Equipment state detection method and system for strip steel trimming shear

Abstract

The invention relates to an equipment state detection method for a strip steel trimming shear. The detection method comprises the steps that S1, strip steel edge quality information on the outlet side of the trimming shear is obtained; and S2, based on the obtained strip steel edge quality information, whether the trimming shear has equipment defects or not is judged, and the equipment defects comprise at least one of improper shear blade clearance, shear blade breakage and shear blade abrasion defects. Correspondingly, the invention further provides an equipment state detection system of the strip steel trimming shear. According to the invention, the judgment result is more accurate and reliable by collecting the strip steel edge quality information, reversely deducing the equipment state of the trimming shear and establishing the direct causal association of the quality defect and the equipment defect, so that the detection accuracy of the equipment state of the trimming shear can be ensured; and various defects such as improper shear blade clearance, shear blade breakage and shear blade abrasion of the trimming shear can be reliably detected, so that the detection reliability is relatively high, clear maintenance guidance can be provided for maintenance personnel, and the troubleshooting time is shortened.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of cold rolling production, and particularly relates to a device state detection method and system for a strip steel trimming shear. BACKGROUND

[0002] On a cold rolling production line, a trimming shear is one of the key processing devices, which is used to accurately cut the two sides of a strip steel to obtain a required width and good edge quality. The device state of the trimming shear, especially the state of the shear blade (such as a gap, wear, fracture, etc.), directly determines the shearing quality of the strip steel. Device defects such as an improper shear blade gap, shear blade wear or fracture can cause the strip steel edge to have defects such as excessively large burrs, tearing, and incomplete cutting (with continuous filaments). These defects not only affect the appearance and dimensional accuracy of the product, but can also cause serious problems such as broken strips and scratches in subsequent processes (such as galvanizing and color coating), affecting production stability and yield.

[0003] Currently, the monitoring of the device state of the trimming shear mainly relies on periodic manual inspection, monitoring of device operating parameters (such as shearing force and current), or using a camera to monitor the device state of the trimming shear. Manual inspection has the disadvantages of strong subjectivity, low frequency, and inability to find problems in real time. The operating parameter monitoring and camera monitoring methods are real-time, but have problems such as insufficient detection accuracy and difficulty in comprehensively detecting various defects of the trimming shear. SUMMARY

[0004] The present application relates to a device state detection method and system for a strip steel trimming shear, which can at least solve some of the defects of the prior art.

[0005] The present application relates to a device state detection method for a strip steel trimming shear, which comprises:

[0006] S1, obtaining strip steel edge quality information on the outlet side of the trimming shear;

[0007] S2, judging whether the trimming shear has a device defect based on the obtained strip steel edge quality information, the device defect including at least one of improper shear blade gap, shear blade fracture, and shear blade wear defect.

[0008] As one of the embodiments, the S1 specifically comprises:

[0009] S11, collecting a real-time image of the strip steel edge on the outlet side of the trimming shear;

[0010] S12, inputting the real-time image of the strip steel edge into a shearing quality detection model to identify whether the strip steel edge has a shearing defect and the type of the shearing defect, the shearing defect including at least one of a defect caused by improper shear blade gap, a defect caused by shear blade fracture, and a defect caused by shear blade wear;

[0011] S13, outputting the information of whether the strip edge has a shearing defect and the type of the shearing defect as the strip edge quality information.

[0012] As one of the embodiments, the shearing quality detection model comprises a target detection model, and the target detection model is used to identify whether the strip edge has a shearing defect and the type of the shearing defect;

[0013] The method for establishing the target detection model comprises:

[0014] obtaining a target detection initial model;

[0015] collecting a plurality of strip edge images, marking the strip edge images to obtain first image samples, and obtaining a first strip edge image dataset based on the first image samples; the marking of the strip edge images comprises marking the type and position of the shearing defect of the strip edge images;

[0016] training the target detection initial model by using the first strip edge image dataset to obtain the target detection model.

[0017] As one of the embodiments, the detection method further comprises:

[0018] calculating the burr and cutting ratio of the strip edge by using the shearing quality detection model, and guiding the shearing operation of the edge shearing machine based on the calculation result.

[0019] As one of the embodiments, the shearing quality detection model comprises a target detection model and a semantic segmentation model, the target detection model is used to identify whether the strip edge has a shearing defect and the type of the shearing defect, and the semantic segmentation model is used to calculate the burr and cutting ratio of the strip edge.

[0020] As one of the embodiments, the method for establishing the semantic segmentation model comprises:

[0021] obtaining a semantic segmentation initial model;

[0022] collecting a plurality of strip edge images, marking the strip edge images to obtain second image samples, and obtaining a second strip edge image dataset based on the second image samples; the marking of the strip edge images comprises marking the position of the burr of the strip edge images;

[0023] training the semantic segmentation initial model by using the second strip edge image dataset to obtain the semantic segmentation initial model.

[0024] The present application also relates to a device state detection system of a strip edge shearing machine, and the detection system comprises:

[0025] A strip steel quality defect acquisition module is configured to acquire strip steel edge quality information at the outlet side of the trimming shear and transmit the information to the state diagnosis and decision module.

[0026] A state diagnosis and decision module is configured to diagnose the shear blade equipment state of the trimming shear according to the strip steel edge quality information and output corresponding operation guidance or early warning information.

[0027] As one of the embodiments, the strip steel quality defect acquisition module comprises:

[0028] An image acquisition unit is configured to acquire a strip steel edge image at the outlet side of the trimming shear.

[0029] A target detection module is configured to acquire the strip steel edge image and identify whether the strip steel edge has a shearing defect and the type of the shearing defect, and transmit the information about whether the strip steel edge has the shearing defect and the type of the shearing defect to the state diagnosis and decision module as the strip steel edge quality information; the shearing defect includes at least one of a defect caused by improper shear blade gap, a defect caused by shear blade fracture and a defect caused by shear blade wear.

[0030] As one of the embodiments, the strip steel quality defect acquisition module further comprises:

[0031] A semantic segmentation module is configured to calculate the burr and cutting ratio of the strip steel edge according to the strip steel edge image information.

[0032] As one of the embodiments, the target detection module and the semantic segmentation module are fused into a shearing quality detection module.

[0033] The present application has at least the following beneficial effects:

[0034] In the present application, the equipment state of the trimming shear is deduced by acquiring the strip steel edge quality information, the direct causal relationship between the quality defect and the equipment defect is established, the judgment result is more accurate and reliable, the detection accuracy of the equipment state of the trimming shear can be ensured, and various defects such as improper shear blade gap, shear blade fracture and shear blade wear defect of the trimming shear can be reliably detected, thus having high detection reliability, clear maintenance guidance can be provided for maintenance personnel, and the troubleshooting time is shortened. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0036] Figure 1A flow chart of a strip edge trimmer equipment state detection method provided by the embodiment of the present application is shown in the figure;

[0037] Figure 2 A strip edge quality information identification result map is shown in the figure. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0039] Embodiment one

[0040] As Figure 1 The embodiment of the present application provides a strip edge trimmer equipment state detection method, which comprises the following steps:

[0041] S1, acquiring strip edge quality information at the outlet side of the edge trimmer;

[0042] S2, judging whether the edge trimmer has equipment defects based on the acquired strip edge quality information, wherein the equipment defects include at least one of improper shear blade gap, shear blade fracture and shear blade wear defects.

[0043] In the embodiment, the equipment state of the edge trimmer is deduced by collecting the strip edge quality information, the direct causal relationship between the quality defects and the equipment defects is established, the judgment result is more accurate and reliable, the detection accuracy of the edge trimmer equipment state can be ensured, and the improper shear blade gap, shear blade fracture and shear blade wear defects and other defects of the edge trimmer can be reliably detected, so that the detection reliability is high, the maintenance personnel can be provided with clear maintenance guidance, and the troubleshooting time is shortened.

[0044] When it is judged that the edge trimmer has equipment defects, corresponding early warning is performed, for example, improper shear blade gap is prompted to guide the adjustment of the shear blade gap, or shear blade replacement is prompted, or the shear blade wear state is given early warning.

[0045] Preferably, the strip edge quality information at the outlet side of the edge trimmer is acquired by visual detection, and the S1 specifically comprises the following steps:

[0046] S11, collecting real-time images of the strip edge at the outlet side of the edge trimmer;

[0047] S12, inputting the strip edge real-time image into a shearing quality detection model, identifying whether the strip edge has a shearing defect and a type of the shearing defect, the shearing defect including at least one of a defect caused by improper shearing blade gap, a defect caused by shearing blade fracture and a defect caused by shearing blade wear;

[0048] S13, outputting the information of whether the strip edge has a shearing defect and a type of the shearing defect as strip edge quality information.

[0049] In S11, preferably, an industrial line array or area array camera is adopted, and a suitable lighting system (such as LED backlight or sidelight) is matched, and the above-mentioned industrial line array or area array camera and other image acquisition devices are installed on a stable section at an exit of the edge shearing machine to obtain clear and stable strip edge images.

[0050] For example, Figure 2 The defect caused by improper shearing blade gap refers to a strip edge defect caused by improper shearing blade gap of the edge shearing machine, such as large burr, double-edge burr and the like; the defect caused by shearing blade fracture refers to a strip edge defect caused by shearing blade fracture of the edge shearing machine, such as periodic notch, local tearing and the like; and the defect caused by shearing blade wear refers to a strip edge defect caused by shearing blade wear of the edge shearing machine, such as uniform but continuously increasing burr, rough cut surface and the like.

[0051] In S13, when the strip edge has no shearing defect, the strip edge quality information is that there is no shearing defect; and when the strip edge has a shearing defect, generally, the type of the shearing defect is directly outputted.

[0052] In one embodiment, the shearing quality detection model includes a target detection model, and the target detection model is used to identify whether the strip edge has a shearing defect and a type of the shearing defect.

[0053] The method for establishing the target detection model includes:

[0054] An initial target detection model is obtained, such as Faster-RCNN, Yolo, Cascade R-CNNs and the like;

[0055] A plurality of strip edge images (which can be historical images or real-time images containing various shearing defects) are collected, the strip edge images are labeled to obtain first image samples, and a first strip edge image data set is obtained based on the first image samples; the labeling of the strip edge images includes labeling a type of shearing defect and a position of the shearing defect in the strip edge images;

[0056] The initial target detection model is trained by using the first strip edge image data set to obtain the target detection model.

[0057] Optionally, sample labeling is performed using EasyDL software or other software, and based on the labeled first image samples, an initial target detection model is trained using PaddlePaddle or other deep learning environment.

[0058] In one of the embodiments, the detection method further comprises:

[0059] The burr-to-cut ratio of the strip edge is calculated by the shearing quality detection model, and based on the calculation result, the edge shearing machine is guided to optimize the shearing operation, such as automatically or prompting the operator to adjust the blade gap, replace the blade, etc.

[0060] Further, to realize the above quantitative analysis function, the shearing quality detection model can adopt a fusion model, specifically including a target detection model and a semantic segmentation model. Among them, based on the target detection model, it is identified whether there is a shearing defect in the strip edge and the type of the shearing defect; based on the semantic segmentation model, the strip edge image is classified at the pixel level, and the "burr area" and "normal cut area" are accurately segmented, and the burr-to-cut ratio is calculated.

[0061] By introducing the semantic segmentation model to calculate the burr-to-cut ratio, quantitative evaluation of the shearing quality is realized, which provides data support for fine and intelligent adjustment of process parameters, and helps to continuously improve product quality.

[0062] Preferably, the method for establishing the semantic segmentation model comprises:

[0063] An initial semantic segmentation model is obtained, such as DeepLab, ICNET, UNET, etc.

[0064] A plurality of strip edge images are collected, the strip edge images are labeled to obtain second image samples, and based on the second image samples, a second strip edge image dataset is obtained; the labeling of the strip edge images includes labeling the burr positions of the strip edge images;

[0065] The initial semantic segmentation model is trained using the second strip edge image dataset to obtain the semantic segmentation model.

[0066] Optionally, sample labeling is performed using EasyDL software or other software, and based on the labeled second image samples, the initial semantic segmentation model is trained using PaddlePaddle or other deep learning environment.

[0067] Embodiment two

[0068] The embodiment of the application provides a device state detection system of a strip edge shearing machine, which comprises:

[0069] The strip steel quality defect acquisition module is configured to acquire the strip steel edge quality information at the outlet side of the trimming shear and transmit the information to the state diagnosis and decision module.

[0070] The state diagnosis and decision module is configured to diagnose the shear blade equipment state of the trimming shear according to the strip steel edge quality information and output corresponding operation guidance or early warning information.

[0071] The detection system provided in the embodiment can be used to implement the detection method provided in the above embodiment one, or the related technical means in the detection method provided in the above embodiment one is applicable to the embodiment. For example, the vision detection method is used to acquire the strip steel edge quality information at the outlet side of the trimming shear, and the strip steel quality defect acquisition module accordingly includes:

[0072] The image acquisition unit is configured to acquire the strip steel edge image at the outlet side of the trimming shear.

[0073] The target detection module is configured to acquire the strip steel edge image, identify whether the strip steel edge has a shearing defect and the type of the shearing defect, and transmit the information about whether the strip steel edge has the shearing defect and the type of the shearing defect to the state diagnosis and decision module as the strip steel edge quality information. The shearing defect includes at least one of a defect caused by improper shear blade gap, a defect caused by shear blade fracture, and a defect caused by shear blade wear.

[0074] Further, the strip steel quality defect acquisition module further includes:

[0075] The semantic segmentation module is configured to calculate the burr and cutting ratio of the strip steel edge according to the strip steel edge image information.

[0076] Preferably, the target detection module and the semantic segmentation module are fused into a shearing quality detection module.

[0077] The image acquisition unit can adopt an industrial linear array or area array camera, cooperate with a proper lighting system (such as LED backlight or sidelight), and install the image acquisition device such as the industrial linear array or area array camera at the stable section of the outlet of the trimming shear to acquire a clear and stable strip steel edge image. The industrial linear array or area array camera can be carried by a motorized translation stage or other equipment to approach or move away from the strip steel edge to focus, so as to ensure clear imaging.

[0078] The above only describes the preferred embodiments of the present application and should not be used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An apparatus state detection method for a strip steel trimming shear, characterized by, The detection method comprises: S1, obtaining strip edge quality information at the outlet side of the trimming shear; S2, judging whether a device defect occurs in the trimming shear based on the obtained strip edge quality information, the device defect comprising at least one of improper shear blade gap, shear blade fracture and shear blade wear defect.

2. The apparatus state detection method of a strip steel slitting shear according to claim 1, characterized by, The S1 specifically comprises: S11, collecting a real-time image of the strip edge at the outlet side of the trimming shear; S12, inputting the real-time image of the strip edge into a shear quality detection model to identify whether the strip edge has a shear defect and the type of the shear defect, the shear defect comprising at least one of a defect caused by improper shear blade gap, a defect caused by shear blade fracture and a defect caused by shear blade wear; S13, outputting the information of whether the strip edge has the shear defect and the type of the shear defect as the strip edge quality information.

3. The apparatus state detection method of a strip steel slitting shear according to claim 2, characterized by, The shear quality detection model comprises a target detection model, and the target detection model is used to identify whether the strip edge has the shear defect and the type of the shear defect; The method for establishing the target detection model comprises: obtaining a target detection initial model; collecting a plurality of strip edge images, marking the strip edge images to obtain first image samples, and obtaining a first strip edge image dataset based on the first image samples, wherein marking the strip edge images comprises marking the type and position of the shear defect of the strip edge image; training the target detection initial model by using the first strip edge image dataset to obtain the target detection model.

4. The apparatus state detecting method of a strip steel slitting shear according to claim 2, characterized by, The detection method further comprises: calculating the burr and cutting-off ratio of the strip edge by using the shear quality detection model, and guiding the trimming shear to optimize the shearing operation based on the calculation result.

5. The apparatus state detecting method of a strip steel slitting shear according to claim 4, characterized by, The shear quality detection model comprises a target detection model and a semantic segmentation model, the target detection model is used to identify whether the strip edge has the shear defect and the type of the shear defect, and the semantic segmentation model is used to calculate the burr and cutting-off ratio of the strip edge.

6. The apparatus state detecting method of a strip steel slitting shear according to claim 5, characterized by, The method for establishing the semantic segmentation model comprises: obtaining a semantic segmentation initial model; collecting a plurality of strip edge images, marking the strip edge images to obtain second image samples, and obtaining a second strip edge image dataset based on the second image samples, wherein marking the strip edge images comprises marking the position of the burr of the strip edge image; training the semantic segmentation initial model by using the second strip edge image dataset to obtain the semantic segmentation initial model.

7. A device status monitoring system for a strip steel edge cutting shear, characterized in that, The detection system comprises: a strip quality defect obtaining module configured to obtain strip edge quality information at the outlet side of the trimming shear and transmit the strip edge quality information to a state diagnosis and decision module; the state diagnosis and decision module configured to diagnose the shear blade device state of the trimming shear according to the strip edge quality information and output corresponding operation guidance or early warning information.

8. The apparatus condition detection system for a strip steel slitting shear as set forth in claim 7, wherein, The strip quality defect obtaining module comprises: an image collection unit configured to collect a strip edge image at the outlet side of the trimming shear; The target detection module is configured to acquire the strip steel edge image and identify whether the strip steel edge has a shearing defect and a type of the shearing defect, and transmit information about whether the strip steel edge has the shearing defect and the type of the shearing defect as strip steel edge quality information to the state diagnosis and decision module; the shearing defect includes at least one of a defect caused by improper shearing blade gap, a defect caused by shearing blade fracture and a defect caused by shearing blade wear.

9. The apparatus condition detection system for a strip steel slitting shear as defined in claim 8 wherein, The strip steel quality defect acquisition module further includes: The semantic segmentation module is configured to calculate burr and cutting-off ratio of the strip steel edge according to the strip steel edge image information.

10. The apparatus condition detection system for a strip steel slitting shear of claim 9, wherein, The target detection module and the semantic segmentation module are fused into a shearing quality detection module.

Images