A steel body weld joint polishing and rust removing intelligent robot control method

Abstract

The application discloses a kind of steel body weld polishing rust with intelligent robot control method, comprising the following steps: S1: by magnetic wheel adsorption mechanism makes robot adhere to steel surface;S2: real-time tracking weld using multispectral vision system;S3: based on Hessian matrix extraction weld ridge line, ridge line direction is determined by the eigenvector corresponding to maximum eigenvalue;S4: generate weld three-dimensional model by line structured light three-dimensional reconstruction and binocular stereo vision fusion;S5: utilize dynamic window method (DWA) and carry out real-time path planning;S6: adopt B spline curve and carry out path smoothing optimization;S7: real-time acquisition MEMS gyroscope attitude angle θ And six-dimensional force sensor feedback data Fmeasured;S8: visual system detects surface finish, triggers automatic repair mechanism, regenerates path;S9: realize multi-sensor data synchronization by DDS middleware.The application belongs to the field of weld polishing machine, specifically refers to a kind of steel body weld polishing rust with intelligent robot control method.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of weld polishing machines, in particular to a steel body weld polishing and rust removal intelligent robot control method. BACKGROUND

[0002] In the field of steel weld processing after welding, traditional polishing and rust removal operations mainly rely on manual operation or semi-automatic equipment, which has problems such as low efficiency, poor precision, and high labor intensity. In recent years, fully automatic walking type polishing robots have been gradually applied in this field, but there are still the following technical defects:

[0003] 1. Insufficient automation: existing equipment lacks a high-precision navigation system and cannot achieve fully autonomous operation, still requiring manual intervention in path planning or remote control (such as CN112847057B). Especially in closed spaces (such as spherical tanks and pipelines), manual operation is difficult and has low safety.

[0004] 2. To meet detection requirements (such as surface finish Ra≤3.2μm), existing equipment needs to reduce walking speed, resulting in low efficiency; the matching of polishing materials and process parameters (such as speed and pressure) is poor, and needs to be repeatedly tested and adjusted (such as the single abrasive used in CN113146532A leading to insufficient adaptability).

[0005] 3. Poor adaptability to dynamic working conditions: when operating at an inclination angle (such as spherical tank welds), the robot's polishing pressure fluctuates (±30%) due to changes in the gravitational component, which cannot guarantee consistency; existing force control systems rely on preset parameters and lack real-time sensor feedback compensation (such as the lack of gyroscopic and force fusion control). SUMMARY

[0006] The technical problem to be solved by the present application is low automation, inability to simultaneously ensure polishing rate and quality, and poor working condition adaptability.

[0007] To solve the above problems, the technical solution adopted by the present application is as follows: the steel body weld polishing and rust removal intelligent robot control method proposed by the present application includes the following steps:

[0008] S1: attaching the robot to the steel body surface through a magnetic wheel adsorption mechanism, with a magnetic attraction force ≥500N;

[0009] S2: using a multi-spectral vision system to track the weld in real time, the vision system including an infrared laser profiler (Basler acA2000-165um), an anti-arc filter (OD6 level), and a binocular camera, performing adaptive double-threshold edge detection through an improved Canny operator, with threshold range T1=0.05×I max , T2=0.15×I max , where I maxThe maximum gray value of the image is determined as follows:

[0010] S3: Extract the weld ridge line based on the Hessian matrix, and the calculation formula is:

[0011] The ridge line direction is determined by the eigenvector corresponding to the maximum eigenvalue;

[0012] S4: Generate a three-dimensional model of the weld by combining line structured light three-dimensional reconstruction and binocular stereo vision, with a reconstruction accuracy of ≤0.1mm;

[0013] S5: Real-time path planning is performed using the dynamic window method (DWA), and the cost function is:

[0014] J(v, ω) = α·heading(v, ω) + β·dist(v, ω) + γ·velocity(v, ω),

[0015] where α, β, and γ are weight coefficients; α is the path orientation weight coefficient, β is the obstacle distance weight coefficient, and γ is the speed weight coefficient; heading(v, ω) is the path orientation evaluation index, where v is the linear speed of the robot and ω is the angular speed of the robot; dist(v, ω) is the obstacle distance evaluation index; and velocity(v, ω) is the speed evaluation index;

[0016] S6: Smooth and optimize the path using B-spline curves, and the parametric equation is:

[0017]

[0018] where P i is the control point, and N i，p is the B-spline basis function;

[0019] S7: Real-time acquisition of MEMS gyroscope attitude angle θ and six-axis force sensor feedback data F measured , and establishment of a dynamic force compensation model based on finite element analysis:

[0020] F comp = K·(G·sinθ + F measured ),

[0021] where K is the stiffness coefficient, which is calibrated through finite element simulation;

[0022] S8: When the vision system detects that the surface finish Ra is greater than 3.2μm, the automatic repair mechanism is triggered, the path is regenerated, and the polishing pressure is adjusted to 3±0.5N;

[0023] S9: Multi-sensor data synchronization is achieved through DDS middleware, with a communication delay of ≤5ms, and OPC UA protocol is used for interaction with the upper computer.

[0024] Further, in the step S2, the multispectral fusion adopts visible light and near-infrared band weighted superposition, and the formula is:

[0025] I fusion = 0.7·I visible + 0.3·I NIR .

[0026] Further, in the step S5, the search space of the dynamic window method is:

[0027] v∈[v min , v max ], ω∈[ω min , ω max ],

[0028] wherein v max = 0.5 m / s, ω max = 1.0 rad / s.

[0029] Further, in the step S7, the sampling frequency of the six-dimensional force sensor is ≥1 kHz, and the noise is eliminated through Kalman filtering, and the state equation is:

[0030] x k = Ax k-1 + Bu k + w k , z k = Hx k + v k .

[0031] wherein k is the discrete time step; x k is the system state vector at time k; x k-1 is the system state vector at time k-1; A is the state transition matrix; u k is the system control input vector at time k; B is the control input matrix; w k is the process noise vector at time k.

[0032] Further, in the step S8, the surface finish detection adopts gray level co-occurrence matrix (GLCM) texture analysis, and the contrast calculation formula is:

[0033]

[0034] wherein P(i,j) is the gray level co-occurrence matrix element.

[0035] Further, the magnetic wheel adsorption mechanism adopts Halbach array permanent magnet, and the magnetic flux density is ≥1.2T, and the adsorption force calculation formula is:

[0036]

[0037] Where B is the magnetic induction intensity, A is the magnetic pole area, and μ0 is the vacuum permeability.

[0038] Further, a polishing mechanism is also included, which is driven by a planetary gear box with a rotation speed of ≥30000 rpm and a gear box efficiency of ≥95%, and the transmission ratio calculation formula of which is:

[0039]

[0040] Where Z sun and Z ring are the number of teeth of the sun gear and the ring gear, respectively.

[0041] Further, the image processing of the visual system adopts a sub-pixel edge fitting algorithm, and the fitting formula is:

[0042] e(x)=ax 2 +bx+c, and the edge position is determined by .

[0043] Where e(x) is the gray value function of the weld edge; x is used to locate the horizontal position of the weld edge in the image; a is used to determine the concave-convexity and steepness of the gray value function e(x); b is used to describe the linear offset trend of the gray value function e(x); and c represents the reference gray level of the gray value function e(x).

[0044] The beneficial effects achieved by the above method are as follows:

[0045] 1. The steel body weld polishing and rust removing intelligent robot control method proposed in the scheme realizes full autonomous operation through multi-modal sensor fusion (vision + force sense + gyroscope), replaces traditional manual remote control operation, and avoids human error and safety hazards.

[0046] 2. The steel body weld polishing and rust removing intelligent robot control method proposed in the scheme adopts an improved Canny operator and a dynamic force compensation model to realize sub-pixel level weld identification (accuracy 0.1 mm) and constant force control (3±0.5N), and solves the problem of uneven polishing caused by changes in the inclination angle of traditional equipment.

[0047] 3. The steel body weld polishing and rust removing intelligent robot control method proposed in the scheme ensures reliable adsorption and walking of the robot on complex curved surfaces (such as inclination angles of 0-90°) through magnetic wheel adsorption (≥500N) and dynamic path planning (DWA+B spline), without the risk of falling. DETAILED DESCRIPTION

[0048] The technical solutions in the embodiments of the present application will be clearly and completely described below. Apparently, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments; based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work belong to the protection scope of the present application.

[0049] The application provides an intelligent robot control method for steel body weld polishing and rust removal, which comprises the following steps:

[0050] S1: the robot is attached to the surface of the steel body through a magnetic wheel adsorption mechanism, and the magnetic attraction force is greater than or equal to 500 N;

[0051] S2: a multi-spectral vision system is used to track the weld in real time, the vision system comprises an infrared laser profiler (Basler acA2000-165um), an arc light resistant filter (OD6 level) and a binocular camera, adaptive double-threshold edge detection is performed through an improved Canny operator, the threshold value range is T1=0.05×I max ,T2=0.15×I max , wherein I max is the maximum gray value of the image, multi-spectral fusion adopts weighted superposition of a visible light band and a near-infrared band, and the formula is:

[0052] I fusion =0.7·I visible +0.3·I NIR ;

[0053] S3: the weld ridge line is extracted based on a Hessian matrix, and the calculation formula is:

[0054] The ridge line direction is determined by the eigenvector corresponding to the maximum eigenvalue;

[0055] S4: a weld three-dimensional model is generated through line structured light three-dimensional reconstruction and binocular stereo vision fusion, and the reconstruction accuracy is less than or equal to 0.1 mm;

[0056] S5: a dynamic window method (DWA) is used for real-time path planning, and the cost function is:

[0057] J(v,ω)=α·heading(v,ω)+β·dist(v,ω)+γ·velocity(v,ω),

[0058] wherein α, β and γ are weight coefficients,

[0059] The search space of the dynamic window method is:

[0060] v∈[v min ,v max ],ω∈[ω min, ω max ],

[0061] where v max = 0.5 m / s, ω max = 1.0 rad / s;

[0062] S6: Smooth and optimize the path by B-spline curve, the parametric equation is:

[0063]

[0064] where P i is the control point, N i,p is the B-spline basis function;

[0065] S7: Real-time collect MEMS gyroscope attitude angle θ and six-axis force sensor feedback data F measured , based on finite element analysis to establish dynamic force compensation model:

[0066] F comp = K·(G·sinθ+F measured ),

[0067] where K is the stiffness coefficient, calibrated by finite element simulation, the sampling frequency of six-axis force sensor is ≥1 kHz, and the noise is eliminated by Kalman filter, the state equation is:

[0068] x k = Ax k-1 + Bu k +w k , z k = Hx k + v k ;

[0069] S8: When the vision system detects that the surface finish Ra>3.2μm, trigger the automatic repair mechanism, regenerate the path and adjust the polishing pressure to 3±0.5N, the surface finish detection uses gray level co-occurrence matrix (GLCM) texture analysis, the contrast calculation formula is:

[0070]

[0071] where P(i, j) is the gray level co-occurrence matrix element;

[0072] S9: Realize multi-sensor data synchronization through DDS middleware, communication delay ≤5ms, and use OPC UA protocol to interact with the host computer.

[0073] The magnetic wheel adsorption mechanism uses Halbach array permanent magnet, magnetic flux density ≥1.2T, and the adsorption force calculation formula is:

[0074]

[0075] Where B is the magnetic induction intensity, A is the magnetic pole area, and μ0 is the vacuum permeability; the polishing mechanism is driven by a planetary gear box, the rotation speed is ≥30000 rpm, the gear box efficiency is ≥95%, and the transmission ratio calculation formula is:

[0076]

[0077] Where Z sun and Z ring are the number of teeth of the sun gear and the ring gear, respectively; the image processing of the vision system uses a sub-pixel edge fitting algorithm, and the fitting formula is:

[0078] e(x) = ax2 + bx + c, and the edge position is determined by .

[0079] Example: polishing of the weld on the inner wall of a spherical tank

[0080] 1. System initialization:

[0081] The magnetic wheel adsorption force is set to 600 N (equation 8);

[0082] The scanning frequency of the vision system is 60 Hz.

[0083] 2. Weld detection:

[0084] Improved Canny operator (T1 = 30, T2 = 100);

[0085] Extraction of ridge lines by Hessian matrix (equation 3).

[0086] 3. Path planning:

[0087] DWA parameters: α = 0.6, β = 0.3, γ = 0.1;

[0088] B-spline control point spacing is 10 mm.

[0089] 4. Force control execution:

[0090] When the inclination angle is 45°, the compensation force F comp = 25 N (equation 7);

[0091] The feedback frequency of the six-axis force sensor is 1 kHz.

[0092] 5. Quality detection:

[0093] The GLCM contrast threshold is set to 150;

[0094] The unqualified area is automatically repaired for 3 times.

[0095] The above describes the present application and its embodiments, which are not limited. In general, if a person skilled in the art is inspired by the above, without departing from the spirit of the present application, similar structural modes and embodiments can be designed without creativity, and should belong to the protection scope of the present application.

Claims

1. A steel body weld joint polishing and rust removing intelligent robot control method, characterized by, Comprise the following steps: S1: The robot is attached to the surface of the steel body by a magnetic wheel adsorption mechanism, with a magnetic attraction force ≥ 500N; S2: Real-time tracking of the weld seam using a multispectral vision system, which includes an infrared laser profilometer, an anti-arc filter and a binocular camera, adaptive double-threshold edge detection is performed by an improved Canny operator, the threshold range T1 = 0.05 x I max , T2 = 0.15 x I max , where I max is the maximum gray value of the image; S3: Extract the weld ridge line based on the Hessian matrix, and the calculation formula is: The ridge direction is determined by the eigenvector corresponding to the largest eigenvalue; S4: Generate a three-dimensional model of the weld by line structure light three-dimensional reconstruction and binocular stereo vision fusion, with a reconstruction accuracy ≤ 0.1mm; S5: Use the dynamic window method (DWA) for real-time path planning, and the cost function is: J(v,ω)=α·heading(v,ω)+β·dist(v,ω)+γ·velocity(v,ω), Wherein α, β, γ are weight coefficients; S6: Smooth and optimize the path using B-spline curves, and the parameterized equation is: where P i is a control point, N i,p is a B-spline basis function; S7: Real-time acquisition of MEMS gyroscope attitude angle θ and six-dimensional force sensor feedback data F measured Based on finite element analysis, a dynamic force compensation model is established: F comp = K - (G - sin θ + F measured ), Wherein K is the stiffness coefficient, which is calibrated by finite element simulation; S8: When the visual system detects that the surface finish Ra>3.2μm, trigger the automatic repair mechanism, regenerate the path and adjust the polishing pressure to 3±0.5N; S9: Realize multi-sensor data synchronization through DDS middleware, with a communication delay ≤5ms, and use OPC UA protocol to interact with the host computer.

2. A control method of an intelligent robot for steel body weld joint polishing and rust removal according to claim 1, characterized in that: In the step S2, the multispectral fusion adopts visible light and near-infrared band weighted superposition, and the formula is: I fusion = 0.7 · I visible + 0.3 · I NIR .

3. A control method of an intelligent robot for steel body weld joint polishing and rust removal according to claim 2, characterized in that: In the step S5, the search space of the dynamic window method is: v e [v m in,v max ], ω e [ω m in,ω max ], where v max = 0.5 m / s, ω max = 1.0 rad / s.

4. A control method of an intelligent robot for steel body weld joint polishing and rust removal according to claim 3, characterized in that: In the step S7, the sampling frequency of the six-dimensional force sensor is ≥1kHz, and the Kalman filter is used to eliminate noise, and the state equation is: x k = Ax k -1 + Bu k + w k , z k = Hx k + v k .

5. A method of controlling an intelligent robot for steel body weld joint polishing and rust removal according to claim 4, characterized in that: In the step S8, the surface finish detection adopts gray level co-occurrence matrix (GLCM) texture analysis, and the contrast calculation formula is: Wherein P(i,j) is the gray level co-occurrence matrix element.

6. A control method of an intelligent robot for steel body weld joint polishing and rust removal according to claim 5, characterized in that: The magnetic wheel adsorption mechanism adopts Halbach array permanent magnet, with a magnetic flux density ≥1.2T, and the adsorption force calculation formula is: Wherein B is the magnetic induction intensity, A is the magnetic pole area, and μ0 is the vacuum permeability.

7. A method of controlling an intelligent robot for steel body weld joint polishing and rust removal according to claim 6, characterized in that: Also includes a polishing mechanism, the polishing mechanism adopts planetary gear box drive, the rotating speed ≥30000rpm, the gear box efficiency ≥95%, and the transmission ratio calculation formula is: where Z sun and Z ring are the number of teeth of the sun gear and ring gear, respectively.

8. A control method of an intelligent robot for steel body weld joint polishing and rust removal according to claim 7, characterized in that: The image processing of the visual system adopts sub-pixel edge fitting algorithm, and the fitting formula is: e(x) = ax + bx + c, the edge position is determined by 2 + bx + c. + bx + c.

Images