An intelligent welding path planning method of welding wire width driving

Abstract

The application discloses a kind of intelligent welding path planning methods of welding wire width drive, with welding wire width as global unique driving parameter, intelligent linkage and dynamically generate key threshold and strategy in path planning whole process. Specifically include: based on the generation of cutting threshold, deduplication tolerance, path point deduplication tolerance, continuous welding judgment threshold based on the threshold to the discretization path point of analytic self DXF drawing is globally endpoint cutting to eliminate interference;Carry out angle filtering and neighborhood deduplication cross point accurate detection;Carry out neighborhood order-preserving path point deduplication optimization;And based on the dynamic configuration of three-section type welding speed and the judgment continuous welding logic based on welding wire width. The application is coupled to path planning algorithm by welding process parameter (welding wire width) in depth, realizes the full-link automatic intelligent optimization from drawing analysis to robot code generation, significantly improves welding quality, precision and efficiency, and is strong in universality.

Description

Technical Field

[0001] This invention relates to the field of industrial robot welding technology, specifically to an intelligent welding path planning method driven by welding wire width. It is particularly suitable for automated welding and marking scenarios of marking on the surface of industrial valve bodies, and can achieve high-precision welding and forming of fine markings such as valve body text and logos. It can also be extended to various two-dimensional welding scenarios such as sheet metal parts and automotive parts. Background Technology

[0002] In the field of industrial automated welding, using DXF format 2D drawings as the input source for welding paths and generating robot motion trajectories by parsing the geometric entities in the drawings is the mainstream technical approach in the industry. Its basic processing flow is as follows: read the DXF drawing → extract geometric entities such as straight lines, circles, arcs, and polylines from the drawing → perform simple discretization of the entities into a series of path points → directly generate robot motion control code.

[0003] However, existing technologies suffer from a series of core defects that are disconnected from the welding process when handling this process, resulting in low final welding quality and efficiency:

[0004] 1. Existing methods treat drawings merely as a collection of geometric lines, completely ignoring the crucial process parameter in actual welding – the welding wire width. Since the welding wire has a physical width, the weld bead formed also has a corresponding width. Directly using the geometric endpoints on the drawing as the start and end points of the welding trajectory inevitably leads to overlap between adjacent or intersecting weld beads (too small weld bead spacing) or the formation of missed welds (too large weld bead spacing). Although some post-processing techniques attempt to alleviate this problem by trimming the endpoints at fixed distances, they cannot adapt to welding scenarios with different welding wire widths, resulting in poor versatility.

[0005] 2. For LWPOLYLINE containing arcs, existing technologies often ignore its bulge parameter and directly discretize the LWPOLYLINE as a straight line segment. This approach results in the welded arc trajectory having a noticeable broken line feel, and the arc trajectory is distorted, with errors reaching the millimeter level, which cannot meet the requirements of high-precision welding (such as valve body marking).

[0006] 3. Existing technologies only perform simple line segment intersection judgment when detecting trajectory intersections, which cannot filter out invalid intersections caused by small included angles (<10°). At the same time, they do not remove duplicates from adjacent intersections, and the same intersection area is marked repeatedly. This causes the robot to frequently switch speeds when welding near the intersection, resulting in rough and uneven weld formation, and even the problem of weld burn-through due to excessive local heat concentration.

[0007] 4. To reduce path point redundancy, existing technologies often use a global deduplication strategy to delete duplicate path points. However, this disrupts the original generation order of the path, causing the welding trajectory to backtrack (e.g., a circular trajectory is incorrectly optimized into a polyline for back-and-forth welding), which severely disrupts the continuity of welding and adds extra idle strokes to the robot.

[0008] 5. Existing technologies typically use a single, fixed welding speed without adapting it to the characteristics of the path (normal segments, intersection segments, transition segments). For example, welding too fast at intersections can easily cause burn-through, while welding too slow reduces efficiency. The speed in transition segments is not optimized, further impacting welding efficiency. Furthermore, during transitional movements between different geometric entities, regardless of the distance between the entities, the robot is forced to rise to a safe height before translating, significantly increasing idle travel time and reducing overall welding efficiency by more than 20%.

[0009] In summary, existing technologies lack an intelligent optimization scheme that can deeply integrate welding process parameters with path planning algorithms, resulting in a disconnect between the welding trajectory and actual process requirements, which has become a bottleneck restricting the development of welding automation and precision. Summary of the Invention

[0010] To address the shortcomings of existing intelligent welding path planning methods, such as trajectory interference, insufficient accuracy, and low efficiency, which fail to meet the demands of high-precision and high-efficiency automated welding, especially in the fine-grained scenarios of valve body welding and marking, this invention provides an intelligent welding path planning method driven by welding wire width. This method eliminates trajectory interference, improves the accuracy of curve welding, optimizes intersection point processing, ensures trajectory continuity, and enhances welding efficiency.

[0011] To achieve the above objectives, the present invention adopts the following technical solution:

[0012] This invention provides an intelligent welding path planning method driven by welding wire width, comprising the following steps:

[0013] Obtain the drawing file of the workpiece to be welded and parse it to obtain several geometric entities;

[0014] The wire width W is obtained, and multiple process thresholds required for path planning are generated using the wire width W as a global driving parameter. The process thresholds include: a trimming threshold for trimming path endpoints, a deduplication tolerance for identifying and merging intersections, a deduplication tolerance for optimizing path points, and a continuous welding judgment threshold for determining whether adjacent entities are continuously welded.

[0015] Discretize the geometric entity to obtain the initial path point set;

[0016] Based on the trimming threshold, the endpoints of the initial path point set are globally trimmed to eliminate trajectory interference caused by the physical width of the welding wire, resulting in a trimmed path point set.

[0017] Based on the intersection deduplication tolerance, the intersections between different geometric entities corresponding to the clipped path point set are accurately detected and deduplicated to obtain a path point set marked with intersection information.

[0018] Based on the path point deduplication tolerance, the trimmed path point set is deduplicated by neighborhood order preservation, so as to strictly maintain the original path order while reducing duplicate points.

[0019] Based on the continuous welding judgment threshold, the transition movement strategy between adjacent geometric entities is determined, and robot control code containing welding speed instructions is generated.

[0020] Furthermore, based on the pruning threshold, global pruning is performed on the endpoints of the initial pathpoint set, specifically including:

[0021] For each endpoint in the initial path point set, project it vertically onto all line segments of other entities besides the geometric entity to which it belongs, and calculate the straight-line distance between the endpoint and each projection point.

[0022] If there are one or more straight-line distances less than the clipping threshold, then the endpoint is determined to be an interference point;

[0023] For an endpoint identified as an interference point, starting from its position, traverse the path points inward along the trajectory direction of the geometric entity to which it belongs, and determine the first path point whose projected distance to all other entity line segments is not less than the clipping threshold as the new endpoint after clipping.

[0024] Furthermore, based on the intersection deduplication tolerance, the intersections between different geometric entities corresponding to the clipped path point set are accurately detected and deduplicated, specifically including:

[0025] Iterate through pairs of line segments on different geometric entities, calculate the angle between the vectors containing the two line segments, and filter out line segment pairs with an angle less than a preset angle threshold.

[0026] For unfiltered line segment pairs, calculate their intersection coordinates and determine whether the intersection falls within the valid interval of the two line segments (the algorithm-level screening requires that the coordinates of the intersection point be on both line segment AB and line segment CD to ensure that it is a valid intersection point, which can also be understood as: determining whether the intersection point is simultaneously within the geometric range of the two line segments). If so, mark it as a valid intersection point.

[0027] Based on the deduplication tolerance of cross points, all valid cross points are clustered and merged, and multiple cross points located in the same neighborhood are merged into a unique cross point label.

[0028] Furthermore, based on the path point deduplication tolerance, neighborhood-preserving deduplication is performed on the clipped path point set, specifically including:

[0029] Iterate through the path points in the clipped path point set in sequence, and compare the distance between the current point and the previous unique path point that was retained (the previous path point that was retained after deduplication).

[0030] If the distance is less than the path point deduplication tolerance, then the current point is determined to be a duplicate point and is removed;

[0031] If the distance is not less than the path point deduplication tolerance, then retain the current point as the new unique path point and continue traversing.

[0032] Furthermore, the trimming threshold, the intersection deduplication tolerance, the path point deduplication tolerance, and the continuous welding judgment threshold are obtained by multiplying the welding wire width W by the corresponding preset coefficient.

[0033] In some preferred embodiments, these thresholds are set as follows:

[0034] Clipping threshold T cut =α×W, where α is the trimming factor, preferably in the range of 0.4-0.6, for example 0.5. This threshold is used to determine how much distance needs to be trimmed inward from the endpoint to ensure that the weld bead spacing matches the wire width and avoids interference.

[0035] Cross-point deduplication tolerance T ntersect =β×W, where β is the deduplication coefficient for crossover points, preferably in the range of 0.3-0.4. This threshold is used to merge multiple adjacent crossover points caused by discretization or minor deviations, avoiding duplicate marking or missed detection.

[0036] Path point deduplication tolerance T point =γ×W, where γ is the path point deduplication coefficient, preferably in the range of 0.2-0.3. This threshold is used to remove duplicate path points while preserving the shape of the trajectory.

[0037] Continuous welding judgment threshold T continuous =δ×W, where δ is the continuous welding coefficient, preferably in the range of 1.5-2.5, for example 2.0. This threshold is used to determine whether two separate entities are close enough to be directly moved for welding without lifting.

[0038] Furthermore, robot control code containing welding speed instructions is generated, specifically including:

[0039] Based on the wire width W, the basic welding speed, intersection overlap speed, and transition movement speed are dynamically calculated.

[0040] Based on the set of path points marked with intersection information, configure the corresponding welding speed for each path point or path segment: the basic welding speed is used for ordinary segments, the intersection point and its neighborhood are used for intersection overlap speed, and the transition segment between entities is used for transition movement speed.

[0041] Furthermore, based on the continuous welding judgment threshold, the transition movement strategy between adjacent geometric entities is determined, specifically including:

[0042] Calculate the distance between the starting point of the current geometric entity and the ending point of the previous geometric entity;

[0043] If the spacing is less than the continuous welding judgment threshold, it is judged as continuous welding. There is no need to raise the safety height. A straight movement command is generated between the start and end points, and the movement speed is configured as the transition movement speed or the basic welding speed.

[0044] If the spacing is not less than the continuous welding judgment threshold, it is judged as discontinuous welding. First, a command to raise to a safe height is generated, then a command to move horizontally is generated, and finally a command to descend to the starting point of the current geometric entity is generated.

[0045] Furthermore, the geometric entity is discretized, specifically including:

[0046] For a polyline entity containing a convexity parameter, its convexity value is analyzed to inversely deduce the center, radius, and starting angle of the arc, thereby generating high-precision discrete points to reconstruct the arc trajectory.

[0047] Specifically, for straight lines, discretization can be performed with a fixed step size (e.g., 0.2-0.5 mm). For arcs, circles, and ellipses, discretization can be performed with a fixed angular step size (e.g., 1-5°). In particular, for polylines containing bulge parameters, the bulge value b is analyzed based on the known conditions of the starting point S(x1,y1) and the ending point E(x2,y2). The true center, radius, and starting angle of the arc are then derived using the formula, thereby generating high-precision discrete points located on the true arc trajectory, fundamentally solving the problem of arc trajectory distortion.

[0048] Furthermore, the drawing file is a DXF format file; geometric entities include one or more of the following: straight lines, arcs, circles, polylines, and ellipses; curved entities are processed first.

[0049] Specifically, the drawing file (e.g., DXF format) of the workpiece to be welded is first obtained. All geometric entities within the model space are analyzed and classified by type into lines, circles, arcs, polylines, ellipses, annotations, and text. To facilitate subsequent processing, non-welding entities such as annotations and text are removed. The entities to be processed are sorted in the order of ellipse → circle → arc → polyline → line, with priority given to processing curved entities to ensure the accuracy of the curve trajectory planning.

[0050] Furthermore, after generating the robot control code, the process also includes: calling the visualization module to draw a welding path preview based on the path point set marked with intersection information and the configured welding speed. The preview should at least mark the trimmed endpoints, valid intersections, and segmented areas for different speeds.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] (1) The present invention uses the width of the welding wire as a reference for global projection cutting, which ensures that the spacing between adjacent welds is precisely matched with the width of the welding wire, eliminates weld overlap and missed welding from the root, and significantly improves the first-pass yield of welding. According to the test, the weld formation pass rate reaches 100%.

[0053] (2) This invention restores the real arc trajectory by analyzing the convexity parameters of multi-segment lines, and controls the discrete error within 0.1mm. Compared with the traditional straight line approximation method, the accuracy of curve welding is improved by 10-20 times, which is perfectly adapted to precision welding scenarios (such as the welding requirements of precision sheet metal parts).

[0054] (3) The present invention adopts the intersection detection strategy of "angle filtering" and "neighborhood deduplication", which greatly improves the accuracy of intersection identification. The dynamic speed adaptation of the intersection area takes into account both welding quality and efficiency, and avoids the problems of weld penetration and false welding caused by repeated lifting and falling. The surface roughness of the intersection area is reduced by 30%, and the overall welding efficiency is increased by 18-30%, which greatly improves production quality and efficiency.

[0055] (4) The neighborhood order preservation and deduplication algorithm of the present invention effectively reduces the redundancy of path points by 10-20% while strictly ensuring the original order of the path, completely eliminating the phenomenon of trajectory reversal, making the robot's movements smoother and the welding continuity better.

[0056] (5) The key thresholds of the core algorithm of the present invention are dynamically generated by the welding wire width W, which can be adapted to the full range of welding wires from 0.8 to 2.0 mm; it supports a variety of geometric entities such as straight lines, circles, arcs, polylines, and ellipses in DXF drawings, and is suitable for automated welding scenarios of various plates, with strong industrial practicality and versatility. Attached Figure Description

[0057] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0058] Figure 1 The image shows the welding effect of the conventional method in Example 2 (left) and the welding effect of the method in this invention (right).

[0059] Figure 2 The image shows the welding effect of the conventional method in Example 3 (left) and the welding effect of the method in this invention (right). Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Embodiments of the present invention provide an intelligent welding path planning method driven by welding wire width. The core logic is "using welding wire width as a global driving parameter, and intelligently optimizing all modules including trimming, detection, speed adjustment, and deduplication." The specific working process is as follows:

[0062] Step 1: DXF Drawing Analysis and Entity Classification

[0063] The system reads the drawing file (e.g., DXF format) of the workpiece to be welded, parses all geometric entities within the model space, and classifies them by type as lines, circles, arcs, polylines, ellipses, annotations, and text. To facilitate subsequent processing, non-welding entities such as annotations and text are removed. The entities to be processed are sorted in the order of ellipse → circle → arc → polyline → line, with priority given to processing curved entities to ensure the accuracy of curve trajectory planning.

[0064] Step 2: Initialize global parameters driven by welding wire width

[0065] Based on the input welding wire width W, the system automatically generates optimized parameters for the entire module, achieving intelligent parameter adaptation.

[0066] The trimming threshold, intersection deduplication tolerance, path point deduplication tolerance, and continuous welding judgment threshold are obtained by multiplying the welding wire width W by the corresponding preset coefficient.

[0067] In some preferred embodiments, these thresholds are set as follows:

[0068] Clipping threshold T cut =α×W, where α is the trimming factor, preferably in the range of 0.4-0.6, for example 0.5. This threshold is used to determine how much distance needs to be trimmed inward from the endpoint to ensure that the weld bead spacing matches the wire width and avoids interference.

[0069] Cross-point deduplication tolerance T ntersect =β×W, where β is the deduplication coefficient for crossover points, preferably in the range of 0.3-0.4. This threshold is used to merge multiple adjacent crossover points caused by discretization or minor deviations, avoiding duplicate marking or missed detection.

[0070] Path point deduplication tolerance T point=γ×W, where γ is the path point deduplication coefficient, preferably in the range of 0.2-0.3. This threshold is used to remove duplicate path points while preserving the shape of the trajectory.

[0071] Continuous welding judgment threshold T continuous =δ×W, where δ is the continuous welding coefficient, preferably in the range of 1.5-2.5, for example 2.0. This threshold is used to determine whether two separate entities are close enough to be directly moved for welding without lifting.

[0072] Step 3: Geometric Entity Discretization and Global Endpoint Clipping

[0073] The geometric entity is discretized, and interference points at the endpoints of the entity are eliminated using the wire width as the core driving parameter. The specific process is as follows:

[0074] (i) High-precision discretization of various geometric entities. For straight lines, discretization can be performed with a fixed step size (e.g., 0.2-0.5 mm). For arcs, circles, and ellipses, discretization can be performed with a fixed angular step size (e.g., 1-5°). In particular, for polylines containing bulge parameters, their bulge values ​​are analyzed, and the true center, radius, and starting angle of the arc are derived from the formula, thereby generating high-precision discrete points located on the true arc trajectory, fundamentally solving the problem of arc trajectory distortion.

[0075] String length ;

[0076] Bow height ;

[0077] radius ;

[0078] center ;

[0079] Central angle ;

[0080] Starting angle .

[0081] (ii) Use the projection method to perform global clipping on the initial path point set obtained after discretization: project each endpoint perpendicularly onto other entity line segments and calculate the straight-line distance between the point to be judged and the projection point.

[0082] (iii) If the distance is less than the clipping threshold, it is determined to be an interference point. Traverse the path inward to find the first non-interference point as the new endpoint.

[0083] (iv) If there are fewer than 2 valid points after clipping, the original point set will be automatically restored to avoid path failure.

[0084] Step 4: Precise detection of intersection points for angle filtering and neighborhood deduplication

[0085] The specific process for achieving accurate identification of cross-entity intersections is as follows:

[0086] (a) Collect all continuous line segments of the trimmed entities, construct a line segment list and associate it with the entity ID.

[0087] (ii) Traverse the line segment pairs of different entities and calculate the angle between the direction vectors of the two line segments. If the angle is less than the preset minimum threshold (10-15°), the two line segments are determined to be nearly parallel, and their "intersection" is an invalid intersection and is directly filtered out.

[0088] (iii) For line segment pairs with valid included angles (filtered by included angle), calculate the coordinates of the intersection point and determine whether the intersection point falls within the valid interval of the line segment. If so, record it as a valid intersection point.

[0089] (iv) Spatial clustering of effective intersections is performed according to the deduplication tolerance, and multiple intersections in the same neighborhood are merged into one unique intersection label to avoid multiple velocity changes in a very small area.

[0090] Step 5: Neighborhood-preserving path deduplication optimization

[0091] To reduce the number of path points while ensuring the original order of the welding trajectories is not disrupted, a neighborhood-preserving deduplication algorithm is used. The specific process is as follows:

[0092] (a) Set the path point deduplication tolerance threshold: Deduplication tolerance = welding wire width × (0.2-0.3).

[0093] (ii) Traverse the set of path points after clipping, compare the current path point with the last retained unique path point and calculate the distance between them.

[0094] (iii) If the distance is less than the deduplication tolerance threshold, it is determined to be a duplicate point and removed; if it is greater than or equal to the deduplication tolerance threshold, the current point is retained and used as a new unique path point for comparison with the next path point.

[0095] (iv) This method relies only on the local continuity of the path and strictly follows the traversal order, thus maintaining the original direction of the trajectory and eliminating the trajectory reversal problem caused by traditional global deduplication.

[0096] Step Six: Three-Segment Dynamic Speed ​​Configuration and Continuous Welding Judgment

[0097] The specific process for calculating the three-segment dynamic speed based on the welding wire width is as follows:

[0098] (a) Dynamic calculation of three-stage velocity:

[0099] a. Basic welding speed: V base = V max- k×(W- 1);

[0100] Where V max The base speed is set to the upper limit (e.g., 0.1 m / s), W is the reference wire width (e.g., 1.0 mm), and k is the calculation coefficient. This formula is used to adjust the base speed according to the wire width, and V... base Limit it to a reasonable range (e.g., 0.05~0.1m / s).

[0101] b. Crossover speed: V overlap = min(V base ×1.5, V overlap-max );

[0102] Where V overlap-max This is the upper limit for the speed at the intersection (e.g., 0.15 m / s). Using a slightly faster speed at and around the intersection can prevent the intersection from bulging and improve the welding quality.

[0103] c. Transition speed: V trans = min(V base ×2.0, V trans-max );

[0104] Where V trans-max This is the upper limit for the transition speed (e.g., 0.2 m / s). A faster speed is used during the inter-entity transition segment to improve efficiency.

[0105] (ii) Path point feature marking: Mark each path point as a normal point, intersection point or repeating point, and calculate the distance of each point to the nearest intersection point.

[0106] (III) Speed ​​switching logic: The basic welding speed is used for the normal section, the intersection and transition area (distance from the intersection < welding wire width × 2) adopts the intersection overlap speed, and the transition section between entities adopts the transition movement speed.

[0107] (iv) Continuous welding judgment: Calculate the distance between the starting point of the current entity and the ending point of the previous entity. If the distance is less than 2 × the width of the welding wire, it is judged as continuous welding, and a straight movement command is directly generated without raising the safety height. Otherwise, it is judged as discontinuous welding, and a complete empty movement command of raising the gun → moving horizontally → lowering the gun needs to be generated to ensure safety.

[0108] Step 7: Robot Code Generation and Visual Verification

[0109] Based on the marked optimized path point set and segmented speed, standardized KRL motion control code adapted to industrial robots is generated, including instructions such as speed definition, safe height movement, welding torch start and stop, and trajectory movement.

[0110] At the same time, the visualization module is invoked to draw a comparison diagram of welding paths on the graphical interface, marking the cutting start point, end point, effective intersection point and speed segmentation area, so as to intuitively verify the path planning effect.

[0111] Example 1

[0112] This embodiment aims to verify the actual effect of the global trimming technology based on the projection method of the welding wire width in eliminating weld overlap and missing welding problems.

[0113] The experimental group used the global trimming algorithm based on the projection method of the welding wire width W described in step S3 of this invention, with a trimming threshold Tcut = 0.5×W (W=1.2mm, Tcut=0.6mm); the control group used the traditional fixed threshold trimming algorithm, with a fixed trimming distance of 0.5mm (independent of the welding wire width).

[0114] Three typical weld patterns were selected: cross welds, zigzag welds, and closed-ring welds. Ten test samples were prepared for each pattern, for a total of 30 samples. Welding process parameters (welding current, voltage, speed reference, etc.) were kept consistent for both groups, and the welding wire diameter was 1.2 mm.

[0115] The welds were tested according to the standard of "no overlapping welds, no porosity or undercut (undercut depth < 0.5 mm)", the number of qualified welds was counted, the forming qualification rate was calculated, and the test results are shown in Table 1.

[0116] Table 1

[0117] Test group Total number / Number of overlaps / items Number of missing welds / pieces Number of qualified items / pieces Molding pass rate / % experimental group 30 0 0 30 100 control group 30 6 5 19 63.3

[0118] Test results show that the global trimming technology based on the projection method of the welding wire width (experimental group) of this invention achieves a 100% weld bead formation qualification rate, completely eliminating common weld bead overlap and missing weld defects in traditional methods. In contrast, the control group, using a fixed threshold trimming, suffered from insufficient trimming (overlap) or excessive trimming (missing weld) in some areas due to the trimming distance not matching the welding wire width, resulting in a qualification rate of only 63.3%. The technical solution of this invention achieves precise matching between weld bead spacing and welding wire width by coupling the trimming threshold with the welding wire width depth, significantly improving the weld formation quality.

[0119] Example 2

[0120] This embodiment aims to verify the effect of the present invention in analyzing the bulge parameter of a multi-segment line on improving the accuracy of circular arc trajectories.

[0121] For welding scenarios involving precision sheet metal parts with curved features, a new curve trajectory accuracy test is added to verify the effectiveness of the multi-segment line bulge parameter analysis technology of this invention. The discrete error is calculated based on the maximum deviation between the actual point and the theoretical point of the trajectory.

[0122] The experimental group used the method described in step S3 of this invention to analyze the bulge parameters of the polyline to deduce the center, radius and starting angle of the arc and generate high-precision discrete points; the control group used the traditional method, ignoring the bulge parameters and directly discretizing the polyline with the arc as a straight line segment (equidistant interpolation).

[0123] Three groups of polyline graphics of precision sheet metal parts containing arcs of different radii (R5, R10, R15) were selected, for a total of nine test objects. The discretization step size was set as follows: the experimental group was discretized with an angular step size of 1°, and the control group was discretized with a linear step size of 0.2mm.

[0124] The coordinates of discrete points of two sets of trajectories were calculated using a computer program. The maximum deviation (discrete error) was calculated by comparing the coordinates with the theoretical coordinates of the circular arc. The test was repeated 3 times and the average value was taken. The test results are shown in Table 2.

[0125] Table 2

[0126] Test metrics Control group (traditional linear approximation) Experimental Group (Analysis of bulge parameters in this invention) Trajectory Discrete Error 1.2-2.0mm ≤0.1mm Circular arc fit 82-88% More than 99.5% Visual observation effect The arc has a distinct zigzag shape, and the corners are abrupt. The arcs are smooth and continuous, with natural transitions.

[0127] From Table 2 and Figure 1 Test results show that the method for analyzing multi-segment line bulge parameters using this invention can control the discrete error of the arc trajectory within 0.1mm, which is 10-20 times higher than the traditional straight-line approximation method (error 1.2-2.0mm). The arc fitting accuracy reaches over 99.5%, perfectly restoring the original geometric features of the design drawings. This is of great significance for high-precision welding scenarios such as valve body text logos and precision sheet metal parts, and can significantly improve the product's aesthetics and sealing performance.

[0128] Example 3

[0129] This embodiment aims to verify the effect of the angle filtering + neighborhood deduplication strategy of the present invention on improving the accuracy of intersection identification, as well as the comprehensive impact of the three-stage dynamic speed configuration on welding quality and efficiency.

[0130] The experimental group used the angle filtering + neighborhood deduplication strategy of this invention to identify intersections, and enabled dynamic speed in the intersection area (approaching segment speed 0.08m / s, intersection segment speed 0.12m / s, and inter-entity movement speed 0.16m / s); the control group used the traditional unfiltered global deduplication strategy, with a single speed of 0.08m / s in the intersection area and forced to rise to a safe height before falling when moving between different entities.

[0131] A complex graphic containing 12 real intersections was selected as the test object. The welding wire diameter was 1.2 mm. Each group was repeated five times, and the average value was taken.

[0132] The computer program statistically analyzed the data of the intersection identification, the roughness (Ra) of the intersection area was measured by a roughness meter, the number of times the robotic arm lifted and fell to complete the welding of the whole picture was counted, and the total time spent to complete the welding of the whole picture was timed. The performance differences between the two groups were compared, and the test results are shown in Table 3.

[0133] Table 3

[0134] Test metrics Control group (traditional method) Experimental group (this invention) Performance improvement effect Correct number of intersections 8 12 Higher recognition accuracy Surface roughness of the intersection area 10.0μm 7.0μm Improved welding quality Number of times the robotic arm lifts and falls 19 times 7 times No need to lift at intersection Total welding time for the whole drawing 26s 20s Improved welding efficiency

[0135] From Table 3 and Figure 2 The test results show that this invention effectively eliminates invalid intersections caused by parallel line segments with small included angles (such as misidentification of collinear overlapping areas) through the included angle filtering strategy; and completely solves the problem of duplicate marking by merging multiple adjacent intersections generated by discretization into a single label through the neighborhood deduplication strategy. The intersection recognition accuracy has been improved from 66.7% (8 / 12) to 100%.

[0136] This invention uses a slightly faster speed of 0.12 m / s in the neighborhood of the intersection point to avoid excessive heat concentration, thereby reducing the surface roughness of the intersection area from 10.0 μm to 7.0 μm and significantly improving the molding quality.

[0137] The continuous welding judgment mechanism of this invention eliminates the need for lifting the welding torch during the movement of numerous close-range entities, reducing the number of times the robotic arm lifts from 19 to 7, thus minimizing ineffective idle travel. Combined with the use of a faster speed (0.16 m / s) in the transition section of the three-stage speed system, the overall welding efficiency is improved by more than 25%.

[0138] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A wire width-driven intelligent welding path planning method, characterized in that, Includes the following steps: Obtain the drawing file of the workpiece to be welded and parse it to obtain several geometric entities; Obtain the wire width W, and use the wire width W as a global driving parameter to generate multiple process thresholds required for path planning; The process thresholds include: a trimming threshold for trimming path endpoints, a crosspoint deduplication tolerance for identifying and merging crosspoints, a pathpoint deduplication tolerance for optimizing path points, and a continuous welding judgment threshold for determining whether adjacent entities are continuously welded. Discretize the geometric entity to obtain the initial path point set; Based on the trimming threshold, the endpoints of the initial path point set are globally trimmed to eliminate trajectory interference caused by the physical width of the welding wire, resulting in a trimmed path point set. Based on the intersection deduplication tolerance, the intersections between different geometric entities corresponding to the clipped path point set are accurately detected and deduplicated to obtain a path point set marked with intersection information. Based on the path point deduplication tolerance, the trimmed path point set is deduplicated by neighborhood order preservation, so as to strictly maintain the original path order while reducing duplicate points. Based on the continuous welding judgment threshold, the transition movement strategy between adjacent geometric entities is determined, and robot control code containing welding speed instructions is generated.

2. The intelligent welding path planning method according to claim 1, characterized in that, Based on the pruning threshold, global pruning is performed on the endpoints of the initial pathpoint set, specifically including: For each endpoint in the initial path point set, project it vertically onto all line segments of other entities besides the geometric entity to which it belongs, and calculate the straight-line distance between the endpoint and each projection point. If there are one or more straight-line distances less than the clipping threshold, then the endpoint is determined to be an interference point; For an endpoint identified as an interference point, starting from its position, traverse the path points inward along the trajectory direction of the geometric entity to which it belongs, and determine the first path point whose projected distance to all other entity line segments is not less than the clipping threshold as the new endpoint after clipping.

3. The intelligent welding path planning method according to claim 1, characterized in that, Based on the intersection deduplication tolerance, the intersection points between different geometric entities corresponding to the clipped path point set are accurately detected and deduplicated, specifically including: Iterate through pairs of line segments on different geometric entities, calculate the angle between the vectors containing the two line segments, and filter out line segment pairs with an angle less than a preset angle threshold. For unfiltered line segment pairs, calculate their intersection coordinates and determine whether the intersection point is simultaneously located within the geometric range of both line segments. If so, mark it as a valid intersection point. Based on the deduplication tolerance of cross points, all valid cross points are clustered and merged, and multiple cross points located in the same neighborhood are merged into a unique cross point label.

4. The intelligent welding path planning method according to claim 1, characterized in that, Based on the path point deduplication tolerance, neighborhood-preserving deduplication is performed on the clipped path point set, specifically including: Iterate through the path points in the clipped path point set in order, and compare the distance between the current point and the previous unique path point that was retained. If the distance is less than the path point deduplication tolerance, then the current point is determined to be a duplicate point and is removed. If the distance is not less than the path point deduplication tolerance, then retain the current point as the new unique path point and continue traversing.

5. The intelligent welding path planning method according to claim 1, characterized in that, The trimming threshold, intersection deduplication tolerance, path point deduplication tolerance, and continuous welding judgment threshold are obtained by multiplying the welding wire width W by the corresponding preset coefficient.

6. The intelligent welding path planning method according to claim 1, characterized in that, Generate robot control code containing welding speed instructions, specifically including: Based on the wire width W, the basic welding speed, intersection overlap speed, and transition movement speed are dynamically calculated. Based on the set of path points marked with intersection information, configure the corresponding welding speed for each path point or path segment: the basic welding speed is used for ordinary segments, the intersection point and its neighborhood are used for intersection overlap speed, and the transition segment between entities is used for transition movement speed.

7. The intelligent welding path planning method according to claim 6, characterized in that, Based on the continuous welding judgment threshold, the transition movement strategy between adjacent geometric entities is determined, specifically including: Calculate the distance between the starting point of the current geometric entity and the ending point of the previous geometric entity; If the spacing is less than the continuous welding judgment threshold, it is judged as continuous welding. There is no need to raise the safety height. A straight movement command is generated between the start and end points, and the movement speed is configured as the transition movement speed or the basic welding speed. If the spacing is not less than the continuous welding judgment threshold, it is judged as discontinuous welding. First, a command to raise to a safe height is generated, then a command to move horizontally is generated, and finally a command to descend to the starting point of the current geometric entity is generated.

8. The intelligent welding path planning method according to claim 1, characterized in that, Discretization of geometric entities specifically includes: For a polyline entity containing a convexity parameter, its convexity value is analyzed to inversely deduce the center, radius, and starting angle of the arc, thereby generating high-precision discrete points to reconstruct the arc trajectory.

9. The intelligent welding path planning method according to claim 1, characterized in that, The drawing file is in DXF format; geometric entities include one or more of the following: straight lines, arcs, circles, polylines, and ellipses; curved entities are processed first.

10. The intelligent welding path planning method according to claim 1, characterized in that, After generating the robot control code, the process also includes: calling the visualization module to draw a welding path preview based on the path point set marked with intersection information and the configured welding speed. The preview should at least mark the trimmed endpoints, valid intersections, and segmented areas for different speeds.

Images