A laser welding method for corrugated steel cylinder

By recording the structural characteristics and deformation sequence on the welding trajectory of corrugated steel cylinders, adjusting the welding trajectory and power, and optimizing the welding parameters using the mapping coefficient, the problem of uneven welding deformation of corrugated steel cylinders is solved, and the reasonable selection of welding sequence and path is achieved. Ensure peak and trough alignment and high-quality welding of the structure.

CN118682286BActive Publication Date: 2024-11-29武汉华源电力设计院有限公司
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202411194072.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-29
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

During the laser welding of corrugated steel cylinders, due to the misalignment of the peak and trough structures and the improper selection of welding sequences, the welding deformation is uneven, affecting the structure's load-bearing capacity and safety.

Method used

By recording the structural characteristics and deformation sequence on the welding trajectory, adjusting the welding trajectory and power, and continuously optimizing welding parameters using mapping coefficients to ensure reasonable selection of welding sequence and paths, reducing welding deformation, and ensuring peaks and trough alignment.

Benefits of technology

Effectively reduce welding deformation, improve the structural alignment and bearing capacity of corrugated steel cylinders, and ensure welding quality and safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118682286B_ABST
    Figure CN118682286B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of laser welding, and in particular to a laser welding method for a corrugated steel tube, comprising: welding along S1 in a welding trajectory; after welding is completed, determining a new local welding trajectory S2 from a position that has not been welded according to the deformation sequence and structural characteristics of each welded position; mapping the historical welding power to the welding power on S2 based on the deformation sequence of the welded position with the greatest similarity on S2 and through a mapping coefficient; adjusting the mapping coefficient according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2; continuing welding according to the welding power of S2 and the adjusted mapping coefficient until all welding positions are welded. The present invention avoids the problem of misalignment of the crests and troughs of the corrugated steel caused by welding deformation to a certain extent.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of laser welding, and in particular to a laser welding method for a corrugated steel cylinder. Background Art

[0002] With the development of cities, it is an inevitable trend for high-voltage overhead lines to be gradually buried underground. Urban power grid construction also needs to adopt the method of burying cables underground. The use of jacking method to build cable tunnels is the main construction method.

[0003] When the cable is in the power tunnel and operates normally, the electromagnetic induction voltage to the surrounding subways, gas pipelines, and oil pipelines is very small. However, when a single-phase grounding fault occurs, the generated induced voltage may exceed the allowable voltage requirement of the subway transponder equipment, causing signal malfunction. The induction voltage to the metal pipes of the gas pipeline and oil pipeline exceeds the requirement. Under the long-term effect of the interference voltage, it will cause AC corrosion of the pipe body, peeling of the anti-corrosion layer, hydrogen embrittlement of the pipeline metal, and in severe cases, major safety accidents.

[0004] In order to eliminate the electromagnetic interference that may be caused by cables in the top pipe tunnel, a circular metal steel plate is embedded in the top pipe wall according to the principle of "Faraday cage" to achieve complete shielding of electromagnetic induction. Therefore, it is urgent to carry out research on new corrugated steel tube concrete electromagnetic shielding pipes and use corrugated steel plates to replace some steel bars in traditional concrete top pipes.

[0005] When corrugated steel plates are used to weld into new corrugated steel cylinders, when the corrugated steel plate wave crests and troughs are accurately welded and the welding stress is evenly distributed, the quality of the corrugated steel plate circle cylinder can be good and the bearing capacity of the structure can be improved. However, in the specific welding process, due to the obvious wave crests and troughs of the corrugated steel plate and the small thickness of the corrugated steel plate, when the laser welding technology is used for welding, when the inappropriate laser power is used for welding, after laser heating or after the welding is completed and cooled, the welding positions with different structural characteristics will have different degrees of local deformation due to different thermal expansion and cooling contraction, which will lead to inaccurate welding of the wave crests and troughs of the corrugated steel plate. On this basis, improper selection of welding sequence and path will further amplify or strengthen the degree of welding deformation. For example, the first welding will constrain the subsequent welding part, which may cause uneven stress and deformation. Summary of the invention

[0006] In order to solve the above problems, the present invention provides a laser welding method for a corrugated steel cylinder.

[0007] The laser welding method of a corrugated steel cylinder of the present invention adopts the following technical scheme:

[0008] An embodiment of the present invention provides a laser welding method for a corrugated steel cylinder, the method comprising the following steps:

[0009] The path formed by the joint positions of several corrugated steel plates is recorded as a welding trajectory; the structural characteristics of each position on the welding trajectory before welding are obtained;

[0010] Welding is performed along a preset local welding track S1 in the welding track;

[0011] After welding is completed, a new local welding trajectory S2 and the welding power of S2 are obtained according to the welding power of S1 and the initial mapping coefficient, including:

[0012] The difference between the structural features of each welded position at all times after welding and the structural features before welding is recorded as the deformation sequence of each welded position; a new local welding trajectory S2 is determined from the position that has not been welded according to the deformation sequence of each welded position and the structural features; based on the deformation sequences of several welded positions with the greatest similarity to the structural features of each position on S2, the historical welding power including the welding power of S1 is mapped to the welding power on S2 through the mapping coefficient;

[0013] Welding is performed along S2 according to the welding power of S2. After welding is completed, the mapping coefficient is adjusted according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2;

[0014] A new local welding trajectory and the welding power on the new local welding trajectory continue to be obtained according to the welding power of S2 and the adjusted mapping coefficient until all welding positions are welded.

[0015] Preferably, the specific steps of obtaining the structural features are as follows:

[0016] For any position of the welding trajectory;

[0017] Obtain the closest distance a from the crest of the corrugated steel plate and the closest distance b from the trough of the corrugated steel plate, and simultaneously obtain the docking offset A of the crest closest to the position and the docking offset B of the trough closest to the position, and use [a, b, A, B] as the structural features of the position.

[0018] Preferably, the method of determining a new local welding trajectory S2 from a position that has not been welded according to the deformation sequence and structural characteristics of each position that has been welded includes the following specific steps:

[0019] For the differences of the structural features in the deformation sequence of each welded position, when the mean value of the differences of the structural features is less than a first preset threshold, each welded position is recorded as a reference position;

[0020] Among all positions that have not been welded, the positions with the greatest similarity to the structural features of each reference position are obtained respectively, and are recorded as the target positions of each reference position; according to the target positions of all reference positions, a local welding trajectory is selected from the welding trajectory, and is recorded as a new local welding trajectory S2.

[0021] Preferably, the deformation sequence of several welded positions with the greatest similarity to the structural features of each position on S2, and mapping the historical welding power including the welding power of S1 to the welding power on S2 through the mapping coefficient, includes the following specific steps:

[0022] The mapping coefficients include: mapping granularity x1, historical welding power acceptance x2;

[0023] The average welding power of the historical welding powers including the welding power of S1 is recorded as P1;

[0024] Several welded positions with the greatest structural feature similarity to each position on S2 are recorded as the power reference positions of each position on S2;

[0025] The mean value of the difference of the structural features in the deformation sequence of each power reference position is recorded as the power impact factor of each power reference position, and the power impact factors of the power reference positions at all positions on S2 are clustered into x1 categories; for any category, the mean value of the power impact factor in the category is recorded as the average power impact factor corresponding to each category;

[0026] The predicted power in each category is obtained according to the average power influencing factor corresponding to each category and the average welding power P1, and the welding power of S2 is obtained according to the predicted power in all categories, the average welding power P1 and the historical welding power acceptance x2.

[0027] Preferably, the step of adjusting the mapping coefficient according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2 includes the following specific steps:

[0028] The structural features of each position on S2 before welding are recorded as the first features; the structural features of each welded position outside S2 before welding are recorded as the second features;

[0029] Among the positions on S2 and the positions outside S2 that have been welded, if there is a pair of positions, denoted as p1 and p2, respectively, satisfying the maximum similarity between the first feature of p1 and the second feature of p2, then the positions p1 and p2 are denoted as a matching position pair, and the difference between the deformation sequences of p1 and p2 is denoted as the welding difference of the matching position pair after the welding power is adjusted;

[0030] The mapping coefficients are adjusted using the mean of the welding differences of all matching position pairs after the welding power is adjusted.

[0031] Preferably, the difference of the structural features is equal to the Euclidean distance between the structural features.

[0032] Preferably, the step of selecting a local welding track from the welding track according to the target positions of all reference positions, recorded as a new local welding track S2, includes the following specific steps:

[0033] For the path composed of unwelded positions on the welding trajectory, obtain any path containing the reference position, and the ratio of the length of any path to the number of reference positions on any path is recorded as the reference position density of any path. The path that is not connected with S1 and has the largest reference position density is recorded as the new local welding trajectory S2.

[0034] Preferably, the method of adjusting the mapping coefficient by using the mean value of the welding differences of all matching positions after the welding power is adjusted includes the following specific steps:

[0035] The mean value of the welding differences of all matching position pairs after welding power adjustment is recorded as F4;

[0036] When F4 is less than the preset lower threshold, the mapping coefficient is not adjusted; when F4 is greater than the preset upper threshold, x2 is reduced by a preset percentage and x1 is reduced by a preset value; when F4 is greater than or equal to the preset lower threshold and less than or equal to the preset upper threshold, x1 and x2 are randomly adjusted.

[0037] Preferably, the predicted power under each category is obtained according to the average power influence factor corresponding to each category and the average welding power P1, and the welding power of S2 is obtained according to the predicted power under all categories, the average welding power P1 and the historical welding power acceptance x2, including the following specific steps:

[0038] The formula for the predicted power P3 under each category is: , w1 represents the average power impact factor corresponding to each category;

[0039] The mean of the predicted power under all categories is recorded as , the calculation formula of welding power P2 of S2 is:

[0040] .

[0041] Preferably, the specific steps for obtaining A and B are as follows:

[0042] The depth map of the surface of several corrugated steel plates is obtained by using laser radar. The depth map contains the depth value of each position. The Hough line detection algorithm is used to obtain all line segments on the depth map. The line segment with the largest average depth value is regarded as the trough, and the line segment with the smallest average depth value is regarded as the peak.

[0043] The average of the distances between all adjacent troughs and peaks is recorded as the average half-period; for the two line segments closest to each position and belonging to the same peak or trough, when the distance between the nearest endpoints of the two line segments is less than the average half-period, the two line segments are recorded as misaligned line segments, and the distance between the misaligned line segments is recorded as the docking offset A or B.

[0044] The beneficial effects of the technical solution of the present invention are:

[0045] The present invention obtains a new local welding track S2 and the welding power thereon through the local welding track S1 and the welding power thereon, and continues to obtain the new local welding track and the welding power on the new local welding track until all welding positions are welded. In this process, the mapping coefficient is adjusted once after welding is completed. This process continuously selects new local welding tracks and welding powers, and at the same time, by continuously adjusting the mapping coefficient, it can ensure that the welding power selected for the next welding can further reduce the welding deformation than the welding power in the history. With the increase in the number of welding times and the increase in the number of welded corrugated steel cylinders, the mapping coefficient will be adjusted more reasonably, and the obtained welding power can ensure that the peaks and troughs of the corrugated steel cylinders are aligned after welding. At the same time, welding according to different welding tracks or welding sequences can further reduce welding deformation and ensure that the peaks and troughs are aligned. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 A flowchart of the steps of a laser welding method for a corrugated steel cylinder provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the laser welding method of a corrugated steel cylinder proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0049] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0050] The specific scheme of the laser welding method for a corrugated steel cylinder provided by the present invention is described in detail below with reference to the accompanying drawings.

[0051] First embodiment:

[0052] See also Figure 1 , which shows a flowchart of a laser welding method for a corrugated steel cylinder provided by an embodiment of the present invention, the method comprising the following steps:

[0053] Step S001: a path formed by the joint positions of a plurality of corrugated steel plates is recorded as a welding trajectory.

[0054] The corrugated steel cylinder is formed by splicing and welding a plurality of (for example, four) corrugated steel plates. The number of corrugated steel plates used in each corrugated steel cylinder is determined by the specific production and use of the corrugated steel cylinder, such as the diameter and length of the corrugated steel cylinder, or the width and height of each corrugated steel plate. This embodiment does not limit the number of corrugated steel plates.

[0055] Before welding, it is first necessary to align multiple corrugated steel plates, especially to align the peaks and valleys of the corrugated steel plates with each other. This embodiment is manually aligned and fixed, and other embodiments can use automated alignment methods, for example: multiple corrugated steel plates are fixed on a movable fixed table using clamps, and each corrugated steel plate image is captured using a camera, and the edge of the peaks and valleys of the corrugated steel plates is obtained using edge detection technology or Hough line detection technology. The peaks and valleys in the images of different corrugated steel plates are overlapped or aligned into a straight line by moving the fixed table, thereby achieving the alignment of the corrugated steel plates.

[0056] This embodiment does not limit the alignment method. It should be noted that during the alignment process, the periodic distribution lengths of the crests and troughs of different corrugated steel plates or the height differences of the crests and troughs may be inconsistent due to manufacturing errors, resulting in the inability to completely align the crests and troughs; or the corrugated steel plates may be slightly deformed during transportation or there are errors in the cutting dimensions of the corrugated steel plates during the previous cutting process, which will also cause the crests and troughs to be unable to be completely aligned. In this case, only the maximum number of crests and troughs need to be aligned.

[0057] In other embodiments, when the alignment result is unacceptable, part of the corrugated steel plate may be cut and corrected, and then realigned.

[0058] After multiple corrugated steel plates are welded, spliced ​​and aligned, the joint positions of different corrugated steel plates are obtained, and the path formed by the joint positions is recorded as a welding trajectory, that is, when laser welding is performed, the welding equipment performs welding along the welding trajectory.

[0059] The welding trajectory can be drawn by manual measurement or obtained through image processing technology or lidar technology.

[0060] Step S002: Obtain the structural features of each position on the welding trajectory before welding.

[0061] Taking into account that in the specific welding process, due to the obvious peaks and troughs of the corrugated steel plate (even the structures of the peaks and troughs on different corrugated steel plates are different), different from welding flat metal components, when welding, after laser heating or after cooling after welding, the welding positions with different structural characteristics will have different degrees of local deformation due to thermal expansion and cooling contraction (for example, the inner right angles and outer right angles on the peaks and troughs have different deformation conditions), which will lead to inaccurate welding of the peaks and troughs of the corrugated steel plate.

[0062] Therefore, before welding, it is first necessary to record the structural features of each position on the welding trajectory before welding, and the structural features are used to describe the structural characteristics of each position on the corrugated steel plate.

[0063] As a preferred example, the method for acquiring the structural features of each position includes:

[0064] For any position of the welding trajectory, obtain the closest distance a from the crest of the corrugated steel plate and the closest distance b from the trough of the corrugated steel plate. At the same time, obtain the docking offset A of the crest closest to the position and the docking offset B of the trough closest to the position, and use [a, b, A, B] as the structural features of the position.

[0065] The butt offset A and B refer to the offset distances of the crests and troughs of the two corrugated steel plates at the joint position. When the corrugated steel plates are completely aligned in step S001, the values ​​of A and B are 0. When there is an incomplete alignment, the crests and troughs of the two corrugated steel plates will be misaligned or offset, resulting in the crests or troughs of the two corrugated steel plates no longer being in a straight line. At this time, the misalignment or offset distance is used as A and B.

[0066] The above-mentioned a, b, A, and B can be obtained by image processing technology or lidar technology.

[0067] As an optional example, the method for obtaining the structural features of each position includes:

[0068] Only the above [A, B] will be taken as the structural feature of the position.

[0069] This method reduces the amount of data and computation by ignoring the distance of each position relative to the peaks and troughs.

[0070] As another example, the method for acquiring the structural features of each position includes:

[0071] Collect the image of the corrugated steel plate after alignment. If the collected image is a grayscale image, it needs to be grayed out. The grayscale distribution of each pixel in the grayscale image and the neighborhood of each pixel can represent the structural characteristics of each position on the corrugated steel plate. Specifically, the grayscale value of any position of the welding track in the preset neighborhood (for example, the grayscale value of all pixels in a 9×9 neighborhood) is obtained in the image, and the vector composed of these grayscale values ​​or the grayscale histogram composed of these grayscale values ​​is used as the structural feature of each position.

[0072] Since the grayscale distribution of pixels can better reflect the structural characteristics of each position on the corrugated steel plate, the structural features in this method are more accurate and specific. If there are defects such as rust or burrs and cracks on the corrugated steel plate (these defects will also affect the metal deformation during welding), the above method of obtaining structural features will be more reliable and accurate. However, the above method requires additional calculations.

[0073] Step S003, welding is performed along a preset local welding trajectory S1 in the welding trajectory.

[0074] Considering that different structural features may deform to different degrees during or after welding, improper selection of welding sequence and path may further amplify or strengthen the degree of welding deformation. Therefore, this embodiment first performs welding along a local welding track.

[0075] In this embodiment, a local welding track S1 is randomly selected on the welding track, and the length of the selected local welding track S1 is one tenth of the total length of the welding track. In other embodiments, other selection methods or length selections may be used, such as manually selecting a local welding track S1. This embodiment is not specifically limited.

[0076] In this embodiment, 3 kW laser power is used to perform welding along S1 and complete welding. In other embodiments, other welding powers may be selected. The specific welding power needs to be selected based on parameters such as the thickness of the corrugated steel plate. One method of selecting power is to obtain the welding power of the same laser welder in the history of welding corrugated steel plates of the same thickness or other components of the same thickness and the same material (the power when welding is unqualified is not considered), and the average of these powers is used as the welding power when welding along S1.

[0077] Step S004, the difference between the structural characteristics of each welded position at all times after welding and the structural characteristics before welding is recorded as the deformation sequence of each welded position; and a new local welding trajectory S2 is determined from the unwelded position according to the deformation sequence and structural characteristics of each welded position.

[0078] After welding along S1 is completed, for each welded position, the structural features at all times after welding are calculated separately from the structural features before welding. All the differences obtained constitute a time series, which is recorded as the deformation sequence of each welded position.

[0079] It should be noted that all moments after welding refer to all moments from welding to complete cooling. Infrared imaging technology can be used to collect the temperature of each position. When the temperature is equal to room temperature, it means that it has cooled down. For greater convenience in this embodiment, one minute after welding is completed is regarded as cooling completion. In addition, in this embodiment, each second is a moment.

[0080] It should be noted that, after obtaining the deformation sequence of each position, this embodiment performs linear normalization on all differences in the deformation sequence of each position to eliminate the dimension and order of magnitude of the data (other normalization methods may be used in other embodiments), and then uses the normalized deformation sequence for analysis and calculation.

[0081] In this embodiment, the above-mentioned difference refers to the Euclidean distance between the structural features. In other embodiments, the cosine distance or DTW distance between the structural features may be used as the difference between the structural features.

[0082] The deformation sequence of each position describes the change of the structural characteristics of each position on the weld of the corrugated steel plate over time after welding. If the above difference contained in the deformation sequence of each position is larger, or the mean of the above difference contained in the deformation sequence is larger, it means that the deformation of each position after welding is more serious, or the welding power is not properly selected.

[0083] Furthermore, after the welding of S1 is completed, it is necessary to reselect a new local welding track S2 from the welding track to continue welding. When reselecting a new local welding track S2 from the welding track, it is necessary to select according to the deformation characteristics (i.e., deformation sequence) and structural characteristics of the position that has been welded, so as to avoid the reappearance of severe deformation (such as that in S1) on the new local welding track S2.

[0084] As a preferred example, the specific steps of obtaining the local welding trajectory S2 are as follows:

[0085] The deformation sequence of each position contains the difference between the structural features at each moment after welding and before welding. The mean of all differences contained in the deformation sequence of each position is obtained. When the mean is less than th1, each welded position is recorded as a reference position; the reference position describes a position that has been welded and has not produced a large deformation. The structural characteristics of these positions are not easy to deform under the corresponding welding power, which is conducive to the alignment of the peaks and troughs of the corrugated steel plate.

[0086] This embodiment is described by taking th1=0.4 as an example. Other embodiments may be set to other values, which are not specifically limited in this embodiment.

[0087] Among all the unwelded positions, the positions with the greatest similarity in structural characteristics with each reference position are obtained respectively, which are recorded as the target positions of each reference position; the target positions describe the unwelded positions with similar structural characteristics to the welded positions that are not prone to serious deformation. These unwelded positions are structurally stable when welded and may not suffer serious deformation.

[0088] The similarity of the structural features refers to the cosine similarity of the structural features. In other embodiments, the position with the same structural features as each reference position, or the position with the smallest difference in structural features from each reference position, can be recorded as the target position of each reference position.

[0089] A local welding trajectory is selected from the welding trajectory according to the target positions of all reference positions and is recorded as a new local welding trajectory S2.

[0090] The above process focuses more on the overall deformation of each position after welding, which can avoid the situation where the selected S2 has a more serious overall deformation after welding.

[0091] As an optional example, the specific steps of obtaining the local welding trajectory S2 are as follows:

[0092] The maximum and minimum values ​​of all differences in the deformation sequence of each welded position are obtained respectively. The difference between the maximum and minimum values ​​can also indicate the structural stability of each position after welding. The smaller the difference, the less deformation will occur at each position after welding, and the more stable the structure is. At this time, the position with a difference less than th2 is recorded as the reference position. A new local welding trajectory S2 is further obtained based on the reference position.

[0093] This embodiment is described by taking th2=0.3 as an example. Other embodiments may be set to other values, which are not specifically limited in this embodiment.

[0094] This method pays more attention to the maximum deformation of each position at different times after welding, which can avoid the selected S2 from having a large deformation during the cooling process after welding.

[0095] As another optional example, the specific steps of obtaining the local welding trajectory S2 are as follows:

[0096] On the basis of the reference position obtained in the above preferred example, the reference position obtained in the above optional example is added, and a new local welding trajectory S2 is further obtained according to the reference position.

[0097] This example is applicable to the case where the number of reference positions obtained in the above preferred example is small (that is, the case where large deformation occurs in most positions after welding).

[0098] There are also some optional examples that can be used to obtain the local welding trajectory S2:

[0099] The difference corresponding to the last moment in the deformation sequence of each position is obtained, and when the difference is less than a first preset threshold, each welded position is recorded as a reference position. A new local welding trajectory S2 is further obtained based on the reference position.

[0100] This optional example mainly focuses on the final structural deformation after welding cooling (or after cooling for a period of time). Although the specific deformation process of the structural characteristics of each position after welding is not considered, it can also describe the stability of each position.

[0101] As an optional example, a local welding trajectory is selected from the welding trajectory according to the target positions of all reference positions and recorded as a new local welding trajectory S2, including the following method:

[0102] On the welding trajectory, a path that contains reference positions and the number of reference positions contained is greater than half of the number of all reference positions is randomly selected as a new local welding trajectory S2; if there is S2 that is not connected to S1, ensure that the new local welding trajectory S2 is not connected to S1, for example, the distance between the two nearest endpoints of S2 and S1 on the welding trajectory is greater than 10 cm.

[0103] As a preferred example, a local welding track is selected from the welding track according to the target positions of all reference positions, and is recorded as a new local welding track S2, including the following method:

[0104] For the path composed of unwelded positions on the welding trajectory, obtain any path containing the reference position, and the ratio of the length of each path to the number of reference positions on the path is recorded as the reference position density of any path. The path that is not connected to S1 (for example, the endpoint distance is greater than 10 cm) and has the largest reference position density is recorded as the new local welding trajectory S2.

[0105] Step S005: Based on the deformation sequences of several welded positions with the greatest similarity to the structural features of each position on S2, and through the mapping coefficients, the historical welding power including the welding power of S1 is mapped to the welding power on S2.

[0106] After S2 is obtained, in order to further ensure that deformation is reduced when welding S2, it is necessary to reselect the welding power for the welding process of S2. The acquisition of this welding power needs to be based on the welding situation under the welding power used in S1 (that is, the deformation sequence of several welded positions with the greatest similarity to the structural characteristics of each position on S2). If there are other corrugated steel cylinder welding processes before welding along S1, the acquisition needs to be based on the welding situation under the historical welding power including the welding power of S1.

[0107] The role of the mapping coefficient is to map the historical welding power including the welding power of S1 to the welding power on S2 according to the welding situation under the welding power used in S1, so as to ensure that the welding power on S2 is obtained on the basis of the welding power of S1 (and in history), while at the same time drawing on or discarding some good or bad welding features (such as the welding deformation features at each position, i.e., deformation sequence) in the welding process of S1 (and in history), thereby improving the welding effect of S2.

[0108] As an optional example, the specific method for obtaining the welding power on S2 includes:

[0109] The N1 welded positions with the greatest structural feature similarity to each position on S2 are recorded as the power reference positions of each position on S2. This embodiment is described by taking N1=2 as an example, and other values ​​can be set in other embodiments, which are not specifically limited in this embodiment.

[0110] It should be noted that the welded positions described in this embodiment include not only the welded positions in S1, but also the historical welded positions if S1 has been welded before. In other embodiments, only the most recent welding process (ie, the welding process of S1) may be considered.

[0111] According to the above, it can be seen that the deformation sequence of the power reference position at each position contains the differences in structural characteristics at different moments;

[0112] It is easy to think that if the deformation degree of the power reference position of each position is small (that is, the difference of the structural features in the deformation sequence is small), the welding power of S1 can be directly applied to S2; if the deformation degree of the power reference position of each position is large (that is, the difference of the structural features in the deformation sequence is large), the welding power of S1 can be fine-tuned to a large extent and then directly applied to S2. In summary, the welding power of S1 needs to be applied (or mapped) to S2 according to the size of the difference in the structural features in the deformation sequence.

[0113] The specific method is:

[0114] The mean of the differences in all structural features in the deformation sequence of each power reference position is recorded as the power influence factor of each power reference position, which is used to describe the influence of the welding power on the deformation at each power reference position under the welding power used in S1 (and in history). The larger the value, the greater the deformation influence of the welding power used in S1 (and in history) on each power reference position.

[0115] The mean of the power influence factors of the power reference positions corresponding to all positions on S2 is taken as the average power influence factor w. This value is used to describe the welding situation of the entire welding process under the welding power used in S1 (and in history). The larger the value, the more serious the overall deformation in the welding process of S1 (and in history), and the more it is necessary to make a larger adjustment to the welding power used in S1 (and in history). That is, based on the average power influence factor w, the average welding power used in S1 (and in history) (denoted as P1) is mapped to the welding power on S2 (denoted as P2). The specific formula is:

[0116]

[0117] Wherein k represents the mapping coefficient, which is used to describe the proportional coefficient when mapping P1 to P2. The role is to obtain P2 based on P1. It means that the welding power on S2 is obtained on the basis of the welding power of S1, while at the same time some good or bad welding features (that is, the average power influence factor w) in the S1 welding process can be borrowed or discarded, thereby improving the welding effect of S2.

[0118] After the first welding (i.e., welding along S1), the mapping coefficient k needs to specify an initial mapping coefficient. This embodiment takes the initial mapping coefficient k=0.8 as an example for description. Other embodiments may set it to other values. In subsequent welding processes, the mapping coefficient will be adjusted based on the initial mapping coefficient.

[0119] In the above optional example, although P2 is obtained based on P1, on the one hand, the average power influence factor w can only describe the overall welding deformation during the welding process of S1, ignoring the specific deformation differences between different power reference positions. For example, during the welding process of S1, some positions (or power reference positions) may have small deformation, while only a small number of positions have large deformation. At this time, the welding power used in the welding process of S1 does not need to be adjusted too much and can be directly applied to S2. For another example, the deformation conditions at different positions during the welding process of S1 are different (for example, there are multiple levels of deformation, such as mild deformation, slightly severe deformation, moderately severe deformation, and very severe deformation). At this time, when the above method maps P1 to P2, the mapping result is unreliable.

[0120] As a preferred example, the specific method for obtaining the welding power on S2 includes:

[0121] The mapping coefficient is divided into two parts, namely: mapping granularity x1 (an integer) and historical welding power acceptance x2; after the first welding (that is, after welding along S1), the initial mapping coefficients are set to x1=5 and x2=0.8, and can be set to other values ​​in other embodiments.

[0122] Several welded positions with the greatest structural feature similarity to each position on S2 are recorded as the power reference positions of each position on S2;

[0123] The mean of the difference in the structural features in the deformation sequence of each power reference position is recorded as the power influence factor of each power reference position. The power influence factors of the power reference positions at all positions on S2 are clustered into x1 categories using the K-Means clustering algorithm. Each category contains several power influence factors. These x1 categories correspond to the deformation conditions of different positions of S1 during the welding process.

[0124] For any category, the mean power impact factor in the category is recorded as the average power impact factor corresponding to each category, expressed as w1. Then, under each category, the formula for mapping P1 is: ;

[0125] P3 represents the predicted power of each category. The mean of the predicted power of all categories is recorded as , then the calculation formula of welding power P2 on S2 is: ,in It means that P2 is obtained based on P1, or it means the reference or acceptance of the welding power used in history. It means that the welding power on S2 is obtained on the basis of the welding power of S1 (and history), while at the same time, some good or bad welding features in the welding process of S1 (and history) can be borrowed or discarded, so as to improve the welding effect of S2.

[0126] As another example, the specific method for obtaining the welding power at S2 includes:

[0127] When corrugated steel cylinders are welded and manufactured multiple times in the historical process, all welded positions on the corrugated steel cylinders in the history are obtained, and the deformation sequences of these positions are used as samples. The welding powers of these positions when welded are used as labels, and the average of the welding powers of these positions when welded is recorded as the historical average power (that is, P1 mentioned above). All samples and labels constitute a data set, and an LSTM neural network is trained using the data set. The loss function used is the mean square error loss function.

[0128] The deformation sequence of each power reference position at each position on S2 is input into the trained neural network respectively to obtain the predicted power of each power reference position at each position on S2, and the mean of the predicted power of the power reference positions at all positions on S2 is obtained, which is recorded as .

[0129] At this time, the calculation method of welding power P2 on S2 is: .

[0130] in Represents the mapping coefficient.

[0131] In addition to requiring a large amount of computation, the above method also requires a large-scale data set. Therefore, when there are not many welding operations in the history, the method in this example cannot be used.

[0132] Step S006: adjusting the mapping coefficient according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2.

[0133] Considering that the welding power on S2 obtained above may not necessarily avoid welding deformation, it is necessary to adjust the mapping coefficient according to the welding situation after welding on S2 (that is, the deformation sequence of each position on S2) to ensure that the welding deformation problem can be further solved on the basis of the welding of S2 during subsequent welding.

[0134] As an optional example, the mapping coefficient is adjusted according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2, including the specific method of:

[0135] The mean of the differences of all structural features in the deformation sequence at each position is recorded as the power influence factor of each position;

[0136] After all positions on S2 are welded, the average of the power influence factors of all positions on S2 is recorded as F1, and the average of the power influence factors of all welded positions outside S2 is recorded as F2. According to the difference between F1 and F2, it is recorded as F3, and the mapping coefficient is adjusted using F3.

[0137] As a preferred example, the mapping coefficient is adjusted according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2, including the specific method of:

[0138] The structural features of each position on S2 before welding are recorded as the first features;

[0139] The structural features of each welded position other than S2 before welding are recorded as the second features;

[0140] Among the positions on S2 and the welded positions outside S2, if there is a pair of positions denoted as p1 and p2 respectively, and the similarity between the first feature of p1 and the second feature of p2 is the greatest, then the positions p1 and p2 are recorded as a matching position pair, which describes a pair of positions having the same or most similar structural features before welding.

[0141] The difference between the mean of the differences of all structural features in the deformation sequence of p1 and the mean of the differences of all structural features in the deformation sequence of p2 is calculated, which is recorded as the welding difference of the matching position pair after the welding power is adjusted, and is simply recorded as the deformation difference of the matching position pair;

[0142] The average of the deformation differences of all matching position pairs is recorded as F4, and F4 is used to adjust the mapping coefficient.

[0143] In this example, the mapping coefficient is adjusted according to the deformation difference after welding of similar structural characteristics, which is beneficial to further eliminate the welding deformation to a certain extent when the mapping coefficient is used to calculate the subsequent welding power.

[0144] As an example, the mapping coefficients are adjusted using F3 or F4, including the following methods:

[0145] When F3 or F4 is less than 0, it means that the current mapping coefficient can ensure that the deformation severity of S2 is less than the deformation severity of other welding positions. At this time, the mapping coefficient does not need to be adjusted. When F3 or F4 is greater than or equal to 0, it means that the current mapping coefficient makes the deformation severity of S2 greater than or equal to the deformation severity of other welding positions. At this time, the mapping coefficient needs to be adjusted.

[0146] Specifically, when F3 or F4 is less than -0.2, the mapping coefficient is not adjusted;

[0147] When F3 or F4 is greater than 0.3:

[0148] One method is to increase the above k by 10% so that the welding power of S2 is closer to the welding power of S1.

[0149] Another method is to reduce x2 or x3 by 10%, which also makes the welding power of S2 closer to that of S1. Reduce x1 by 1 (if x1 is less than or equal to 2, do not reduce it). The purpose is to reduce x1 and divide the welding power of S2 into fewer categories (that is, roughly divide the deformation in S2) to avoid introducing too much interference (for example, avoid overfitting).

[0150] When F3 or F4 belongs to [-0.2, 0.3], the welding power on S2 obtained does not significantly reduce the deformation, nor does it increase the deformation. At this time, k, x2 or x3 are randomly reduced by 10% or increased by 10%, and x1 is randomly increased or decreased by 1 to ensure that welding can be performed under more welding power or mapping coefficients, which is helpful for obtaining welding power that can further avoid deformation. In other embodiments, the mapping coefficient can also be kept unchanged.

[0151] The above adjustment range of the mapping coefficient is only an example, and can be set to other values ​​in other embodiments, and is not specifically limited in this embodiment.

[0152] Step S007, continue to obtain a new local welding trajectory and the welding power on the new local welding trajectory according to the welding power in S2 and the adjusted mapping coefficient until all welding positions are welded.

[0153] When the welding on S2 is completed and the mapping coefficient is adjusted, a new local welding track S3 and the welding power on S3 are obtained again. The specific process is the same as step S004 to step S006, that is:

[0154] A new local welding trajectory S3 is determined from the positions that have not been welded according to the deformation sequence of each welded position and the structural features, and the historical welding power including the welding power of S2 is mapped to the welding power on S3 based on the deformation sequences of several welded positions that have the greatest similarity with the structural features of each position on S3, and the mapping coefficient is adjusted after the S3 welding is completed, and the mapping coefficient is adjusted again according to the deformation sequence of each position on S3 and the deformation sequence of each welded position outside S3.

[0155] A new local welding trajectory S4 is determined from the positions that have not been welded according to the deformation sequence of each welded position and the structural features, and the historical welding power including the welding power of S3 is mapped to the welding power on S4 based on the deformation sequences of several welded positions that have the greatest similarity with the structural features of each position on S4, and the mapping coefficient is adjusted after the S4 welding is completed, and the mapping coefficient is adjusted again according to the deformation sequence of each position on S4 and the deformation sequence of each welded position other than S4.

[0156] And so on until the welding is completed.

[0157] It should be noted that the above process continuously selects new local welding trajectories and welding powers, and by continuously adjusting the mapping coefficient, it can be ensured that the welding power selected for the next welding can further reduce welding deformation than the welding power in history. As the number of welding times increases and the number of welded corrugated steel tubes increases, the mapping coefficient will be adjusted more reasonably, and the obtained welding power can ensure that the peaks and troughs of the corrugated steel tubes are aligned after welding. At the same time, in this embodiment, welding according to different welding tracks or welding sequences can further reduce welding deformation and ensure that the peaks and troughs are aligned.

[0158] It should also be noted that, in the above welding process, some intermittent shorter tracks will be formed at the unwelded positions. In this embodiment, when the length of the unwelded track is less than or equal to 10 cm, it is directly used as a new local welding track.

[0159] Second embodiment:

[0160] This embodiment provides a method for obtaining a welding trajectory, including:

[0161] Use a camera to capture images of several aligned corrugated steel plates, convert the images into grayscale images, and use the Canny edge detection algorithm on the grayscale image to obtain the edges on the grayscale image. Considering that the grayscale value of the seam position is low, the edge with the mean grayscale value less than 10 is the seam position, that is, the welding trajectory.

[0162] In other embodiments, the Hough line detection algorithm is used to obtain a straight line on the grayscale image, and the pixel points with a grayscale value less than 10 and on the straight line are taken as seam positions, and the continuous curve formed by all seam positions is recorded as the welding trajectory.

[0163] Further considering that the seams of the corrugated steel plates are not smooth but have depth information, the camera used in this embodiment is a binocular camera to obtain the depth value of each position on the above-mentioned welding track, and the laser welding equipment performs welding along the welding track with depth information.

[0164] In other examples, the depth pixels can be replaced by lidar, and the depth value of each position on the welding trajectory can be obtained through the point cloud data output by the lidar.

[0165] It should be noted that, considering that the depth value of each position on the welding trajectory may not be collected, the average of the depth values ​​within the 8-neighborhood of each position on the welding trajectory is taken as the depth value of each position on the welding trajectory.

[0166] In this embodiment, the method for obtaining the structural features [a, b, A, B] includes:

[0167] Use a camera to capture images of several aligned corrugated steel plates, convert the images into grayscale images, and use the Hough line detection algorithm to obtain the lines on the grayscale images. For parallel lines, each line will be divided into line segments due to the presence of seams; the average grayscale value of each line segment contained in each line is counted, and the line segments with the maximum or minimum average grayscale values ​​represent the peaks or troughs respectively. The distance between each pair of adjacent peaks and troughs is recorded as a half-period, and the average of the half-periods of all adjacent peaks and troughs is recorded as the average half-period of the corrugated steel plate.

[0168] The closest distance of each position to the crest of the corrugated steel plate is recorded as a, and the closest distance to the trough is recorded as b.

[0169] When the distance between the nearest endpoints of the two line segments closest to each position and belonging to the same wave crest is less than the average half period, these two line segments are recorded as misaligned line segments, and the distance between the misaligned line segments is recorded as the docking offset A.

[0170] When the distance between the nearest endpoints of the two line segments closest to each position and belonging to the same trough is less than the average half period, these two line segments are recorded as misaligned line segments, and the distance between the misaligned line segments is recorded as the docking offset B.

[0171] It should be noted that due to the complex specific welding environment, such as the complex lighting environment, the above-mentioned peak or trough acquisition results have errors, but in this embodiment, the difference in peaks or troughs before and after welding is mainly analyzed (which can offset the error), so this error has little effect on this embodiment.

[0172] A preferred example obtains the structural features [a, b, A, B], including:

[0173] The depth map of the surface of several corrugated steel plates is obtained by using laser radar. The depth map contains the depth value of each position. The Hough line detection algorithm is used to obtain the above line segments on the depth map (that is, the line segments in the above parallel straight lines). The line segment with the largest average depth value is regarded as the trough, and the line segment with the smallest average depth value is regarded as the peak.

[0174] Then the structural features [a, b, A, B] are obtained.

[0175] It should be noted that, on a corrugated steel plate, if there are some adjacent parallel line segments, and the average depth values ​​of these line segments are the same and are both maximum or minimum (for example, when the trough or peak area of ​​some corrugated steel plates is flat and there is no height fluctuation), the middle line segment of these line segments is taken as the trough or peak.

[0176] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A laser welding method for a corrugated steel cylinder, characterized in that: The method comprises the following steps: The path formed by the joint positions of several corrugated steel plates is recorded as a welding trajectory; the structural characteristics of each position on the welding trajectory before welding are obtained; Welding is performed along a preset local welding track S1 in the welding track; After welding is completed, a new local welding trajectory S2 and the welding power of S2 are obtained according to the welding power of S1 and the initial mapping coefficient, including: The difference between the structural features of each welded position at all times after welding and the structural features before welding is recorded as the deformation sequence of each welded position; a new local welding trajectory S2 is determined from the position that has not been welded according to the deformation sequence of each welded position and the structural features; based on the deformation sequences of several welded positions with the greatest similarity to the structural features of each position on S2, the historical welding power including the welding power of S1 is mapped to the welding power on S2 through the mapping coefficient; Welding is performed along S2 according to the welding power of S2. After welding is completed, the mapping coefficient is adjusted according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2; Continue to obtain a new local welding trajectory and the welding power on the new local welding trajectory according to the welding power of S2 and the adjusted mapping coefficient until all welding positions are welded; The initial mapping coefficient is an initial value of the mapping coefficient, and the initial value is a preset value; the mapping coefficient represents a parameter when the historical welding power mapping is transformed into the welding power on S2; The step of adjusting the mapping coefficient according to the deformation sequence of each position on S2 and the deformation sequence of each welded position outside S2 includes the following specific steps: The structural features of each position on S2 before welding are recorded as the first features; the structural features of each welded position outside S2 before welding are recorded as the second features; Among the positions on S2 and the positions outside S2 that have been welded, if there is a pair of positions, denoted as p1 and p2, respectively, satisfying the maximum similarity between the first feature of p1 and the second feature of p2, then the positions p1 and p2 are denoted as a matching position pair, and the difference between the deformation sequences of p1 and p2 is denoted as the welding difference of the matching position pair after the welding power is adjusted; The mapping coefficients are adjusted using the mean of the welding differences of all matching position pairs after the welding power is adjusted.

2. The laser welding method for a corrugated steel cylinder according to claim 1, characterized in that: The specific steps for obtaining the structural features are as follows: For any position of the welding trajectory; Obtain the closest distance a from the crest of the corrugated steel plate and the closest distance b from the trough of the corrugated steel plate, and simultaneously obtain the docking offset A of the crest closest to the position and the docking offset B of the trough closest to the position, and use [a, b, A, B] as the structural features of the position.

3. The laser welding method for a corrugated steel cylinder according to claim 1, characterized in that: The specific steps of determining a new local welding trajectory S2 from a position that has not been welded according to the deformation sequence and structural characteristics of each position that has been welded include the following: For the differences of the structural features in the deformation sequence of each welded position, when the mean value of the differences of the structural features is less than a first preset threshold, each welded position is recorded as a reference position; Among all positions that have not been welded, the positions with the greatest similarity to the structural features of each reference position are obtained respectively, and are recorded as the target positions of each reference position; according to the target positions of all reference positions, a local welding trajectory is selected from the welding trajectory, and is recorded as a new local welding trajectory S2.

4. The laser welding method for a corrugated steel cylinder according to claim 1, characterized in that: The deformation sequence of several welded positions with the greatest similarity to the structural features of each position on S2 is used, and the historical welding power including the welding power of S1 is mapped to the welding power on S2 through the mapping coefficient, and the specific steps include the following: The mapping coefficients include: mapping granularity, historical welding power acceptance x2; the mapping granularity is an integer, denoted as x1; The average welding power of the historical welding powers including the welding power of S1 is recorded as P1; Several welded positions with the greatest structural feature similarity to each position on S2 are recorded as the power reference positions of each position on S2; The mean value of the difference of the structural features in the deformation sequence of each power reference position is recorded as the power impact factor of each power reference position, and the power impact factors of the power reference positions at all positions on S2 are clustered into x1 categories; for any category, the mean value of the power impact factor in the category is recorded as the average power impact factor corresponding to each category; The predicted power in each category is obtained according to the average power influencing factor corresponding to each category and the average welding power P1, and the welding power of S2 is obtained according to the predicted power in all categories, the average welding power P1 and the historical welding power acceptance x2.

5. A laser welding method for a corrugated steel cylinder according to any one of claims 1, 3 and 4, characterized in that: The difference of the structural features is equal to the Euclidean distance between the structural features.

6. The laser welding method for a corrugated steel cylinder according to claim 3, characterized in that: The specific steps of selecting a local welding track from the welding track according to the target positions of all reference positions, which is recorded as a new local welding track S2, are as follows: For the path composed of unwelded positions on the welding trajectory, obtain any path containing the reference position, and the ratio of the length of any path to the number of reference positions on any path is recorded as the reference position density of any path. The path that is not connected with S1 and has the largest reference position density is recorded as the new local welding trajectory S2.

7. The laser welding method for a corrugated steel cylinder according to claim 1, characterized in that: The method of adjusting the mapping coefficient by using the mean value of the welding differences of all matching positions after the welding power is adjusted includes the following specific steps: The mean value of the welding differences of all matching position pairs after welding power adjustment is recorded as F4; When F4 is less than the preset lower threshold, the mapping coefficient is not adjusted; when F4 is greater than the preset upper threshold, x2 is reduced by a preset percentage and x1 is reduced by a preset value; when F4 is greater than or equal to the preset lower threshold and less than or equal to the preset upper threshold, x1 and x2 are randomly adjusted.

8. The laser welding method for a corrugated steel cylinder according to claim 4, characterized in that: The predicted power of each category is obtained according to the average power influence factor corresponding to each category and the average welding power P1, and the welding power of S2 is obtained according to the predicted power of all categories, the average welding power P1 and the historical welding power acceptance x2, including the following specific steps: The formula for the predicted power P3 under each category is: , w1 represents the average power impact factor corresponding to each category; The mean of the predicted power under all categories is recorded as , the calculation formula of welding power P2 of S2 is: 。 9. The laser welding method for a corrugated steel cylinder according to claim 2, characterized in that: The specific steps for obtaining A and B are as follows: The depth map of the surface of several corrugated steel plates is obtained by using laser radar. The depth map contains the depth value of each position. The Hough line detection algorithm is used to obtain all line segments on the depth map. The line segment with the largest average depth value is regarded as the trough, and the line segment with the smallest average depth value is regarded as the peak. The average of the distances between all adjacent troughs and peaks is recorded as the average half-period; for the two line segments closest to each position and belonging to the same peak or trough, when the distance between the nearest endpoints of the two line segments is less than the average half-period, the two line segments are recorded as misaligned line segments, and the distance between the misaligned line segments is recorded as the docking offset A or B.

Citation Information

Patent Citations

  • Laser power self-matching method and device based on welding material

    CN111098032A