An intelligent quality control system and method for manufacturing a full-weld pressure-bearing plate heat exchanger

Abstract

The application discloses an intelligent quality control system and method for manufacturing of full-welding pressure plate heat exchangers, and relates to the field of intelligent manufacturing, and comprises: a detection module for detecting heat exchanger surface structure parameters and constructing a heat exchanger three-dimensional model based on the heat exchanger surface structure parameters; and a derivation module for receiving the heat exchanger three-dimensional model, picking up incomplete models on the surface of the heat exchanger three-dimensional model, and respectively deriving bounding boxes of the incomplete models based on the picked-up incomplete models. The application realizes high-precision digital restoration of the heat exchanger structure by means of infrared distance measuring sensors densely arranged on a ring guide rail, combination of position and time stamp information, conversion of a large number of distance measuring results into three-dimensional space line segments, and splicing of the three-dimensional space line segments into a closed three-dimensional model, so as to provide quality control services for the heat exchanger.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, specifically to an intelligent quality control system and method for manufacturing fully welded pressure plate heat exchangers. Background Technology

[0002] Welded pressure plate heat exchangers employ a fully welded process and a gasket-free design, enabling them to withstand high pressure and high temperature, making them suitable for harsh operating conditions. They offer high heat transfer efficiency, a compact structure, corrosion resistance, and low maintenance costs, and are widely used in heat exchange applications in chemical, energy, and other fields.

[0003] Patent application No. 202510322055.2 discloses a visual inspection method for heat exchanger welding quality, comprising the following steps: collecting raw image data from welding images; preprocessing the raw image data, including noise removal and brightness adjustment, to obtain a preprocessed image; applying an iterative deconvolution method to the preprocessed image to obtain a super-resolution image through continuous image reconstruction and detail optimization; based on the super-resolution image, performing multi-level decomposition of the image using a Laplacian pyramid to extract image features at each scale level, generating a multi-scale feature set; analyzing the information in the multi-scale feature set; and generating a comprehensive analysis result by comparing image details at different levels. Based on the comprehensive analysis results, the types and severity of welding defects are classified and quantitatively analyzed. The size and shape of the defect features are calculated to obtain defect feature data. The defect feature data is compared with known defects to determine the defect type and generate welding defect identification results. The image acquisition frequency is adjusted in real time according to the welding defect identification results to generate dynamic sampling parameters. This application aims to solve the problem that "in the prior art, the ability to remove high-frequency noise pixels in the image is insufficient, resulting in insufficient reliability of subsequent processing results. Moreover, there is a lack of effective super-resolution reconstruction operations, making it difficult to effectively optimize the texture details of the welding area image, which often leads to misjudgment in the defect identification process due to blurred image details."

[0004] However, on the heat exchanger assembly line, due to the large number of parts in the heat exchanger assembly, the overall quality control of the assembly is relatively more difficult, and there are few existing technologies for quality control and testing of the heat exchanger as a whole.

[0005] To this end, we propose an intelligent quality control system and method for the manufacture of fully welded pressure plate heat exchangers. Summary of the Invention

[0006] In view of the above-mentioned shortcomings of the prior art, the present invention provides an intelligent quality control system and method for manufacturing fully welded pressure plate heat exchangers, which can effectively solve the problems of the prior art.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions;

[0008] This invention discloses an intelligent quality control system for the manufacture of fully welded pressure plate heat exchangers, comprising:

[0009] The system comprises the following modules: a detection module for detecting surface structural parameters of the heat exchanger and constructing a 3D model of the heat exchanger based on these parameters; a derivation module for receiving the 3D model of the heat exchanger, picking incomplete models from its surface, and deriving bounding boxes for each incomplete model; a comparison module for receiving the bounding box groups derived by the derivation module, comparing them with a reference bounding box group, and identifying the overall similarity between the two groups; a judgment module for setting a pass / fail threshold, obtaining the overall similarity recognition result from the comparison module, and determining whether the heat exchanger corresponding to the source 3D model of the derived bounding box group is qualified based on the comparison result and the pass / fail threshold; and an indication module for triggering operation when the judgment module's judgment result is negative, indicating the model faces corresponding to each bounding box in the derived bounding box group on the 3D model of the heat exchanger.

[0010] In this design, all bounding boxes are set to be cuboids.

[0011] Furthermore, the detection module is integrated with a detection field and an infrared ranging module. The detection field is a ground with a circular guide rail. The infrared ranging module is integrated with an inverted L-shaped frame. The bottom end of the inverted L-shaped frame is embedded in the circular guide rail and moves in a circle along the circular guide rail by electric drive. The length of the top horizontal bar of the inverted L-shaped frame is the radius of the inscribed circle of the circular guide rail. The infrared ranging module is integrated with several sets of infrared ranging sensors. The several sets of infrared ranging sensors are equidistantly deployed on the two inner sides of the inverted L-shaped frame at the concave corner position. The distance between adjacent infrared ranging sensors is no more than 0.1 cm.

[0012] The detection module is equipped with a camera module and a transmission module. During the circular motion of the inverted L-shaped frame rod in the annular guide rail driven by electricity, all infrared ranging sensors on the inverted L-shaped frame rod continuously and synchronously operate based on a specified frequency until the inverted L-shaped frame rod completes one revolution. Then all moving and running components stop moving and running to obtain a ranging result data packet.

[0013] Each ranging result in the ranging result data packet is marked with the location information of its source infrared ranging sensor and a ranging timestamp. The ranging frequency of the infrared ranging sensor is set to run at least once for every 0.1cm displacement.

[0014] Furthermore, the camera module is deployed directly above the output end of the heat exchanger assembly line. When the assembled heat exchanger is output from the output end of the heat exchanger assembly line, the camera module collects top-view image data of the heat exchanger and identifies the center point of the image data of the area where the heat exchanger is located. The transmission module is integrated by several robotic arms. After the center point is identified, the robotic arms grab the heat exchanger and transfer it to the testing field and place it. The placement position follows the following: draw a perpendicular line from the vertical corresponding point of the center point on the top surface of the heat exchanger to the ground of the testing field, and the perpendicular line intersects the center of the circle on the ground of the testing field.

[0015] After the heat exchanger completes the transfer and placement to the detection field, it then performs the operation of acquiring the ranging result data packet.

[0016] Furthermore, the 3D model of the heat exchanger in the detection module is constructed based on the detection result data package:

[0017] Each ranging result in the data packet is marked with the corresponding infrared ranging sensor position information. The corresponding points of each position information are picked in three-dimensional space. Based on each picked corresponding point, line segments are drawn along the ranging direction and combined with the ranging results. The ends of the line segments closest to the center of the ground of the detection field are connected to each other to obtain a closed three-dimensional model composed of multiple surfaces, which is called the three-dimensional model of the heat exchanger.

[0018] Furthermore, during the operation of the derived module, when picking up an incomplete model on the surface of the heat exchanger's 3D model, it follows the following rules:

[0019] Select at least three adjacent model faces on the surface of the heat exchanger 3D model. The edges of each selected model face are spliced ​​together, and the combination of selected model faces forms at least one corner. For the unclosed positions of the selected model faces, draw a plane based on the edge lines of the selected model faces to close the unclosed positions, thus obtaining an incomplete model.

[0020] The number of incomplete models picked up by the derivative module is based on a preset logic. When selecting a model face, it is based on the user's manual selection on the heat exchanger 3D model received by the derivative module, and the positional relationship of the selected incomplete model relative to the heat exchanger 3D model includes at least directly above, directly to the left, directly to the right, directly in front, and directly behind.

[0021] Furthermore, the preset logic used to determine the number of incomplete models to be picked is as follows:

[0022] ;

[0023] In the formula: This represents the final number of incomplete models picked. The number of incomplete models picked up based on preset logic;

[0024] ;

[0025] In the formula: This is the accuracy requirement coefficient; This represents the total number of faces on the 3D model of the heat exchanger. Let be the area of ​​the i-th face; The average surface area; This represents the total number of edges in the 3D model of the heat exchanger. Let j be the length of the j-th edge; The average side length;

[0026] Among them, the accuracy requirement coefficient ∈[0.5, 1], its value is defined by the system user, and follows the principle that the higher the heat exchanger quality control requirements, the larger the value, and vice versa. Indicates the redundancy of the model's face count. To indicate irregularities in surface shape. The sum of the three factors, representing the difference in side length, is used to indicate the overall complexity of the three-dimensional model of the heat exchanger.

[0027] Furthermore, the comparison module is interactively connected to a priori database, which is used to store standard heat exchanger 3D models. The construction method of the standard heat exchanger 3D model is consistent with the construction method of the heat exchanger 3D model in the detection module. During the operation of the comparison module, the standard heat exchanger 3D model in the priori database performs the same operation as the derivation module during the operation of the derivation module, picking up incomplete models on the surface of the heat exchanger 3D model to obtain standard bounding boxes. All standard bounding boxes are recorded as a reference bounding box group.

[0028] Both the derived bounding box group and the reference bounding box group are represented in three-dimensional space, and both retain the relative positional relationship and orientation of each bounding box in the bounding box group.

[0029] The combined similarity recognition logic of the two sets of bounding box groups is expressed as follows:

[0030] ;

[0031] In the formula: This represents the total number of combinations of any two bounding boxes in the derived bounding box group; The straight-line distance between the centroids of the two bounding boxes in the derived bounding box group is the vth combination. The straight-line distance between the centroids of the two bounding boxes in the derived bounding box group and the corresponding two bounding boxes in the reference bounding box group is given by the vth combination method. This represents the total number of bounding boxes in the derived bounding box group; Let r be the structural similarity between the r-th bounding box in the derived bounding box group and its corresponding bounding box in the reference bounding box group.

[0032] Furthermore, the aforementioned The quantization logic is as follows:

[0033] Align the centroids of the two bounding boxes in three-dimensional space;

[0034] ;

[0035] In the formula: This represents the overlapping area of ​​the two bounding box spatial projections; Let be the angle between the normal vector of model r at point x and the normal vector of model r′ at the corresponding point y; These are the curvature densities of models r and r′ at corresponding positions, respectively; Let r and r′ be the volumes of models r and r′, respectively. These are the total surface areas of models r and r′, respectively; This is the adjustment coefficient;

[0036] Among them, the adjustment coefficient >0, and obeys: The initial setting is 0.5. The more emphasis is placed on local detail matching during similarity calculation, the better. The smaller the value, the more emphasis is placed on overall scale matching. The larger the value;

[0037] When the indicator module indicates the model face corresponding to each bounding box in the derived bounding box group on the 3D model of the heat exchanger, it performs a rendering operation of a specified color on the model face corresponding to each bounding box to complete the indication.

[0038] Furthermore, the detection module is interconnected with a camera module and a transmission module via a wireless network. The detection module is also interconnected with a derivative module and a comparison module via a wireless network. The comparison module is interconnected with a priori database via a wireless network. The comparison module is also interconnected with a judgment module and an indication module via a wireless network.

[0039] On the other hand, an intelligent quality control method for manufacturing fully welded pressure plate heat exchangers includes:

[0040] The heat exchanger is transferred to the testing field, where its surface structural parameters are detected using an infrared ranging module. A 3D model of the heat exchanger is then constructed based on these parameters. Incomplete models are picked up on the surface of the 3D model, and bounding boxes for each incomplete model are derived, referred to as the derived bounding box group. The same operation as the detection of structural parameters before the 3D model construction is performed on a standard heat exchanger to further obtain standard bounding boxes, referred to as the reference bounding box group. The similarity between the derived bounding box group and the reference bounding box group is comprehensively evaluated at the spatial distribution and structural levels. A pass / fail threshold is set, and the pass / fail threshold is compared with the comprehensive similarity evaluation results to determine whether the heat exchanger corresponding to the 3D model of the heat exchanger from which the derived bounding box group originates is qualified. When a heat exchanger is determined to be unqualified, the model faces corresponding to each bounding box in the derived bounding box group are indicated on the surface of the 3D model of the heat exchanger.

[0041] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects:

[0042] This invention provides an intelligent quality control system and method for manufacturing fully welded pressure plate heat exchangers. During the execution of the system and method, infrared ranging sensors densely deployed on the annular guide rail, combined with position and timestamp information, measure the distance at least once every 0.1cm displacement, convert a large number of ranging results into three-dimensional spatial line segments and splice them into a closed three-dimensional model, thereby achieving high-precision digital reconstruction of the heat exchanger structure.

[0043] When picking incomplete models on a 3D model, the number of models to be picked is dynamically determined by a formula with a precision coefficient based on parameters such as the number of models, area, number of sides, and length, taking into account the corner features formed by adjacent model faces. This covers multiple directions and makes subsequent comparisons more consistent with the actual complexity of the model.

[0044] The comparison comprehensively considers the centroid distance relationship between bounding boxes and the structural similarity of individual bounding boxes. The structural similarity calculation integrates details such as spatial overlap, normal vector angle, and curvature density. Furthermore, the adjustment coefficient can balance the emphasis on local and overall matching, thereby improving the accuracy and flexibility of the judgment. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0046] Figure 1 This is a schematic diagram of an intelligent quality control system for the manufacture of fully welded pressure plate heat exchangers.

[0047] Figure 2 This is a flowchart illustrating an intelligent quality control method for manufacturing fully welded pressure plate heat exchangers.

[0048] Figure 3 This is a schematic diagram illustrating the structure and distribution posture of the infrared ranging module for detecting rain in the detection field in this invention.

[0049] Figure 4 This is a schematic diagram illustrating an example of the construction process of an incomplete model in this invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0051] The present invention will be further described below with reference to embodiments.

[0052] Example 1:

[0053] This embodiment describes an intelligent quality control system for the manufacture of fully welded pressure plate heat exchangers, such as... Figure 1 As shown, it includes:

[0054] The detection module is used to detect the surface structure parameters of the heat exchanger and construct a three-dimensional model of the heat exchanger based on the surface structure parameters.

[0055] The detection module integrates a detection field and an infrared ranging module. The detection field is a ground with a circular guide rail. The infrared ranging module is integrated with an inverted L-shaped frame. The bottom end of the inverted L-shaped frame is embedded in the circular guide rail and moves in a circle along the circular guide rail by electric drive. The length of the top horizontal bar of the inverted L-shaped frame is the radius of the inscribed circle of the circular guide rail. The infrared ranging module is integrated with several sets of infrared ranging sensors. Several sets of infrared ranging sensors are equally spaced on the two inner sides of the inverted L-shaped frame at the concave corners. The distance between adjacent infrared ranging sensors is no more than 0.1cm.

[0056] The detection module is equipped with a camera module and a transmission module. During the circular motion of the inverted L-shaped frame rod in the ring guide rail by electric drive, all infrared ranging sensors on the inverted L-shaped frame rod continuously and synchronously operate based on a specified frequency until the inverted L-shaped frame rod completes one revolution. Then all moving and running components stop moving and running to obtain a ranging result data packet.

[0057] Each ranging result in the ranging result data packet is marked with the location information of its source infrared ranging sensor and a ranging timestamp. The ranging frequency of the infrared ranging sensor is set to run at least once for every 0.1cm displacement.

[0058] The camera module is deployed directly above the output end of the heat exchanger assembly line. When the assembled heat exchanger is output from the output end of the heat exchanger assembly line, the camera module collects top view image data of the heat exchanger and identifies the center point of the image data of the area where the heat exchanger is located. The transmission module is integrated by several robotic arms. After the center point is identified, the robotic arms grab the heat exchanger and transfer it to the inspection field and place it. The placement position follows the following: draw a perpendicular line from the vertical corresponding point of the center point on the top surface of the heat exchanger to the ground of the inspection field, and the perpendicular line intersects the center of the circle on the ground of the inspection field.

[0059] After the heat exchanger completes the transfer and placement to the detection field, it then performs the operation of acquiring the ranging result data packet.

[0060] The 3D model of the heat exchanger in the detection module is constructed based on the detection result data package:

[0061] Each ranging result in the data packet is marked with the corresponding infrared ranging sensor position information. The corresponding points of each position information are picked in three-dimensional space. Based on each picked corresponding point, line segments are drawn along the ranging direction and combined with the ranging results. The ends of the line segments closest to the center of the ground of the detection field are connected to each other to obtain a closed three-dimensional model composed of multiple spliced ​​surfaces, which is called the three-dimensional model of the heat exchanger.

[0062] The derivation module is used to receive the 3D model of the heat exchanger, pick out the incomplete model on the surface of the 3D model of the heat exchanger, and derive the bounding box of the incomplete model based on the picked incomplete model.

[0063] During the derivation module's operation, when picking an incomplete model on the surface of the heat exchanger's 3D model, the following rules apply:

[0064] Select at least three adjacent model faces on the surface of the heat exchanger 3D model. The edges of each selected model face are spliced ​​together, and the combination of selected model faces forms at least one corner. For the unclosed positions of the selected model faces, draw a plane based on the edge lines of the selected model faces to close the unclosed positions, thus obtaining an incomplete model.

[0065] The number of incomplete models picked by the derivative module is based on a preset logic. When selecting a model face, it is based on the user's manual selection on the heat exchanger 3D model received by the derivative module. The positional relationship of the selected incomplete model relative to the heat exchanger 3D model includes at least the top, left, right, front, and back.

[0066] The default logic used when determining the number of incomplete models to pick is as follows:

[0067] ;

[0068] In the formula: This represents the final number of incomplete models picked. The number of incomplete models picked up based on preset logic;

[0069] ;

[0070] In the formula: This is the accuracy requirement coefficient; This represents the total number of faces on the 3D model of the heat exchanger. Let be the area of ​​the i-th face; The average surface area; This represents the total number of edges in the 3D model of the heat exchanger. Let j be the length of the j-th edge; The average side length;

[0071] Among them, the accuracy requirement coefficient ∈[0.5, 1], its value is defined by the system user, and follows the principle that the higher the heat exchanger quality control requirements, the larger the value, and vice versa. Indicates the redundancy of the model's face count. To indicate irregularities in surface shape. The average of the three factors, representing the difference in side length, is used to indicate the overall complexity of the three-dimensional model of the heat exchanger.

[0072] The above formula dynamically adjusts the number of samples based on the overall complexity of the heat exchanger's 3D model. The formula introduces a precision requirement coefficient K (which users can set according to the stringency of quality control), and calculates the average of the model's face redundancy, face shape irregularity, and edge length differences, multiplying this average by K to obtain the final number of samples. This design allows the number of samples to flexibly change according to the model's complexity and precision requirements, ensuring that sufficient key feature areas are covered during subsequent comparisons, thus improving the targeting of quality control.

[0073] The comparison module is used to receive the bounding box groups derived by the derivation module, compare them with the reference bounding box group, and identify the overall similarity between the two bounding box groups.

[0074] The comparison module has an interactive connection with a priori database, which stores standard heat exchanger 3D models. The construction method of the standard heat exchanger 3D model is the same as that of the heat exchanger 3D model in the detection module. During the operation of the comparison module, the standard heat exchanger 3D model in the priori database performs the same operation as the derivation module during the operation of the comparison module, picking up incomplete models on the surface of the heat exchanger 3D model to obtain standard bounding boxes. All standard bounding boxes are recorded as the reference bounding box group.

[0075] Both the derived bounding box group and the reference bounding box group are represented in three-dimensional space, and both retain the relative positional relationship and orientation of each bounding box in the bounding box group;

[0076] The logical representation of the comprehensive similarity recognition of two sets of bounding boxes is as follows:

[0077] ;

[0078] In the formula: This represents the total number of combinations of any two bounding boxes in the derived bounding box group; The straight-line distance between the centroids of the two bounding boxes in the derived bounding box group is the vth combination. The straight-line distance between the centroids of the two bounding boxes in the derived bounding box group and the corresponding two bounding boxes in the reference bounding box group is given by the vth combination method. This represents the total number of bounding boxes in the derived bounding box group; Let r be the structural similarity between the r-th bounding box in the derived bounding box group and its corresponding bounding box in the reference bounding box group;

[0079] The above formula comprehensively measures the degree of matching from both relative position and structural details. In the formula, on the one hand, the difference in the centroid distance between the two bounding boxes under different combinations is calculated, and on the other hand, the structural similarity of each bounding box is included. This design takes into account both the correspondence of the overall spatial position and the matching of the local structure, so that the comprehensive similarity result can fully reflect the overall fit of the model.

[0080] The quantization logic is as follows:

[0081] Align the centroids of the two bounding boxes in three-dimensional space;

[0082] ;

[0083] In the formula: This represents the overlapping area of ​​the two bounding box spatial projections; Let be the angle between the normal vector of model r at point x and the normal vector of model r′ at the corresponding point y; These are the curvature densities of models r and r′ at corresponding positions, respectively; Let r and r′ be the volumes of models r and r′, respectively. These are the total surface areas of models r and r′, respectively; This is the adjustment coefficient;

[0084] Among them, the adjustment coefficient >0, and obeys: The initial setting is 0.5. The more emphasis is placed on local detail matching during similarity calculation, the better. The smaller the value, the more emphasis is placed on overall scale matching. The larger the value;

[0085] The above formula is used to quantify the structural similarity between the derived bounding box and the reference bounding box. It integrates matching information of local details and overall scale. In the formula, the overlapping area of ​​spatial projection reflects the overall positional coincidence, the product of the normal vector angle and curvature density reflects the consistency of local microstructures, and the difference in volume and surface area is introduced to measure the degree of matching of the overall scale. An adjustment coefficient is also used. The emphasis on local details or overall scale can be adjusted according to needs. This design can flexibly adapt to the needs of different quality control scenarios, making the calculation of structural similarity more in line with actual quality control standards.

[0086] When the indicator module indicates the model face corresponding to each bounding box in the derived bounding box group on the 3D model of the heat exchanger, it performs a color-specified rendering operation on the model face corresponding to each bounding box to complete the indication.

[0087] The judgment module is used to set the pass / fail judgment threshold, obtain the comprehensive similarity recognition result from the comparison module, and determine whether the heat exchanger corresponding to the 3D model of the heat exchanger from the source of the derived bounding box group is qualified based on the comparison of the recognition result and the pass / fail judgment threshold.

[0088] The indicator module is used to trigger operation when the determination module determines the result as negative, and to indicate the model face of each bounding box in the derived bounding box group on the 3D model of the heat exchanger.

[0089] In this case, all bounding boxes are set to be cuboids;

[0090] The detection module has a camera module and a transmission module connected to it via a wireless network. The detection module also has a derivative module and a comparison module connected to it via a wireless network. The comparison module has a judgment module and an indication module connected to it via a wireless network.

[0091] In this embodiment, the detection module detects the surface structure parameters of the heat exchanger and constructs a three-dimensional model of the heat exchanger based on these parameters. The derivation module then receives the three-dimensional model of the heat exchanger, picks up incomplete models on the surface of the three-dimensional model, and derives bounding boxes for each incomplete model based on the picked incomplete models. The comparison module further receives the bounding box groups derived by the derivation module and compares them with a reference bounding box group to identify the overall similarity between the two bounding box groups. The judgment module then sets a pass / fail threshold and obtains the overall similarity recognition result from the comparison module. Based on the recognition result and the pass / fail threshold, it determines whether the heat exchanger corresponding to the source three-dimensional model of the heat exchanger from the derived bounding box group is qualified. Finally, the indication module triggers operation when the judgment module's judgment result is negative, indicating the model faces corresponding to each bounding box in the derived bounding box group on the three-dimensional model of the heat exchanger.

[0092] The system in the above embodiments constructs a three-dimensional model of the heat exchanger through precise detection, generates corresponding bounding boxes by combining model parts selected from multiple angles, and compares them with standard data to determine whether they are qualified or not. If they are unqualified, the system can also clearly indicate the problem area. This can significantly improve the quality control efficiency and accuracy of all-welded pressure plate heat exchangers, reduce human judgment errors, lower the risk of defective products flowing out, and at the same time help to quickly locate defects, providing a basis for production improvement and ensuring product quality stability.

[0093] See Figure 3 As shown, further explanation is provided regarding the following: "The detection field is a ground surface with a circular guide rail. The infrared ranging module is integrated with an inverted L-shaped frame. The bottom end of the inverted L-shaped frame is embedded in the circular guide rail. It moves in a circle along the circular guide rail via electric drive. The length of the top horizontal bar of the inverted L-shaped frame is the radius of the inscribed circle of the circular guide rail. The infrared ranging module is integrated with several sets of infrared ranging sensors. Several sets of infrared ranging sensors are equidistantly deployed on the two inner sides of the inverted L-shaped frame at the location of the concave corner. The distance between adjacent infrared ranging sensors is no greater than 0.1cm." This explanation is intended to assist users implementing this technical solution in understanding and correctly deploying the infrared ranging module in the detection field.

[0094] See Figure 4 As shown, Figure (A) displays the three model faces used to determine the incomplete model, distinguished by the reference numerals 1, 2, and 3. Figure (B) shows the inner faces of the three model faces in their combined state through different types of fill. Figure (C) shows how to "draw a plane based on the selected model face edges to close the unclosed positions, thus obtaining the incomplete model" using another fill mode. In other words, the model represented by Figure (C) is the "incomplete model" after the unclosed positions have been closed.

[0095] Example 2:

[0096] At the implementation level, based on Example 1, this example refers to... Figure 2 A further detailed description is provided of the intelligent quality control system for manufacturing all-welded pressure plate heat exchangers in Example 1:

[0097] A smart quality control method for manufacturing all-welded pressure plate heat exchangers includes the following steps:

[0098] The heat exchanger is transferred to the testing field to detect the surface structure parameters of the heat exchanger using an infrared ranging module, and a three-dimensional model of the heat exchanger is constructed based on the surface structure parameters of the heat exchanger.

[0099] Pick incomplete models on the surface of the heat exchanger 3D model, derive bounding boxes for each incomplete model, and call them the derived bounding box group;

[0100] Perform the same operations as those performed on the standard heat exchanger before constructing the 3D model of the heat exchanger to further obtain the standard bounding box, denoted as the reference bounding box group;

[0101] The similarity between the derived bounding box group and the reference bounding box group is comprehensively evaluated at the spatial distribution and structural levels.

[0102] Set a pass / fail threshold, and compare the pass / fail threshold with the comprehensive similarity evaluation results to determine whether the heat exchanger corresponding to the 3D model of the heat exchanger from the source of the derived enclosing box group is qualified.

[0103] When a heat exchanger is deemed unqualified, the corresponding model faces of each bounding box in the derived bounding box group are indicated on the surface of the heat exchanger's 3D model.

[0104] In summary, during the execution of the system and method in the above embodiments, the infrared ranging sensors densely deployed on the circular guide rail, combined with position and timestamp information, measure distances at least once every 0.1cm displacement, converting a large number of ranging results into three-dimensional spatial line segments and splicing them into a closed three-dimensional model, thereby achieving high-precision digital reconstruction of the heat exchanger structure. When picking incomplete models on the three-dimensional model, the number of models is dynamically determined by a formula containing a precision coefficient, based on the angular features formed by adjacent model faces and parameters such as the number of model faces, area, number of sides, and length, and covers multiple directions, making subsequent comparisons more consistent with the actual complexity of the model. During comparison, the relationship between the centroids of bounding boxes and the structural similarity of individual bounding boxes are comprehensively considered. The structural similarity calculation integrates details such as spatial overlap, normal vector angle, and curvature density. Furthermore, the emphasis on local and overall matching can be balanced by adjusting the coefficients, improving the accuracy and flexibility of the judgment.

[0105] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent quality control system for manufacturing all-welded pressure plate heat exchangers, characterized in that, include: The detection module is used to detect the surface structure parameters of the heat exchanger and construct a three-dimensional model of the heat exchanger based on the surface structure parameters. The derivation module is used to receive the 3D model of the heat exchanger, pick out the incomplete model on the surface of the 3D model of the heat exchanger, and derive the bounding box of the incomplete model based on the picked incomplete model. The comparison module is used to receive the bounding box groups derived by the derivation module, compare them with the reference bounding box group, and identify the overall similarity between the two bounding box groups. The judgment module is used to set the pass / fail judgment threshold, obtain the comprehensive similarity recognition result from the comparison module, and determine whether the heat exchanger corresponding to the 3D model of the heat exchanger from the source of the derived bounding box group is qualified based on the comparison of the recognition result and the pass / fail judgment threshold. The indicator module is used to trigger operation when the determination module determines the result as negative, and to indicate the model face of each bounding box in the derived bounding box group on the 3D model of the heat exchanger. In this case, all bounding boxes are set to be cuboids; During the operation of the derived module, when picking an incomplete model on the surface of the heat exchanger's 3D model, the following applies: Select at least three adjacent model faces on the surface of the heat exchanger 3D model. The edges of each selected model face are spliced ​​together, and the combination of selected model faces forms at least one corner. For the unclosed positions of the selected model faces, draw a plane based on the edge lines of the selected model faces to close the unclosed positions, thus obtaining an incomplete model. The number of incomplete models picked up by the derivative module is based on a preset logic. When selecting a model face, it is based on the user's manual selection on the heat exchanger 3D model received by the derivative module. The positional relationship of the selected incomplete model relative to the heat exchanger 3D model includes at least the top, left, right, front, and back. The preset logic used to determine the number of incomplete models to be picked is as follows: ； In the formula: This represents the final number of incomplete models picked. The number of incomplete models picked up based on preset logic; ； In the formula: This is the accuracy requirement coefficient; This represents the total number of faces on the 3D model of the heat exchanger. Let be the area of ​​the i-th face; The average surface area; This represents the total number of edges in the 3D model of the heat exchanger. Let j be the length of the j-th edge; The average side length; Among them, the accuracy requirement coefficient ∈[0.5, 1], its value is defined by the system user, and follows the principle that the higher the heat exchanger quality control requirements, the larger the value, and vice versa. Indicates the redundancy of the model's face count. To indicate irregularities in surface shape. The average of the three factors, representing the difference in side length, is used to indicate the overall complexity of the three-dimensional model of the heat exchanger. The comparison module is interactively connected to a priori database, which is used to store standard heat exchanger 3D models. The construction method of the standard heat exchanger 3D model is the same as that of the heat exchanger 3D model in the detection module. During the operation of the comparison module, the standard heat exchanger 3D model in the priori database performs the same operation as the derivation module during the operation of the derivation module, picking up incomplete models on the surface of the heat exchanger 3D model to obtain standard bounding boxes. All standard bounding boxes are recorded as a reference bounding box group. Both the derived bounding box group and the reference bounding box group are represented in three-dimensional space, and both retain the relative positional relationship and orientation of each bounding box in the bounding box group. The combined similarity recognition logic of the two sets of bounding box groups is expressed as follows: ； In the formula: This represents the total number of combinations of any two bounding boxes in the derived bounding box group; The straight-line distance between the centroids of the two bounding boxes in the derived bounding box group is the vth combination. The straight-line distance between the centroids of the two bounding boxes in the derived bounding box group and the corresponding two bounding boxes in the reference bounding box group is given by the vth combination method. This represents the total number of bounding boxes in the derived bounding box group; Let r be the structural similarity between the r-th bounding box in the derived bounding box group and its corresponding bounding box in the reference bounding box group; The The quantification logic is as follows: Align the centroids of the two bounding boxes in three-dimensional space; ； In the formula: This represents the overlapping area of ​​the two bounding box spatial projections; Let be the angle between the normal vector of model r at point x and the normal vector of model r′ at the corresponding point y; These are the curvature densities of models r and r′ at corresponding positions, respectively; Let r and r′ be the volumes of models r and r′, respectively. These are the total surface areas of models r and r′, respectively; This is the adjustment coefficient; Among them, the adjustment coefficient >0, and obeys: The initial setting is 0.

5. The more emphasis is placed on local detail matching during similarity calculation, the better. The smaller the value, the more emphasis is placed on overall scale matching. The larger the value; When the indicator module indicates the model face corresponding to each bounding box in the derived bounding box group on the 3D model of the heat exchanger, it performs a rendering operation of a specified color on the model face corresponding to each bounding box to complete the indication.

2. The intelligent quality control system for manufacturing all-welded pressure plate heat exchangers according to claim 1, characterized in that, The detection module is integrated with a detection field and an infrared ranging module. The detection field is a ground with a circular guide rail. The infrared ranging module is integrated with an inverted L-shaped frame. The bottom end of the inverted L-shaped frame is embedded in the circular guide rail and moves in a circle along the circular guide rail by electric drive. The length of the top horizontal bar of the inverted L-shaped frame is the radius of the inscribed circle of the circular guide rail. The infrared ranging module is integrated with several sets of infrared ranging sensors. The several sets of infrared ranging sensors are equally spaced on the two inner sides of the inverted L-shaped frame at the concave corner position. The distance between adjacent infrared ranging sensors is no more than 0.1 cm. The detection module is equipped with a camera module and a transmission module. During the circular motion of the inverted L-shaped frame rod in the annular guide rail driven by electricity, all infrared ranging sensors on the inverted L-shaped frame rod continuously and synchronously operate based on a specified frequency until the inverted L-shaped frame rod completes one revolution. Then all moving and running components stop moving and running to obtain a ranging result data packet. Each ranging result in the ranging result data packet is marked with the location information of its source infrared ranging sensor and a ranging timestamp. The ranging frequency of the infrared ranging sensor is set to run at least once for every 0.1cm displacement.

3. The intelligent quality control system for manufacturing all-welded pressure plate heat exchangers according to claim 2, characterized in that, The camera module is deployed directly above the output end of the heat exchanger assembly line. When the assembled heat exchanger is output from the output end of the heat exchanger assembly line, the camera module collects top-view image data of the heat exchanger and identifies the center point of the image data of the area where the heat exchanger is located. The transmission module is integrated by several robotic arms. After the center point is identified, the robotic arms grab the heat exchanger and transfer it to the testing field and place it. The placement position follows the following: draw a perpendicular line from the vertical corresponding point of the center point on the top surface of the heat exchanger to the ground of the testing field, and the perpendicular line intersects the center of the circle on the ground of the testing field. After the heat exchanger completes the transfer and placement to the detection field, it then performs the operation of acquiring the ranging result data packet.

4. The intelligent quality control system for manufacturing all-welded pressure plate heat exchangers according to claim 2, characterized in that, The 3D model of the heat exchanger in the detection module is constructed based on the detection result data package. Each ranging result in the data packet is marked with the corresponding infrared ranging sensor position information. The corresponding points of each position information are picked in three-dimensional space. Based on each picked corresponding point, line segments are drawn along the ranging direction and combined with the ranging results. The ends of the line segments closest to the center of the ground of the detection field are connected to each other to obtain a closed three-dimensional model composed of multiple surfaces, which is called the three-dimensional model of the heat exchanger.

5. The intelligent quality control system for manufacturing all-welded pressure plate heat exchangers according to claim 1, characterized in that, The detection module is connected to a camera module and a transmission module via a wireless network. The detection module is also connected to a derivative module and a comparison module via a wireless network. The comparison module is connected to a priori database via a wireless network. The comparison module is also connected to a judgment module and an indication module via a wireless network.

6. An intelligent quality control method for manufacturing fully welded pressure plate heat exchangers, wherein the method is an implementation method of the intelligent quality control system for manufacturing fully welded pressure plate heat exchangers as described in any one of claims 1-5, characterized in that, Includes the following steps: The heat exchanger is transferred to the testing field to detect the surface structure parameters of the heat exchanger using an infrared ranging module, and a three-dimensional model of the heat exchanger is constructed based on the surface structure parameters of the heat exchanger. Pick incomplete models on the surface of the heat exchanger 3D model, derive bounding boxes for each incomplete model, and call them the derived bounding box group; Perform the same operations as those performed on the standard heat exchanger before constructing the 3D model of the heat exchanger to further obtain the standard bounding box, denoted as the reference bounding box group; The similarity between the derived bounding box group and the reference bounding box group is comprehensively evaluated at the spatial distribution and structural levels. Set a pass / fail threshold, and compare the pass / fail threshold with the comprehensive similarity evaluation results to determine whether the heat exchanger corresponding to the 3D model of the heat exchanger from the source of the derived enclosing box group is qualified. When a heat exchanger is deemed unqualified, the corresponding model faces of each bounding box in the derived bounding box group are indicated on the surface of the heat exchanger's 3D model.

Images