fusing the spatial feature and the temporal feature to obtain a fused feature comprises the steps of: Splicing the spatial features and the temporal features in feature dimensions; inputting the spliced features to a full-connection layer of the feature fusion sub-network for nonlinear combination; And executing regularization processing on the features combined by the full connection layers to generate the fusion features.
7. A vibrating screen spring fault diagnosis device, characterized by comprising: the acquisition module is used for acquiring an original vibration signal of the vibrating screen spring through the vibration sensor; the signal extraction module is used for extracting a time-frequency characteristic signal from the original vibration signal; The feature extraction module is used for respectively extracting spatial features and time features from the time-frequency feature signals by utilizing a pre-trained feature fusion neural network model; The fusion module is used for fusing the spatial features and the time features to obtain fusion features; And the prediction module is used for inputting the fusion characteristics into a pre-trained fault classifier and outputting a fault diagnosis result.
8. An electronic device comprising a processor and a memory storing computer readable instructions that, when executed by the processor, perform the method of any of claims 1-6.
9. A storage medium having stored thereon a computer program which, when executed by a processor, performs the method of any of claims 1-6.
10. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method of any of claims 1-6.

Description

Fault diagnosis method and device for vibrating screen spring Technical Field The invention relates to the field of mechanical fault diagnosis, in particular to a fault diagnosis method and device for a vibrating screen spring. Background The vibrating screen spring is used as key equipment in coal preparation plants, and the stable operation of the vibrating screen spring plays a critical role in production efficiency and product quality. However, under the working environment of high-speed rotation and strong vibration, the vibrating screen is subjected to frequent change of alternating load and unexpected impact, and the factors easily cause the faults of parts such as springs of the vibrating screen. The fault diagnosis of the vibrating screen spring has important significance for guaranteeing the safe and stable operation of the coal preparation production system. At present, research on a vibrating screen spring fault diagnosis technology mainly focuses on improving accuracy of a diagnosis result by using an intelligent diagnosis method, especially under the condition that a fault sample is limited. The fault diagnosis technology of the vibrating screen spring is developing towards intelligentization and precision, and has important practical significance for improving the manufacturing level and optimizing the production flow. However, the comprehensive influence of multiple factors on the fault diagnosis of the vibrating screen spring is not fully considered in the prior art, and particularly the problem of joint noise interference generated when a plurality of vibrating screens are operated simultaneously is solved. Such disturbances may mask the true fault signature, resulting in erroneous or missed diagnosis of the diagnostic model, reducing the accuracy of fault identification. Disclosure of Invention The invention provides a fault diagnosis method and device for a vibrating screen spring, which are used for solving the problem of nonlinear and non-stationary fault signals caused by external noise and other factors in the process of coal dressing of the vibrating screen spring, and realizing effective extraction and accurate diagnosis of uncertain fault characteristics. In a first aspect, the present invention provides a method for diagnosing faults of a vibrating screen spring, comprising: Collecting an original vibration signal of a vibrating screen spring through a vibration sensor; Extracting a time-frequency characteristic signal from the original vibration signal; Respectively extracting spatial features and time features from the time-frequency feature signals by utilizing a pre-trained feature fusion neural network model; Fusing the spatial features and the temporal features to obtain fusion features; and inputting the fusion characteristics into a pre-trained fault classifier, and outputting a fault diagnosis result. Optionally, extracting a time-frequency characteristic signal from the original vibration signal includes: performing continuous wavelet transformation on the original vibration signal, and extracting corresponding first time-frequency characteristics; performing characteristic modal decomposition on the original vibration signal to obtain a series of eigen-modal functions as second time-frequency characteristics; And performing short-time Fourier transform on the original vibration signal to obtain a third time-frequency characteristic, wherein the time-frequency characteristic signal comprises at least one of the first time-frequency characteristic, the second time-frequency characteristic and the third time-frequency characteristic. Optionally, the feature fusion neural network model comprises a dynamic graph convolution network and a long-term and short-term memory network, and the method for extracting the spatial features and the temporal features from the time-frequency feature signals by utilizing the pre-trained feature fusion neural network model comprises the following steps of: Patterning the time-frequency characteristic signal into space-time diagram data; Extracting the spatial features from the space-time diagram data by using the dynamic diagram convolution network; and extracting the time characteristic from the time-frequency characteristic signal by using the long-period and short-period memory network. Optionally, extracting the spatial feature from the space-time diagram data using the dynamic diagram convolution network includes: processing the space-time diagram data through the dynamic diagram convolution network, performing EdgeConv operations to dynamically update the diagram structure and capture local geometric features; introducing a channel attention mechanism to adaptively adjust weights of different channel features in the local geometric features; and reducing the dimension of the weighted local geometric features through a maximum pooling layer and a global average pooling layer, and integrating global space information to obtain the space