A path planning method and welding system for suppressing tube sheet group seam welding deformation

Abstract

The present application belongs to the technical field related to welding processing, and discloses a path planning method for suppressing deformation of tube sheet group seam welding, the method comprising: obtaining initial point cloud data of the tube sheet component; performing point cloud registration, point cloud rotation, straight-through filtering and weld fitting on the initial point cloud data to obtain weld position information; defining a fitness function for the object to be processed based on the inherent thermal coupling equation of welding and in combination with the weld position information; and using a genetic algorithm to calculate and plan the optimal welding path with the goal of minimizing the total deformation and total path of the welding path. The present invention also discloses a corresponding welding system. Through the present invention, high-efficiency and low-deformation welding of various complex tube sheet group seams can be achieved, and compared with the existing technology, both welding quality and welding efficiency can be significantly improved, and therefore it is particularly suitable for modern manufacturing and processing of various complex tube sheet heat exchangers.

Description

Technical Field

[0001] The present application belongs to the technical field related to welding processing, and more specifically, relates to a path planning method and a welding system for suppressing deformation of tube sheet group seam welding. Background Art

[0002] Tube-sheet heat exchangers, condensers, steam generators, and other pressure vessels are the primary industrial products for heat exchange and play an important role in fields such as metallurgy, nuclear power, shipbuilding, and petrochemicals. The pipe joints in tube-sheet heat exchangers are all circumferential welds, and the number of pipe joints ranges from a few hundred to a few thousand. The welds are densely packed at the joints, and the quality of the welds directly determines the service life and performance of the heat exchanger. Currently, tube-sheet welding mostly relies on manual instruction, which is inefficient and can lead to severe deformation of the tube sheet due to improper welding sequence. Using machine vision for automatic welding positioning, constructing a welding path planning model to determine the optimal welding path, and monitoring tube sheet deformation during the welding process are effective measures to achieve automated welding of tube-sheet seams and suppress tube sheet deformation.

[0003] A search revealed that some related solutions have been proposed in the prior art, but they still have their own limitations or shortcomings. For example, patent CN112958973A, which invented a robot welding vision positioning device based on structured light 3D vision, uses structured light 3D surface scanning technology to scan welds. By segmenting and fitting the 3D point cloud, it extracts the weld trajectory and width, effectively helping robots achieve automated tube-sheet welding. However, this solution does not address measures for thermal deformation during tube-sheet welding.

[0004] Patent CN114571154A discloses a tube-to-sheet welding method and apparatus that uses a monocular area array camera to capture images of all steel tubes to be welded. The system then plans the welding path, achieving continuous welding of the tubes. This vision-based dual-channel tube-to-sheet welding method improves welding efficiency by welding two tubes simultaneously. However, it also fails to address measures to mitigate welding deformation.

[0005] For example, patent CN113399883A describes a method for suppressing weld deformation in heat exchanger tube sheets. This method involves preparing a weld deformation suppression tool before welding. After welding, an elastic sleeve is quickly inserted into the tube end of the heat exchange tube, ensuring that the sleeve's shoulder is in close contact with the tube end. The nut is then quickly tightened, and the weld deformation suppression tool is removed when the weld temperature drops to room temperature. However, this invention does not involve automated welding path planning, and its effectiveness in addressing weld deformation needs further improvement.

[0006] Therefore, there is an urgent need in this field to develop a path automatic planning method and system that can more effectively suppress the welding deformation of tube sheet group seams. Summary of the Invention

[0007] In response to the defects of the existing technology, the purpose of this application is to provide a path planning method and welding system for suppressing the deformation of tube sheet group seam welding. By redesigning its main process steps and functional modules, it is possible to achieve better process effects in suppressing the deformation of tube sheet group seam welding and obtaining automatic planning of the optimal welding path. It can realize high-efficiency and low-deformation welding of various complex tube sheet group seams. Compared with the existing technology, it can significantly improve both welding quality and welding efficiency. Therefore, it is particularly suitable for modern manufacturing and processing of various complex structure tube sheet heat exchangers.

[0008] To achieve the above objectives, according to one aspect of the present invention, the present application provides a path planning method for suppressing tube sheet group seam welding deformation, characterized in that the method comprises:

[0009] S1. Acquisition of point cloud data of tube sheet components

[0010] For the tube sheet component to be processed, the tube sheet seam is photographed in three dimensions at multiple positions and initial point cloud data is obtained.

[0011] S2. Acquisition of weld position information

[0012] The initial point cloud data is processed by point cloud registration, point cloud rotation, straight-through filtering, and weld fitting to obtain weld position information. This step includes the following sub-steps:

[0013] S21. Feature points are collected from the initial point cloud data, and the point clouds at different locations are roughly overlapped by matching the feature points to complete coarse registration. Then, based on the results of the coarse registration, the point cloud data are aligned to the same coordinate system to obtain a complete three-dimensional point cloud model of the tube sheet.

[0014] S22, rotating the tube sheet three-dimensional point cloud model so that the flat plate plane is parallel to the XOY plane in the coordinate system; fitting the flat plate plane and obtaining the corresponding rotation matrix; then multiplying the initial point cloud data obtained in step S1 by the rotation matrix to obtain a tube sheet three-dimensional point cloud model in which the flat plate plane is parallel to the XOY plane;

[0015] S23, performing a through-filter process on the tube sheet three-dimensional point cloud model obtained in step S22 to obtain a three-dimensional model of only the pipes, i.e., a three-dimensional pipe model;

[0016] S24, performing fitting processing on the pipeline three-dimensional model obtained in step S23 to obtain corresponding weld position information;

[0017] Step S3: Define fitness function

[0018] Based on the inherent thermal-mechanical coupling equation of welding and combined with the weld position information obtained above, a fitness function is defined for the object to be processed;

[0019] Step S4: Planning of optimal welding path

[0020] The goal is to minimize the total deformation and total path of the welding path, and a genetic algorithm is used to calculate and plan the optimal welding path.

[0021] As a further preference, in sub-step S21, the accelerated robust feature method (SURF) is preferably used to obtain point cloud feature points, and the initial transformation matrix is ​​estimated by matching the feature points so that the point clouds at different positions basically overlap and complete the coarse alignment; in addition, the iterative closest point method (ICP) is preferably used to align the point cloud data to the same coordinate system to obtain a complete tube sheet three-dimensional point cloud model.

[0022] As a further preferred embodiment, in sub-step S22, the least square method is preferably used to fit the flat plane, and the normal vector n1 of the plane is obtained, and then the angle between n1 and the normal vector n2 (0,0,1) of the XOY plane is obtained. And the rotation axis n3 = n1 × n2; then according to the rotation axis n3 and the angle Use the Rodrigues rotation formula to obtain the rotation matrix R:

[0023]

[0024] Where I is the 3×3 identity matrix and K is the antisymmetric matrix constructed by the rotation axis n3.

[0025] As a further preferred embodiment, in sub-step S23, the height value of the flat plane of the tube sheet three-dimensional point cloud model is preferably used as the minimum allowable value limit_Z of the straight-through filtering. min , while allowing a maximum value limit_Z max =limit_Z min +3mm, thereby obtaining the three-dimensional model of the pipeline.

[0026] As a further preferred embodiment, in sub-step S23, a statistical filtering process is preferably further included to remove noise points and outliers in the three-dimensional pipeline model.

[0027] As a further preferred embodiment, in sub-step S24, a density-based clustering algorithm (DBSCAN) is preferably used to segment the pipeline point cloud model by setting a cluster distance threshold and a minimum number of points in each cluster, and the Hough circle variation method is used to detect circular pipelines to obtain the center coordinates of the pipeline. i (x i ,y i); then calculate the corresponding pipe center coordinates O in the welding process coordinate system i '(x i ',y i '), thereby obtaining the weld position information.

[0028] As a further preferred embodiment, in step S3, according to the pipeline center coordinate O i '(x i ',y i '), calculate the temperature distribution, stress distribution and total distance of the welding path during the entire welding process, and define a fitness function for the processing object based on the total deformation and total distance of the welding path.

[0029] As further preferred, in step S4, the following sub-steps are preferably included:

[0030] S41, randomly generating a number of welding paths and using them as the initial population;

[0031] S42, calculating the fitness value of each welding path based on the fitness function defined in step S3;

[0032] S43, using a roulette wheel method to select a better welding path from each fitness value;

[0033] S44, using a partial matching intersection method, randomly select two welding paths from the result of sub-step S43, interchange the intersection points of the two paths, and adjust the remaining parts to form a new optimized welding path;

[0034] S45, performing a mutation operation to introduce a new welding path by randomly changing a path point in the welding path;

[0035] S46. Repeat the operations of the above sub-steps S44 and S45 to iteratively optimize the welding path until the fitness value reaches the fitness threshold, thereby obtaining the optimal solution of the welding path.

[0036] According to another aspect of the present invention, a corresponding welding system is also provided, characterized in that the system comprises:

[0037] The data acquisition unit includes a scanning robot and a surface structure 3D measuring instrument. The surface structure 3D measuring instrument is installed at the end of the scanning robot's mechanical arm. The scanning robot carries the surface structure 3D measuring instrument to a designated location to take pictures of the tube sheet component and obtain initial point cloud data.

[0038] A data processing unit, which obtains weld position information through image processing based on the initial point cloud data; then uses a genetic algorithm to calculate and plan the optimal welding path with the goal of minimizing the total deformation and total path of the welding path;

[0039] The welding unit includes a welding robot, a welding machine, a welding gun, and a wire feeder. The welding gun is connected to the end flange of the welding robot's mechanical arm. The welding robot is responsible for driving the welding gun to perform welding actions and path planning. The welding machine provides stable power supply and parameter control, and the wire feeder ensures a continuous supply of welding wire.

[0040] A welding control unit compiles industrial control instructions online based on the calculation results of the data processing unit, generates a robot program and sends it to the welding robot, thereby completing the entire welding process of the tube plate component.

[0041] In general, the above technical solutions conceived by this application have the following technical advantages compared with the existing technologies:

[0042] 1. This invention closely meets the welding requirements of various complex tube-sheet components. By optimizing the design of the method for obtaining welding position information, it can more accurately capture the weld position based on point cloud processing methods, thereby ensuring high-precision automated welding of tube-sheet components and improving welding efficiency.

[0043] 2. The present invention also uses a genetic algorithm to plan the optimal welding path, which can minimize the welding deformation while minimizing the total welding distance. It also suppresses deformation by optimizing the welding sequence, thereby further improving welding efficiency and quality.

[0044] 3. The present invention can achieve high-efficiency, low-deformation welding of various complex tube-sheet seams, and compared with the existing technology, it can significantly improve both welding quality and welding efficiency. Therefore, it is particularly suitable for modern manufacturing and processing of various complex tube-sheet heat exchangers. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flow chart of a path planning method for suppressing tube sheet group seam welding deformation according to the present invention;

[0046] Figure 2 This is a schematic diagram of the overall structure of an automated welding system according to a preferred embodiment of the present application;

[0047] Figure 3 is a flow chart for obtaining weld position information according to a preferred embodiment of the present application;

[0048] Figure 4 is a flow chart of a method for generating an optimal welding path based on a genetic algorithm according to a preferred embodiment of the present application;

[0049] Throughout the drawings, the same reference numerals are used to denote the same elements or structures, wherein:

[0050] 1-Welding robot; 2-Welding machine; 3-Welding gun; 4-Wire feeding mechanism; 5-Scanning robot; 6-Surface structure three-dimensional measuring instrument; 7-Data processing unit; 8-Welding control unit; 71-Display A; 72-Industrial computer A; 81-Display B; 82-Industrial computer B. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0052] It should be understood that expressions such as "include" and "may include" used in this application indicate the existence of the disclosed functions, operations, or constituent elements, and do not limit one or more additional functions, operations, and constituent elements. In this application, terms such as "include" and / or "have" may be interpreted as indicating a specific characteristic, number, operation, constituent element, component, or combination thereof, but may not be interpreted as excluding the existence or possibility of adding one or more other characteristics, numbers, operations, constituent elements, components, or combinations thereof.

[0053] It should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0054] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0055] In this application, unless otherwise specified or limited, the terms "mounted," "connected," "connect," "fixed," etc. should be understood broadly. For example, they can refer to fixed connection, detachable connection, or integral connection; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on specific circumstances.

[0056] Figure 1 FIG. 1 is a flow chart of a path planning method for suppressing tube sheet group seam welding deformation according to the present invention. Figure 1 As shown, the process mainly includes the following multiple processing steps, which will be combined with Figure 1 Let’s explain it in more detail.

[0057] Step 1: Obtaining point cloud data of tube sheet components.

[0058] In this step, for the tube sheet component to be processed, the tube sheet group seams are photographed in three dimensions at multiple positions, and initial point cloud data is obtained.

[0059] Step 2: Acquisition of weld position information.

[0060] In this step, the initial point cloud data is processed by point cloud registration, point cloud rotation, through filtering and weld fitting to obtain the weld position information. This step includes the following sub-steps:

[0061] First, feature points are collected from the initial point cloud data, and by matching the feature points, the point clouds at different positions are basically overlapped to complete the coarse registration. Then, based on the results of the coarse registration, the point cloud data are aligned to the same coordinate system to obtain a complete three-dimensional point cloud model of the tube sheet.

[0062] More specifically, according to a preferred embodiment of the present invention, for example, the accelerated robust feature method (SURF) can be used to obtain point cloud feature points, and the initial transformation matrix can be estimated by matching the feature points so that the point clouds at different positions basically overlap and rough alignment is completed; in addition, the iterative closest point method (ICP) is preferably used to align the point cloud data to the same coordinate system to obtain a complete three-dimensional point cloud model of the tube sheet.

[0063] Next, the tube sheet three-dimensional point cloud model is rotated so that the flat plate plane is parallel to the XOY plane in the above coordinate system; the flat plate plane is fitted and the corresponding rotation matrix is ​​obtained; then, the initial point cloud data obtained in step S1 is multiplied by the rotation matrix to obtain a tube sheet three-dimensional point cloud model in which the flat plate plane is parallel to the XOY plane.

[0064] More specifically, according to a preferred embodiment of the present invention, for example, the least squares method can be used to fit the flat plane, and the normal vector n1 of the plane can be obtained, and then the angle between n1 and the normal vector n2 (0,0,1) of the XOY plane can be obtained. And the rotation axis n3 = n1 × n2; then according to the rotation axis n3 and the angle Use the Rodrigues rotation formula to obtain the rotation matrix R:

[0065]

[0066] Where I is the 3×3 identity matrix and K is the antisymmetric matrix constructed by the rotation axis n3.

[0067] Next, the obtained tube sheet three-dimensional point cloud model is subjected to a straight-through filtering process to obtain a three-dimensional model of only the pipes, that is, a three-dimensional pipe model.

[0068] More specifically, according to a preferred embodiment of the present invention, the height value of the plate plane of the tube sheet three-dimensional point cloud model is used as the minimum allowable value limit_Z of the straight-through filtering. min , while allowing a maximum value limit_Z max =limit_Z min +3mm, thereby obtaining the three-dimensional model of the pipeline.

[0069] Next, the obtained three-dimensional pipeline model is fitted to obtain the corresponding weld position information.

[0070] More specifically, according to another preferred embodiment of the present invention, a density-based clustering algorithm (DBSCAN) is preferably used to segment the pipeline point cloud model by setting a cluster distance threshold and a minimum number of points in each cluster, and the Hough circle variation method is used to detect circular pipelines to obtain the center coordinates of the pipeline. i (x i ,y i ); then calculate the corresponding pipe center coordinates O in the welding process coordinate system i '(x i ',y i '), thereby obtaining the weld position information.

[0071] Step 3: Define the fitness function.

[0072] In this step, based on the inherent thermal-mechanical coupling equation of welding and combined with the weld position information obtained above, a fitness function is defined for the object to be processed.

[0073] As an illustrative example, the definition process of the above fitness function can be explained as follows. However, it should be pointed out that the definition of the fitness function can be flexibly set according to the needs of actual application scenarios. Its principles and related formulas are well known to those skilled in the art, so there is no need to elaborate on them.

[0074] Specifically, some relevant parameters can be defined first. For example, S = {s1, s2, ..., s n} represents the welding sequence, where s i is the number of the i-th pipeline, s i The coordinates of the center of the pipeline are (x i ,y i ). T(s i ) indicates welded pipes i Temperature distribution, σ(s i ) indicates welded pipes i The stress distribution of the weld seam is given by the weld seam, and k is the deformation coefficient, which is determined based on the welding material and welding process.

[0075] The relevant temperature distribution equation can be expressed based on the welding thermal-mechanical coupling equation and the Gaussian heat source model:

[0076]

[0077] Among them, T o is the ambient temperature; Q is the heat input; ρ is the material density; c p is the material specific heat capacity; α is the material thermal diffusion coefficient; t is time; (x i ,y i ) is the current welding pipeline s i The center position of the circle, (x j ,y j ) is the previous welded pipe s j The center position of the circle.

[0078] The relevant stress distribution equation σ(s i ) can be expressed as follows:

[0079] σ(s i )=Eα t (T(s i )-To)(1-v)

[0080] Where E is the elastic modulus of the material; αt is the linear expansion coefficient of the material; v is the Poisson's ratio of the material; T(s i ) is the welded pipe s i temperature.

[0081] The relevant deformation calculation can be expressed as follows:

[0082] d(s i)=k·(T(s i )+σ(s i ))

[0083] The associated total deformation D(S) can be expressed as follows:

[0084]

[0085] Furthermore, given that the number, size, and distribution of pipes within a given tubesheet are fixed, the total weld path distance only needs to account for the distance the welding torch travels when switching from the current welded pipe to the next target pipe. The safe height to which the welding torch is lifted during idle travel is consistent, so calculating the total weld path distance is equivalent to calculating the distance between the centers of the i-th welded pipe and the i+1-th welded pipe. Therefore, the weld path distances of adjacent welded pipes can be expressed as:

[0086]

[0087] The total distance of the relevant welding path L(S) can be expressed as follows:

[0088]

[0089] In summary, a fitness function F(s) for exemplary explanation can be expressed as follows:

[0090] F(s)=k1·L(P i )+k2·D(P i ) or F(s)=D(P i )

[0091] Where k1 and k2 are weight coefficients, representing the importance of path length and thermal deformation. Preferably, k2 should be weighted more heavily than k1 (quality outweighs efficiency). In some embodiments, the values ​​of k1 and k2 can be adjusted based on the quality stability coefficient of the production site. For example, if the finished product pass rate is consistently greater than 95%, the value of k1 can be appropriately increased.

[0092] Step 4: Planning the optimal welding path.

[0093] In this step, the goal is to minimize the total deformation and total path of the welding path, and a genetic algorithm is used to calculate and plan the optimal welding path.

[0094] According to another preferred embodiment of the present invention, the process of planning the optimal welding path can be preferably designed as follows:

[0095] First, several welding paths are randomly generated and used as the initial population;

[0096] Next, based on the fitness function defined above, the fitness value of each welding path is calculated;

[0097] Then, a roulette wheel method is used to select the optimal welding path from each fitness value;

[0098] Next, using a partial matching intersection method, two welding paths are randomly selected from the results of sub-step S43, the intersection points of the two paths are interchanged, and the remaining parts are adjusted to form a new optimized welding path;

[0099] Then, a mutation operation is performed to introduce a new welding path by randomly changing a path point in the welding path;

[0100] Finally, the above partial matching crossover and mutation operations are repeated to iteratively optimize the welding path until the fitness value reaches the fitness threshold, thereby obtaining the optimal solution of the welding path.

[0101] Figure 2 This is a schematic diagram of the overall structure of the automated welding system according to the preferred embodiment of the present application. Figure 2 The welding system will be explained in more detail.

[0102] like Figure 2 As shown in , the system includes a data acquisition unit, a data processing unit, a welding unit and a welding control unit, among which:

[0103] The data acquisition unit includes a scanning robot and a surface structure 3D measuring instrument. The surface structure 3D measuring instrument is installed at the end of the scanning robot's mechanical arm. The scanning robot carries the surface structure 3D measuring instrument to a designated location to take pictures of the tube sheet component and obtain initial point cloud data.

[0104] The data processing unit obtains weld position information through image processing based on the initial point cloud data. Then, with the goal of minimizing the total deformation and total path of the welding path, a genetic algorithm is used to calculate and plan the optimal welding path.

[0105] The welding unit includes a welding robot 1, a welding machine 2, a welding gun 3, and a wire feeder 4. The welding gun 3 is flange-connected to the end of the robot arm of the welding robot 1. The welding robot 1 is responsible for driving the welding gun 3 to perform welding actions and path planning. The welding machine 2 provides stable power supply and parameter control, and the wire feeder 4 ensures a continuous supply of welding wire.

[0106] The welding control unit compiles industrial control instructions online based on the calculation results of the data processing unit, and generates a robot program and sends it to the welding robot, thereby completing the entire welding process of the tube plate component.

[0107] The following is a specific processing example to more clearly illustrate the present invention. In this example, a small tube-sheet heat exchanger is welded. The tubesheet has 37 weld joints. The plate thickness is 30 mm, the plate diameter is 350 mm, the pipe outer diameter is 29 mm, and the wall thickness is 2 mm. The pipes are numbered from left to right and from top to bottom according to the natural number pattern: 1, 2, 3, 4, ..., 37.

[0108] Before welding, the data acquisition unit is started, and the scanning robot drives the surface structure 3D measuring instrument to take pictures at the positions just above the tube plate component, 45° to the left, 45° to the right, 45° to the front, and 45° to the back, collecting point cloud data at these five positions.

[0109] Next, the data acquisition unit transmits the collected point cloud data to the welding control unit, which processes the point cloud data based on the image processing algorithm. The SURF method is used to make the five groups of point clouds roughly overlap and complete the rough registration; the iterative closest point method (ICP) is used for iterative optimization to accurately align the five groups of point cloud data to the same coordinate system to obtain a complete and accurate tube sheet component point cloud model; the obtained point cloud model circular plate plane is rotated to be parallel to the XOY plane by left multiplication of the rotation matrix; according to the tube sheet size, the straight-through filtering method is used to set the limit_Z min =540mm, limit_Z min =543mm, and the pipeline point cloud data is obtained; statistical filtering is used to remove noise points and discrete points in the point cloud and smooth the point cloud data; finally, the center of the pipeline fitting contour circle is obtained based on the density clustering algorithm (DBSCAN) and the Hough circle change method, as shown in Figure 4 According to the calibration results of the surface structure 3D measuring instrument and the welding robot hand-eye, the obtained weld center coordinates are converted into coordinates in the welding robot coordinate system.

[0110] Next, the fitness function is defined based on the inherent thermal-mechanical coupling equation of welding. In this specific example, the fitness function can be expressed as:

[0111] F(s)=0.2·L(P i )+0.8·D(P i )

[0112] For example, the weight coefficients k1 and k2 of the total path length and total deformation in the fitness function can be set to 0.2 and 0.8 respectively. Other parameters are set as follows: ambient temperature T0 = 25°C; input heat Q = 2000J; material density ρ = 7850kg / m 3 ; Material specific heat capacity c p =500J / (kg·K); Material thermal diffusion coefficient α = 1.172e-5m 2 / s; elastic modulus E = 210 GPa; linear expansion coefficient α t =12e-6 / k; Poisson's ratio v=0.3; time t=10s; deformation coefficient k=0.1.

[0113] Then, the genetic algorithm is used to solve the optimal path of the welding robot.

[0114] 1. Define the goal. Determine the optimal welding path, taking the shortest welding path from the starting point to the end point of the welding robot and the minimum deformation during the welding process as the goal.

[0115] 2. Initialize the population. Assume that the size of the initial population is 10, and each path consists of a random arrangement of 1-37 weld points: S1 = {1, 2, 3, ..., 37}, S2 = {37, 36, 35, ..., 1}, S3 = {1, 3, 4, ..., 36, 2} ... S 10 ={1,3,5,…,37}.

[0116] 3. Based on the fitness function defined in S3, the fitness of each welding path in the population is calculated. The calculation results are shown in Table 1:

[0117] Initial path editing Total path length Total deformation Fitness value 1 4.183 2.3 2.6766 2 4.183 2.3 2.6766 3 5.463 1.8 2.5326 4 4.874 1.6 2.2548 5 4.783 1.4 2.0766 6 5.863 1.2 2.1326 7 5.342 1.8 2.5084 8 6.871 1.8 2.8142 9 6.793 1.6 2.6386 10 7.233 1.5 2.6466

[0118] Table 1

[0119] 4. Selection operation. Use the roulette wheel selection method to select the better individual according to the fitness value. According to the fitness value calculated in step 3, its reciprocal is the selection probability. The results are shown in Table 2:

[0120] Initial path number Fitness value Selection probability 1 2.6766 0.373608 2 2.6766 0.373608 3 2.5326 0.394851 4 2.2548 0.443498 5 2.0766 0.481556 6 2.1326 0.468911 7 2.5084 0.398661 8 2.8142 0.355341 9 2.6386 0.378989 10 2.6466 0.377843

[0121] Table 2

[0122] 5. Crossover operation. Use the partial matching crossover method to randomly select two welding paths and swap the intersection points of the two paths. For example, select path S1 and path S3 to cross, and select the intersection point {5,10} to obtain the child path: S 11 =[1,2,3,4,5,10,11,…,9,6,7,8], S 12 =[1,3,4,5,6,7,…,2].

[0123] 6. Mutation operation. Mutate the new individuals, assuming a mutation rate of 3%, and mutate one point for each individual. For example, after the path S1 = [1, 2, 3, 4, 5…, 37] is mutated, S 13 =[1,2,3,5,4,6,7…,37].

[0124] 7. Iterative optimization. Repeat the selection, crossover, and mutation operations, and the fitness value changes gradually decrease until the change is small, and the algorithm stops. After 200 generations of iterative optimization, the optimal path is obtained:

[0125] S = [19, 20, 13, 12, 18, 25, 26, 27, 21, 14, 8, 7, 6, 11, 17, 16, 10, 5, 1, 2, 3, 4, 9, 15, 22, 28, 33, 32, 31, 30, 24, 23, 29, 34, 35, 36, 37]. The total length of the path is L = 6.468 m, the total deformation is 0.8 mm, and the fitness value is f = 1.9336.

[0126] Finally, the welding control unit generates a welding program based on the above calculation results and sends it to the welding robot for execution to complete the welding of the tube-sheet components.

[0127] In summary, the path planning method and welding system for suppressing tube sheet group seam welding deformation designed in this application can achieve better process effects in suppressing tube sheet group seam welding deformation and obtaining automatic planning of the optimal welding path by redesigning its main process steps and functional modules. It can realize high-efficiency and low-deformation welding of various complex tube sheet group seams. Compared with the existing technology, it can significantly improve both welding quality and welding efficiency. Therefore, it is particularly suitable for modern manufacturing and processing of tube sheet heat exchangers with various complex structures.

[0128] It is easy for those skilled in the art to understand that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A path planning method for suppressing tube sheet group seam welding deformation, characterized in that: The method includes: S1. Acquisition of point cloud data of tube sheet components For the tube sheet component to be processed, the tube sheet seam is photographed in three dimensions at multiple positions and initial point cloud data is obtained. S2. Acquisition of weld position information The initial point cloud data is processed by point cloud registration, point cloud rotation, straight-through filtering, and weld fitting to obtain weld position information. This step includes the following sub-steps: S21. Feature points are collected from the initial point cloud data, and the point clouds at different locations are roughly overlapped by matching the feature points to complete coarse registration. Then, based on the results of the coarse registration, the point cloud data are aligned to the same coordinate system to obtain a complete three-dimensional point cloud model of the tube sheet. S22, rotating the tube sheet three-dimensional point cloud model so that the flat plate plane is parallel to the XOY plane in the coordinate system; fitting the flat plate plane and obtaining the corresponding rotation matrix; then multiplying the initial point cloud data obtained in step S1 by the rotation matrix to obtain a tube sheet three-dimensional point cloud model in which the flat plate plane is parallel to the XOY plane; S23, performing a through-filter process on the tube sheet three-dimensional point cloud model obtained in sub-step S22 to obtain a three-dimensional model of only the pipes, i.e., a three-dimensional pipe model; S24, performing fitting processing on the pipeline three-dimensional model obtained in sub-step S23 to obtain corresponding weld position information; Step S3: Define fitness function Based on the inherent thermal-mechanical coupling equation of welding and combined with the weld position information obtained above, a fitness function is defined for the object to be processed; Step S4: Planning of optimal welding path The goal is to minimize the total deformation and total path of the welding path, and a genetic algorithm is used to calculate and plan the optimal welding path.

2. The method according to claim 1, wherein In sub-step S21, the accelerated robust feature method (SURF) is used to obtain point cloud feature points, and the initial transformation matrix is ​​estimated by matching the feature points so that the point clouds at different positions are basically overlapped, thus completing the coarse registration; In addition, the iterative closest point (ICP) method is used to align the point cloud data to the same coordinate system to obtain a complete 3D point cloud model of the tube sheet.

3. The method according to claim 2, wherein In sub-step S22, the least squares method is used to fit the flat plane and the normal vector n1 of the plane is obtained. Then, the angle φ between n1 and the normal vector n2 (0,0,1) of the XOY plane and the rotation axis n3 = n1 × n2 are obtained. Then, the rotation matrix is ​​obtained using the Rodrigues rotation formula based on the rotation axis n3 and the angle φ. R : in, I is a 3×3 identity matrix, and K is an antisymmetric matrix constructed by the rotation axis n3.

4. The method according to claim 3, wherein In sub-step S23, the height value of the plate plane of the tube sheet three-dimensional point cloud model is used as the minimum allowable value limit_Z of the straight-through filtering. min , while allowing a maximum value limit_Z max =limit_Z min +3 mm, thereby obtaining the three-dimensional model of the pipeline.

5. The method according to claim 4, wherein In sub-step S23, a statistical filtering process is also included to remove noise points and outliers in the three-dimensional pipeline model.

6. The method according to claim 5, wherein In sub-step S24, the density-based clustering algorithm (DBSCAN) is used to segment the pipeline point cloud model by setting the cluster distance threshold and the minimum number of points in each cluster, and the Hough circle variation method is used to detect circular pipelines to obtain the center coordinates of the pipeline. i (x i ,y i ); Then calculate the corresponding pipe center coordinates O in the welding processing coordinate system i '(x i ',y i '), thereby obtaining the weld position information.

7. The method according to claim 6, wherein In step S3, according to the pipeline center coordinate O i '(x i ',y i '), calculate the temperature distribution, stress distribution and total distance of the welding path during the entire welding process, and define a fitness function for the processing object based on the total deformation and total distance of the welding path.

8. The method according to any one of claims 1 to 7, wherein: In step S4, the following sub-steps are included: S41, randomly generating a number of welding paths and using them as the initial population; S42, calculating the fitness value of each welding path based on the fitness function defined in step S3; S43, using a roulette wheel method to select a better welding path from each fitness value; S44, using a partial matching intersection method, randomly select two welding paths from the result of sub-step S43, interchange the intersection points of the two paths, and adjust the remaining parts to form a new optimized welding path; S45, performing a mutation operation to introduce a new welding path by randomly changing a path point in the welding path; S46. Repeat the operations of the above sub-steps S44 and S45 to iteratively optimize the welding path until the fitness value reaches the fitness threshold, thereby obtaining the optimal solution of the welding path.

9. A welding system using the method according to any one of claims 1 to 8, characterized in that: The system includes: The data acquisition unit includes a scanning robot and a surface structure 3D measuring instrument. The surface structure 3D measuring instrument is installed at the end of the scanning robot's mechanical arm. The scanning robot carries the surface structure 3D measuring instrument to a designated location to take pictures of the tube sheet component and obtain initial point cloud data. A data processing unit, which obtains weld position information through image processing based on the initial point cloud data; then uses a genetic algorithm to calculate and plan the optimal welding path with the goal of minimizing the total deformation and total path of the welding path; The welding unit includes a welding robot, a welding machine, a welding gun, and a wire feeder. The welding gun is connected to the end flange of the welding robot's mechanical arm. The welding robot is responsible for driving the welding gun to perform welding actions and path planning. The welding machine provides stable power supply and parameter control, and the wire feeder ensures a continuous supply of welding wire. A welding control unit compiles industrial control instructions online based on the calculation results of the data processing unit, generates a robot program and sends it to the welding robot, thereby completing the entire welding process of the tube plate component.

Images