Gear wear monitoring method and system based on sensor data

Abstract

The invention discloses a gear wear monitoring method and system based on sensor data, and the method comprises the steps: setting a first infrared visual sensor, obtaining a gray image of the end face of a target gear in real time, and generating a first gray image; a second infrared vision sensor is arranged, the gray level image of the tooth surface of the target gear is obtained in real time, and a second gray level image is generated; inputting the first grayscale image into a pre-trained target detection model, outputting a detection frame of a normal gap at a gear meshing position in the first grayscale image, and generating a first judgment parameter based on the detection frame; inputting the second grayscale image into a pre-trained convolutional neural network model, outputting a tooth surface wear confidence coefficient of the target gear, and taking the tooth surface wear confidence coefficient as a second judgment parameter; if the first judgment parameter and the second judgment parameter both exceed the corresponding judgment threshold values, it is judged that the target gear is seriously abraded, and real-time automatic abrasion degree monitoring on the gear is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of gear wear monitoring, and particularly relates to a gear wear monitoring method and system based on sensor data. Background Art

[0002] Gears are widely used in various industries and can transmit motion and power through continuous meshing. Gears are usually made of metal and have a certain hardness and wear resistance. When the gear surface is worn excessively, it will affect the normal transmission of motion and power of the gear, thereby reducing the efficiency of the equipment. To evaluate the wear degree of the gear, it is usually judged by artificial vision to determine the wear marks and wear amount on the tooth surface. This method has certain subjectivity and cannot achieve real-time and automatic monitoring of the wear degree of the gear.

[0003] CN116642691A and CN111751108A disclose methods for monitoring the wear degree of gears based on the vibration generated by gear meshing. However, these methods have the problem that the vibration sensor is easily affected by external interference, resulting in a relatively high monitoring error rate. Summary of the Invention

[0004] Aiming at the above-mentioned technical deficiencies, the purpose of the present invention is to provide a gear wear monitoring method and system based on sensor data, aiming to solve the problems that the existing methods for detecting the wear degree of gears are usually judged by artificial vision to determine the wear marks and wear amount on the tooth surface, which has certain subjectivity and cannot achieve real-time and automatic monitoring of the wear degree of gears, and the method for monitoring the wear degree of gears based on the vibration generated by gear meshing has the problem that the vibration sensor is easily affected by external interference, resulting in a relatively high monitoring error rate.

[0005] In view of the above problems, the present application provides a gear wear monitoring method and system based on sensor data.

[0006] In the first aspect disclosed in the present application, a gear wear monitoring method based on sensor data is provided. The method includes: Set a first infrared vision sensor with the shooting direction parallel to the axis of the target gear, and obtain the grayscale image of the end face of the target gear in real time to generate a first grayscale image. Among them, the first infrared vision sensor takes the direction of the tangent at the gear meshing point as the horizontal direction of shooting; Set a second infrared vision sensor with the shooting direction perpendicular to the axis of the target gear, and obtain the grayscale image of the tooth surface of the target gear in real time to generate a second grayscale image. Among them, the second infrared vision sensor takes the direction of the gear axis as the horizontal direction of shooting; Input the first grayscale image into a pre-trained object detection model. The object detection model outputs a detection frame of the normal clearance at the gear meshing position in the first grayscale image. Based on the detection frame, obtain the width of the normal clearance at the gear meshing position, and generate a first determination parameter; Input the second grayscale image into a pre-trained convolutional neural network model. The convolutional neural network model outputs the tooth surface wear confidence of the target gear, and use the tooth surface wear confidence as the second determination parameter; If both the first determination parameter and the second determination parameter exceed their corresponding determination thresholds, then determine that the target gear is severely worn, and control to send an alarm signal.

[0007] Preferably, the step of inputting the first grayscale image into a pre-trained object detection model, where the object detection model outputs a detection frame of the normal clearance at the gear meshing position in the first grayscale image, and based on the detection frame, obtain the width of the normal clearance at the gear meshing position, and generate a first determination parameter, specifically includes: Input the first grayscale image into a pre-trained object detection model. Among them, the object detection model is the YOLOv4 model, which includes an input layer, a backbone network, a neck network, and a head network. The input layer is used to perform normalization operations on the input first grayscale image. The backbone network and the neck network are used to perform feature extraction and feature fusion on the first grayscale image that has completed the normalization operation. The head network is used to output the object detection result of the target of the normal clearance at the gear meshing position in the first grayscale image; Decode the object detection result to generate a detection frame of the normal clearance at the gear meshing position in the first grayscale image. Among them, the detection frame is represented by the width, height, and center point position of the detection frame; Use the width of the detection frame as the width of the normal clearance at the gear meshing position, and generate a first determination parameter.

[0008] Preferably, the step of inputting the second grayscale image into a pre-trained convolutional neural network model, where the convolutional neural network model outputs the tooth surface wear confidence of the target gear, and use the tooth surface wear confidence as the second determination parameter, specifically includes: Input the second grayscale image into a pre-trained convolutional neural network model. Among them, the convolutional neural network model includes a convolutional layer, a pooling layer, and a fully connected layer. The convolutional layer is used to extract image features through convolutional calculations. The pooling layer is used for downsampling. The fully connected layer is used to output a determination result through non-linear calculations; Input the determination result into the sigmod activation function to generate the tooth surface wear confidence of the target gear, and use the tooth surface wear confidence as the second determination parameter.

[0009] Preferably, if both the first determination parameter and the second determination parameter exceed their respective determination thresholds, the target gear is determined to be severely worn, and an alarm signal is controlled to be issued. Specifically, it includes: A first determination threshold is preset, and the numerical range of the first determination threshold is greater than 0 and less than 20% of the tooth thickness of the target gear. A second determination threshold is preset, and the numerical range of the second determination threshold is greater than 0.5 and less than 1. If the first determination parameter exceeds the first determination threshold and the second determination parameter exceeds the second determination threshold, the target gear is determined to be severely worn, and an alarm signal is controlled to be issued.

[0010] In a second aspect disclosed in the present application, a gear wear monitoring system based on sensor data is provided. The system is used for the above-mentioned gear wear monitoring method based on sensor data. The system includes: A first acquisition module, which is used to set a first infrared vision sensor with the shooting direction parallel to the axis of the target gear, and obtain the grayscale image of the end face of the target gear in real time to generate a first grayscale image. Among them, the first infrared vision sensor takes the direction of the tangent at the gear meshing position as the horizontal direction of shooting; A second acquisition module, which is used to set a second infrared vision sensor with the shooting direction perpendicular to the axis of the target gear, and obtain the grayscale image of the tooth surface of the target gear in real time to generate a second grayscale image. Among them, the second infrared vision sensor takes the direction of the gear axis as the horizontal direction of shooting; A first determination module, which is used to input the first grayscale image into a pre-trained target detection model. The target detection model outputs the detection frame of the normal clearance at the gear meshing position in the first grayscale image, and obtains the width of the normal clearance at the gear meshing position based on the detection frame to generate a first determination parameter; A second determination module, which is used to input the second grayscale image into a pre-trained convolutional neural network model. The convolutional neural network model outputs the tooth surface wear confidence of the target gear, and uses the tooth surface wear confidence as the second determination parameter; An alarm module, which is used to determine that the target gear is severely worn if both the first determination parameter and the second determination parameter exceed their respective determination thresholds, and control the issuance of an alarm signal.

[0011] In a third aspect disclosed in the present application, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned gear wear monitoring method based on sensor data is implemented.

[0012] The fourth aspect disclosed in the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the above-mentioned gear wear monitoring method based on sensor data.

[0013] The fifth aspect disclosed in the present application provides a computer program product, including a computer program or instruction, and when the computer program or instruction is executed by a processor, it implements the above-mentioned gear wear monitoring method based on sensor data.

[0014] The beneficial effects of the present invention are as follows: (1) It realizes real-time and automatic monitoring of the wear degree of gears, and solves the problems that manual detection of tooth surface wear marks and wear amounts cannot achieve real-time and automatic monitoring and is subjective.

[0015] (2) By using an infrared vision sensor to detect the wear degree of gears based on computer vision, it solves the problem that vibration sensors are easily affected by external interference, resulting in a high monitoring error rate. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] Figure 1 It is the overall flowchart of a gear wear monitoring method based on sensor data.

[0018] Figure 2 It is the overall structure diagram of a gear wear monitoring system based on sensor data. Detailed Embodiments

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0020] As Figure 1 shown, the embodiments of the present application provide a gear wear monitoring method based on sensor data, and the method includes: Set a first infrared vision sensor with its shooting direction parallel to the axis of the target gear, and obtain in real time the grayscale image of the end face of the target gear to generate a first grayscale image. Among them, the first infrared vision sensor takes the direction of the tangent at the gear meshing position as the horizontal shooting direction; Set a second infrared vision sensor with its shooting direction perpendicular to the axis of the target gear, and obtain in real time the grayscale image of the tooth surface of the target gear to generate a second grayscale image. Among them, the second infrared vision sensor takes the direction of the gear axis as the horizontal shooting direction; Input the first grayscale image into a pre-trained target detection model. The target detection model outputs the detection frame of the normal clearance at the gear meshing position in the first grayscale image, and based on the detection frame, obtain the width of the normal clearance at the gear meshing position to generate a first determination parameter; Input the second grayscale image into a pre-trained convolutional neural network model. The convolutional neural network model outputs the tooth surface wear confidence of the target gear, and take the tooth surface wear confidence as the second determination parameter; If both the first determination parameter and the second determination parameter exceed their corresponding determination thresholds, then determine that the target gear is severely worn and control to send an alarm signal.

[0021] Furthermore, the step of inputting the first grayscale image into a pre-trained target detection model, where the target detection model outputs the detection frame of the normal clearance at the gear meshing position in the first grayscale image, and based on the detection frame, obtain the width of the normal clearance at the gear meshing position to generate a first determination parameter specifically includes: Input the first grayscale image into a pre-trained target detection model. Among them, the target detection model is the YOLOv4 model, which includes an input layer, a backbone network, a neck network, and a head network. The input layer is used to perform normalization operations on the input first grayscale image. The backbone network and the neck network are used to perform feature extraction and feature fusion on the first grayscale image that has completed the normalization operation. The head network is used to output the target detection result of the target of the normal clearance at the gear meshing position in the first grayscale image; Decode the target detection result to generate the detection frame of the normal clearance at the gear meshing position in the first grayscale image. Among them, the detection frame is represented by the width, height, and center point position of the detection frame; Take the width of the detection frame as the width of the normal clearance at the gear meshing position to generate a first determination parameter.

[0022] Specifically, before the training of the YOLOv4 model starts, grayscale images containing the normal clearance at the gear meshing are collected. At the same time, the grayscale images are processed by rotation and mirroring to ensure the diversity of the grayscale images, thereby constructing an image dataset. The annotation tool is used to annotate the normal clearance at the gear meshing in the grayscale images in the image dataset. The annotation format is the YOLO format, that is, the annotation file of each grayscale image contains the center point position, width, and height of the normal clearance at the gear meshing, and the position and size are both proportional values relative to the width and height of the image. The annotated image dataset is divided into a training set, a validation set, and a test set, with a division ratio of 7:2:1. The YOLOv4 model is trained using the training set, the validation set, and the test set. During the training of the YOLOv4 model, the terminal will output the loss value. When the loss value tends to converge, the YOLOv4 model is completed training.

[0023] Furthermore, the step of inputting the second grayscale image into the pre-trained convolutional neural network model, and the convolutional neural network model outputs the confidence level of the tooth surface wear of the target gear, and taking the confidence level of the tooth surface wear as the second determination parameter specifically includes: Input the second grayscale image into the pre-trained convolutional neural network model. Among them, the convolutional neural network model includes a convolutional layer, a pooling layer, and a fully connected layer. The convolutional layer is used to extract image features through convolutional calculations. The pooling layer is used for downsampling. The fully connected layer is used to output the determination result through non-linear calculations; Input the determination result into the sigmod activation function to generate the confidence level of the tooth surface wear of the target gear, and take the confidence level of the tooth surface wear as the second determination parameter.

[0024] Specifically, before the training of the convolutional neural network model starts, tooth surface images of gears with different wear degrees are collected to ensure that the images cover various possible wear conditions, including slight wear, moderate wear, and severe wear. At the same time, different external conditions are ensured to be covered, including lighting conditions, shooting angles, and gear materials. The collected images are labeled whether there is tooth surface wear in the form of a binary classification label, and the labeling format is in the form of a one-hot vector, thereby constructing an image dataset. The labeled image dataset is divided into a training set, a validation set, and a test set, with a division ratio of 7:2:1. The convolutional neural network model is trained using the training set, the validation set, and the test set, and the binary cross-entropy loss function is used as the loss function. When the loss value of the loss function tends to converge, the convolutional neural network model is completed training.

[0025] Furthermore, the step of determining that the target gear is severely worn and controlling the issuance of an alarm signal if both the first determination parameter and the second determination parameter exceed the corresponding determination thresholds specifically includes: A first determination threshold is preset, and the value range of the first determination threshold is greater than 0 and less than 20% of the target gear tooth thickness. A second determination threshold is preset, and the value range of the second determination threshold is greater than 0.5 and less than 1; If the first determination parameter exceeds the first determination threshold and the second determination parameter exceeds the second determination threshold, the target gear is determined to be severely worn, and an alarm signal is controlled to be sent.

[0026] In summary, a gear wear monitoring method based on sensor data provided by an embodiment of the present application has the following technical effects: (1) Realize real-time automatic monitoring of the wear degree of the gear, and solve the problem that manual detection of the wear marks and wear amount on the tooth surface cannot achieve real-time automatic monitoring and has subjectivity.

[0027] (2) Use an infrared vision sensor to detect the wear degree of the gear based on computer vision, and solve the problem that vibration sensors are easily affected by external interference, resulting in a relatively high monitoring error rate.

[0028] Based on the same inventive concept as a gear wear monitoring method based on sensor data in the foregoing embodiment, as Figure 2 shown, the present application provides a gear wear monitoring system based on sensor data, and the system includes: A first acquisition module, which is used to set a first infrared vision sensor, the shooting direction is parallel to the axis of the target gear, and the grayscale image of the end face of the target gear is obtained in real time to generate a first grayscale image. Among them, the first infrared vision sensor takes the tangent direction of the gear meshing part as the horizontal direction of shooting; A second acquisition module, which is used to set a second infrared vision sensor, the shooting direction is perpendicular to the axis of the target gear, and the grayscale image of the tooth surface of the target gear is obtained in real time to generate a second grayscale image. Among them, the second infrared vision sensor takes the axis direction of the gear as the horizontal direction of shooting; A first determination module, which is used to input the first grayscale image into a pre-trained target detection model, and the target detection model outputs the detection frame of the normal clearance at the gear meshing part in the first grayscale image, and the width of the normal clearance at the gear meshing part is obtained based on the detection frame to generate a first determination parameter; A second determination module, which is used to input the second grayscale image into a pre-trained convolutional neural network model, and the convolutional neural network model outputs the tooth surface wear confidence of the target gear, and uses the tooth surface wear confidence as the second determination parameter; An alarm module, which is used to determine that the target gear is severely worn if both the first determination parameter and the second determination parameter exceed the corresponding determination thresholds, and control to send an alarm signal.

[0029] Through the foregoing detailed description of a gear wear monitoring method based on sensor data in this specification, those skilled in the art can clearly know a gear wear monitoring system based on sensor data in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For related parts, reference may be made to the description in the method section.

[0030] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the above-mentioned gear wear monitoring method based on sensor data is implemented.

[0031] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned gear wear monitoring method based on sensor data is implemented.

[0032] In one embodiment, a computer program product is provided, including a computer program or instruction. When the computer program or instruction is executed by a processor, the above-mentioned gear wear monitoring method based on sensor data is implemented.

[0033] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0034] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A gear wear monitoring method based on sensor data, characterized in that: The method comprises: A first infrared vision sensor is set, and the shooting direction is parallel to the axis of the target gear, and a grayscale image of the end face of the target gear is acquired in real time to generate a first grayscale image, wherein the first infrared vision sensor uses the direction of the tangent line at the meshing position of the gear as the horizontal direction of shooting; A second infrared vision sensor is set up, and the shooting direction is the vertical direction of the target gear axis, and the grayscale image of the tooth surface of the target gear is obtained in real time to generate a second grayscale image, wherein the second infrared vision sensor uses the direction of the gear axis as the horizontal direction of shooting; The first grayscale image is input into a pre-trained target detection model, and the target detection model outputs a detection frame of the normal gap at the gear meshing position in the first grayscale image, and the width of the normal gap at the gear meshing position is obtained based on the detection frame to generate a first determination parameter; Inputting the second grayscale image into a pre-trained convolutional neural network model, the convolutional neural network model outputs the tooth surface wear confidence of the target gear, and using the tooth surface wear confidence as the second judgment parameter; If both the first determination parameter and the second determination parameter exceed the corresponding determination thresholds, the target gear is determined to be severely worn, and the control sends an alarm signal.

2. A gear wear monitoring method based on sensor data as claimed in claim 1, characterized in that: The first grayscale image is input into a pre-trained target detection model, the target detection model outputs a detection frame of the normal gap at the gear meshing position in the first grayscale image, and the width of the normal gap at the gear meshing position is obtained based on the detection frame to generate a first determination parameter, specifically including: Input the first grayscale image into a pre-trained target detection model, wherein the target detection model is a YOLOv4 model, including an input layer, a backbone network, a neck network, and a head network, the input layer is used to perform a normalization operation on the input first grayscale image, the backbone network and the neck network are used to perform feature extraction and feature fusion on the first grayscale image that has completed the normalization operation, and the head network is used to output a target detection result of a target of a normal clearance at a gear meshing position in the first grayscale image; Decoding the target detection result to generate a detection frame of the normal gap at the meshing position of the gears in the first grayscale image, wherein the detection frame is represented by the width, height and center point position of the detection frame; The width of the detection frame is used as the width of the normal gap at the gear meshing position to generate the first determination parameter.

3. A gear wear monitoring method based on sensor data as claimed in claim 1, characterized in that: The step of inputting the second grayscale image into a pre-trained convolutional neural network model, wherein the convolutional neural network model outputs the tooth surface wear confidence of the target gear, and the tooth surface wear confidence is used as the second determination parameter, specifically includes: Inputting the second grayscale image into a pre-trained convolutional neural network model, wherein the convolutional neural network model includes a convolutional layer, a pooling layer, and a fully connected layer, the convolutional layer is used to extract image features through convolution calculation, the pooling layer is used to perform downsampling, and the fully connected layer is used to output a determination result through nonlinear calculation; The judgment result is input into the sigmoid activation function to generate the tooth surface wear confidence of the target gear, and the tooth surface wear confidence is used as the second judgment parameter.

4. A gear wear monitoring method based on sensor data as claimed in claim 1, characterized in that: If both the first determination parameter and the second determination parameter exceed the corresponding determination thresholds, the target gear is determined to be severely worn, and an alarm signal is issued, specifically including: A first determination threshold is preset, and the numerical range of the first determination threshold is greater than 0 and less than 20% of the target gear tooth thickness; a second determination threshold is preset, and the numerical range of the second determination threshold is greater than 0.5 and less than 1; If the first determination parameter exceeds the first determination threshold value, and the second determination parameter exceeds the second determination threshold value, the target gear is determined to be severely worn, and the control sends an alarm signal.

5. A gear wear monitoring system based on sensor data, the system comprising: A first acquisition module, wherein the first acquisition module is used to set a first infrared visual sensor, the shooting direction of which is parallel to the axis of the target gear, to obtain a grayscale image of the end face of the target gear in real time, and to generate a first grayscale image, wherein the first infrared visual sensor uses the direction of the tangent at the meshing position of the gear as the horizontal direction of shooting; A second acquisition module, wherein the second acquisition module is used to set a second infrared vision sensor, the shooting direction of which is the vertical direction of the target gear axis, to obtain a grayscale image of the tooth surface of the target gear in real time, and to generate a second grayscale image, wherein the second infrared vision sensor uses the direction of the gear axis as the horizontal direction of shooting; A first determination module, wherein the first determination module is used to input the first grayscale image into a pre-trained target detection model, the target detection model outputs a detection frame of the normal gap at the gear meshing position in the first grayscale image, and the width of the normal gap at the gear meshing position is obtained based on the detection frame to generate a first determination parameter; a second determination module, wherein the second determination module is used to input the second grayscale image into a pre-trained convolutional neural network model, the convolutional neural network model outputs a tooth surface wear confidence of the target gear, and uses the tooth surface wear confidence as a second determination parameter; An alarm module is used to determine that the target gear is severely worn and control the issuance of an alarm signal if both the first determination parameter and the second determination parameter exceed corresponding determination thresholds.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of a gear wear monitoring method based on sensor data as claimed in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a gear wear monitoring method based on sensor data as claimed in any one of claims 1 to 4 are implemented.

8. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of a gear wear monitoring method based on sensor data described in any one of claims 1 to 4 are implemented.

Images