Automation equipment frame inspection method and system

Abstract

The present invention relates to a method and system for inspecting frames for automated equipment. Specifically, the frame inspection system for automated equipment may include an information input unit that receives actual measurement information for each inspection item for a target frame used in automated equipment; an error analysis unit that compares the actual measurement information for each inspection item with design reference information and calculates error information for each inspection item based on the comparison result; a quality diagnosis unit that diagnoses the quality status of the target frame as either a normal state or a defective state based on the error information for each inspection item; and a calibration management unit that, when the quality status is diagnosed as a defective state, applies the error information for each inspection item to a preset artificial intelligence-based calibration solution selection model to derive a customized calibration task list.

Description

Technology Field

[0001] The present invention relates to a method and system for inspecting frames of automated equipment. It falls within the technical field capable of providing correction solutions by precisely inspecting frames of automated equipment. Background Technology

[0002] Generally, the frame for automation equipment serves as a foundational structure on which various drive modules, sensors, and transfer rails are mounted, and it is a key element that determines the driving precision and durability of the equipment.

[0003] In particular, for semiconductor equipment performing nano-scale processes or display equipment requiring high-speed transfer, frame flatness, parallelism, and positional precision of pinholes and tap holes require tolerances in the micrometer range.

[0004] Conventional frame quality inspection mainly involved manual measurement by an operator using vernier calipers, height gauges, etc., or obtaining numerical data using a three-dimensional measuring machine (CMM).

[0005] However, conventional technology had limitations in that it could only verify the error value between the data measured by the measuring tool and the design drawing, but could not provide a correction or calibration solution to resolve the error value.

[0006] As a result, when a defective frame occurs, grinding or reassembly must be performed relying on the intuition or experience of skilled workers. During this process, problems arise such as the frames being discarded due to excessive grinding or variations in the quality of correction depending on the worker. Additionally, there is a problem of wasting labor and costs by attempting to correct frames despite severe, irreparable warping or substandard dimensions.

[0007] Accordingly, we aim to provide a frame inspection method and system for automated equipment that not only precisely analyzes frame errors but also proposes optimal correction measures through artificial intelligence and simulates work feasibility in advance to provide visualized work instruction guides to the operator. Prior art literature

[0008] (01) Korean Published Patent No. 10-2908465 (December 31, 2025) The problem to be solved

[0009] The present invention aims to solve the aforementioned problems, and the objective of the present invention is to provide a frame inspection method and system for automated equipment capable of diagnosing the quality status based on error information for each inspection item and deriving a customized correction task list through an artificial intelligence model based on the defect status.

[0010] In addition, the purpose is to provide a frame inspection method and system for automated equipment capable of outputting a calibration simulation graphic that visualizes the before and after of the calibration work by using a customized calibration variable detected according to the characteristics of the defect type and a target frame 3D modeling drawing.

[0011] In addition, the invention aims to provide a frame inspection method and system for automated equipment that can selectively generate a work instruction guide based on whether a calibration operation is possible, determined according to a calibration simulation graphic. means of solving the problem

[0012] A frame inspection system for automated equipment according to an embodiment of the present invention for achieving the above-mentioned purpose comprises: an information input unit that receives actual measurement information for each inspection item for a target frame used in automated equipment; an error analysis unit that compares the actual measurement information for each inspection item with design reference information and calculates error information for each inspection item based on the comparison result; a quality diagnosis unit that diagnoses the quality state of the target frame as either a normal state or a defective state based on the error information for each inspection item; and a calibration management unit that, when the quality state is diagnosed as a defective state, applies the error information for each inspection item to a preset artificial intelligence-based calibration solution selection model to derive a customized calibration work list. The inspection items include dimensions by part, design angle, parallelism and flatness, slot shape alignment, rail surface roughness and pinhole inner diameter precision, tap hole fastening alignment, and welding cleanliness. The customized calibration work list may be a table in which at least one of grinding, polishing, reassembly, and forming operations is listed by part.

[0013] In an embodiment, the calibration management unit includes a type feature analysis unit that analyzes error information for each inspection item and derives defect type features as the quality condition is diagnosed as a defective state, a correction variable detection unit that detects a customized correction variable from a preset correction variable for each defect type feature based on the defect type features, and a work order management unit that virtually applies the customized correction variable to a target frame 3D modeling drawing drawn according to the actual measurement information for each inspection item and outputs a calibration simulation graphic, wherein the calibration simulation graphic is a before and after image of a calibration operation visualized by matching the customized correction variable to the object coordinates for each part detected in the target frame 3D modeling drawing, and at least one of the location coordinates, processing values, and post-processing methods for the customized calibration operation list may be tagged and displayed.

[0014] In an embodiment, the work instruction management unit includes a work judgment unit that determines whether a correction operation is possible by applying the correction simulation graphic to a preset artificial intelligence-based correction operation capability diagnostic model, and a guide generation unit that generates a work instruction guide for the target frame based on whether a correction operation is possible, and the work instruction guide may include the correction simulation graphic, the customized correction variable, and the customized correction operation list.

[0015] In an embodiment, the work judgment unit may include a thickness calculation unit that calculates the remaining thickness of the area to be ground or cut by analyzing the correction simulation graphic; a grade determination unit that compares the remaining thickness with a preset minimum safety thickness and determines a structural safety grade for the target frame based on the comparison result; a risk judgment unit that determines the target frame to be in a structurally dangerous state when the structural safety grade is less than the preset safety grade; and a reinforcement recommendation unit that searches for and recommends specifications and attachment location information for a reinforcement plate that can be attached to the area when the target frame is determined to be unable to perform structural correction work.

[0016] In an embodiment, the calibration management unit further includes a model learning unit that models the calibration solution selection model by learning through machine learning, which takes collected error information samples as input and outputs a list of calibration tasks for each error information sample. The model learning unit includes a standard derivation unit that derives standard numerical information for each inspection item by analyzing the calibration simulation graphic; an information receiving unit that receives actual calibration measurement information for each inspection item, in which the target frame is calibrated and re-measured according to the work instruction guide; an accuracy calculation unit that compares the actual calibration information and the standard numerical information for each inspection item and calculates a calibration accuracy index based on the comparison result; and a model management unit that individually adjusts each weight for the calibration solution selection model based on the calibration accuracy index. The calibration accuracy index may be a similarity value for each inspection item between the actual calibration information and the standard numerical information.

[0017] In an embodiment, the error analysis unit may include an information detection unit that detects the values ​​of pinholes and tapholes of the target frame from error information for each inspection item; a diameter calculation unit that calculates each diameter machining value for pinholes and tapholes when the values ​​of pinholes and tapholes are smaller than each standard specification; a pinhole management unit that calculates the capacity of repair material for hole filling when the pinhole is larger than the standard specification; a taphole management unit that calculates the helicoil specification when the taphole is larger than the standard specification; and a guide update unit that classifies at least one of the diameter machining value, the capacity of repair material for hole filling, and the helicoil specification into a fastening correction guide and updates it to the work instruction guide.

[0018] In an embodiment, the guide update unit applies the parallelism and flatness information of the target frame detected from the error information for each inspection item to a preset artificial intelligence-based torsion diagnosis model to determine the torsion level for the target frame, and based on the torsion level, derives a disassembly and reassembly guide for the target frame from a preset disassembly and assembly design drawing and updates the work instruction guide.

[0019] In an embodiment, the guide update unit can classify the thickness of the surface coating agent corresponding to the specifications of the shield for masking the grounding terminal area identified from the design reference information and the real-time temperature and humidity information into a post-processing management guide and update it in the work instruction guide. Effects of the invention

[0020] According to an embodiment of the present invention, a frame inspection method and system for automated equipment can support frame correction of uniform quality regardless of the operator's skill level by deriving a customized correction task list through an artificial intelligence model according to the defect status.

[0021] In addition, the frame inspection method and system for automated equipment can prevent errors when an operator performs calibration work by outputting a calibration simulation graphic that visualizes the before and after of the calibration work.

[0022] In addition, the frame inspection method and system for automated equipment can increase work efficiency by selectively generating a work instruction guide based on whether a correction operation is possible, thereby blocking resources consumed on frames that cannot be restored. Brief explanation of the drawing

[0023] FIG. 1 is a schematic diagram showing a frame inspection system (1000) for automated equipment according to one embodiment of the present invention. FIG. 2 is a block diagram specifically illustrating an embodiment of the calibration management unit (400) of FIG. 1. FIG. 3 is a block diagram showing an example of the work instruction management unit (430) of FIG. 2. FIG. 4 is a block diagram showing an example of the work judgment unit (431) of FIG. 3. FIG. 5a is a block diagram showing another embodiment of the correction management unit (400_1) of FIG. 2, and FIG. 5b is a block diagram showing a more specific embodiment of the model learning unit (440) of FIG. 5a. FIG. 6 is a block diagram showing an example of the error analysis unit (200) of FIG. 1. FIG. 7 is a diagram showing the operation process for the frame inspection system (1000) for automated equipment of FIG. 1. Specific details for implementing the invention

[0024] The present invention will be described in detail below with reference to the embodiments and drawings. These embodiments are presented merely as examples to explain the invention more specifically, and it will be obvious to those skilled in the art that the scope of the invention is not limited by these embodiments.

[0025] Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention pertains, and in the event of a conflict, the description in this specification, including the definitions, shall prevail.

[0026] To clearly explain the proposed invention in the drawings, parts unrelated to the description have been omitted, and similar parts throughout the specification have been given similar reference numerals. Furthermore, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Additionally, the term "part" as described in the specification refers to a single unit or block that performs a specific function.

[0027] In each step, identification codes (1st, 2nd, etc.) are used for convenience of explanation and do not describe the order of the steps; the steps may be performed differently from the specified order unless a specific order is clearly indicated in the context. That is, the steps may be performed in the same order as specified, substantially simultaneously, or in the reverse order.

[0028] FIG. 1 is a schematic diagram showing a frame inspection system (1000) for automated equipment according to one embodiment of the present invention.

[0029] Referring to FIG. 1, the frame inspection system (1000) for automated equipment may include an information input unit (100), an error analysis unit (200), a quality diagnosis unit (300), and a calibration management unit (400).

[0030] First, the information input unit (100) can receive actual measurement information for each inspection item of the target frame used in the automation equipment.

[0031] Here, the target frame may be a structure for forming at least one of the cover, rail, base, and support of the automation equipment.

[0032] In this case, automation equipment may refer to industrial machinery including semiconductor manufacturing equipment, display panel transfer and bonding equipment, secondary battery assembly lines, surface mount equipment, CNC machine tools, and industrial robot arms. Such automation equipment can ensure stability for micro-processes by utilizing a frame that supports loads and absorbs vibrations. In addition,

[0033] The inspection items according to the embodiment may include dimensions by part, design angle, parallelism and flatness, slot shape alignment, rail surface roughness, pinhole inner diameter precision, tap hole fastening alignment, and welding cleanliness.

[0034] For example, dimensions include the length, width, thickness, and height of the part; design angle is the perpendicularity or inclination angle required by the design of the part; parallelism is an indicator indicating whether the surface of the part is flat; parallelism is an indicator indicating the parallel state between the part and a pre-set part; slot shape compatibility refers to whether the shape of the slot to which a specific part is to be mounted matches the part; rail surface roughness is an indicator indicating the degree of smoothness of the rail mounting surface to which the drive motor is attached; pinhole inner diameter precision refers to the inner diameter precision of the pinhole into which the reference pin is inserted; tap hole fastening compatibility refers to whether it corresponds to tap hole standard specifications, such as the degree of thread damage, verticality, and effective depth; and weld cleanliness may be an indicator indicating the uniformity and contamination of the weld bead shape.

[0035] According to an embodiment, the information input unit (100) can receive the shape and use of the target frame and extract and provide a customized inspection item list corresponding to the target frame from pre-set inspection items for each shape and use.

[0036] Next, the error analysis unit (200) can compare actual measurement information for each inspection item with design reference information and calculate error information for each inspection item based on the comparison result.

[0037] Here, error information for each inspection item may refer to data that quantifies the deviation between the actual value and the reference value. In this case, design reference information may refer to the target value for each inspection item.

[0038] Next, the quality diagnosis unit (300) can diagnose the quality status of the target frame as either a normal state or a defective state based on error information for each inspection item.

[0039] For example, if the error information for each inspection item falls within a preset allowable range, the history management unit (300) diagnoses the quality status of the target frame as normal, and if the error information for each inspection item falls outside the preset allowable range for each inspection item, the history management unit (300) can diagnose the quality status of the target frame as defective.

[0040] According to an embodiment, the quality diagnosis unit (300) can diagnose the assembly interference status between the frame and the equipment by using 3D scan data of the target frame and preset automated equipment 3D scan data.

[0041] Next, when the quality condition is diagnosed as poor, the calibration management unit (400) can apply error information for each inspection item to a preset artificial intelligence-based calibration solution selection model to derive a customized calibration task list for the target frame.

[0042] Here, the customized straightening work list may be a table in which at least one of the grinding, polishing, reassembly, and forming operations is listed by part.

[0043] At this time, the pre-configured artificial intelligence-based correction solution selection model may be an artificial neural network algorithm that is trained by a model learning unit (440 in FIG. 5 below) and receives error information for each inspection item and outputs a customized correction task list.

[0044] According to one embodiment, the correction management unit (400) can check whether to perform a correction operation on a target frame based on a correction operation image uploaded according to a customized correction operation list.

[0045] For example, the calibration work images may include dimension inspection images using a tape measure, parallelism and flatness inspection images using a right-angle ruler, straight edge and gap gauge, shape inspection images using a shape matching jig, surface inspection images using a sliding jig, pinhole and screw hole inspection images, etc.

[0046] According to another embodiment, the calibration management unit (400) can identify the type of calibration work by applying a calibration work video of a target frame monitored through a provided camera (not shown) to a preset artificial intelligence-based calibration work identification model. Here, the calibration work identification model may be an artificial neural network algorithm modeled by learning through machine learning that takes a plurality of calibration work video samples as input and outputs the type of calibration work for each calibration work video sample. At this time, the calibration management unit (400) can output a work guide projection image on the surface of the target frame by determining it as an error event based on whether there is a difference between the type of calibration work and the customized calibration work list.

[0047] According to another embodiment, the calibration management unit (400) can calculate a thermal deformation compensation value based on the temperature and humidity information detected through the provided temperature and humidity sensor and the thermal expansion coefficient of the target frame to correct the actual measurement information for each inspection item.

[0048] According to another embodiment, the calibration management unit (400) may change the customized fixing operation for the area to be ground or cut to a laser calibration method when the target frame is intended for entry into a cleanroom. Here, the laser calibration method may refer to a local melting and evaporation method using a high-power laser to prevent the scattering of fine dust.

[0049] According to another embodiment, the calibration management unit (400) can merge an insulating tape attachment operation into a customized calibration operation list when the target frame is used as a battery fixing frame.

[0050] According to another embodiment, the calibration management unit (400) can inspect natural frequency information measured through a vibration detection sensor (not shown) attached to the target frame when the target frame is intended for use in semiconductor lithography equipment. At this time, the calibration management unit (400) can merge mass calibration work for the target frame into a customized calibration work list based on the difference between the natural frequency information and the reference frequency range confirmed from the design reference information.

[0051] Specifically, the calibration management unit (400) can generate a mass calibration task to increase the mass of the target frame and merge it into the calibration task list when the natural frequency information is higher than the reference frequency range. Additionally, the calibration management unit (400) can generate a mass calibration task to decrease the mass of the target frame and merge it into the calibration task list when the natural frequency information is lower than the reference frequency range.

[0052] Hereinafter, the structure of the present invention and the resulting effects are to be explained in more detail through specific embodiments and comparative examples. However, these embodiments are intended to explain the present invention more specifically, and the scope of the present invention is not limited to these embodiments.

[0053] FIG. 2 is a block diagram specifically illustrating an embodiment of the calibration management unit (400) of FIG. 1.

[0054] Referring to FIGS. 1 and FIGS. 2, the calibration management unit (400) may include a type feature analysis unit (410), a correction variable detection unit (420), and a work instruction management unit (430).

[0055] First, the type characteristic analysis unit (410) can derive defect type characteristics by analyzing error information for each inspection item as the quality condition is diagnosed as defective.

[0056] Here, defect type characteristics may refer to defect classification criteria indicating the type of defect and correction direction of the target frame. These defect type characteristics may include dimensional defect types, shape distortion types, surface defect types, and fastening defect types.

[0057] Next, the correction variable detection unit (420) can detect a customized correction variable from the correction variables for each defect type characteristic already collected in the database (500) based on the defect type characteristics.

[0058] Here, the correction variable may refer to a processing control variable for restoring the target frame to a normal state according to the defect type characteristics.

[0059] Next, the work order management unit (430) can virtually apply customized correction variables to the target frame 3D modeling drawing drawn according to the actual measurement information for each inspection item, and output a correction simulation graphic.

[0060] The calibration simulation graphic according to the embodiment may be an image of the calibration operation before and after visualization by matching customized correction variables to the object coordinates of each part detected in the target frame 3D modeling drawing. This calibration simulation graphic may display at least one of the location coordinates, processing values, and post-processing methods for the customized calibration operation list tagged.

[0061] FIG. 3 is a block diagram showing an example of the work instruction management unit (430) of FIG. 2.

[0062] Referring to FIGS. 2 and FIGS. 3, the work instruction management unit (430) may include a work judgment unit (431) and a guide generation unit (432).

[0063] First, the work judgment unit (431) can determine whether a correction work is possible by applying a correction simulation graphic to a preset artificial intelligence-based correction work capability diagnosis model.

[0064] Here, the AI-based diagnostic model for correction feasibility may be an artificial neural network algorithm modeled by learning through machine learning that takes multiple graphic samples as input and outputs whether correction feasibility is possible for each graphic sample.

[0065] Next, the guide generation unit (432) can generate and provide a work instruction guide for the target frame based on whether a calibration operation is possible. Here, the work instruction guide may include a calibration simulation graphic, a customized calibration variable, and a customized calibration operation list.

[0066] This guide generation unit (432) may also provide a target frame check sheet that is generated by applying the work instruction guide through a pre-set check sheet creation template.

[0067] FIG. 4 is a block diagram showing an example of the work judgment unit (431) of FIG. 3.

[0068] Referring to FIGS. 3 and 4, the work judgment unit (431) may include a thickness calculation unit (431_1), a grade determination unit (431_2), a risk judgment unit (431_3), and a reinforcement recommendation unit (431_4).

[0069] First, the thickness calculation unit (431_1) can calculate the remaining thickness of the area to be ground or cut by analyzing the correction simulation graphic.

[0070] Next, the grade determination unit (431_2) can compare the remaining thickness with the preset minimum safety thickness and determine the structural safety grade for the target frame based on the comparison result.

[0071] Next, the risk assessment unit (431_3) can determine that the target frame is in a structurally dangerous state if the structural safety rating is lower than a preset safety rating.

[0072] Next, when the target frame is determined to be in a structurally dangerous state, the reinforcement recommendation unit (431_4) can search for and recommend specifications and attachment location information for a reinforcement plate that can be attached to the relevant part.

[0073] FIG. 5a is a block diagram showing another embodiment of the correction management unit (400_1) of FIG. 2, and FIG. 5b is a block diagram showing a more specific embodiment of the model learning unit (440) of FIG. 5a.

[0074] Referring to FIGS. 2, FIGS. 3, FIGS. 5a and FIGS. 5b, the calibration management unit (400_1) may include a type feature analysis unit (410), a correction variable detection unit (420), a work instruction management unit (430), and a model learning unit (440).

[0075] Hereinafter, the redundant description of the type feature analysis unit (410), correction variable detection unit (420), and work instruction management unit (430) of the same member number described in FIG. 2 is omitted.

[0076] First, the model learning unit (440) can model an artificial intelligence-based correction solution selection model by learning through machine learning, which takes collected error information samples as input and outputs a list of correction tasks for each error information sample, as shown in FIG. 5a.

[0077] Next, as illustrated in FIG. 5b, the model learning unit (440) may include a standard derivation unit (441), an information receiving unit (442), an indicator calculation unit (443), and a model management unit (444).

[0078] Specifically, the standard derivation unit (441) can derive standard numerical information for each inspection item by analyzing the calibration simulation graphic. At this time, the information receiving unit (442) can receive calibration actual measurement information for each inspection item, in which the target frame is calibrated and re-measured according to the work instruction guide generated through the guide generation unit (432). Then, the index calculation unit (443) can compare the re-measured calibration actual measurement information and the standard numerical information for each inspection item, and calculate a calibration accuracy index based on the comparison result. Here, the calibration accuracy index may be a similarity value for each inspection item between the calibration re-measured information and the standard numerical information. Subsequently, the model management unit (444) can individually adjust each weight for the calibration solution selection model based on the calibration accuracy index.

[0079] FIG. 6 is a block diagram showing an example of the error analysis unit (200) of FIG. 1.

[0080] Referring to FIGS. 3 and FIGS. 6, the error analysis unit (200) may include a diameter calculation unit (210), a pinhole management unit (220), a tap hole management unit (230), and a guide update unit (240).

[0081] First, the diameter calculation unit (210) can calculate the diameter processing value when the pinhole and tap hole of the target frame, identified from the error information for each inspection item, are smaller than each standard specification.

[0082] Next, the pinhole management unit (220) can calculate the volume of repair material for filling holes according to the pinhole value when the pinhole is larger than the standard specification.

[0083] Next, the tap hole management unit (230) can calculate the helicoil specifications according to the tap hole value when the tap hole is larger than the standard specification.

[0084] Next, the guide update unit (240) can classify at least one of the diameter processing value, the capacity of the repair material for hole filling, and the helicoil specification as a fastening correction guide and update it in the work instruction guide generated through the calibration management unit (400).

[0085] According to one embodiment, the guide update unit (240) can determine the torsion level for the target frame by applying parallelism and flatness information of the target frame detected from error information for each inspection item to a preset artificial intelligence-based torsion diagnosis model. Here, the artificial intelligence-based composite torsion diagnosis model may be an artificial neural network algorithm modeled by learning through machine learning that takes the previously collected parallelism and flatness information for each frame as input and outputs the torsion level for each parallelism and flatness information. At this time, the guide update unit (240) can derive a disassembly and reassembly guide for the target frame from a preset disassembly and assembly design drawing based on the torsion level and update it in the work instruction guide.

[0086] According to another embodiment, the guide update unit (240) can classify the thickness of the surface coating agent corresponding to the specifications of the shield for masking the grounding terminal area identified from the design reference information and the real-time temperature and humidity information into a post-processing management guide and update it in the work instruction guide.

[0087] FIG. 7 is a diagram showing the operation process for the frame inspection system (1000) for automated equipment of FIG. 1.

[0088] Referring to FIGS. 1 and FIGS. 7, first, in step S110, the information input unit (100) can receive actual measurement information for each inspection item of the target frame used in the automation equipment.

[0089] Then, in step S120, the error analysis unit (200) can compare actual measurement information for each inspection item with design reference information and calculate error information for each inspection item based on the comparison result.

[0090] Then, in step S130, the quality diagnosis unit (300) can diagnose the quality status of the target frame as either a normal state or a defective state based on error information for each inspection item.

[0091] At this time, in step S140, if the quality condition is diagnosed as poor, the calibration management unit (400) can apply error information for each inspection item to a preset artificial intelligence-based calibration solution selection model to derive a customized calibration task list for the target frame.

[0092] Then, at step S150, the calibration management unit (400) can output a calibration simulation graphic that visualizes the calibration work before and after using a customized calibration variable analyzed from the error information for each inspection item and a target frame 3D modeling drawing drawn according to the actual measurement information for each inspection item.

[0093] Subsequently, at step S160, the calibration management unit (400) may selectively generate and provide a work instruction guide for the target frame based on whether calibration work is possible, determined according to the calibration simulation graphic.

[0094] In this specification, only a few examples among the various embodiments performed by the inventors are described; however, the technical concept of the present invention is not limited or restricted thereto, and it is obvious that it can be modified and implemented in various ways by those skilled in the art. Explanation of the symbols

[0095] 100: Information Input Section 200: Error Analysis Department 300: Quality Diagnosis Department 400: Correction Management Department 1000: Frame Inspection System for Automated Equipment

Claims

Claim 1 An information input unit that receives actual measurement information for each inspection item for a target frame used in automated equipment; an error analysis unit that compares the actual measurement information for each inspection item with design reference information and calculates error information for each inspection item based on the comparison result; and a quality diagnosis unit that diagnoses the quality status of the target frame as either a normal state or a defective state based on the error information for each inspection item. The calibration management unit includes, when the quality condition is diagnosed as defective, an error information for each inspection item, and applies it to a preset AI-based calibration solution selection model to derive a customized calibration work list; the inspection items include dimensions by part, design angle, parallelism and flatness, slot shape alignment, rail surface roughness and pinhole inner diameter precision, tap hole fastening alignment, and welding cleanliness; the customized calibration work list is a table in which at least one of grinding, polishing, reassembly, and forming operations is listed by part; the calibration management unit includes a type feature analysis unit that analyzes the error information for each inspection item to derive defect type characteristics as the quality condition is diagnosed as defective; and a correction variable detection unit that detects customized correction variables from preset correction variables for each defect type characteristic based on the defect type characteristics. and includes a work order management unit that virtually applies the customized correction variable to a target frame 3D modeling drawing drawn according to the actual measurement information for each inspection item and outputs a calibration simulation graphic, wherein the calibration simulation graphic is a before-and-after image of a calibration operation visualized by matching the customized correction variable to the object coordinates of each part detected in the target frame 3D modeling drawing, and at least one of the position coordinates, processing values, and post-processing plan for the customized calibration operation list is tagged and displayed, and the work order management unit applies the calibration simulation graphic to a preset AI-based calibration operation capability diagnostic model to determine whether a calibration operation is possible;and includes a guide generation unit that generates a work instruction guide for the target frame based on whether the above calibration work is possible, wherein the work instruction guide includes the calibration simulation graphic, the customized correction variable, and the customized calibration work list, and the work judgment unit includes a thickness calculation unit that analyzes the calibration simulation graphic to calculate the remaining thickness of the area to be ground or cut; a grade determination unit that compares the remaining thickness with a preset minimum safety thickness and determines a structural safety grade for the target frame based on the comparison result; and a risk judgment unit that determines the target frame to be in a structural risk state if the structural safety grade is less than a preset safety grade. and includes a reinforcement recommendation unit that searches for and recommends specifications and attachment location information for a reinforcement plate that can be attached to the area when the target frame is determined to be unfit for structural correction work; the correction management unit further includes a model learning unit that models the correction solution selection model by learning through machine learning, which takes collected error information samples as input and outputs a list of correction work for each error information sample; the model learning unit includes a standard derivation unit that analyzes the correction simulation graphic to derive standard numerical information for each inspection item; an information receiving unit that receives actual correction measurement information for each inspection item, in which the target frame is corrected and remeasured according to the work instruction guide; and an accuracy calculation unit that compares the remeasured actual correction measurement information and the standard numerical information for each inspection item and calculates a correction accuracy index based on the comparison result. and includes a model management unit that individually adjusts each weight for the correction solution selection model based on the calibration accuracy indicator, wherein the calibration accuracy indicator is a similarity value for each inspection item between the re-measured calibration actual information and the reference numerical information, and the error analysis unit includes an information detection unit that detects the numerical values ​​of pinholes and tapholes of the target frame from the error information for each inspection item;A diameter calculation unit that calculates the respective diameter machining values ​​for the pinhole and tap hole when the values ​​of the pinhole and tap hole are smaller than the respective standard specifications; a pinhole management unit that calculates the capacity of the repair material for hole filling when the pinhole is larger than the standard specifications; and a tap hole management unit that calculates the helicoil specifications when the tap hole is larger than the standard specifications. and includes a guide update unit that classifies at least one of the diameter processing value, the capacity of the repair material for hole filling, and the helicoil specification as a fastening correction guide and updates it to the work instruction guide; the guide update unit applies the parallelism and flatness information of the target frame detected from the error information for each inspection item to a preset AI-based torsion diagnosis model to determine the torsion level for the target frame, and based on the torsion level, derives a disassembly and reassembly guide for the target frame from a preset disassembly and assembly design drawing and updates it to the work instruction guide; the guide update unit classifies the shield specification for masking the grounding terminal area confirmed from the design drawing reference information and the application thickness of the surface coating agent corresponding to real-time temperature and humidity information as a post-processing management guide and updates it to the work instruction guide; and the calibration management unit applies the calibration work video of the target frame monitored through the provided camera to a preset AI-based calibration work identification model to identify the type of calibration work, and determines it as an error event based on whether there is a difference between the identified type of calibration work and the customized calibration work list, and the A frame inspection system for automated equipment that outputs a work guide projected image on the surface of a target frame, and the calibration management unit calculates a thermal deformation compensation value based on temperature and humidity information detected through a provided temperature and humidity sensor and the thermal expansion coefficient of the target frame to correct actual measurement information for each inspection item. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete Claim 8 delete

Images