Pipeline welding operation position deviation real-time monitoring and correcting system and method

Abstract

The invention belongs to the technical field of pipeline welding, and relates to a pipeline welding operation position deviation real-time monitoring and correcting system and method. The method comprises the following steps: fitting a central axis through a pipe end centering module to calculate a residual error and converting the residual error into a welding gun tool coordinate system to obtain a pose compensation amount; constructing a displacement change track, and screening a low-frequency axis offset coordinate sequence based on a pipeline mechanical motion inertia characteristic setting direction and acceleration constraint condition; and after the space offset is superposed with the compensation amount, a correction instruction is generated through a PID control algorithm. The problems of welding gun shaking caused by high-frequency noise interference in suspended pipeline welding, error accumulation and amplification caused by initial fixed deviation and dynamic deviation mixed compensation, and control overshoot or slow convergence caused by large rigid response difference of pipelines of different specifications are solved. Static residual error pre-compensation and dynamic offset decoupling are achieved, synchronous following of the real displacement trend is ensured, misjudgment interference is effectively avoided, and the stability and trajectory tracking precision of the welding process under the complex working condition are improved.

Description

Technical Field

[0001] This invention belongs to the field of pipeline welding technology and relates to a real-time monitoring and correction system and method for pipeline welding operation position deviation. Background Technology

[0002] In municipal engineering projects such as underground water plants, there are numerous process pipelines with high welding quality requirements. Furthermore, the underground space structure limits the density of supports and hangers, restricting the operational space for high-altitude pipeline hoisting. Under these conditions, pipelines are often suspended and not securely fixed, making them susceptible to swaying during welding due to gravity, thermal stress, and external disturbances. When the pipeline shifts or the diameters differ at cross-sections, the central axis is prone to deviation, leading to eccentricity errors in the relative position of the welding torch and the weld, thus affecting the quality of the weld formation.

[0003] Existing technologies for welding suspended pipes suffer from the following main drawbacks: First, the sensor-collected data contains high-frequency noise caused by mechanical vibration and the true trend signal generated by the overall mechanical displacement of the pipe. Existing filtering techniques struggle to effectively remove noise while preserving the true displacement characteristics, leading the control system to misinterpret high-frequency interference as actual deviations, causing frequent welding torch vibrations; conversely, excessive smoothing results in a delayed response to the true displacement, making synchronous tracking impossible.

[0004] Secondly, during the pipe end alignment stage, factors such as fixture accuracy and pipe ellipticity limit the initial deviation between the two pipe axes, even after mechanical adjustment. Existing technologies typically mix these fixed and dynamic deviations and use a unified feedback control strategy for compensation. This results in the initial error persisting throughout the welding process and may even be amplified due to the integral accumulation effect, making it difficult to achieve high-precision alignment.

[0005] Finally, pipes of different specifications and wall thicknesses exhibit varying rigidity response characteristics, resulting in significant differences in their mechanical drift patterns. Existing control schemes often employ fixed-parameter algorithms, which cannot adaptively adjust the correction intensity based on physical characteristics such as pipe mass and moment of inertia. This can easily lead to system overshoot or slow convergence, making it difficult to maintain stable welding quality under complex operating conditions. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background art, a real-time monitoring and correction system and method for pipeline welding operation position deviation is proposed.

[0007] The first aspect of the present invention proposes a real-time monitoring and correction system for position deviation in pipeline welding operations, characterized in that it includes the following modules: Pipe end alignment module: acquiring point cloud data of the cross-sections of the two pipe sections to be welded, fitting and calculating the actual central axis and spatial relative position to obtain the coaxiality deviation of the pipe end, generating clamping and alignment commands to drive the device to adjust the posture, and converting the adjusted residual to the welding torch tool coordinate system to calculate the welding torch posture compensation amount.

[0008] Offset Extraction Module: During the welding process, the coordinate data of weld feature points are acquired in real time and the displacement change trajectory is constructed. Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertial threshold are set. Continuous coordinate subsequences with consistent direction and acceleration feature values ​​within the inertial threshold range are selected from the displacement change trajectory as low-frequency axis offset coordinate sequences.

[0009] The correction control module calculates the spatial offset between the welding torch center point and the actual center axis of the pipeline in real time from the low-frequency axis offset coordinate sequence, and adds the welding torch pose compensation amount. It calculates the correction step size and generates correction control commands through the PID control algorithm.

[0010] The method for obtaining the coaxiality deviation of the pipe ends is as follows: the point cloud data of the cross sections of the two pipe sections are divided into layers at equal intervals along the pipe axis. The coordinates of the center of the cross section are calculated by the least squares circle fitting algorithm for each layer of point cloud. The actual center axis is obtained by fitting the coordinates of the center of the cross section with a straight line using the least squares method.

[0011] The vertical distance between the actual center axes of the two pipe sections at the weld joint is calculated as the positional deviation value, and the angle between the direction vectors of the two actual center axes is calculated as the angular deviation value. The positional deviation value and the angular deviation value together constitute the pipe end coaxiality deviation.

[0012] The method for obtaining the clamping and centering command is as follows: obtain the position deviation value and angle deviation value in the coaxiality deviation of the tube end, construct the pose difference matrix based on the position deviation value and angle deviation value, and decompose it into three-dimensional translation component and three-dimensional rotation component.

[0013] The three-dimensional translation and rotation components are mapped to the drive quantities of each actuator of the automatic fixing and centering device to generate clamping and centering commands.

[0014] The method for obtaining the welding torch pose compensation is as follows: after the equipment has adjusted its pose, the vision sensor is controlled to scan again to obtain the cross-sectional point cloud data of the two pipe sections, and the adjusted actual center axis is calculated.

[0015] The straightness error of each axis segment is calculated based on the adjusted actual center axis, and the segment with the smaller straightness error is selected as the reference segment.

[0016] Using the actual center axis of the reference pipe as a reference, calculate the angle vector between the vertical distance vector and the direction vector of the actual center axis of another pipe segment relative to the reference axis, and use it as the original residual.

[0017] A local coordinate system for the pipeline is constructed with the actual central axis of the reference pipeline as the Z-axis and the center point of the weld joint between the two pipeline sections as the origin. The original residual is projected onto the local coordinate system to obtain the projected residual.

[0018] Construct a homogeneous transformation matrix from the local coordinate system of the pipeline to the coordinate system of the welding torch tool. Use the homogeneous transformation matrix to transform the projection residual to the coordinate system of the welding torch tool to obtain the welding torch pose compensation amount.

[0019] The specific method for constructing the displacement change trajectory is as follows: during the welding process, the vision sensor is controlled to collect the original point cloud data of the weld area in real time.

[0020] A statistical filtering algorithm is used to remove noise points from the original point cloud data. Then, background points with Z-axis coordinate values ​​less than a preset height threshold are removed from the original point cloud data, and the remaining point cloud data is used as the point cloud of the weld area.

[0021] The geometric centroid of the point cloud in the weld area is calculated as the coordinates of the weld feature points. The coordinates of the weld feature points are then transformed to the world coordinate system to obtain the coordinate data of the weld feature points.

[0022] Mark the acquisition timestamps for the weld feature point coordinate data, and sort the weld feature point coordinate data according to the order of acquisition timestamps.

[0023] The displacement change trajectory is constructed by connecting the coordinate data of adjacent weld feature points after sorting through linear interpolation.

[0024] The method for obtaining the low-frequency axis offset coordinate sequence is as follows: a sliding window is used to traverse the displacement change trajectory, and the displacement vector and displacement vector unit vector between adjacent sampling points within the window are calculated.

[0025] The dot product of adjacent displacement vectors is calculated as the directional consistency coefficient.

[0026] The difference in magnitude between adjacent displacement vectors is calculated as the displacement increment amplitude. The sampling period is determined based on the preset sampling frequency. The displacement increment amplitude is divided by the sampling period to obtain the average velocity within that time period. The average velocities of adjacent time periods are differentiated to obtain the average velocity change. The average velocity change is then divided by the sampling period to obtain the acceleration characteristic value.

[0027] Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertia threshold are set to filter coordinate points whose directional consistency coefficient is greater than the directional consistency threshold and whose acceleration characteristic value is within the range of the acceleration inertia threshold.

[0028] The set of points with a preset number of consecutively occurring coordinate points is taken as a coordinate subsequence, and they are arranged in order to obtain the low-frequency axis offset coordinate sequence.

[0029] The setting of directional consistency thresholds and acceleration inertia thresholds based on the mechanical motion inertia characteristics of the suspended pipe specifically involves: acquiring the mass parameters and rotational inertia parameters of the suspended pipe, and calculating the maximum allowable angular acceleration and maximum allowable linear acceleration of the pipe based on the mechanical motion laws by combining a preset sampling frequency.

[0030] The product of the maximum permissible angular acceleration and the square of the sampling period is determined as the maximum permissible angle change, and the cosine value of the maximum permissible angle change is calculated as the directional consistency threshold.

[0031] The absolute value of the maximum permissible linear acceleration is determined as the acceleration-inertia threshold.

[0032] The method for obtaining the spatial offset is as follows: for each coordinate point in the low-frequency axis offset coordinate sequence, it is used as the coordinate of the welding gun center point, and the perpendicular vector from the point to the corresponding pipeline actual center axis is calculated as the position deviation vector.

[0033] The angle between the welding torch axis and the actual center axis is calculated as the attitude deviation vector. The position deviation vector and the attitude deviation vector are combined to construct a six-dimensional pose error vector as the spatial offset.

[0034] The method for obtaining the correction control command is as follows: the spatial offset and the welding torch pose compensation are vector-superimposed to obtain the total error value.

[0035] Obtain the total error value at the current time and historical time, and use the discrete incremental PID control algorithm to calculate the sum of the proportional, integral and derivative terms as the correction step size.

[0036] The kinematic transmission ratio of the welding torch actuator is used to convert the correction step size into a drive signal to generate correction control commands.

[0037] The second aspect of the present invention proposes a method for real-time monitoring and correction of position offset in pipeline welding operations, characterized by the following steps: S1, acquiring point cloud data of the cross-sections of two pipeline segments to be welded, fitting and calculating the actual central axis and spatial relative position to obtain the coaxiality deviation of the pipe ends, generating a clamping and centering command to drive the device to adjust the posture, and converting the adjusted residual to the welding torch tool coordinate system to calculate the welding torch posture compensation amount.

[0038] S2. During the welding process, the coordinate data of the weld feature points are acquired in real time and the displacement change trajectory is constructed. Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertial threshold are set. Continuous coordinate subsequences with consistent direction and acceleration feature values ​​within the inertial threshold range are selected from the displacement change trajectory as low-frequency axis offset coordinate sequences.

[0039] S3. The low-frequency axis offset coordinate sequence is solved in real time as the spatial offset between the welding torch center point and the actual center axis of the pipeline section, and the welding torch posture compensation is superimposed. The correction step size is calculated by the PID control algorithm and the correction control command is generated.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention uses the pipe end centering module to scan and calculate the remaining residual after mechanical adjustment, and converts it to the welding torch tool coordinate system to calculate the welding torch pose compensation amount, thereby achieving decoupling between fixed deviation and dynamic deviation. This method solves the problem that the prior art mixes small initial deviation and dynamic deviation for compensation, which leads to continuous amplification of error and difficulty in achieving high-precision centering. It enables the welding torch to directly eliminate the residual error after mechanical centering based on the pre-compensation amount during the welding process, thereby improving the initial centering accuracy.

[0041] (2) This invention uses an offset extraction module to set directional consistency thresholds and acceleration inertia thresholds based on the mechanical motion inertial characteristics of the suspended pipe, and uses physical constraints to filter out low-frequency axis offset coordinate sequences from the trajectory data. This method solves the problem that existing filtering techniques are unable to effectively remove high-frequency noise while retaining the true displacement characteristics, avoids frequent welding torch jitter caused by misjudging high-frequency interference in the control system, and eliminates response lag caused by excessive smoothing, thus achieving synchronous tracking of the true displacement trend.

[0042] (3) This invention uses a correction control module to calculate the low-frequency axis offset coordinate sequence into a spatial offset, and then superimposes the welding torch pose compensation amount before using a PID control algorithm to calculate the correction step size and generate correction control commands. This method solves the problems of large differences in the rigid response characteristics of pipes of different specifications and the tendency of fixed parameter algorithms to cause overshoot or slow convergence. It realizes dynamic correction based on the actual motion characteristics of the pipe, ensuring the stability of the welding process and the accuracy of trajectory tracking under complex working conditions. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1This is a schematic diagram showing the connection of each module in a real-time monitoring and correction system for pipeline welding operation position offset according to the present invention.

[0045] Figure 2 This is a flowchart of the method for obtaining the low-frequency axis offset coordinate sequence in this invention.

[0046] Figure 3 This is a flowchart illustrating the steps of a real-time monitoring and correction method for position offset in pipeline welding operations according to the present invention. Detailed Implementation

[0047] 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.

[0048] Unless otherwise defined, 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 pertains.

[0049] The following description, in conjunction with the accompanying drawings, details the specific scheme of a real-time monitoring and correction system and method for pipeline welding operation position deviation provided by the present invention.

[0050] Example 1.

[0051] Please see Figure 1 As shown, the present invention provides a real-time monitoring and correction system for pipeline welding operation position deviation, including: a pipe end alignment module, a deviation extraction module and a correction control module. The connection relationship between the modules is as follows: the pipe end alignment module and the correction control module are connected, and the deviation extraction module and the correction control module are connected.

[0052] Pipe end alignment module: acquires point cloud data of the cross-sections of the two pipe sections to be welded, fits and calculates the actual center axis and spatial relative position to obtain the pipe end coaxiality deviation, generates clamping and alignment commands to drive the equipment to adjust the posture, and converts the adjusted residual to the welding torch tool coordinate system to calculate the welding torch posture compensation amount.

[0053] Considering that pipelines in underground water treatment plants and similar applications are often suspended and subject to limitations in hoisting error and clamp precision, microscopic deviations may still remain between the axes of the two pipes after mechanical adjustment. Directly compensating for these fixed and dynamic deviations together would amplify the initial error due to the integral accumulation effect. Therefore, it is necessary to quantify the coaxiality deviation at the pipe ends to drive the automatic fixing and centering equipment to eliminate macroscopic deviations. At the same time, the microscopic residuals that cannot be completely eliminated should be pre-compensated to the welding torch coordinate system, thereby decoupling the fixed and dynamic deviations and laying the foundation for subsequent dynamic correction.

[0054] In one specific embodiment, the method for obtaining the coaxiality deviation of the pipe ends is as follows: the point cloud data of the cross sections of the two pipes are divided into equal-interval layers along the pipe axis, the coordinates of the center of the cross section are calculated by the least squares circle fitting algorithm for each layer of point cloud, and the actual center axis is obtained by the least squares line fitting of all the center coordinates of the cross sections.

[0055] The layer spacing is set to 5mm to 10mm, and the specific value can be determined according to the pipe diameter and accuracy requirements. A smaller value is used when the diameter is larger or the accuracy requirement is higher. The iterative convergence threshold in the least squares fitting process is set by the algorithm's default accuracy, for example, 0.01mm. The implementer can set it according to the specific situation.

[0056] The vertical distance between the actual center axes of the two pipe sections at the weld joint is calculated as the positional deviation value, and the angle between the direction vectors of the two actual center axes is calculated as the angular deviation value. The positional deviation value and the angular deviation value together constitute the pipe end coaxiality deviation. This deviation is used to quantitatively describe the spatial misalignment state of the two pipe sections at the joint, where the positional deviation reflects the straight-line distance of the axis offset, and the angular deviation reflects the degree of inclination of the intersecting axes.

[0057] Furthermore, since the obtained pipe end coaxiality deviation reflects the actual physical misalignment, direct welding cannot meet the process requirements. Therefore, it needs to be converted into action commands that the equipment can execute to eliminate macroscopic errors. Specifically, the method for obtaining the clamping and centering command is as follows: obtain the position deviation value and angle deviation value in the pipe end coaxiality deviation, construct a 6-DOF pose difference matrix based on these two deviation values, which describes the transformation relationship required to adjust from the current actual pose of the pipe end to the ideal centering state, and then decompose the matrix into three-dimensional translational components and three-dimensional rotational components.

[0058] Finally, the three-dimensional translation and rotation components are mapped to the driving quantities of each actuator, such as the pulse count of the servo motor, according to the kinematic model of the automatic fixing and centering device, to generate a clamping and centering command. This command is a driving signal used to control the motor of the centering device to achieve macroscopic adjustment of the physical position, so that the two pipe sections reach a preliminary alignment state.

[0059] Furthermore, considering that mechanical adjustment has a limit to accuracy and the pipeline itself may be bent and deformed, clamping equipment alone cannot completely eliminate all errors. The remaining micro-deviations still need to be compensated by the movement of the welding torch. Therefore, it is necessary to further obtain the welding torch position compensation amount to achieve micro-deviation compensation.

[0060] Specifically, the method for obtaining the welding torch pose compensation is as follows: After the equipment has adjusted its pose, the vision sensor is controlled to scan again to obtain the cross-sectional point cloud data of the two pipe segments, and the adjusted actual center axis is calculated. Then, the straightness error of each axis segment is calculated, and the pipe segment with the smaller straightness error is selected as the reference pipe and used as the compensation benchmark.

[0061] Using the actual center axis of the reference pipe as a reference, calculate the angle vector between the vertical distance vector and the direction vector of the actual center axis of another pipe segment relative to the reference axis. These two vectors constitute the original residual.

[0062] Next, a local coordinate system for the pipeline is constructed with the actual central axis of the reference pipeline as the Z-axis and the center point of the weld joint between the two pipeline sections as the origin. The original residual is projected onto this coordinate system to obtain the projected residual that is directly related to the geometric features of the pipeline.

[0063] Obtain the fixed transformation relationship between the vision sensor and the welding torch tool coordinate system, which is predetermined through hand-eye calibration. Combine this with the pose of the local pipeline coordinate system relative to the vision sensor coordinate system calculated in real time, and construct a homogeneous transformation matrix from the local pipeline coordinate system to the welding torch tool coordinate system.

[0064] The projection residual is transformed to the welding torch tool coordinate system using a homogeneous transformation matrix to obtain the welding torch pose compensation amount. This compensation amount is a six-dimensional data vector, which is used to pre-superimpose the welding torch control command during the welding process, thereby eliminating the micro-deviations remaining after mechanical alignment.

[0065] Offset Extraction Module: During the welding process, the coordinate data of weld feature points are acquired in real time and the displacement change trajectory is constructed. Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertial threshold are set. Continuous coordinate subsequences with consistent direction and acceleration feature values ​​within the inertial threshold range are selected from the displacement change trajectory as low-frequency axis offset coordinate sequences.

[0066] During the welding of suspended pipelines, the pipeline experiences low-frequency mechanical oscillations due to gravity, thermal stress, and external disturbances, while welding spatter, arc light interference, and sensor noise introduce high-frequency random jitter. Existing filtering techniques struggle to effectively eliminate noise while preserving true displacement characteristics, easily leading to response lag or misjudgment of jitter. Therefore, it is necessary to establish physical constraints based on the pipeline's inherent mechanical inertial characteristics to distinguish between true offset and noise, thereby extracting a low-frequency axial offset sequence that accurately reflects the pipeline's motion.

[0067] In one specific embodiment, the method for constructing the displacement change trajectory is as follows: during the welding process, the vision sensor is controlled to collect the original point cloud data of the weld area in real time at a fixed frequency, such as 50Hz.

[0068] A statistical filtering algorithm is used to calculate the average distance between each point and its neighbors, and outlier noise points with distances greater than the average are removed. Then, background points with Z-axis coordinate values ​​less than a preset height threshold are removed from the original point cloud data, and the remaining point cloud data is used as the weld area point cloud. The preset height threshold can be set by collecting background point clouds without pipes, statistically analyzing their Z-axis distribution range, and setting the threshold to 2mm to 5mm above the highest point of the background.

[0069] Calculate the geometric centroid of the point cloud in the weld area. The coordinates of this centroid are the coordinates of the weld feature points at that moment. Transform the coordinates of the weld feature points to the world coordinate system to obtain the weld feature point coordinate data.

[0070] The coordinate data of weld feature points are marked with acquisition timestamps, and the coordinate data of weld feature points are sorted according to the order of acquisition timestamps. The sorted adjacent coordinate data of weld feature points are then connected by linear interpolation to construct a displacement change trajectory. This trajectory represents the continuous movement path of the welding torch relative to its initial position during the welding process, and intuitively reflects the dynamic changes in the weld position caused by pipe swaying, thermal deformation, and external disturbances.

[0071] Furthermore, the displacement trajectory contains not only the actual offset caused by the low-frequency mechanical motion of the pipeline, but also high-frequency noise and random disturbances. If it is directly used for correction control, it will cause frequent welding torch vibration or response lag, making it impossible to achieve stable and accurate following. Therefore, it is necessary to select a low-frequency axis offset coordinate sequence from the trajectory that can truly reflect the physical motion of the pipeline, based on the mechanical motion inertial characteristics of the pipeline itself.

[0072] For details, please refer to Figure 2 As shown, the method for obtaining the low-frequency axis offset coordinate sequence is as follows: a sliding window of fixed length is used to traverse the displacement change trajectory in time order. For adjacent sampling points within the window, the displacement vector formed by them is calculated, and the unit vector of the vector is obtained.

[0073] Calculate the dot product of two adjacent unit displacement vectors as the directional consistency coefficient. The closer this coefficient is to 1, the smaller the change in the direction of motion.

[0074] Simultaneously, the difference in magnitude between two adjacent displacement vectors is calculated to obtain the displacement increment amplitude; the sampling period is determined based on a preset sampling frequency, and the displacement increment amplitude is divided by the sampling period to obtain the average velocity within that time period. The preset sampling frequency can be set to 50Hz to 100Hz.

[0075] The average velocity over adjacent time periods is differentiated to obtain the average velocity change; this average velocity change is then divided by the sampling period to obtain an acceleration characteristic value with acceleration dimensions. This value can approximately reflect the pipeline's average acceleration capability within that time window.

[0076] Based on the mechanical motion inertial characteristics of the suspended pipeline, a directional consistency threshold and an acceleration inertia threshold are preset. Specifically, the mass parameters and rotational inertia parameters of the suspended pipeline are obtained, and these parameters can be obtained from pipeline design drawings or material manuals.

[0077] Since the displacement of a real pipeline is constrained by physical inertia, its acceleration and rate of change of direction have physical limits, while high-frequency noise usually exhibits abrupt motion characteristics that exceed these physical limits. Therefore, by combining a preset sampling frequency and based on the laws of mechanical motion and the support constraints of the pipeline, including but not limited to the fixed stiffness of the pipeline clamps and the cantilever length of the pipeline, the maximum allowable angular acceleration and maximum allowable linear acceleration that the pipeline may generate when disturbed are estimated using finite element simulation software.

[0078] In the specific estimation, the geometric dimensions, material density, elastic modulus and equivalent stiffness of the pipe are first obtained; then, a pipe-fixture coupled model is established in the finite element simulation software, the pipe is simplified into beam elements or shell elements, and the contact area of ​​the fixture is set as elastic constraint.

[0079] Since the pipeline satisfies the principle of linear superposition within its elastic range, a unit disturbance force is applied to the end of the pipeline to simulate the external disturbance during the welding process, and transient dynamic analysis is performed. Finally, the acceleration response time history curve of the pipeline end node is extracted, and the peak value in the curve is taken as the maximum allowable linear acceleration. At the same time, the angular acceleration time history curve of the pipeline around the support point is extracted, and the peak value is taken as the maximum allowable angular acceleration.

[0080] Secondly, the sampling period is calculated based on the preset sampling frequency. The product of the maximum allowable angular acceleration and the square of the sampling period is determined as the maximum allowable angle change. The cosine value of the maximum allowable angle change is calculated as the direction consistency threshold. This threshold represents the minimum allowable similarity of motion directions at adjacent moments. The larger the threshold, the higher the requirement for motion smoothness and the more direction changes are filtered out. The smaller the threshold, the more directional fluctuations are allowed, but some noise may be retained.

[0081] Simultaneously, the absolute value of the maximum permissible linear acceleration is defined as the acceleration-inertia threshold. This threshold has an acceleration dimension and represents the maximum physical limit of instantaneous acceleration allowed during pipe motion. A larger threshold indicates a faster dynamic response allowed by the system; a smaller threshold indicates a more rigid system and a stronger ability to suppress abrupt signals.

[0082] All coordinate points that simultaneously satisfy the conditions of having a directional consistency coefficient greater than the directional consistency threshold and an acceleration characteristic value within the acceleration-inertia threshold range are selected. These coordinate points represent displacement points where the direction of motion remains basically unchanged and the acceleration is limited by the inertial characteristics of the pipe. The acceleration-inertia threshold range is set as a symmetrical interval [-threshold, +threshold].

[0083] A set of consecutive coordinate points, such as five consecutive points, is extracted and used as a coordinate subsequence. These subsequences are then arranged in chronological order to obtain a low-frequency axis offset coordinate sequence. This sequence eliminates high-frequency noise and random disturbances while preserving the stable offset trend caused by the low-frequency oscillation of the pipeline itself.

[0084] It should be noted that the screening conditions based on the mechanical motion inertia characteristics of suspended pipelines are not simply adjustments to conventional filtering parameters. Instead, a simplified pipeline dynamics model is established to transform the pipeline mass distribution, rotational inertia, and support constraints into physical constraints on the displacement trajectory. Through progressive logical judgment, only continuous coordinate subsequences that simultaneously satisfy the directional change rate constraint and the acceleration physical constraint are retained, thereby achieving an essential distinction between real mechanical displacement and high-frequency interference.

[0085] The correction control module calculates the spatial offset between the welding torch center point and the actual center axis of the pipeline in real time from the low-frequency axis offset coordinate sequence, and adds the welding torch pose compensation amount. It calculates the correction step size and generates correction control commands through the PID control algorithm.

[0086] Considering the significant differences in the rigidity response characteristics of pipes with different specifications and wall thicknesses, control algorithms with fixed parameters struggle to adaptively adjust the correction intensity, easily leading to overshoot or slow convergence. Furthermore, if initial error and dynamic deviation are compensated together, the error may be amplified due to integral accumulation. Therefore, it is necessary to convert the extracted offset sequence into a six-dimensional pose error, superimpose the welding torch pose compensation, and use a PID control algorithm to generate a smooth drive signal, achieving adaptive adjustment and avoiding error accumulation.

[0087] In one specific embodiment, the method for obtaining the spatial offset is as follows: for each coordinate point in the low-frequency axis offset coordinate sequence, it is used as the coordinate of the welding gun center point, and the perpendicular vector from the point to the corresponding pipeline actual center axis is calculated as the position deviation vector.

[0088] The angle between the welding torch axis and the actual central axis is calculated as the attitude deviation vector. The position deviation vector and the attitude deviation vector are combined to construct a six-dimensional pose error vector as the spatial offset. This offset characterizes the degree of deviation of the welding torch from the actual central axis of the pipe at the current moment, including lateral, longitudinal, and vertical deviations in position, and tilt angle deviations in attitude.

[0089] Furthermore, considering that relying solely on real-time dynamic offset for correction cannot eliminate the initial system error remaining after mechanical alignment, and that the continued existence of this initial error will affect welding accuracy, it is necessary to fuse the welding torch pose compensation amount with the current spatial offset to obtain the total error value used for control.

[0090] Specifically, the method for obtaining the correction control command is as follows: the spatial offset and the welding torch pose compensation are vector-superimposed to obtain the total error value, wherein the welding torch pose compensation is the negative vector of the remaining static residual after mechanical alignment, and the spatial offset includes a static residual component and a dynamic offset component; through the vector superposition, the static residual component in the total error value is canceled out, so that the error signal input to the PID control algorithm mainly includes the dynamic offset component of the welding process.

[0091] Subsequently, the correction step size is calculated and correction control commands are generated using a PID control algorithm. Since the static component in the total error value has been canceled out by feedforward compensation, the integral term of the PID control algorithm does not need to be cumulatively adjusted for the fixed deviation, thus avoiding response lag or overshoot caused by integral saturation.

[0092] Specifically, the process involves obtaining the total error value at the current moment and historical moments such as the previous moment and the two moments before that, and then calculating the sum of the proportional, integral, and derivative terms using a discrete incremental PID control algorithm as the correction step size. The proportional term is used for rapid response to the current error, the integral term is used to eliminate steady-state accumulated error, and the derivative term is used to predict error trends and suppress overshoot. The proportional, integral, and derivative coefficients can be determined using the Ziegler-Nichols rule or through trial and error.

[0093] The kinematic transmission ratio of the welding torch actuator, such as a cross slide or robot joint, is used to convert the correction step size into drive signals, such as the speed and direction of a servo motor or the number of pulses, thereby generating the final correction control command. The kinematic transmission ratio is obtained from the robot or linear module's technical manual.

[0094] The correction control command is sent to the servo driver, which drives the welding torch to adjust its position and posture in real time according to the calculated step size, so that the welding torch is always aligned with the actual center axis of the pipeline, thereby realizing dynamic correction of axis deviation during the welding process.

[0095] Example 2.

[0096] Please see Figure 3As shown, the present invention proposes a real-time monitoring and correction method for pipeline welding operation position offset, including the following steps: S1, acquire point cloud data of two pipeline sections to be welded, fit and calculate the actual central axis and spatial relative position to obtain the pipe end coaxiality deviation, generate clamping and centering command to drive the device to adjust the posture, and convert the adjusted residual to the welding torch tool coordinate system to calculate the welding torch posture compensation amount.

[0097] S2. During the welding process, the coordinate data of the weld feature points are acquired in real time and the displacement change trajectory is constructed. Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertial threshold are set. Continuous coordinate subsequences with consistent direction and acceleration feature values ​​within the inertial threshold range are selected from the displacement change trajectory as low-frequency axis offset coordinate sequences.

[0098] S3. The low-frequency axis offset coordinate sequence is solved in real time as the spatial offset between the welding torch center point and the actual center axis of the pipeline section, and the welding torch posture compensation is superimposed. The correction step size is calculated by the PID control algorithm and the correction control command is generated.

[0099] In summary, this invention decouples fixed and dynamic deviations through the pipe end alignment module, solving the problem of decreased accuracy caused by mixed compensation of initial and dynamic deviations; extracts the true displacement trend based on mechanical inertia characteristics through the offset extraction module, solving the control jitter problem caused by high-frequency noise interference; and achieves dynamic correction by combining PID and pre-compensation through the correction control module, solving the problem of insufficient adaptive capability caused by differences in the rigidity response of different pipelines.

[0100] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0101] Those skilled in the art will recognize that the algorithmic steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0102] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0103] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0104] Finally, 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 spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A real-time monitoring and correction system for position deviation in pipeline welding operations, characterized in that, Includes the following modules: Pipe end alignment module: acquire point cloud data of the two pipe sections to be welded, fit and calculate the actual center axis and spatial relative position to obtain the pipe end coaxiality deviation, generate clamping and alignment command to drive the device to adjust the posture, and convert the remaining residual after adjustment to the welding torch tool coordinate system to calculate the welding torch posture compensation amount. Offset Extraction Module: During the welding process, the coordinate data of weld feature points are acquired in real time and the displacement change trajectory is constructed. Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertial threshold are set. Continuous coordinate subsequences with consistent direction and acceleration feature values ​​within the inertial threshold range are selected from the displacement change trajectory as low-frequency axis offset coordinate sequences. The correction control module calculates the spatial offset between the welding torch center point and the actual center axis of the pipeline in real time from the low-frequency axis offset coordinate sequence, and adds the welding torch pose compensation amount. It calculates the correction step size and generates correction control commands through the PID control algorithm.

2. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 1, characterized in that, The method for obtaining the coaxiality deviation of the pipe end is as follows: The point cloud data of the cross sections of the two pipes are divided into layers at equal intervals along the pipe axis. The coordinates of the center of the cross section are calculated by the least squares circle fitting algorithm for each layer of point cloud. The actual center axis is obtained by fitting the coordinates of the center of the cross section with a straight line using the least squares method. The vertical distance between the actual center axes of the two pipe sections at the weld joint is calculated as the positional deviation value, and the angle between the direction vectors of the two actual center axes is calculated as the angular deviation value. The positional deviation value and the angular deviation value together constitute the pipe end coaxiality deviation.

3. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 2, characterized in that, The method for obtaining the clamping centering command is as follows: Obtain the positional deviation and angular deviation values ​​in the coaxiality deviation of the pipe end, construct the pose difference matrix based on the positional deviation and angular deviation values, and decompose it into three-dimensional translational components and three-dimensional rotational components. The three-dimensional translation and rotation components are mapped to the drive quantities of each actuator of the automatic fixing and centering device to generate clamping and centering commands.

4. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 1, characterized in that, The method for obtaining the welding torch pose compensation amount is as follows: After the device adjusts its pose, the vision sensor is controlled to scan again to obtain the point cloud data of the cross-sections of the two pipes and to calculate the actual center axis after adjustment. The straightness error of each axis segment is calculated based on the adjusted actual center axis, and the segment with the smaller straightness error is selected as the reference pipe. Using the actual center axis of the reference pipe as a reference, calculate the angle vector between the vertical distance vector and the direction vector of the actual center axis of another pipe segment relative to the reference axis, and use it as the original residual. Construct a local coordinate system for the pipeline with the actual central axis of the reference pipeline as the Z-axis and the center point of the weld joint between the two pipeline sections as the origin. Project the original residual onto the local coordinate system to obtain the projected residual. Construct a homogeneous transformation matrix from the local coordinate system of the pipeline to the coordinate system of the welding torch tool. Use the homogeneous transformation matrix to transform the projection residual to the coordinate system of the welding torch tool to obtain the welding torch pose compensation amount.

5. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 1, characterized in that, The specific method for constructing the displacement change trajectory is as follows: During the welding process, the vision sensor is controlled to collect raw point cloud data of the weld area in real time. Statistical filtering algorithm is used to remove noise points from the original point cloud data. Then, background points with Z-axis coordinate values ​​less than a preset height threshold are removed from the original point cloud data. The remaining point cloud data is used as the point cloud of the weld area. The geometric centroid of the point cloud in the weld area is calculated as the coordinates of the weld feature points. The coordinates of the weld feature points are then transformed to the world coordinate system to obtain the coordinate data of the weld feature points. Mark the acquisition timestamps for the weld feature point coordinate data, and sort the weld feature point coordinate data according to the order of acquisition timestamps; The displacement change trajectory is constructed by connecting the coordinate data of adjacent weld feature points after sorting through linear interpolation.

6. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 1, characterized in that, The method for obtaining the low-frequency axis offset coordinate sequence is as follows: A sliding window is used to traverse the displacement change trajectory, and the displacement vector and the unit vector of the displacement vector between adjacent sampling points within the window are calculated. Calculate the dot product of adjacent unit displacement vectors as the directional consistency coefficient; The difference in magnitude between adjacent displacement vectors is calculated as the displacement increment amplitude. The sampling period is determined based on the preset sampling frequency. The displacement increment amplitude is divided by the sampling period to obtain the average velocity within that time period. The average velocities of adjacent time periods are differentiated to obtain the average velocity change. The average velocity change is then divided by the sampling period to obtain the acceleration characteristic value. Based on the mechanical motion inertial characteristics of the suspended pipe, a directional consistency threshold and an acceleration inertia threshold are set, and coordinate points with a directional consistency coefficient greater than the directional consistency threshold and an acceleration characteristic value within the range of the acceleration inertia threshold are selected. The set of points with a preset number of consecutively occurring coordinate points is taken as a coordinate subsequence, and they are arranged in order to obtain the low-frequency axis offset coordinate sequence.

7. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 6, characterized in that, The mechanical motion inertial characteristics based on the suspended pipe are used to set directional consistency thresholds and acceleration inertia thresholds, specifically as follows: The mass and moment of inertia parameters of the suspended pipe are obtained, and the maximum allowable angular acceleration and maximum allowable linear acceleration of the pipe are calculated based on the mechanical motion law by combining the preset sampling frequency. The product of the maximum permissible angular acceleration and the square of the sampling period is determined as the maximum permissible angle change, and the cosine value of the maximum permissible angle change is calculated as the directional consistency threshold. The absolute value of the maximum permissible linear acceleration is determined as the acceleration-inertia threshold.

8. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 1, characterized in that, The method for obtaining the spatial offset is as follows: For each coordinate point in the low-frequency axis offset coordinate sequence, it is used as the coordinate of the welding gun center point, and the perpendicular vector from the point to the corresponding actual center axis of the pipe is calculated as the position deviation vector. The angle between the welding torch axis and the actual center axis is calculated as the attitude deviation vector. The position deviation vector and the attitude deviation vector are combined to construct a six-dimensional pose error vector as the spatial offset.

9. The real-time monitoring and correction system for pipeline welding operation position deviation as described in claim 1, characterized in that, The method for obtaining the correction control command is as follows: The total error value is obtained by vector superposition of the spatial offset and the welding torch pose compensation. Obtain the total error value at the current moment and historical moments, and use the discrete incremental PID control algorithm to calculate the sum of the proportional, integral and derivative terms as the correction step size; The kinematic transmission ratio of the welding torch actuator is used to convert the correction step size into a drive signal to generate correction control commands.

10. A method for real-time monitoring and correction of positional deviation during pipeline welding operations, characterized in that, Includes the following steps: S1. Obtain point cloud data of the two pipe sections to be welded, fit and calculate the actual center axis and spatial relative position to obtain the pipe end coaxiality deviation, generate clamping and centering command to drive the device to adjust the posture, and convert the remaining residual after adjustment to the welding gun tool coordinate system to calculate the welding gun posture compensation amount. S2. During the welding process, the coordinate data of the weld feature points are acquired in real time and the displacement change trajectory is constructed. Based on the mechanical motion inertial characteristics of the suspended pipe, the direction consistency threshold and acceleration inertia threshold are set. Continuous coordinate subsequences with consistent direction and acceleration feature values ​​within the inertia threshold range are selected from the displacement change trajectory as low-frequency axis offset coordinate sequences. S3. The low-frequency axis offset coordinate sequence is solved in real time as the spatial offset between the welding torch center point and the actual center axis of the pipeline section, and the welding torch posture compensation is superimposed. The correction step size is calculated by the PID control algorithm and the correction control command is generated.

Images