6460 results about "Weld bead" patented technology

The invention relates to the technical field of data processing, and discloses a steel structure welding seam quality ultrasonic detection method and system. The method comprises the following steps: scanning a welding seam to obtain digital parameters, constructing a virtual model to generate an ultrasonic response database, optimizing a detection parameter forming scheme, executing distributed acquisition to obtain a signal matrix, extracting defect features by using a neural network, and outputting a digital detection report through transfer learning. According to the method, automatic optimization of detection parameters and accurate extraction and intelligent analysis of defect features are realized, the accuracy, efficiency and reliability of detection are improved, and the dependence on artificial experience is reduced.

The invention provides a teaching-free welding robot path planning system based on visual autonomous learning, and relates to the technical field of industrial welding, and the teaching-free welding robot path planning system comprises a pose compensation module which is used for arranging three infrared reflective mark points on the surface of a clamp base in an L shape, collecting original point clouds of the mark points and a workpiece by using a stereo camera, and preprocessing the point clouds; generating a pose compensation matrix based on the offset of the actual coordinate and the theoretical coordinate of the mark point, and outputting a de-noising point cloud data set and a compensation matrix; and the pose correction module is used for fitting a workpiece plane equation by adopting a random sampling consistency algorithm on the basis of the de-noised point cloud data set, acting a pose compensation matrix on the workpiece plane equation, correcting a pose error caused by clamp drifting, calculating a welding seam track end point coordinate and outputting a corrected welding seam pose parameter and a scanning pose instruction. According to the method, clamp drift can be compensated, weld joint poses can be corrected, three-dimensional features are extracted, a collision-free path is planned, and the welding precision, efficiency and automation degree are improved.

The invention relates to the technical field of welding, in particular to a corrosion steel welding cooperative control method and system. Comprising the following steps that welding seam geometric parameters, molten pool dynamic characteristic parameters and welding heat input parameters in the corrosion steel welding process are collected in real time through a multi-dimensional sensor array; constructing a corroded steel welding seam feature space model based on the welding seam geometric parameters, and determining material corrosion grade distribution and mechanical property parameters of a welding seam area in combination with a preset corroded steel material database; a molten pool form evolution prediction model is established through an adaptive Kalman filtering algorithm by utilizing the dynamic characteristic parameters of the molten pool and the welding heat input parameters, and the solidification behavior and the welding seam forming trend of the molten pool are predicted in real time; according to the material corrosion grade distribution, the mechanical property parameters and the molten pool forming trend, a dynamic adjustment strategy of the welding process parameters is generated through a multi-objective optimization algorithm; the reliability and safety of the corrosion steel welding joint can be improved.

This application relates to the technical field of battery packs, and a battery pack lower box body, a battery pack, an apparatus, and an assembly method of battery pack lower box body are provided. The battery pack lower box body includes a box frame, a base plate, a stiffening beam, and a stiffener. The base plate is arranged at the bottom of the box frame; the stiffening beam is arranged inside a space enclosed by the box frame and on top of the base plate; and the stiffening beam is fixed to the box fame via the stiffener, and the stiffener is welded to the box frame to form a weld seam, where a length of the weld seam is kept greater than a sum of side lengths of a projection of the stiffening beam onto the box frame.

The invention discloses a distributed reinforcement learning control method and system for multi-robot cooperative welding. A heterogeneous multi-robot task allocation algorithm, a multi-agent reinforcement learning model and welding parameters are combined. The method comprises the steps that a state space containing welding current, voltage, speed, welding gun posture, welding seam position and environment obstacle distribution is constructed, and an initial feasible operation area of the robot is determined; the overall welding task is decomposed into sub-tasks, and candidate robots are distributed; configuring a local decision module for each robot, and outputting an action decision containing a welding gun track adjustment amount and a parameter correction value; setting a global coordination module to generate coordination factors; each robot updates the decision and converts the decision into a driving instruction to execute operation, and molten pool temperature and weld size parameters are collected; the parameters are fed back to the model to adjust the state space characterization and decision weight. According to the method, decision collaborative precision is improved in combination with process parameters, self-adaptive optimization is achieved through real-time feedback, and collaborative welding efficiency and quality are improved.

The invention relates to the technical field of welding automation and robot visual perception, and discloses a robot intelligent welding method and system.The method comprises the steps that a three-dimensional point cloud image is constructed through a laser triangulation method; filtering and enhancing the point cloud image, and extracting a weld point cloud set by adopting a U-Net semantic segmentation model; carrying out space path identification and attitude regression based on the set to obtain a track element of a position-attitude pair; fitting a path by adopting a cubic B-spline curve and interpolating to generate a continuous welding track; and in the welding process, visual feedback is combined, and proportional differential control logic is introduced for closed-loop track error compensation. Compared with the technical problem that in the prior art, weld joint path recognition and accurate tracking cannot be stably achieved under complex structures such as curved surfaces and variable cross sections, a perception control closed loop based on visual servo is constructed, high-precision automatic welding of the complex weld joint structures is achieved, and the track fitting precision is improved.

The invention relates to the technical field of laser welding control, and discloses a laser welding head integrated with a push-pull wire and a control system of the laser welding head. The system comprises a push-pull wire action monitoring module, a laser energy modulation module, a welding seam trajectory tracking module, a welding quality feedback module, a parameter collaborative optimization module and a dynamic deviation correction execution module. Wherein the push-pull wire action monitoring module collects motion parameters in real time and generates a cooperative state index; the laser energy modulation module analyzes the laser power requirement according to the workpiece parameters and outputs a regulation and control instruction; the welding seam trajectory tracking module identifies welding seam characteristics and offset through a multispectral sensor and generates a dynamic tracking path; the welding quality feedback module evaluates defects and generates an abnormal index set; the parameter collaborative optimization module generates an optimization control parameter set based on the abnormal index set and the tracking path adjustment parameters; and the dynamic deviation correction execution module is used for correcting the track and generating a closed-loop control signal according to the matching action and the energy output time sequence.

The invention relates to the technical field of electronic packaging, in particular to a welding process applied to a double-mesa device. The technology comprises the following steps that firstly, a heat conduction model of a double-table-board device is built, a temperature field prediction system is built in combination with measured data, secondly, an asymmetric double-pulse-current welding strategy is adopted, differential pulse waveforms are loaded to an upper table board and a lower table board respectively, independent temperature control over all welding areas is achieved, a temperature difference closed-loop feedback mechanism is introduced in the process, and the temperature field prediction system is obtained. Pulse parameters are dynamically adjusted according to real-time temperature deviation, a stress prediction and weld penetration monitoring module is used for assistance, the welding spot forming quality and heat stress balanced release are ensured, a staged annealing type cooling mode is adopted after welding, and the structural stability is improved. The method can be widely applied to MEMS modules, SiC power devices and heterogeneous chip laminated packaging double-mesa structure devices, and has the advantages of being high in welding precision, high in reliability and wide in process adaptability.

The invention discloses a gradient welding strength control method and system for a welded galvanized steel pipe for fire fighting, particularly relates to the technical field of welding automation control, and aims to solve the problem of abnormal welding seam strength gradient caused by instantaneous unbalance of an electric field and a thermal field during dynamic parameter switching in the prior art. Electromagnetic distortion characteristic quantity is extracted through frequency domain energy analysis, and thermodynamic offset state quantity is deduced in combination with a thermal diffusion trend; on the basis of frequency band correlation mapping of electromagnetic and thermodynamic parameters, thermoelectric cooperative imbalance levels are divided through mutual information entropy evolution, and a self-adaptive harmonic attenuation weight tuning instruction is generated; analyzing a mismatching relation between transient disturbance of an environment magnetic field and arc voltage modulation distortion in real time, dynamically generating a magnetic field compensation factor and reversely superposing the magnetic field compensation factor to a control instruction; and in combination with the matching degree of molten pool oscillation energy and a harmonic frequency spectrum, correcting wire feeding rate calibration, reconstructing a multi-band harmonic energy proportion spectrum, outputting an anti-interference welding control signal, and realizing dynamic balance between arc stability and molten pool heat input.

The invention belongs to the technical field of nondestructive testing, and particularly relates to a detection method for quantitatively evaluating fatigue damage in a pipeline through a nonlinear ultrasonic technology. 316 stainless steel and SA335GRP-11 ferrite steel samples are adopted, a welding seam area is formed through V-shaped groove welding, fatigue loading is conducted, and actual pipeline fatigue damage is simulated; for the samples with different fatigue cycles, nonlinear response signals of the samples are obtained; fourier transform is carried out on the obtained signal to obtain an energy spectrum of the signal, and energy changes of the signal on a fundamental frequency band and a second harmonic frequency band are analyzed; and calculating a nonlinear coefficient, and analyzing a change rule of the nonlinear coefficient along with fatigue cycles, so that the evolution process of fatigue damage in the pipeline is evaluated. According to the method, energy is used for replacing the amplitude value, on the basis of the formula principle, integration is used for replacing the amplitude value to improve the intensity of the characterization nonlinear ultrasonic signal, the early stage of damage and the later stage of crack initiation are detected, and reliable guarantee is provided for safe operation of an industrial pipeline system.

The invention discloses an automatic dovetail welding seam grinding control method based on robot visual positioning, and relates to the technical field of visual positioning. The method comprises the steps that after welding is completed, a dovetail welding seam image is collected and converted into a three-dimensional coordinate through a vision algorithm, and a three-dimensional point cloud is generated; smooth interpolation is performed on the three-dimensional point cloud through a B spline curve method to obtain a parameterized curve, and a control point set is optimized through a genetic algorithm to generate a global optimal polishing path; the polishing robot executes a task according to a path, a tail end sensor collects a real-time path and calculates a deviation value with a global path, and when the deviation value exceeds a preset threshold value, inverse kinematics is triggered to solve and correct the path; the hardness of the dovetail weld is measured through laser-induced breakdown spectroscopy, a comprehensive hardness value is obtained in combination with a matrix hardness database, meanwhile, a visual sensor collects the surface state, and polishing process parameters are dynamically adjusted according to the surface state; after the task is completed, the welding seam angle and flatness are detected. According to the method, the optimal path is generated through visual positioning, and automatic grinding of the dovetail welding seam is achieved.

The invention discloses a pulsed eddy current detection method and system for a thermal power pipeline with a coating layer, and relates to the technical field of pulsed eddy current nondestructive detection.The pulsed eddy current detection method comprises the steps that pulsed eddy current detection simulation and experiment are conducted under the condition of a three-dimensional simulation model and a test piece, and excitation current waveform, amplitude and frequency parameters are determined; an induced voltage signal is collected, peak values, time points and attenuation characteristics are extracted, the corresponding relation between the wall thickness, the coating layer thickness and the welding seam structure and detection signals is constructed, a parameterized database is constructed, and temperature correction is completed; and optimizing a sensor structure and a signal acquisition scheme according to a parameterized database, obtaining detection data in a pipeline operation state, completing two-dimensional imaging of wall thickness distribution and welding seam positioning, and forming detection information covering a straight pipe, an elbow and a welding seam area. According to the method, wall thickness detection and welding seam positioning of the thermal power pipeline under the conditions of a thick wall and a thick coating layer are achieved through the full-process design of test piece modeling, simulation experiment and imaging collection.

The invention provides a parameter control method and system in the friction stir welding process, and relates to the technical field of welding equipment of metal materials. The method comprises the steps that before friction stir welding is started, working condition information of a to-be-welded workpiece is input into a welding process simulation model for virtual test welding; screening out a plurality of groups of standard process parameter combinations meeting preset welding quality requirements; in the friction stir welding process, the change trend of all the technological parameters is determined according to the real-time technological parameter combination within the first preset time period; and if it is determined that the deviation risk exists in the current welding state within the second preset time period in the future according to the change trend, the standard technological parameter combination with the maximum similarity with the real-time technological parameter combination in the standard technological parameter library is determined as the target technological parameter combination to conduct parameter control adjustment on the welding equipment. By implementing the method, the welding defects caused by sudden change of the process window can be overcome, and the stability of the welding process and the consistency of the welding seam quality are improved.

The invention belongs to the field of robot control, and particularly relates to a robot self-adaptive impeller welding control method and system fused with multi-modal data, and the method comprises the steps: obtaining an impeller workpiece image and three-dimensional point cloud data through an industrial vision subsystem, and extracting a three-dimensional welding seam position point set representing a welding seam track; generating a candidate welding path set based on welding seam geometric feature analysis; in combination with welding seam morphology and process requirements, optimal welding process parameters are predicted through a regression model, and a welding gun attitude angle is calculated through quaternion conversion according to a curved surface normal vector; in the welding process, the relative position deviation of a robot and a workpiece and the parameter deviation of welding current / voltage and a planned value are monitored in real time, when the deviation exceeds a preset tolerance interval, a PID control algorithm is adopted to dynamically adjust robot motion parameters or welding process parameters, and closed-loop control over the welding process is achieved; according to the method, precise planning of the welding path and self-adaptive regulation and control of the process are achieved, and the welding quality and the forming consistency are remarkably improved.

The invention relates to the technical field of welding platforms, in particular to a multifunctional welding process experiment platform and a process optimization system.The process optimization method comprises the following steps that a welding process setting module is constructed, and a welding mode, a groove type and a heat input scheme are preset; outputting the initial process configuration parameter set to drive a welding execution unit to perform automatic welding operation according to the initial process configuration parameter set; extracting a multi-dimensional process response vector including a stability index, metal fusion depth distribution and a heat affected zone expansion degree; outputting a welding seam grade result according to a preset quality grading standard, and generating a feedback identifier; and after updating, inputting to the step S2 again to form closed-loop optimization control. According to the method, digital modeling and evaluation of key indexes such as the welding stability, the fusion depth and the heat affected zone are achieved, and the problem that a traditional process depends on visual inspection and experience judgment is solved.

The invention discloses a non-standard weldment-oriented robot efficient collision-free path planning method and system, and belongs to the technical field of robot welding, and the method comprises the following steps: S1, obtaining three-dimensional model data of a non-standard weldment and an operation environment, and extracting welding seam point cloud data; s2, carrying out feature extraction and downsampling on the welding seam point cloud data to generate a group of key path points; s3, global optimization is conducted on the welding path through an improved genetic algorithm, candidate paths are generated, and the welding sequence of the multiple sections of welding seams and the welding gun postures at the key path points are optimized at the same time through the improved genetic algorithm; s4, evaluating the candidate paths generated by the improved genetic algorithm through a multi-target fitness function including path length, attitude stability and collision risk; and S5, outputting an optimal collision-free welding path according to an evaluation result, and converting the optimal collision-free welding path into a smooth track which can be executed by the robot. According to the invention, the high-efficiency, high-quality and collision-free welding gun motion trail can be automatically generated.

The invention discloses a pipeline all-position multi-layer multi-channel TIG welding bead morphology measuring device and method, and relates to the technical field of intelligent welding. A molten pool image collecting module is used for collecting morphology information of a molten pool, and the light emitting end of a laser transmitter is used for facing a welding bead; the active and passive vision module is used for collecting an image of light emitted by the laser emitter projected on a weld pool groove and identifying a weld bead corner opening position and a weld bead width, and the active and passive vision module and the weld pool image collecting module are both electrically connected with the control module; the control module can fuse image features collected by the active and passive vision module and the molten pool image collection module, the axes of the molten pool image collection module, the welding module and the active and passive vision module are located in the same plane, and the molten pool image collection module and the active and passive vision module are located on the two sides of the welding module respectively. According to the method, manual participation in the welding process can be reduced, the welding automation degree is improved, and meanwhile, the welding bead morphology measurement accuracy is improved.

The invention relates to the technical field of pipeline welding automation, and discloses a pipeline multi-layer and multi-pass welding track self-adaptive planning method. The method comprises the steps that weld joint track poses, welding electrical parameters and temperature information are collected in real time, and a sliding time window data sequence is constructed; establishing an interlayer constraint model and a trajectory prediction model based on the sequence, and generating a plurality of trajectory deviation predictions through short-term recursive prediction and long-term sequence prediction; a residual error reciprocal weighted fusion strategy is combined with a welding seam forming quality evaluation function, fusion track deviation is obtained, and uncertainty is evaluated; the fusion deviation is superposed to an original planned trajectory, and a self-adaptive correction trajectory is generated through a multi-objective optimization model; and the welding robot is controlled to execute track correction and real-time feedback updating. According to the method, through dual-time scale prediction and trajectory-process parameter collaborative optimization, the problem of insufficient welding seam forming precision caused by lack of dynamic correction in traditional static planning is effectively solved, and the welding quality and efficiency are remarkably improved.

The invention provides a welding path planning method based on multi-modal fusion and deep learning, and relates to the technical field of welding automation, the method comprises the following steps: synchronously acquiring three-dimensional coordinate information of a welding workpiece through a laser ranging module and a CCD visual sensor, and fusing laser height parameters and visual plane feature points; an initial path is generated based on the dynamic welding adaptation coefficient and the welding environment complexity, and laser errors, visual matching deviation and environment interference intensity are quantified; laser and visual data are dynamically weighted by using multi-modal feature fusion weights, spatial-temporal features are extracted through a deep learning model, and a welding track is optimized; and the six-axis mechanical arm is controlled to execute welding operation, weld quality is monitored in real time, and path parameters are adjusted in a closed-loop mode. According to the method, the problem of insufficient precision of a single sensor is solved through a multi-modal data complementation mechanism, sudden obstacles and temperature fluctuation are dealt with in combination with dynamic parameter calculation and real-time path correction, and the local path optimization capability is improved by using the convolutional neural network and a self-attention mechanism.

The invention discloses a weld joint three-dimensional structure detection device based on vortex laser, and relates to the field of laser welding, and the weld joint three-dimensional structure detection device comprises a vortex laser generation system used for generating and emitting vortex laser; the image acquisition unit is used for collecting returned laser data projected to the welding seam and carrying out dynamic scanning and data acquisition; the motion control and feedback system is used for adjusting the relative position of the laser welding head and the workpiece in real time and dynamically adjusting parameters such as welding power and scanning speed according to the collected data result; and the data processing and analyzing platform is used for carrying out image processing and feature extraction on the collected multi-dimensional data and modeling the morphology of the welding sample. The vortex laser serves as a light source for welding seam detection, high-precision, high-depth, multi-parameter and strong-interference-resistance three-dimensional welding seam modeling is achieved, the problems that in the long and thick plate welding process, welding seam detection parameters are few, precision is low, and real-time adjustment cannot be achieved are solved, and meanwhile the device has the performance characteristics of being high in detection precision, high in efficiency and comprehensive in structural parameter.

The invention relates to the technical field of automatic detection, and discloses a cantilever crane welding defect detection method and system. The method comprises the following steps: establishing a detection model by adopting historical boom data to position plate positions and angles; establishing a welding seam detection model by utilizing historical welding data to evaluate the welding quality; collecting and preprocessing six-view-angle image data of the boom; inputting the preprocessed images into the two models for respective detection; if the output of the two models is qualified, the welding is passed, otherwise, an alarm is activated. The system comprises a data acquisition module, an image preprocessing module, a cantilever crane detection module, a welding seam detection module and a result judgment module. According to the invention, the automatic and intelligent detection of the welding quality of the cantilever crane is realized, the detection efficiency and accuracy are improved, the welding quality of the cantilever crane is ensured, and the equipment stability and service life are improved.

A friction stir additive manufacturing device configured to join a first work-piece and a second work-piece is provided. In one aspect, the device includes a rotating spindle configured to deposit a filler material over a weld line as the device is advanced along an interface between the first work-piece and the second work-piece. The device also includes a deposition head configured to receive at least a portion of the spindle, the deposition head configured to remain stationary relative to the rotating spindle. The deposition head includes a first semi-cylindrical portion having an inner radius and an outer radius relative to a first axis, and a second semi-cylindrical portion having an inner radius and an outer radius relative to a second axis that is perpendicular to the first axis. The second semi-cylindrical portion can include a chamfered inner surface configured to define a weld profile.

The precise welding control method for vacuum brazing comprises the steps that the geometrical shape and material characteristics of a workpiece are obtained, modeling processing is conducted in combination with the geometrical shape and the material characteristics, and a high-dimensional tensor and a high-dimensional time sequence are compared and calculated to obtain a deviation matrix; based on the deviation matrix, the brazing heating power and the brazing filler metal supply rate are adjusted, and a real-time control instruction set is generated; constructing a deep reinforcement learning model, performing iterative optimization on the real-time control instruction set based on the deep reinforcement learning model, extracting weld defect characteristics of the workpiece, generating a weld quality score according to the weld defect characteristics, and obtaining a difference value between the weld quality score and a set threshold value; and performing comprehensive optimization on the optimization parameter vector according to the difference value to generate a global optimization parameter set. According to the method, systematic dynamic control over complex geometrical shapes and multi-material workpieces is achieved, the defects of response lag and insufficient regulation and control are overcome, and therefore the process stability and production efficiency of vacuum brazing are greatly improved.

The invention provides an aluminum alloy battery shell automatic laser welding method and system.The method comprises the steps that firstly, three-dimensional geometric data and contact position coordinates of the surface of a shell are obtained through a laser scanner, a digital model is constructed, then contact material characteristics are extracted, and initial welding parameters are determined; a real-time visual system and an ultrasonic sensor are used for correcting the position of a welding head, stable contact is established, energy distribution is optimized according to the heat conductivity and the geometrical shape of a material, preheating treatment is executed, in the welding process, temperature changes are monitored through infrared thermal imaging, laser parameters are adjusted, stable penetration depth is kept, and finally the quality of a welding seam is evaluated through air tightness detection. By means of multi-sensor fusion and self-adaptive control, high-precision and high-quality laser welding of the battery shell with the complex shape is achieved, and welding reliability and production efficiency are improved.

The invention relates to the technical field of industrial CT, discloses an automatic light beam centering method and system based on industrial CT, and aims at solving the technical defects that in the welding process, due to thermal deformation, the centering precision of a light beam and a welding seam is low, and the reliability is poor. The method comprises the steps that before welding, three-dimensional CT volume data of a workpiece to be welded are obtained to generate a nominal welding seam trajectory; establishing a thermal-mechanical coupling finite element analysis model, and simulating a welding process to generate a predicted thermal teratogenic field map; superimposing the predicted thermal teratogenesis to a nominal trajectory to generate a dynamic welding trajectory; in the welding process, a laser contour measurement sensor is driven to collect three-dimensional contour data of a welding seam area in real time and extract the actual welding seam center position; and establishing a Kalman filtering state observer. By the adoption of the technical scheme, continuous and high-precision tracking of the center position of the welding seam can be achieved, the influence of thermal teratogenesis is overcome, and the welding quality and the yield of complex structural parts are remarkably improved.

The invention relates to the field of welding pressure control, in particular to a friction stir welding pressure control method, system and equipment and a storage medium. A three-axis acceleration sensor is used for monitoring the vibration acceleration of a main shaft to start a pressure compensation algorithm, a thermal infrared imager is used for collecting temperature field distribution of a welding area to dynamically correct welding pressure, and after welding is completed, control parameters are reversely optimized through ultrasonic flaw detection to form a self-learning control loop. The invention further comprises a corresponding control system, electronic equipment and a computer readable storage medium. The technical effects that the friction stir welding pressure is accurately controlled, the pressure can be adjusted in real time according to factors such as material characteristics, vibration and temperature, the welding process is optimized, the welding seam quality is improved, and the control parameters are continuously improved through self-learning are achieved.

The invention provides a pipeline welding all-position welding seam identifying and tracking system and method, and belongs to the technical field of pipeline welding automation. In order to solve the problems of strong arc light interference, low tracking precision, slow response speed and insufficient robustness in the existing pipeline all-position welding tracking technology, the system comprises a sensing layer, a processing layer, a decision-making layer and an execution layer. The welding seam identifying and tracking method comprises the following steps: S1, calibrating the welding seam identifying and tracking system; s2, all-position welding tracking is carried out; and S3, performing adaptive adjustment on different scenes. According to the method, dynamic weighted fusion and fuzzy PID control are adopted, multi-posture welding such as flat welding, vertical position welding, horizontal position welding and overhead position welding can be automatically adapted, parameters do not need to be manually adjusted, and the operation threshold is lowered.

The invention relates to the technical field of process optimization, in particular to a collaborative optimization analysis method for multi-station tower drum welding. In order to solve the problem that cooperative control of tower drum multi-station welding is difficult, firstly, multi-mode sensing data of welding seam vision, arc sensing, thermal field distribution and welding voiceprint are loaded, a fusion evaluation model is constructed to obtain a welding seam deviation value and a sensor confidence score, and the health state of a sensor is judged according to a threshold value; and a corresponding fault-tolerant reconstruction control mode is triggered to obtain welding tracking control parameters, then a multi-station welding heat source model is loaded, thermal interference is decoupled in combination with the control parameters and thermal field data, multiple stations are cooperatively regulated and controlled, a cooperative control instruction is output and executed, and welding process optimization is achieved.

The invention discloses a welding defect detection method based on deep learning. The method comprises the following steps: collecting a welding defect image; carrying out image preprocessing on the image, carrying out classified marking on welding defects in the image and forming a data set; dividing the data set; constructing a weld defect detection model; training the model by using the training set, verifying by using the verification set, calculating the gradient of a loss function and a loss value of the model in each training to model parameters by using training and verification results, and updating the parameters of the model according to the gradient and the learning rate; judging whether the model performance reaches the expectation or reaches the maximum number of training times; and outputting the trained model and evaluating the performance of the trained welding defect detection model by using the verification set. According to the multi-scale convolution attention module, the calculation amount is reduced, the model can more effectively capture the characteristics of the strip-shaped welding defects, and the multi-scale defect detection performance of the model under the complex background is improved.

The invention provides an auxiliary device for steel pipe construction welding, and relates to the technical field of pipeline welding, and the auxiliary device comprises two arc-shaped bearing bodies, an auxiliary main frame, an aligning mechanism and a rotary driving system. The central angle of the arc-shaped bearing body is larger than 180 degrees, automatic centering of the steel pipes is achieved through the outer pressing supporting rod assembly and the inner positioning supporting rod assembly, and it is guaranteed that the two sections of steel pipes are concentric. And the welding head mounting seat is connected with the rotary driving piece through the arc-shaped movable clamping body to drive the welding head to continuously rotate around a welding seam, so that high-quality girth welding is completed. The device solves the problems of low centering precision and poor efficiency in traditional welding, has the advantages of high automation degree, high stability, convenience in assembly and disassembly and the like, and is suitable for quick welding of municipal steel pipe engineering.