A control method, electronic device and system of a sheet metal punching equipment of an energy storage cabinet

Abstract

The application discloses a control method, an electronic device and a system of a sheet metal punching equipment of an energy storage cabinet. The method comprises the following steps: S3, acquiring first thermal image information of a hole peripheral area at a first temperature interval by using a thermal imaging image acquisition device; S4, acquiring a first quality score; S5, when the first quality score does not meet the processing requirements, moving the energy storage cabinet to a secondary detection station by using a conveying mechanism, and adjusting the hole peripheral area to a second temperature interval by using a temperature adjusting device during the moving process; S6, acquiring second thermal image information of the hole peripheral area of the energy storage cabinet to be punched at the secondary detection station, and acquiring a comparison result including a second quality score; and S7, controlling an automatic feeding mechanism to supply a next energy storage cabinet to be punched, and controlling a punching execution mechanism to process the next energy storage cabinet to be punched according to the comparison result. The application realizes accurate determination of the punching quality and closed-loop optimization of the process parameters through a double-temperature-interval comparison and detection mechanism.

Description

Technical Field

[0001] This invention belongs to the technical field of punching equipment, specifically relating to a control method, electronic equipment and system for sheet metal punching equipment for energy storage cabinets. Background Technology

[0002] Sheet metal processing is a key process in manufacturing, widely used in automotive, aerospace, construction, and electrical appliance industries. Among these processes, punching is a crucial step, as the quality of the punching directly impacts product performance and lifespan. Related technologies have enabled sheet metal punching equipment to achieve a high degree of automation, featuring automatic feeding, punching, and unloading functions.

[0003] Referring to patent document CN116890059A, a sheet metal punching device for power distribution cabinets is disclosed. This device, designed with automatic feeding, punching, and unloading, utilizes components such as a motor drive, cam, slide bar, drill bit, and blower to solve the problems of low punching accuracy and efficiency in existing technologies, achieving efficient punching of power distribution cabinet panels and waste removal. However, some problems remain to be solved. The main issue is the punching control method for sheet metal punching equipment in related technologies. This method suffers from high subjectivity, low efficiency, and a high misjudgment rate in quality control, failing to detect issues such as hole deviation in a timely manner, leading to difficulties or even inability to assemble in subsequent processes. Furthermore, the processing of each workpiece is typically treated as an independent event, with manufacturers only being contacted for solutions when problems accumulate to a severe level.

[0004] Therefore, there is an urgent need to provide a control method, electronic equipment, and system for sheet metal punching equipment for energy storage cabinets. Summary of the Invention

[0005] The main objective of this invention is to provide a control method, electronic device and system for sheet metal punching equipment for energy storage cabinets, so as to overcome the shortcomings of the prior art.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0007] An embodiment of the present invention provides a control method for a sheet metal punching device for energy storage cabinets. The sheet metal punching device for energy storage cabinets includes a thermal imaging image acquisition device, an automatic feeding mechanism, a punching execution mechanism, and a conveying mechanism. The method includes the following steps:

[0008] S3, after the punching process of the energy storage cabinet to be punched, the thermal imaging image acquisition device is used to obtain the first thermal image information of the area around the hole in the first temperature range.

[0009] S4, obtain the first quality score based on the actual quality features corresponding to the first image information and the preset quality features;

[0010] S5, when the first quality score does not meet the processing requirements, the conveyor mechanism is used to move it to the secondary inspection station, and the temperature adjustment device is used to adjust the hole periphery area to the second temperature range during the movement.

[0011] S6, acquire the second thermal image information of the area around the hole of the energy storage cabinet to be punched at the secondary inspection station, and obtain a comparison result including the second quality score based on the first thermal image information and the second thermal image information;

[0012] S7, control the automatic feeding mechanism to supply the next energy storage cabinet to be punched, and control the punching execution mechanism to process the next energy storage cabinet to be punched according to the comparison result;

[0013] The first thermal image information and the second thermal image information are used to indicate the temperature distribution of the area around the hole in the energy storage cabinet to be punched in the first temperature range and the second temperature range, respectively.

[0014] In a preferred embodiment, the method further includes:

[0015] S1, before the current energy storage cabinet to be punched is punched, obtain its feedforward information;

[0016] S2, construct the current process timing state of the energy storage cabinet to be punched based on the historical processing sequence; input the feedforward information and the process timing state into the pre-trained timing prediction model to predict the predicted quality characteristics of the current energy storage cabinet to be punched under preset process parameters, and use them as preset quality characteristics.

[0017] In a preferred embodiment, each of the energy storage cabinets to be punched has at least one periphery region according to the punching position, and each periphery region includes multiple punches; when there are multiple periphery regions, the preset quality characteristics include the predicted quality characteristics of each periphery region.

[0018] Step S3 includes: using the thermal imaging image acquisition device to acquire infrared feature images of multiple hole periphery areas of the energy storage cabinet to be punched in the first temperature range, and associating them with their respective tag data as the first thermal image information;

[0019] Step S4 includes: extracting actual quality features from the infrared feature image of each aperture periphery region; obtaining the similarity between the actual quality features and the predicted quality features of the corresponding aperture periphery region; and obtaining a first quality score based on the similarity between each aperture periphery region.

[0020] In a preferred embodiment, obtaining the first quality score based on the similarity of each periapical region includes:

[0021] Obtain the correspondence between the similarity of the periapical region and the quality score, wherein the correspondence includes the weight coefficient of each periapical region and the similarity-score mapping function;

[0022] Based on the correspondence, the similarity of each periapical region is transformed by the similarity-score mapping function, and then weighted and fused according to the weight coefficient to obtain the first quality score;

[0023] When the similarity of each hole periphery area meets its corresponding similarity threshold, the first quality score is considered to meet the processing requirements, and the conveying mechanism is controlled to move the current energy storage cabinet to be punched to the good product unloading station; otherwise, step S5 is executed.

[0024] In a preferred embodiment, step S6 includes:

[0025] The area around the hole with a similarity lower than its corresponding similarity threshold is taken as the middle area. The infrared feature image of the middle area of ​​the energy storage cabinet to be punched is obtained at the secondary inspection station, and then associated with its respective label data as the second thermal image information.

[0026] The infrared feature images indicating the same label data in the first thermal image information and the second thermal image information are compared to obtain a comparison result including a second quality score.

[0027] In a preferred embodiment, step S7 includes:

[0028] The quality level of the middle area of ​​the current energy storage cabinet to be punched is determined based on the second quality score.

[0029] The preset process parameters corresponding to the intermediate region are selectively adjusted according to the quality level, and the punching actuator is controlled to process the next energy storage cabinet to be punched based on the adjusted process parameters.

[0030] When the quality grade is unqualified, at least one of the punching pressure, punching speed and holding time in the preset process parameters is adjusted according to the deviation between the actual quality characteristics and the predicted quality characteristics.

[0031] An embodiment of the present invention also provides an electronic device, including one or more processors and a memory; one or more programs are stored in the memory and configured to be executed by the one or more processors according to any of the above methods.

[0032] An embodiment of the present invention also provides a sheet metal punching system for energy storage cabinets, the sheet metal punching system for energy storage cabinets including sheet metal punching equipment and electronic equipment.

[0033] In a preferred embodiment, the energy storage cabinet sheet metal punching equipment includes a frame, an automatic feeding mechanism, a punching execution mechanism, and a conveying mechanism. The punching execution mechanism includes a servo drive device. The conveying mechanism is equipped with an online quality inspection station for setting up a thermal imaging image acquisition device to obtain thermal images of the workpiece hole periphery area.

[0034] An embodiment of the present invention also provides a computer-readable storage medium storing a computer program that, when executed by at least one processor, implements the steps of any of the methods described above.

[0035] An embodiment of the present invention also provides a computer program product, the computer program product including a computer program, which, when executed by at least one processor, implements the steps of the method described above.

[0036] Compared with existing technologies, the advantages of this invention are as follows: By constructing a dual-temperature-range comparison detection mechanism, it effectively solves the problems of missed and false detections in online detection in existing technologies. The temperature adjustment device actively intervenes thermally in the peripheral area of ​​the hole, acquiring dual thermal image information of the same area under different temperature conditions. Utilizing the difference in thermal response characteristics between the defective and normal areas, interference is effectively suppressed through comparison of the dual image information, reducing the false judgment and missed detection rates. Hole deviation defects may only manifest as a weak temperature gradient anomaly in the instantaneous temperature field after punching, making them difficult to reliably identify using traditional detection methods. This technical solution, through an active temperature adjustment process, significantly amplifies the difference in the defective area in the second temperature range. Combined with dual-image comparison analysis, it achieves effective detection of slight hole deviation defects, preventing defective workpieces from flowing into subsequent processes. Simultaneously, the temperature adjustment device is located on the conveyor line between the punching station and the secondary detection station, completing the temperature adjustment operation during workpiece movement without requiring additional detection dwell time. Workpieces that meet the processing requirements in the first quality score are directly released to the next process; the secondary detection process is triggered only for suspected defective workpieces, achieving optimized allocation of detection resources. By using the comparison results of the current workpiece to control the processing parameters of the next energy storage cabinet to be punched, a closed-loop control system of detection, analysis, feedback, and adjustment is formed. This allows the process parameters to be adaptively adjusted according to the actual processing quality, improving the stability of the control process. The technical solution itself relies on non-contact thermal imaging detection technology, which eliminates the need for physical contact with the workpiece, avoiding the risk of damage to the workpiece surface caused by traditional contact detection. This makes it suitable for applications where the surface quality requirements of sheet metal parts for energy storage cabinets are high. Furthermore, thermal imaging detection has a fast response speed, which can adapt to the cycle time requirements of automated production lines. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a flowchart illustrating a control method for a sheet metal punching device for an energy storage cabinet according to one embodiment of this application.

[0039] Figure 2 This is a flowchart illustrating another control method for a sheet metal punching device for an energy storage cabinet according to one embodiment of this application. Detailed Implementation

[0040] The invention will be more fully understood through the following detailed description, which should be read in conjunction with the accompanying drawings. Detailed embodiments of the invention are disclosed herein; however, it should be understood that the disclosed embodiments are merely exemplary of the invention, which may be embodied in various forms. Therefore, the specific functional details disclosed herein should not be construed as limiting, but rather as the basis for the claims and as intended to teach those skilled in the art to employ the representative basis of the invention in different ways in any suitable detailed embodiment.

[0041] In related technologies, thermal imaging inspection technology can be applied to improve the accuracy of punching control in sheet metal parts and reduce labor costs. Specifically, during the sheet metal punching process, heat is generated due to the plastic deformation and friction of the material. Thermal imaging inspection uses an infrared thermal imager to capture the temperature distribution changes in the punching area and analyzes the temperature field characteristics to determine the punching quality. When punching defects occur, material defects are formed, creating abnormal hot spots or cold spots.

[0042] Referring to patent CN104923653A, a sheet metal punching device that uses thermal imaging detection feedback adjustment automatically adjusts the punching pressure and sets up a proximity switch, solving the problems of low punching quality and safety in existing punching machines, and achieving high-quality processing and a safe punching process. However, it mainly relies on real-time online detection during the punching process (or capturing the highest temperature at the moment of punching or scanning for residual heat within 0.5-1 seconds after punching). In this case, detecting the proportion of cracks and the amount of deformation in the finished workpiece using a thermal imaging detector, and automatically adjusting the punching pressure of the punching head by the host machine, can only reduce the occurrence of burrs on the hole edge. Online detection has high requirements for the processing of the acquired information. If it is used as a quality control step in the punching control process, it is easy to miss or misdetect.

[0043] This application provides a control method, electronic equipment, and system for sheet metal punching equipment for energy storage cabinets. It achieves accurate judgment of punching quality and closed-loop optimization of process parameters through a dual-temperature range comparison detection mechanism, solving the problems of strong subjectivity, low efficiency, and high misjudgment rate in the existing technology.

[0044] Example 1

[0045] See Figure 1 This embodiment provides a control method for a sheet metal punching device for energy storage cabinets. The sheet metal punching device for energy storage cabinets includes a thermal imaging image acquisition device, an automatic feeding mechanism, a punching execution mechanism, and a conveying mechanism. The method includes the following steps:

[0046] S3, after the punching process of the energy storage cabinet to be punched, the thermal imaging image acquisition device is used to obtain the first thermal image information of the area around the hole in the first temperature range.

[0047] S4, obtain the first quality score based on the actual quality features corresponding to the first image information and the preset quality features;

[0048] S5, when the first quality score does not meet the processing requirements, the conveyor mechanism is used to move it to the secondary inspection station, and the temperature adjustment device is used to adjust the hole periphery area to the second temperature range during the movement.

[0049] S6, acquire the second thermal image information of the area around the hole of the energy storage cabinet to be punched at the secondary inspection station, and acquire a comparison result including the second quality score based on the first thermal image information and the second thermal image information; wherein, the first quality score and the second quality score can be numerical scores, or they can be character sequences such as A, B, Jia, Bing, Ding to represent different quality ranges.

[0050] S7, control the automatic feeding mechanism to supply the next energy storage cabinet to be punched, and control the punching execution mechanism to process the next energy storage cabinet to be punched according to the comparison result;

[0051] The first thermal image information and the second thermal image information are used to indicate the temperature distribution of the area around the hole in the energy storage cabinet to be punched in the first temperature range and the second temperature range, respectively.

[0052] If the first temperature range is lower than the second temperature range, the existing automated punching line can be upgraded by replacing the original cam or ordinary motor drive with a high-precision servo electric cylinder as the punching power source. Its controllable parameters include pressing speed, holding pressure, holding time, and return speed. A temperature control device (e.g., an infrared heating device) is installed on the conveyor line between the punching station and the secondary inspection station to uniformly heat the area around the hole before secondary inspection. An industrial control computer (host, electronic equipment) is installed on-site, connected to all sensors, servo drives, thermal imaging image acquisition devices, etc., via a data bus, and runs industrial control software.

[0053] After the punching process is completed, a thermal image of the area around the hole is acquired using a thermal imaging device. This thermal image contains rich quality information because the mechanical stress, frictional heat, and internal material defects generated during punching cause the area around the hole to exhibit specific temperature distribution characteristics. The actual quality characteristics of the target workpiece are extracted from the thermal image. These characteristics include, but are not limited to, the temperature gradient at the hole edge, the area of ​​the heat-affected zone, and the distribution of abnormal hot spots, which can reflect quality defects such as burrs at the hole edge, microcracks, and workpiece deformation.

[0054] The extracted actual quality features are compared with preset quality features, such as predicted quality features, so that this comparison process enables a quantitative assessment of the prediction accuracy.

[0055] The technical solution provided in this embodiment is based on the inherent correlation between the temperature distribution around the hole and the processing quality during the punching process. It constructs a dual-temperature range comparison detection mechanism to realize the judgment of punching quality and the feedback optimization of subsequent stamping parameters.

[0056] Specifically, during the punching process, the sheet metal material of the energy storage cabinet undergoes plastic shear deformation, and friction exists between the punch and the sheet metal. This mechanical work is converted into heat energy, forming a specific temperature distribution in the area around the hole. This temperature distribution characteristic has a clear physical correspondence with the punching quality: when the punching quality is qualified, the temperature gradient at the hole edge is uniform, and the heat-affected zone has a regular shape; when defects such as hole deviation, burrs, and microcracks occur, due to uneven material thickness, stress concentration, or changes in the heat conduction path, the area around the hole will exhibit an abnormal temperature distribution, manifested as local hot spots, cold spots, or temperature gradient distortion.

[0057] Utilizing the aforementioned physical mechanism, immediately after punching, a thermal imaging image acquisition device acquires the initial thermal image information of the area around the hole (within the first temperature range). This initial thermal image information reflects the original state of the transient thermal field after punching, containing the initial thermal characteristics of processing stress, frictional heat, and potential defects. Based on this initial thermal image information, actual quality characteristics are extracted and compared with preset quality characteristics to obtain a first quality score, thereby achieving a rapid quantitative assessment of the current processing quality. The first temperature range can be considered as the natural temperature range of the production environment, i.e., the temperature range without direct interference from temperature control devices.

[0058] When the initial quality score fails to meet processing requirements, the workpiece is considered to have a quality defect risk and requires further confirmation. At this point, the workpiece is moved to a secondary inspection station using a conveyor mechanism. During this movement, a temperature adjustment device actively intervenes in the peripheral area of ​​the hole, adjusting it to a second temperature range. This temperature adjustment process is based on the principle of thermal diffusion difference. Normal areas and defective areas (such as cracks) within the peripheral area have different thermal capacities and thermal resistance characteristics. During active temperature adjustment, normal and defective areas will exhibit differentiated thermal response rates, thereby amplifying the identifiability of the thermal characteristics of potential defects.

[0059] At the secondary inspection station, a second thermal image of the peripheral area at the second temperature range is acquired again using a thermal imaging image acquisition device. The first and second thermal images are then compared and analyzed to obtain a comparison result including a second quality score. This comparison result, based on the thermal image differences of the same workpiece and the same peripheral area at different temperature conditions, effectively eliminates interference from material batch differences and significantly improves the accuracy of defect determination.

[0060] In this case, the automatic feeding mechanism is controlled to supply the next energy storage cabinet to be punched, and the punching execution mechanism is controlled to process the next energy storage cabinet to be punched based on the above comparison results, so as to realize the feedback adjustment of process parameters.

[0061] Therefore, compared to traditional sheet metal punching equipment that mostly uses manual sampling or offline inspection, which suffers from problems such as inspection lag and low efficiency, the technical solution provided in this embodiment relies on non-contact thermal imaging inspection technology. This eliminates the need for physical contact with the workpiece, avoiding the risk of damage to the workpiece surface caused by traditional contact inspections. It is suitable for applications requiring high surface quality of sheet metal parts for energy storage cabinets. Furthermore, thermal imaging inspection has a fast response speed, which can adapt to the cycle time requirements of automated production lines.

[0062] Compared to online inspection, which relies solely on a single-frame thermal image taken immediately after punching for quality assessment, making it difficult to distinguish between genuine and false defects due to uneven emissivity of the material surface, environmental thermal radiation interference, and noise in the inspection system, this technical solution effectively addresses the problems of missed and false detections in existing online inspection technologies by constructing a dual-temperature-range comparison detection mechanism. The temperature adjustment device actively intervenes thermally in the perforated area, acquiring dual thermal image information of the same area under different temperature conditions. By utilizing the difference in thermal response characteristics between the defective and normal areas, the comparison of the dual image information effectively suppresses interference, reducing the false positive and missed detection rates.

[0063] Hole misalignment defects may only manifest as a weak temperature gradient anomaly in the instantaneous temperature field after punching, making them difficult to reliably identify using traditional detection methods. This technical solution, through an active temperature adjustment process, significantly amplifies the difference in the defect area within a second temperature range. Combined with dual-image comparative analysis, it achieves effective detection of minor hole misalignment defects, preventing defective workpieces from flowing into subsequent processes. Simultaneously, the temperature adjustment device is located on the conveyor line between the punching station and the secondary inspection station, completing the temperature adjustment operation during workpiece movement without requiring additional inspection dwell time. Workpieces that meet the processing requirements in the first quality score are directly released to the next process; the secondary inspection process is triggered only for suspected defective workpieces, achieving optimized allocation of inspection resources.

[0064] By using the comparison results of the current workpiece to control the processing parameters of the next energy storage cabinet to be punched, a closed-loop control of detection, analysis, feedback, and adjustment is formed, which enables the process parameters to be adaptively adjusted according to the actual processing quality, thereby improving the stability of the control process.

[0065] See Figure 2 In one embodiment, the method further includes:

[0066] S1, before the current energy storage cabinet to be punched is punched, obtain its feedforward information;

[0067] S2, construct the current process timing state of the energy storage cabinet to be punched based on the historical processing sequence; input the feedforward information and the process timing state into the pre-trained timing prediction model to predict the predicted quality characteristics of the current energy storage cabinet to be punched under preset process parameters, and use them as preset quality characteristics.

[0068] In other words, a feedforward prediction mechanism is introduced before the punching process. By constructing the process time sequence state and using the time sequence prediction model, the processing quality of the energy storage cabinet to be punched can be pre-assessed, thereby providing a dynamic benchmark for subsequent quality inspection.

[0069] In continuous sheet metal punching production, the processing quality of each workpiece (energy storage cabinet to be punched) is not independent but is influenced by the gradual evolution of the process system's state. Factors such as die wear, punch dulling, material batch fluctuations, and ambient temperature changes all lead to a temporal correlation in processing quality, meaning that the historical processing sequence contains information about the current state of the process system. Based on this, this embodiment first acquires the feedforward information of the energy storage cabinet to be punched before punching it. This feedforward information includes inherent attribute information such as the current workpiece's material specifications, sheet thickness, and punching position layout, as well as the setting parameters of the current punching actuator, which reflect the initial conditions of the current processing task.

[0070] The current process timeline state of the energy storage cabinet to be punched is constructed based on historical processing sequences. This process timeline state can be formed by time-series encoding of quality inspection data, process parameters, and equipment status information of recently processed workpieces, representing the dynamic evolution trend of the process system at the current moment. The construction of historical processing sequences avoids the traditional approach of treating each workpiece processing as an independent event, establishing the temporal correlation of the processing process, so that the processing quality prediction of the current workpiece can inherit and continue the systematic patterns in historical processing.

[0071] The aforementioned feedforward information, along with the process timing status, is input into a pre-trained timing prediction model. This model, built on a machine learning algorithm, establishes a (non-linear) mapping between the evolution of the process system's state and the processing quality output. The timing prediction model integrates the individual characteristics of the current workpiece with the historical evolution trend of the process system to predict the predicted quality characteristics of the energy storage cabinet to be punched under preset process parameters. These predicted quality characteristics can be represented in the form of thermal image features, including the expected temperature distribution pattern around the hole, the range of the heat-affected zone, and temperature gradient characteristics, serving as preset quality characteristics for subsequent actual quality inspection.

[0072] In practical applications, the aforementioned feedforward prediction mechanism enables the dynamic generation of quality benchmarks. This preset quality characteristic can be a non-fixed threshold standard, adaptively adjusted based on the real-time status of the process system and the current characteristics of the workpiece, thereby improving the targeting and accuracy of quality control.

[0073] Therefore, traditional testing methods often employ fixed quality judgment thresholds, which are prone to subjectivity and rigid pre-set standards, failing to adapt to dynamic changes in the process system state. This leads to increased missed detection rates in the early stages of mold wear and increased false detection rates in the severe wear stages. This embodiment introduces a feedforward prediction mechanism, constructing a process time-series state based on historical processing sequences. A time-series prediction model dynamically generates pre-set quality characteristics that match the current process state, enabling the quality judgment criteria to adaptively adjust as the process system evolves, thus improving the robustness and adaptability of the testing process.

[0074] The established quality data correlation for the processing sequence avoids the shortcomings of treating each workpiece's processing as an independent event and lacking process correlation in related technologies. By constructing the process temporal state, a mathematical correlation can be established between the current workpiece's processing quality prediction and the quality performance of historical processing sequences, forming a temporal transmission chain of quality data. This correlation mechanism enables the gradual degradation of the process system to be detected early, avoiding the passive situation of waiting for faults to accumulate to a severe level before taking action, and achieving proactive prevention in quality control.

[0075] In practical applications, the timing prediction model is a recurrent neural network model built based on a long short-term memory network or a gated recurrent unit. Feedforward information and the process timing state are input together into the pre-trained timing prediction model. This model typically employs a neural network structure with memory capabilities, enabling it to capture temporal dependencies in historical sequences. The model output is the predicted quality characteristics of the current energy storage cabinet (workpiece to be punched) under preset process parameters; that is, a quantitative prediction of the quality that will be obtained if processed according to the current settings.

[0076] In one embodiment, each of the energy storage cabinets to be punched has at least one periphery region according to the punching position, and each periphery region includes multiple punches; when there are multiple periphery regions, the preset quality characteristics include the predicted quality characteristics of each periphery region.

[0077] Step S3 includes: using the thermal imaging image acquisition device to acquire infrared feature images of multiple hole periphery areas of the energy storage cabinet to be punched in the first temperature range, and associating them with their respective tag data as the first thermal image information;

[0078] Step S4 includes: extracting actual quality features from the infrared feature image of each aperture periphery region; obtaining the similarity between the actual quality features and the predicted quality features of the corresponding aperture periphery region; and obtaining a first quality score based on the similarity between each aperture periphery region.

[0079] The control schemes for punching holes in energy storage cabinets in related technologies mainly refer to conventional sheet metal processing, without taking into account the diverse distribution of punching positions in the sheet metal parts of energy storage cabinets. However, the technical solution provided in this embodiment adopts a zone detection and independent comparison strategy to achieve a refined evaluation of the punching quality in multiple areas.

[0080] Specifically, sheet metal parts of energy storage cabinets typically have a large surface area, and the punching of holes in different functional areas varies significantly in terms of location, hole diameter, material thickness, and heat dissipation requirements. For example, punching areas in different locations such as the front door panel, side heat dissipation area, and rear wiring area of ​​the energy storage cabinet include punching areas with the same or similar process attributes. Based on the above structural characteristics, this embodiment divides the energy storage cabinet to be punched into at least one peripheral area according to the punching location, and each peripheral area contains multiple punches with the same or similar process attributes. When there are multiple peripheral areas, the preset quality features correspondingly include the predicted quality features of each peripheral area, that is, a quality prediction model matching its location characteristics is independently established for each area.

[0081] A thermal imaging image acquisition device was used to scan multiple areas around the holes in the punched energy storage cabinet, acquiring infrared characteristic images of each area within a first temperature range. These infrared characteristic images reflect the local temperature field distribution after punching in a specific location. To establish a correspondence between spatial location and thermal image features, each infrared characteristic image was associated with its corresponding tag data. This tag data, for example, identifies the spatial location, area type, and process attributes of the area around the hole. The associated image data serves as the first thermal image information. Through this tag data association mechanism, a precise mapping between thermal image features and physical location was achieved, ensuring the location traceability for subsequent quality assessments.

[0082] Feature extraction is performed independently on the infrared feature image of each perforation region to obtain actual quality features characterizing the actual processing state of that region. In specific applications, this feature extraction process analyzes the local features of the thermal image of a specific perforation region, including parameters such as the temperature gradient distribution, the geometry of the heat-affected zone, and the density of local hot spots within that region. Subsequently, the extracted actual quality features are compared with the predicted quality features of the same perforation region determined through label data, and the similarity between the two is calculated. In this case, the predicted quality features also include parameters such as the predicted temperature gradient distribution, the geometry of the heat-affected zone, and the density of local hot spots corresponding to the actual quality features. The obtained similarity measure quantifies the degree of deviation between the actual processing state and the expected processing state of that region. A higher similarity indicates that the processing quality of that region is closer to the ideal state, while a lower similarity indicates that there may be processing defects in that region.

[0083] After obtaining the similarity of each hole perimeter region, a first quality score is obtained by comprehensively calculating the similarity of each hole perimeter region. The comprehensive acquisition process may consider, for example, the process importance weights and quality tolerance differences of different hole perimeter regions, forming a quantitative evaluation of the punching quality of the entire energy storage cabinet through the fusion calculation of the similarity of multiple regions; or it may be a simple summation calculation to reduce the amount of data processing. The first quality score reflects the overall processing quality level.

[0084] Therefore, to avoid treating the energy storage cabinet as a whole for thermal image acquisition and quality assessment, it is difficult to distinguish the temperature field differences in different locations, leading to an inability to accurately identify the specific area where defects occur. The technical solution provided in this embodiment divides the workpiece into multiple peripheral regions according to the punching position, and establishes a precise correlation between thermal images and spatial locations through tag data, achieving precise defect localization. This facilitates subsequent regional repair or process adjustment, significantly improving the targeted nature of quality control. Actual quality features are independently extracted for each peripheral region and compared with their corresponding predicted quality features, allowing each region to be independently assessed based on its specific process conditions. This avoids mutual interference and misjudgment caused by process differences between regions, improving the accuracy of detection results. In overall detection mode, local defects in a single peripheral region may contribute little to the overall temperature field and be difficult for the detection algorithm to identify. Therefore, by independently extracting features and calculating similarity for each region, the quality status of each peripheral region is independently evaluated. Local defects are not averaged by signals from other normal regions, improving detection capability.

[0085] Furthermore, when obtaining the first quality score based on the similarity of each hole periphery region, different weighting coefficients can be assigned according to the functional importance of each region. This allows defects in critical regions to have a greater impact on the overall score, while minor deviations in non-critical regions can be tolerated, thus achieving a quality level assessment that better meets actual engineering needs. The weighting coefficients can be set by technicians before processing.

[0086] In one embodiment, obtaining the first quality score based on the similarity corresponding to each periapical region includes:

[0087] Obtain the correspondence between the similarity of the periapical region and the quality score, wherein the correspondence includes the weight coefficient of each periapical region and the similarity-score mapping function;

[0088] Based on the correspondence, the similarity of each periapical region is transformed by the similarity-score mapping function, and then weighted and fused according to the weight coefficient to obtain the first quality score;

[0089] When the similarity of each hole periphery area meets its corresponding similarity threshold, the first quality score is considered to meet the processing requirements, and the conveying mechanism is controlled to move the current energy storage cabinet to be punched to the good product unloading station; otherwise, step S5 is executed.

[0090] The similarity-rating mapping function establishes a nonlinear transformation relationship between the similarity of a single periapical region and the regional quality score. This mapping function can take the form of a piecewise linear function, dividing the similarity interval into discrete levels such as excellent, good, acceptable, and unacceptable, and mapping them to the corresponding score values. Alternatively, it can take the form of a membership function, converting the similarity into linguistic variables (such as highly consistent, basically acceptable, and significantly deviating) through fuzzification, and then calculating a quantitative score through declarative calculation. This mapping mechanism realizes the transformation from the similarity dimension (representing the degree of conformity between the actual and predicted state) to the quality score dimension (representing the absolute quality level), and at the same time, the nonlinear mapping amplifies the score attenuation in low similarity regions, enhancing the sensitivity to the identification of serious defects.

[0091] After obtaining the regional quality scores for each hole periphery area, a weighted fusion is performed based on the aforementioned weighting coefficients to calculate the first quality score. Simultaneously, a similarity threshold matching each hole periphery area (process characteristics, etc.) is set. This similarity threshold, for example, considers differences in material thickness, hole diameter, and punching density between different areas, and is pre-set by technicians, achieving regional adaptability of the judgment criteria. In terms of judgment logic, a strict "full satisfaction is pass" strategy is adopted: only when the similarity of all hole periphery areas meets their respective similarity thresholds is the first quality score considered to meet the processing requirements. This judgment mechanism ensures that defects in any critical area are not masked by the good performance of other normal areas.

[0092] Based on the above judgment results, differentiated logistics control is implemented. When the first quality score meets the processing requirements, it indicates that all areas around the holes are in good processing condition. The control conveyor directly transfers the punched energy storage cabinet to the good product unloading station without entering the secondary inspection process, achieving rapid transfer. When the similarity of any area around the hole is lower than its corresponding threshold, step S5 is triggered, and the workpiece is transferred to the secondary inspection station for in-depth verification to prevent potential defects from flowing into downstream processes. In this case, it can be considered that the secondary inspection verification is fine, and it can be transferred to the good product unloading station; otherwise, it is transferred to the defective product station for further processing.

[0093] Therefore, considering that using a uniform similarity threshold and averaging process cannot reflect the differences in functional importance of different periphery areas, it may lead to minor defects in critical areas not being detected in a timely manner, or normal fluctuations in non-critical areas being excessively amplified. The technical solution provided in this embodiment can match the quality evaluation results with the actual engineering requirements of the energy storage cabinet through differentiated weight configuration, thereby improving the pertinence and rationality of quality control. By designing piecewise functions or membership functions, rapid decay of scores is achieved in the low similarity range (areas with obvious defects), while a gradual change in scores is achieved in the high similarity range (normal fluctuation range). This ensures strict control over serious defects while avoiding overreaction to normal process fluctuations, thus improving the discrimination and robustness of the scoring system.

[0094] In one embodiment, step S6 includes:

[0095] The area around the hole with a similarity lower than its corresponding similarity threshold is taken as the middle area. The infrared feature image of the middle area of ​​the energy storage cabinet to be punched is obtained at the secondary inspection station, and then associated with its respective label data as the second thermal image information.

[0096] The infrared feature images indicating the same label data in the first thermal image information and the second thermal image information are compared to obtain a comparison result including a second quality score.

[0097] The technical solution provided in this embodiment is based on a graded detection and targeted re-inspection strategy. It performs precise secondary verification on the suspected defect areas identified in the initial inspection, which optimizes the allocation of detection resources while ensuring the comprehensiveness of the detection.

[0098] It can be considered that the similarity of each hole periphery region reflects the degree of conformity between the actual processing state and the expected state. When the similarity of a certain hole periphery region is lower than its corresponding similarity threshold, it indicates that there is a potential quality defect risk in that region, but it is not enough to directly determine it as a non-conforming product. Such regions are defined as intermediate regions, that is, in a pending state between conformity and clear defects, requiring further verification and inspection to confirm the authenticity of the defect. At the same time, infrared feature image comparison can be considered as treating one or more regions included in the intermediate region as a whole and comparing them as a whole with the corresponding regions in the first thermal image information.

[0099] The technical solution provided in this embodiment employs a targeted detection strategy for the aforementioned intermediate region, rather than performing an indiscriminate secondary scan of the entire periphery area. During the process of the energy storage cabinet to be punched being moved to the secondary detection station by the conveying mechanism, infrared feature images are selectively acquired only for the area marked as the intermediate region at the secondary detection station. This targeted detection mechanism significantly reduces the amount of data collected and the processing load for secondary detection by reducing unnecessary detection areas.

[0100] Similar to acquiring the first thermal image information, a spatial correspondence is established using tag data when acquiring the second thermal image information. The infrared feature image of the intermediate region acquired during the second detection is associated with its tag data, which is consistent with the tag data of the corresponding region in the first thermal image information, ensuring that both detections target the same physical aperture periphery area. Using the indexing function of the tag data, infrared feature images indicating the same tag data are retrieved from the first and second thermal image information for comparative analysis.

[0101] It can be assumed that the above comparison is based on the difference in thermal images of the same periphery region in different temperature ranges (the first temperature range and the second temperature range). If there is indeed a physical defect in the region (such as a microcrack or hole misalignment), then during temperature adjustment, the difference in thermal response between the defective region and the normal region will exhibit stable and identifiable characteristics. If the anomaly in the initial inspection is only caused by accidental noise or temporary thermal interference, then the difference between the dual-temperature zone images will not conform to the physical characteristics of the defect. Based on the above comparison results, a second quality score is obtained to achieve the final determination of the authenticity of the defect in the intermediate region.

[0102] Therefore, considering that repeatedly scanning the entire area of ​​the workpiece during the re-inspection stage would generate a large amount of redundant data, increasing the workload of the thermal imaging image acquisition device and the computational pressure of subsequent image processing, the introduction of an intermediate area screening and targeted re-inspection mechanism effectively solves the inefficiency problem caused by comprehensive secondary inspection in existing technologies and adapts to the strict cycle time requirements of automated production lines.

[0103] Meanwhile, the initial inspection stage may be prone to misjudgment due to interference from factors such as transient thermal noise and uneven emissivity of the material surface. This embodiment re-images the intermediate area selected in the initial inspection in a second temperature range and compares it with the initial inspection image for verification. By utilizing the physical law that real defects have stable thermal characteristics under different temperature conditions, while random noise does not have this stability, it effectively distinguishes real defects from false signals and significantly reduces the misjudgment rate.

[0104] By employing a tag data association mechanism, the accuracy of regional correspondence in dual-temperature zone image comparison is ensured. In the inspection of complex workpieces involving multi-hole periphery regions, the lack of precise location markers can lead to misalignment of image regions from two inspections, rendering the comparison results meaningless. This embodiment utilizes tag data to establish a precise mapping between the first and second thermal image information for the same periphery region, ensuring spatial consistency in dual-temperature zone comparative analysis. This allows the comparison of temperature evolution characteristics to be based on accurate physical locations, improving the reliability of the inspection results.

[0105] In one embodiment, step S7 includes: determining the quality level of the middle area of ​​the current energy storage cabinet to be punched based on the second quality score;

[0106] The preset process parameters corresponding to the intermediate region are selectively adjusted according to the quality level, and the punching actuator is controlled to process the next energy storage cabinet to be punched based on the adjusted process parameters.

[0107] When the quality grade is unqualified, at least one of the punching pressure, punching speed and holding time in the preset process parameters is adjusted according to the deviation between the actual quality characteristics and the predicted quality characteristics.

[0108] After completing the dual-temperature zone comparison detection in step S6, a second quality score for the intermediate region has been obtained. This score characterizes the true quality status after temperature adjustment verification. This embodiment classifies the intermediate region into quality grades based on this second quality score, such as excellent, qualified, and unqualified, or uses a more granular multi-level classification. This quality grade determination considers the differences in thermal response characteristics revealed in the dual-temperature zone comparison. Unlike related technologies that adjust the global process parameters of the entire workpiece, this embodiment adopts a regional selective adjustment strategy. Specifically, only the preset process parameters corresponding to the specific hole periphery area identified as the intermediate region are adjusted, while the process parameters of other normal areas remain unchanged. This targeted adjustment achieves precise mapping based on the label data of the hole periphery area: by identifying the label data of the intermediate region, the specific position of this region in the working coordinate system of the punching actuator and the corresponding set of process parameters are determined, and then the processing parameters of only this position area are corrected.

[0109] The adjustment range of process parameters is quantitatively determined by the deviation between actual and predicted quality characteristics. This deviation quantifies the degree of deviation between the current processing state and the ideal state, including differences in characteristics such as temperature distribution morphology deviation, heat-affected zone range deviation, and abnormal temperature gradient amplitude. When the quality grade is determined to be unqualified, it indicates that there are substantial processing defects (such as hole deviation or microcracks) in the intermediate area. At this time, based on the magnitude and direction of the above deviation, at least one of the preset process parameters—punching pressure, punching speed, and holding time—is selectively adjusted. For example, when the deviation indicates excessive deformation in the hole periphery area (manifested as abnormal thermal diffusion), the punching pressure can be appropriately reduced or the holding time increased to stabilize material flow; when the deviation indicates severe burrs at the hole edge (manifested as local hot spot concentration), the punching speed can be adjusted to optimize the shearing process.

[0110] Therefore, by using regionalized quality level determination and targeted process adjustment, the problem of overcorrection caused by global parameter adjustments in related technologies is effectively solved. Specifically, by calculating the deviation between actual and predicted quality characteristics, the severity of quality anomalies is converted into specific process parameter correction values, achieving objective quantification of the adjustment range. When the quality level is unqualified, the adjustment amount of punching pressure, speed, or holding time is automatically determined based on the magnitude of the deviation, avoiding over- or under-adjustment phenomena in manual experience judgment, and improving the accuracy and repeatability of process optimization.

[0111] As an example, taking the punching process of the side heat dissipation area of ​​the energy storage cabinet as an example, the energy storage cabinet has two periphery areas: the left heat dissipation area (Zone-A) and the front wiring area (Zone-B). The similarity of Zone-B (front wiring area) is higher than the threshold, and it is judged as a normal area; after a second inspection, Zone-A (left heat dissipation area) obtained a second quality score of 62 points, which is lower than the preset threshold of 70 points, and is judged as unqualified. Through the label data of Zone-A "Area ID: Side-L-01, Coordinate Origin: (150,200), Array: 4×6, Hole Diameter: 10mm", it is determined that the middle area corresponds to the No. 1 punching head module of the punching actuator. The current preset process parameters of this module are: punching pressure 8.5MPa, punching speed 120mm / s, and holding time 0.3s. Comparing the actual quality characteristics of Zone-A with the predicted quality characteristics, the deviation indicates that there is hole deviation accompanied by microcracks in this area, and selective adjustment is performed according to the magnitude of the deviation.

[0112] Based on the deviation between the actual quality characteristics and the predicted quality characteristics, at least one of the preset process parameters—punching pressure, punching speed, and holding time—is adjusted, for example, including:

[0113] This is achieved through a pre-defined mapping table containing a defect type index, deviation level, process parameter correction matrix, and adjustment strategy. The mapping table stores the corresponding rules for different quality defect patterns and process parameter adjustments. After determining the quality defect characteristics in the intermediate region, it automatically retrieves and loads the corresponding process parameter correction matrix and activates the appropriate adjustment strategy.

[0114] In practical applications, the actual quality characteristics of the current energy storage cabinet to be punched, the second quality score, and the corresponding process parameters can be added to the historical processing sequence to update the time-series prediction model.

[0115] As an example, a control method for a sheet metal punching device for energy storage cabinets is provided. The sheet metal punching device for energy storage cabinets includes a thermal imaging image acquisition device, an automatic feeding mechanism, a punching execution mechanism, and a conveying mechanism. Each of the energy storage cabinets to be punched has at least one periphery area according to the punching position, and each periphery area includes multiple punches. When there are multiple periphery areas, the preset quality characteristics include the predicted quality characteristics of each periphery area.

[0116] The method includes the following steps:

[0117] R1, obtain its feedforward information before the current energy storage cabinet to be punched is punched;

[0118] R2, construct the current process timing state of the energy storage cabinet to be punched based on the historical processing sequence; input the feedforward information and the process timing state into the pre-trained timing prediction model to predict the predicted quality characteristics of the current energy storage cabinet to be punched under the preset process parameters, and use them as the preset quality characteristics.

[0119] R3, after the punching process of the energy storage cabinet to be punched, the infrared feature images of multiple hole periphery areas of the energy storage cabinet to be punched are obtained in the first temperature range using the thermal imaging image acquisition device, and then associated with their respective tag data as the first thermal image information.

[0120] R4: For the infrared feature image of each aperture periphery region, perform feature extraction to obtain the actual quality features; obtain the similarity between the actual quality features and the predicted quality features of the corresponding aperture periphery region; obtain the correspondence between aperture periphery region similarity and quality score, the correspondence including the weight coefficient of each aperture periphery region and the similarity-score mapping function;

[0121] Based on the correspondence, the similarity of each periapical region is transformed by the similarity-score mapping function, and then weighted and fused according to the weight coefficient to obtain the first quality score;

[0122] When the similarity of each hole periphery region meets its corresponding similarity threshold, the first quality score is considered to meet the processing requirements, and the conveying mechanism is controlled to move the current energy storage cabinet to be punched to the good product unloading station; otherwise, step R5 is executed.

[0123] R5, when the first quality score does not meet the processing requirements, the conveyor mechanism is used to move it to the secondary inspection station, and the temperature adjustment device is used to adjust the hole periphery area to the second temperature range during the movement.

[0124] R6, obtain the second thermal image information of the hole periphery area of ​​the energy storage cabinet to be punched at the secondary inspection station, take the hole periphery area with similarity lower than its corresponding similarity threshold as the middle area, obtain the infrared feature image of the middle area of ​​the energy storage cabinet to be punched at the secondary inspection station, and associate it with their respective tag data as the second thermal image information.

[0125] The infrared feature images indicating the same label data in the first thermal image information and the second thermal image information are compared to obtain the comparison results including the second quality score;

[0126] R7 controls the automatic feeding mechanism to supply the next energy storage cabinet to be punched, and determines the quality level of the middle area of ​​the current energy storage cabinet to be punched based on the second quality score.

[0127] The preset process parameters corresponding to the intermediate region are selectively adjusted according to the quality level, and the punching actuator is controlled to process the next energy storage cabinet to be punched based on the adjusted process parameters.

[0128] When the quality grade is unqualified, at least one of the punching pressure, punching speed and holding time in the preset process parameters is adjusted according to the deviation between the actual quality characteristics and the predicted quality characteristics.

[0129] The first thermal image information and the second thermal image information are used to indicate the temperature distribution of the area around the hole in the energy storage cabinet to be punched in the first temperature range and the second temperature range, respectively.

[0130] The technical solution provided in this example includes feedforward prediction and dynamic benchmark generation, multi-region initial inspection and weighted quality evaluation, targeted re-inspection and dual-temperature zone verification, and regionalized process feedback and parameter optimization steps. Before the punching process is executed, a dynamic quality benchmark for the current processing task is first established. By acquiring the feedforward information of the energy storage cabinet to be punched (including inherent attributes such as material specifications, plate thickness, and punching layout), combined with the process time-series state constructed based on historical processing sequences, a time-series prediction model is used to quantitatively describe the current evolution trend of the process system. This process time-series state can capture the cumulative impact of time-varying factors such as die wear on processing quality, making the predicted quality characteristics of the current workpiece not a static threshold, but a dynamic benchmark that inherits historical processing patterns and reflects the real-time state of the system. The time-series prediction model integrates feedforward information and process time-series state to output the predicted quality characteristics of each hole periphery area (including expected temperature distribution pattern, heat-affected zone range, etc.), providing a regionalized reference standard for subsequent inspection.

[0131] After the punching process is completed, the initial inspection stage begins. The thermal imaging image acquisition device scans multiple areas around the holes in the energy storage cabinet, acquiring infrared feature images of each area within the first temperature range (natural cooling state), and establishes a precise mapping between the thermal images and physical locations using tag data. For each area around the holes, the system independently extracts actual quality features (temperature gradient, heat-affected zone geometry, etc.) and calculates their similarity to the corresponding predicted quality features.

[0132] A differentiated evaluation mechanism is introduced during the quality scoring stage: First, the similarity of each region is converted into a regional quality score through a similarity-score mapping function. This mapping function employs a non-linear design, amplifying score attenuation in low similarity intervals to enhance defect sensitivity. Subsequently, weighted fusion is performed based on weight coefficients, which are configured according to the functional importance of each hole periphery region (e.g., high weight for electrical installation areas and low weight for heat dissipation areas), ensuring that defects in key areas can have a decisive impact on the overall evaluation. Only when the similarity of all hole periphery regions meets their respective similarity thresholds is the first quality score deemed to meet the processing requirements, and the workpiece is directly transferred to the good product unloading station; if any region fails to meet the threshold, a secondary inspection process is triggered to ensure that local defects are not masked by the overall average effect.

[0133] For the intermediate areas (suspected defect areas with similarity below the threshold) selected in the initial inspection, targeted re-inspection is initiated. During the process of transferring the workpiece to the secondary inspection station using a conveyor mechanism, a temperature adjustment device actively intervenes in the workpiece thermally, adjusting the periphery area to the second temperature range. This temperature adjustment process is based on the principle of thermal diffusion difference: defective areas (such as microcracks or hole deviations) have different thermal capacities and thermal resistance characteristics compared to normal areas, exhibiting differentiated thermal response rates during temperature adjustment, thereby amplifying the identifiability of the thermal features of potential defects. The second temperature range is, for example, generally 10°C higher than the first temperature range. At the secondary inspection station, infrared feature images of the intermediate areas in the second temperature range are selectively acquired as second thermal image information. Through the indexing effect of tag data, the infrared feature images indicating the same physical area in the first and second thermal image information are compared. This dual-temperature zone comparison is based on the temperature evolution characteristics of the same area under different thermal states. Obtaining a second quality score verifies the suspected defects from the initial inspection.

[0134] After determining the quality level of the intermediate area based on the second quality score, selective process adjustments are performed. Unlike the global adjustment strategy, this approach only corrects the preset process parameters corresponding to specific perimeter areas identified as intermediate areas, while other normal areas retain their original parameters. This targeted adjustment uses tag data to locate the corresponding process parameter groups (such as the punching pressure, speed, and holding time of the No. 1 punching head die) in the punching actuator. When the quality level is determined to be unqualified, the process parameter correction value is quantitatively calculated based on the deviation between the actual and predicted quality characteristics (including temperature distribution morphology deviation, heat-affected zone range deviation, etc.). Through a preset mapping table (containing the corresponding rules of defect type index - deviation level - process parameter correction matrix), the corresponding correction coefficients are automatically retrieved and loaded, selectively adjusting at least one of the punching pressure, punching speed, and holding time. The adjusted process parameters are applied to the corresponding area processing of the next energy storage cabinet to be punched, forming a feedforward control closed loop of detection-analysis-correction-prevention, achieving early intervention for quality problems.

[0135] Example 2

[0136] This embodiment also provides an electronic device, including one or more processors and a memory; one or more programs are stored in the memory and configured to be executed by the one or more processors according to any of the methods described above.

[0137] Example 3

[0138] This embodiment also provides a sheet metal punching system for energy storage cabinets, which includes sheet metal punching equipment for energy storage cabinets and the electronic equipment described in Embodiment 2.

[0139] In one embodiment, the energy storage cabinet sheet metal punching equipment includes a thermal imaging image acquisition device, a frame, an automatic feeding mechanism, a punching execution mechanism, and a conveying mechanism, wherein the punching execution mechanism includes a servo drive device.

[0140] Example 4

[0141] This application also provides a computer-readable storage medium, the specific embodiments of which are consistent with the embodiments described above and the technical effects achieved, and some contents will not be repeated.

[0142] The computer-readable storage medium stores a computer program that, when executed by at least one processor, implements the steps of any of the above methods or the functions of any of the above electronic devices.

[0143] A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. In embodiments of this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0144] Computer-readable storage media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable storage medium may also be any computer-readable medium capable of sending, propagating, or transmitting a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, or any suitable combination thereof. Program code for performing operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar programming languages. The program code may be executed entirely on a user computing device, partially on a user device, as a standalone software package, partially on a user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to user computing devices via any type of network, including local area networks (LANs) or wide area networks (WANs), or they can be connected to external computing devices (e.g., via the Internet using an Internet service provider).

[0145] Example 5

[0146] This application also provides a program product embodiment, wherein the computer program product includes a computer program, and when the computer program is executed by at least one processor, it implements the steps of the method described in any one of the method embodiments. Its specific embodiments are consistent with the embodiments described above and the technical effects achieved are the same, and some details will not be repeated.

[0147] Although the invention has been described with reference to illustrative embodiments, those skilled in the art will understand that various other changes, omissions, and / or additions can be made without departing from the spirit and scope of the invention, and that elements of the described embodiments can be substituted with substantially equivalents. Furthermore, many modifications can be made without departing from the scope of the invention to adapt particular situations or materials to the teachings of the invention. Therefore, this document is not intended to limit the invention to the specific embodiments disclosed for carrying out the invention, but rather to include all embodiments falling within the scope of the appended claims.

Claims

1. A control method for a sheet metal punching device for an energy storage cabinet, the sheet metal punching device for the energy storage cabinet comprising a thermal imaging image acquisition device, an automatic feeding mechanism, a punching execution mechanism, and a conveying mechanism; characterized in that, The method includes the following steps: S3, after the punching process of the energy storage cabinet to be punched, the thermal imaging image acquisition device is used to obtain the first thermal image information of the area around the hole in the first temperature range. S4, obtain the first quality score based on the actual quality features corresponding to the first image information and the preset quality features; S5, when the first quality score does not meet the processing requirements, the conveyor mechanism is used to move it to the secondary inspection station, and the temperature adjustment device is used to adjust the hole periphery area to the second temperature range during the movement. S6, acquire the second thermal image information of the area around the hole of the energy storage cabinet to be punched at the secondary inspection station, and obtain a comparison result including the second quality score based on the first thermal image information and the second thermal image information; S7, control the automatic feeding mechanism to supply the next energy storage cabinet to be punched, and control the punching execution mechanism to process the next energy storage cabinet to be punched according to the comparison result; The first thermal image information and the second thermal image information are used to indicate the temperature distribution of the area around the hole in the energy storage cabinet to be punched in the first temperature range and the second temperature range, respectively.

2. The control method according to claim 1, characterized in that, The method further includes: S1, before the current energy storage cabinet to be punched is punched, obtain its feedforward information; S2, construct the current process timing state of the energy storage cabinet to be punched based on the historical processing sequence; input the feedforward information and the process timing state into the pre-trained timing prediction model to predict the predicted quality characteristics of the current energy storage cabinet to be punched under preset process parameters, and use them as preset quality characteristics.

3. The control method according to claim 1, characterized in that, Each of the energy storage cabinets to be punched has at least one periphery region according to the punching position, and each periphery region includes multiple punches; when there are multiple periphery regions, the preset quality characteristics include the predicted quality characteristics of each periphery region. Step S3 includes: using the thermal imaging image acquisition device to acquire infrared feature images of multiple hole periphery areas of the energy storage cabinet to be punched in the first temperature range, and associating them with their respective tag data as the first thermal image information; Step S4 includes: extracting actual quality features from the infrared feature image of each aperture periphery region; obtaining the similarity between the actual quality features and the predicted quality features of the corresponding aperture periphery region; and obtaining a first quality score based on the similarity between each aperture periphery region.

4. The control method according to claim 3, characterized in that, The step of obtaining the first quality score based on the similarity of each periapical region includes: Obtain the correspondence between the similarity of the periapical region and the quality score, wherein the correspondence includes the weight coefficient of each periapical region and the similarity-score mapping function; Based on the correspondence, the similarity of each periapical region is transformed by the similarity-score mapping function, and then weighted and fused according to the weight coefficient to obtain the first quality score; When the similarity of each hole periphery area meets its corresponding similarity threshold, the first quality score is considered to meet the processing requirements, and the conveying mechanism is controlled to move the current energy storage cabinet to be punched to the good product unloading station; otherwise, step S5 is executed.

5. The control method according to claim 3, characterized in that, Step S6 includes: The area around the hole with a similarity lower than its corresponding similarity threshold is taken as the middle area. The infrared feature image of the middle area of ​​the energy storage cabinet to be punched is obtained at the secondary inspection station, and then associated with its respective label data as the second thermal image information. The infrared feature images indicating the same label data in the first thermal image information and the second thermal image information are compared to obtain a comparison result including a second quality score.

6. The control method according to claim 5, characterized in that, Step S7 includes: The quality level of the middle area of ​​the current energy storage cabinet to be punched is determined based on the second quality score. The preset process parameters corresponding to the intermediate region are selectively adjusted according to the quality level, and the punching actuator is controlled to process the next energy storage cabinet to be punched based on the adjusted process parameters. When the quality grade is unqualified, at least one of the punching pressure, punching speed and holding time in the preset process parameters is adjusted according to the deviation between the actual quality characteristics and the predicted quality characteristics.

7. An electronic device, characterized in that, It includes one or more processors and memory; one or more programs are stored in the memory and configured to be executed by the one or more processors according to any one of claims 1-6.

8. A sheet metal punching system for an energy storage cabinet, characterized in that, The energy storage cabinet sheet metal punching system includes an energy storage cabinet sheet metal punching device and the electronic device described in claim 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by at least one processor, implements the steps of the method according to any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by at least one processor, implements the steps of the method according to any one of claims 1-6.

Images