Intelligent laser welding method based on CCD visual positioning

Abstract

The application discloses an intelligent laser welding method based on CCD visual positioning, and comprises the following steps: S1) collecting workpiece images through a CCD visual positioning module to generate an initial welding track; S2) an intelligent welding planning module divides a linear welding area and a swing welding area, and matches corresponding welding parameters; the welding area is divided into the linear welding area and the swing welding area according to weld characteristics; S3) a multi-mode cooperative control module optimizes welding parameters of a transition area, and adjusts a welding sequence; and S4) an adaptive path optimization module smoothes a welding path, and controls a laser welding device to perform welding. 2 The maximum acceleration of the transition area of the linear welding area and the swing welding area is controlled within 0.3 m / s The path fitting error is less than 0.05 mm, and the precision welding requirement is met.

Description

Technical Field

[0001] This invention relates to the field of laser welding technology, and in particular to an intelligent laser welding method based on CCD vision positioning. Background Technology

[0002] Traditional laser welding systems often require manual intervention to switch welding modes (such as wire welding and oscillating welding) when welding complex structures, resulting in low welding efficiency and unstable weld quality. Particularly in the transition zones between different welding modes, problems such as uneven welding and thermal deformation are prone to occur. Current technologies lack intelligent collaborative control for multi-mode welding, making it difficult to achieve automatic optimization of welding parameters and smooth path transitions. Summary of the Invention

[0003] To overcome the shortcomings and deficiencies of existing technologies, the present invention aims to provide an intelligent laser welding method based on CCD vision positioning. This method is suitable for high-precision automated welding of shell-type workpieces.

[0004] The objective of this invention is achieved through the following technical solution: an intelligent laser welding method based on CCD vision positioning, comprising the following steps;

[0005] S1) Acquire workpiece images through the CCD vision positioning module, extract workpiece contours and weld positions through image processing algorithms, and generate initial welding trajectories;

[0006] S2) The intelligent welding planning module divides the line welding area and the oscillating welding area and matches the corresponding welding parameters; it divides the welding area into the line welding area and the oscillating welding area according to the weld characteristics; the weld characteristics include width, curvature, and material;

[0007] The wire welding zone uses high-power focused welding, which is suitable for straight or narrow weld seams;

[0008] The oscillating welding zone uses wide-amplitude oscillating welding, which is suitable for wide welds or heat-sensitive areas;

[0009] The junction between the wire bonding area and the oscillating bonding area is designated as a transition zone.

[0010] S3) The multi-mode collaborative control module optimizes the welding parameters in the transition zone and adjusts the welding sequence; it automatically calculates the optimal overlap parameters in the mode transition zone to achieve a smooth transition of power, speed and oscillation amplitude; and it dynamically adjusts the welding sequence through the heat-affected zone prediction algorithm to avoid material deformation caused by local overheating.

[0011] S4) The adaptive path optimization module smooths the welding path and controls the laser welding equipment to perform welding. The module uses a cubic uniform B-spline curve algorithm to achieve smooth optimization of the welding path. B-spline curve fitting technology is used to optimize the welding path, ensuring seamless connection between different trajectories; eliminating mechanical vibration and improving the stability of the welding process.

[0012] As an improvement to the intelligent laser welding method based on CCD vision positioning of the present invention, the specific method of the cubic uniform B-spline curve algorithm in step S4) is as follows:

[0013] Control point extraction:

[0014] Control points P0, P1, ..., P are extracted from the initial welding trajectory obtained by the CCD vision positioning module. n ,

[0015] Increase the sampling density of control points in the transition area between the wire bonding area and the oscillating bonding area, with a minimum of 5 control points in the transition area;

[0016] Node vector generation:

[0017] The node vector U=[u0,u1,...,u] is constructed using the uniform parameterization method. m The nodes are distributed at equal intervals, and the length of the node interval is Δu = u. i+1 - u i

[0018] Define the node interval length as Δu = 0.05 mm;

[0019] Curve fitting calculation:

[0020] Establish B-spline basis functions N i,k (u), where k=3, and K is the number of splines.

[0021] Calculate the coordinates of the trajectory points using a recursive formula:

[0022]

[0023] A weighting factor ωi is introduced to adjust the curve approximation degree, and ωi=1.2 is set at the welding inflection point;

[0024] Transition region processing:

[0025] In the mode switching interval [u trans1 , u trans2 Using dual control point overlay:

[0026] P' trans =αP i +(1-α)P j α∈[0,1]

[0027] Set the curvature constraint |C"(u)|≤0.8mm -1 .

[0028] As an improvement to the intelligent laser welding method based on CCD vision positioning of this invention, a real-time correction mechanism is also included:

[0029] The actual weld position is fed back by a laser displacement sensor.

[0030] The control point coordinates are updated online using the least squares method, with an error correction range of ±0.02mm;

[0031] The mathematical form of the least squares method is: min Σ ||C(u) - Q actual ||²,

[0032] Among them, Q actual The coordinates of the actual weld position fed back by the laser displacement sensor.

[0033] ||C(u) - Q actual ||² represents the squared Euclidean distance between the theoretical and actual points, used to quantify the error.

[0034] Σ represents the sum of errors over all sampling points.

[0035] min represents minimizing the total error by optimizing control points Pi.

[0036] The beneficial effects of this invention are as follows: It achieves continuity (curvature continuity) between the straight segment of wire welding and the curved segment of oscillating welding; the maximum acceleration in the transition zone between the wire welding zone and the oscillating welding zone is controlled within 0.3 m / s², and the path fitting error is <0.05 mm, meeting the requirements of precision welding. Through intelligent zoning and multi-mode collaborative control, it solves the problem of uneven weld seams caused by process switching in traditional welding; by employing heat-affected zone prediction and parameter optimization, it significantly reduces thermal deformation and improves welding quality; welding efficiency is increased by more than 30%, making it suitable for precision welding fields such as new energy vehicle battery casings and electronic device casings. Attached Figure Description

[0037] Figure 1 This is a flowchart of the cubic uniform B-spline curve algorithm of the present invention. Detailed Implementation

[0038] 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 a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0039] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of the components in a specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0040] Furthermore, the use of terms such as "first" and "second" in this invention is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. If the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed by this invention.

[0041] A smart laser welding method based on CCD vision positioning includes the following steps;

[0042] S1) Acquire workpiece images through the CCD vision positioning module, extract workpiece contours and weld positions through image processing algorithms, and generate initial welding trajectories;

[0043] S2) The intelligent welding planning module divides the line welding area and the oscillating welding area and matches the corresponding welding parameters; it divides the welding area into the line welding area and the oscillating welding area according to the weld characteristics; the weld characteristics include width, curvature, and material;

[0044] The wire welding zone uses high-power focused welding, which is suitable for straight or narrow weld seams;

[0045] The oscillating welding zone uses wide-amplitude oscillating welding, which is suitable for wide welds or heat-sensitive areas;

[0046] The junction between the wire bonding area and the oscillating bonding area is designated as a transition zone.

[0047] S3) The multi-mode collaborative control module optimizes the welding parameters in the transition zone and adjusts the welding sequence; it automatically calculates the optimal overlap parameters in the mode transition zone to achieve a smooth transition of power, speed and oscillation amplitude; and it dynamically adjusts the welding sequence through the heat-affected zone prediction algorithm to avoid material deformation caused by local overheating.

[0048] S4) The adaptive path optimization module smooths the welding path and controls the laser welding equipment to perform welding. The module uses a cubic uniform B-spline curve algorithm to achieve smooth optimization of the welding path. B-spline curve fitting technology is used to optimize the welding path, ensuring seamless connection between different trajectories; eliminating mechanical vibration and improving the stability of the welding process.

[0049] Preferably, the specific method of the cubic uniform B-spline curve algorithm described in step S4) is as follows: Figure 1 As shown,

[0050] Control point extraction:

[0051] Control points P0, P1, ..., P are extracted from the initial welding trajectory obtained by the CCD vision positioning module. n , used to define the shape of the B-spline curve;

[0052] Increase the density of control points in the transition area between the wire bonding area and the oscillating bonding area, with a minimum of 5 control points in the transition area.

[0053] Enhancing sampling: Increase the density of control points (≥5) in the transition area (the section where wire welding and oscillating welding meet) to ensure a smooth transition.

[0054] Node vector generation:

[0055] The node vector U=[u0,u1,...,u] is constructed using the uniform parameterization method. m The nodes are distributed at equal intervals, and the length of the node interval is Δu = u. i+1 - u i

[0056] Define the node interval length as Δu = 0.05 mm;

[0057] The node vector determines the parameterization range of the B-spline curve;

[0058] Curve fitting calculation:

[0059] Establish B-spline basis functions N i,k (u), where k=3, and K is the number of splines.

[0060] Calculate the coordinates of the trajectory points using a recursive formula:

[0061]

[0062] A weighting factor ωi is introduced to adjust the curve approximation degree. ωi=1.2 is set at the welding inflection point to enhance the curve's fit to the inflection point.

[0063] Zero-order basis function (piecewise constant):

[0064]

[0065] Higher-order basis functions (recurrence formulas):

[0066]

[0067] Function: Weighting function, representing control point P i The degree of influence on the curve points at parameter u.

[0068] Transition zone handling:

[0069] In the mode switching interval [u trans1 , u trans2 Using dual control point overlay:

[0070] P' trans =αP i +(1-α)P j α∈[0,1]

[0071] Set the curvature constraint |C"(u)|≤0.8mm -1 .

[0072] Function: In the mode switching interval [u trans1 , u trans2 Within the range, transition control points are generated through linear interpolation to achieve a smooth transition.

[0073] P' trans The apostrophe (') in P is a mathematical symbol typically used to indicate a "corrected" or "newly generated" control point, distinguishing it from the original control point P. i and P j P' trans : Indicates the transition interval during mode switching [u trans1 , u trans2 Within ], new control points are generated through interpolation (i.e., the result of "double control point overlay"). This is achieved by blending adjacent control points (P... i and P j The coordinates of the curve are used to generate additional control points in the transition zone, so that the curve can be smoothly connected when switching welding modes (such as wire welding → oscillating welding).

[0074] Curvature constraint: The second derivative (curvature) is restricted to |C"(u)| ≤ 0.8 mm⁻¹.

[0075] The second derivative is directly constrained to have a modulus |C''(u)| ≤ 0.8 mm⁻¹, which is an engineering simplification (assuming that the curvature is mainly dominated by the second derivative when the velocity is approximately uniform).

[0076] Curvature constraint function: Prevents vibration of the robotic arm due to sudden changes in curvature, ensuring smooth movement.

[0077] As an improvement to the intelligent laser welding method based on CCD vision positioning of this invention, a real-time correction mechanism is also included:

[0078] The actual weld position is fed back by a laser displacement sensor.

[0079] The control point coordinates are updated online using the least squares method, with an error correction range of ±0.02mm;

[0080] The mathematical form of the least squares method is: min Σ ||C(u) - Q actual ||²,

[0081] Among them, Q actual The coordinates of the actual weld position fed back by the laser displacement sensor.

[0082] ||C(u) - Q actual ||² represents the squared Euclidean distance between the theoretical and actual points, used to quantify the error.

[0083] Σ represents the sum of errors over all sampling points.

[0084] min represents minimizing the total error by optimizing control points Pi.

[0085] Parameter comparison table for each welding zone:

[0086]

[0087] The CCD vision positioning module is used to acquire surface images of the shell cover workpiece in real time; it extracts the workpiece contour and weld position through image processing algorithms to generate a high-precision welding trajectory.

[0088] Intelligent welding planning module: intelligently divides the welding area into line welding zone and oscillating welding zone according to weld characteristics;

[0089] The wire welding zone uses high-power focused welding, which is suitable for straight or narrow weld seams;

[0090] The oscillating welding zone uses wide-amplitude oscillating welding to improve heat distribution and is suitable for wide welds or heat-sensitive areas.

[0091] Multi-mode collaborative control module; automatically calculates the optimal overlap parameters in the mode transition region to achieve smooth transition of power, speed and swing amplitude; dynamically adjusts the welding sequence through heat-affected zone prediction algorithm to avoid material deformation caused by local overheating;

[0092] The adaptive path optimization module uses B-spline curve fitting technology to optimize the welding path. The adaptive path optimization module uses a cubic uniform B-spline curve algorithm to achieve smooth optimization of the welding path.

[0093] 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 structure of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent laser welding method based on CCD vision positioning, characterized in that, Comprise the following steps; S1) collect workpiece image by CCD vision positioning module, generate initial welding trajectory; S2) intelligent welding planning module divides line welding area and swing welding area, and matches corresponding welding parameters; according to the weld characteristics, the welding area is divided into line welding area and swing welding area; Line welding area adopts high-power focused welding, which is suitable for straight line or narrow weld; Swing welding area adopts wide swing welding, which is suitable for wide weld or heat sensitive area; The junction of line welding area and swing welding area is divided into transition area; S3) multi-mode cooperative control module optimizes the welding parameters of transition area and adjusts the welding sequence; S4) adaptive path optimization module smoothes the welding path and controls the laser welding equipment to execute welding, and the adaptive path optimization module realizes the smoothing optimization of the welding path by using cubic uniform B spline curve algorithm; The specific mode of the cubic uniform B spline curve algorithm in step S4) is as follows: Control point extraction: The control points P0, P1,..., Pn in the initial welding trajectory obtained from the CCD visual positioning module are extracted n , Increase the sampling density of control points for the junction of line welding area and swing welding area, and the transition area contains at least 5 control points; Node vector generation: The node vector U = [u0, u1,..., un] is constructed by using the uniform parameterization method, the nodes are distributed at equal intervals, and the node interval length Δu = u m - u i+1 . i Define node interval length Δu=0.05mm; Curve fitting calculation: Establishing B-spline base function N i,k (u), where k = 3, K is the number of splines Calculate the trajectory point coordinates by recursive formula: Introduce weight factor ωi to adjust the curve approximation degree, set ωi=1.2 at the welding inflection point; Transition area processing: In the mode switching interval [u trans1 , u trans2 ] a double control point superposition is employed: P' trans =αP i +(1-α)P j , α∈[0,1], P' trans is a new control point generated by interpolation in the mode switching transition interval [u trans1 , u trans2 ]. Setting curvature constraint condition |C"(u)|≤0.8mm -1 ; C"(u) is the second derivative of the trajectory curve with respect to the parameter u, reflecting the rate of change of the trajectory curvature.

2. The intelligent laser welding method based on CCD vision positioning according to claim 1, characterized in that, It also includes real-time correction mechanism: Through the feedback of laser displacement sensor actual weld position, Using least square method to update control point coordinates online, the error range is ±0.02mm; The mathematical representation of the least square method is: min∑||C(u) - Q actual ||², wherein Q actual is the actual weld position coordinate fed back by the laser displacement sensor, ||C(u) - Q actual ||² is the squared Euclidean distance between the theoretical and actual points, used to quantify the error, ∑ is the sum of the errors for all sampling points; min is to minimize the total error by optimizing the control points Pi.

Images