A cable joint digital acceptance system

Abstract

The application provides a cable joint digital acceptance system, and relates to the fields of cable accessory construction detection and point cloud processing. The system receives original point clouds of a cable joint, scanning device calibration parameters, a cable structure template, a stripping and cutting process template and an acceptance rule template, performs coordinate unification and reliable weight calculation on multiple frames of point clouds, establishes a curved engineering reference axis and a non-torsion cross-section coordinate system; generates a cross-section sequence along the engineering reference axis, identifies the structure area and processing boundary of a conductor, an insulating layer, a semi-conductive layer and a sheath; constructs a virtual caliper with stable contact band constraints at a specified back-off position, calculates diameter, ovality, length, slope and defect size; forms a judgment result according to measurement uncertainty and acceptance rules, and binds original point clouds, algorithm parameters, measurement cross-sections, artificial correction and report fields into an evidence chain. Micro-loading detection and contact pressure inversion are set as external extension data interfaces.

Description

Technical Field

[0001] This invention relates to the fields of cable accessory construction inspection, three-dimensional point cloud processing, and digital measurement technology, specifically to a digital acceptance system for cable joints. Background Technology

[0002] The cable joint manufacturing process includes conductor treatment, insulation shielding layer break processing, insulation surface cutting, insulation beveling, surface grinding, and accessory installation. The quality of cable joint manufacturing is affected by conductor eccentricity, remaining insulation layer thickness, insulation surface ellipticity, local depressions, semi-conductive layer break location, surface scratches, and grinding uniformity.

[0003] Current on-site acceptance mainly relies on measuring conductor exposed length, insulation break location, insulation outer diameter, and surface finish using measuring tapes, rulers, vernier calipers, and visual photographs. 3D scanning technology converts the cable joint surface into a point cloud, obtaining geometric dimensions and surface defects through cross-section fitting, virtual calipers, and image recognition. However, the general point cloud processing workflow lacks a unified data relationship between the cable joint's structural layers, stripping process sequence, and acceptance fields.

[0004] Cable joints in the field exhibit bending, oblique cuts, gentle slope transitions, partial gaps, and scan obstruction. When using a whole straight axis or a single cylindrical fitting to establish the cross-section, local cross-sections become tilted; when using point cloud projection extrema to calculate the diameter, isolated noise points, missing points, and low-coverage areas appear in the dimensional results; when using a single color threshold to identify structural areas, black semiconductive layers, shadows, and reflective areas cause confusion.

[0005] In existing software, structure recognition, dimensional measurement, manual correction, and report output are all generated as independent data sets. Measurement results are not linked to corresponding point cloud regions, measurement sections, algorithm parameters, rule entries, and manual modification records. Acceptance personnel find it difficult to verify the spatial location and calculation process of the same measurement value, and the dimensional fields in the report lack a continuous traceability relationship with the 3D evidence. Summary of the Invention

[0006] Technical problems to be solved

[0007] The technical problem to be solved by this invention is that existing digital acceptance of cable joints lacks a unified measurement weight for the quality difference of multi-frame point clouds, dimensional deviations occur when the cross-section of a bent cable is established using an overall straight axis, and the identification of structural boundaries is unstable due to bevels, gentle slopes, and local residues. Point cloud extreme value measurement cannot express the stable contact process of a physical caliper, and there is a lack of fixed correlation between measurement results, rule judgments, and three-dimensional evidence.

[0008] Technical solution

[0009] To achieve the above objectives, the present invention provides the following technical solution: a digital acceptance system for cable joints, comprising a task and template management module, a point cloud data access module, a point cloud quality review module, an engineering reference coordinate construction module, a structure identification module, a dimension measurement module, an acceptance judgment module, and a report evidence chain module; the task and template management module stores scanning equipment calibration parameters, cable structure templates, stripping process templates, and acceptance rule templates; the point cloud data access module receives multiple frames of original point clouds of the cable joint and the acquisition parameters corresponding to each scanning frame; the point cloud quality review module converts the multiple frames of original point clouds to a unified engineering coordinate system and determines the measurement confidence weight of each point based on point density, normal consistency, scanning incident direction, multi-frame registration residual, and equipment acquisition confidence; the engineering reference coordinate construction module establishes an engineering reference axis in the form of a spatial curve based on the weighted point cloud, and establishes an axial arc length along the engineering reference axis. The system comprises a non-torsion curve coordinate system consisting of the cross-sectional orientation angle and radial position; the structure identification module generates a cross-sectional sequence along the engineering reference axis, and determines the structural regions and processing boundaries of the conductor, insulation layer, semi-conductive layer and sheath based on the cross-sectional radial envelope, outer point ratio, normal variation, cross-sectional coverage, material type probability and axial continuity; the dimension measurement module establishes a measurement cross-section at the specified retraction position corresponding to the processing boundary, constructs mutually parallel virtual cards with stable contact band constraints, and calculates the axial length, transverse diameter, longitudinal diameter, ellipticity, slope and defect size based on the normal distance between the virtual cards; the acceptance judgment module maps the measurement results, measurement uncertainty and acceptance rule template to form a qualified, reviewed or unqualified status; the report evidence chain module establishes an associated record of the original point cloud, point cloud version, algorithm version, template version, measurement cross-section, manual correction record, 3D screenshot and acceptance field.

[0010] Preferably, the point cloud quality review module converts the points in each scan frame to the engineering coordinate system according to the calibration parameters of the scanning equipment. It performs rigid registration and local non-rigid correction through the overlapping area of ​​adjacent scan frames, and normalizes the point density, normal consistency, scanning incident direction, registration residual, and equipment acquisition confidence to mass components from 0 to 1. It also calculates the measurement confidence weight according to the five weights in the acceptance rule template. Points with measurement confidence weights lower than the effective threshold retain their original index and are marked as non-metric points. Non-metric points do not participate in the processing boundary fitting and virtual caliper contact calculation.

[0011] Preferably, the engineering reference coordinate construction module uses weighted principal component analysis to obtain the initial axial direction, extracts the center of continuous cross sections along the initial axial direction, fits the center trajectory of a cubic B-spline with the center of the cross sections as the control point, and establishes a parallel moving frame along the center trajectory of the cubic B-spline; the parallel moving frame includes a tangential unit vector of the center trajectory and a first radial unit vector and a second radial unit vector located within the local cross sections, and the first radial unit vector and the second radial unit vector are transferred between adjacent cross sections according to the minimum rotation angle.

[0012] Preferably, the structure recognition module forms a cross-section sequence according to a set axial step size, calculates the radial upper quantile envelope, radial lower quantile envelope, outer layer point ratio, normal dispersion, curvature distribution, cross-section coverage, and material category probability for each cross-section, and forms a processing boundary response based on the characteristic difference between adjacent cross-sections; the structure recognition module performs continuous interval filtering on the processing boundary response according to the structural order of conductor, insulating layer, semiconductive layer, and sheath arranged along the axial direction, and outputs the axial range of each structural region and the axial position of the processing boundary at different circumferential angles.

[0013] Preferably, the dimension measurement module establishes a transverse virtual caliper and a longitudinal virtual caliper within a series of consecutive measurement sections at a specified retraction position. Each virtual caliper includes a first virtual caliper plate and a second virtual caliper plate that move toward each other along the same normal direction. The first virtual caliper plate and the second virtual caliper plate are respectively stopped at positions where the number of contact points, the axial span of the contact points, the circumferential span of the contact points, and the contact point measurement confidence weight satisfy the stable contact zone threshold. The normal distance between the two stop positions is determined as the measurement diameter in the corresponding direction.

[0014] Preferably, the structure recognition module unfolds the cable surface into a two-dimensional surface grid according to the axial arc length and circumferential angle, calculates the height residual, curvature abrupt change and normal change based on the locally fitted surface, and determines the three-dimensional regions of semiconductive layer residue, scratches, protrusions and depressions by combining color features or reflection intensity features; the size measurement module records the axial length, circumferential angle range, maximum height difference, maximum depth and surface area of ​​each three-dimensional region.

[0015] Preferably, the acceptance judgment module generates uncertainty intervals for each measurement item based on scanning calibration error, multi-frame registration error, measurement section positioning deviation, and continuous section repeat measurement deviation, and compares the measurement item and its uncertainty interval with the lower limit, upper limit, and verification zone in the acceptance rule template; the report evidence chain module saves the corresponding point cloud region index, section number, virtual card position, rule entry identifier, algorithm parameter summary, and differences before and after manual correction for each measurement item.

[0016] Preferably, a digital acceptance method for cable joints includes the following steps: acquiring multi-frame original point clouds of the cable joint, scanning equipment calibration parameters, cable structure templates, stripping process templates, and acceptance rule templates; converting the multi-frame original point clouds to a unified engineering coordinate system; and determining measurement confidence weights based on point density, normal consistency, scanning incident direction, multi-frame registration residuals, and equipment acquisition confidence; establishing an engineering reference axis in the form of a spatial curve based on the weighted point cloud; establishing a torsion-free curve coordinate system along the engineering reference axis and generating a cross-section sequence; and determining the measurement confidence weights based on the cross-section radial envelope, outer layer point ratio, and normal variation. The structural region and processing boundary are determined by the cross-sectional coverage, material type probability, and axial continuity. A measurement section is established at the specified retraction position corresponding to the processing boundary. Dimensional measurement results are obtained using transverse and longitudinal virtual calipers with stable contact zone constraints. Defect measurement results are obtained based on the height residual, curvature abrupt change, and normal change of the unfolded surface. An acceptance status is formed based on the dimensional measurement results, defect measurement results, measurement uncertainty, and acceptance rule template. The original point cloud, algorithm version, template version, measurement section, manual correction record, and acceptance fields are written into the acceptance report and evidence chain.

[0017] Preferably, when determining the processing boundary, the radial upper quantile envelope difference, outer layer point ratio difference, normal dispersion difference, curvature distribution difference, and material category probability difference of the continuous cross section are normalized and weighted. The continuous boundary interval is screened according to the structural arrangement order defined by the cable structure template, and the position with the largest boundary response within the continuous boundary interval is determined as the processing boundary of the corresponding circumferential angle. When performing virtual caliper measurement, points with measurement confidence weights lower than the effective threshold are excluded from contact calculation. The median diameter of multiple consecutive measurement cross sections is used as the measurement diameter, and the dispersion of the diameter of multiple consecutive measurement cross sections is used as the repeatability measurement deviation.

[0018] Preferably, after completing the point cloud-based acceptance processing, the surface load matrix output by the external micro-loading detection device is received. and displacement response matrix ,according to The load transfer operator is derived under displacement reconstruction residuals, reciprocity, spatial smoothing, and positive semidefinite constraints. The load transfer operator The Line number Column elements represent the first When the surface unit bears a unit load, the first Radial displacement generated by each surface unit; based on the load transfer operator of the cable processing area. Attachment load transfer operator Initial gap vector and contact load vector Establish and satisfy , and ; a diagonal matrix formed by the areas of surface units Calculate the contact pressure vector , the contact pressure vector The corresponding extended review results are written into the acceptance record.

[0019] Beneficial effects

[0020] This invention provides a digital acceptance system for cable joints based on point cloud processing and digital metering, which has the following advantages:

[0021] 1. This invention establishes measurement confidence weights for different scan frames and different points, and uses a space curve-shaped engineering reference axis and a non-torsion curve coordinate system to form a continuous normal section, thereby reducing the impact of scan quality differences and cable bending on structural positioning and dimensional measurement.

[0022] 2. The present invention constrains the axial arrangement of the structural area according to the cable structure template and the stripping process template, determines the processing boundary with the continuous features of multiple sections, and uses a virtual caliper with stable contact band constraint to eliminate isolated noise points and local defects.

[0023] 3. This invention incorporates scanning calibration error, multi-frame registration error, cross-section positioning deviation, and repeated measurement deviation into the acceptance criteria, and binds each measurement field to the point cloud region, cross-section position, rule entry, and manual correction record.

[0024] 4. This invention retains micro-load detection, load transfer operator and accessory contact pressure analysis through an independent extended data interface. The basic software system completes point cloud acquisition, point cloud processing, size measurement, structure identification, acceptance judgment and report output without relying on extended hardware. Attached Figure Description

[0025] Figure 1 This is the overall software flowchart of the digital acceptance system for cable joints of the present invention;

[0026] Figure 2 This is a diagram showing the layered architecture and module relationships of the software system of this invention;

[0027] Figure 3 This is a flowchart of the multi-frame point cloud quality review and coordinate unification process of the present invention;

[0028] Figure 4 A flowchart is established for the engineering reference axis, the non-torsional curve coordinate system, and the cross-sectional sequence of this invention.

[0029] Figure 5 This is a flowchart illustrating the identification process of structural regions, processing boundaries, and surface defects in this invention.

[0030] Figure 6 This is a flowchart illustrating the virtual caliper dimension measurement and measurement uncertainty formation process of the present invention.

[0031] Figure 7 This is a flowchart for determining the acceptance rules and generating the evidence chain for the report of this invention;

[0032] Figure 8 This is a diagram showing the extended data interface for external micro-loading detection and contact pressure analysis in this invention.

[0033] Figure 9 This is a screenshot of the main interface of the system of the present invention;

[0034] Figure 10 This is a screenshot of the system functions of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. The described embodiments include basic software embodiments and extended hardware embodiments, wherein the basic software embodiments constitute the implementation basis of the current software product of the present invention, and the extended hardware embodiments are accessed through independent data interfaces. The numerical values ​​in this embodiment are used to provide directly implementable software parameters and extended device configurations, and do not constitute an additional limitation on the scope of protection of the claims. To avoid the same mathematical symbol corresponding to different meanings, the mathematical symbols in this specification are explained one by one after the first appearance of the formula.

[0036] Specific Implementation Example 1: Basic Software Implementation Example.

[0037] This embodiment is deployed on an industrial computer or server. The software consists of a task and template management module, a point cloud data access module, a point cloud quality review module, an engineering reference coordinate construction module, a structure identification module, a dimension measurement module, an acceptance judgment module, and a report evidence chain module. Each module uses the task number as the business primary key and the point cloud version, surface unit number, and section number as spatial data indexes.

[0038] The task input includes cable type, voltage rating, conductor cross-sectional area, nominal insulation outer diameter, scanning equipment calibration parameters, cable structure template, stripping process template, and acceptance rule template. The point cloud data adopts a file structure that includes three-dimensional coordinates, color, reflection intensity, normal vector, scan frame number, scan pose, and equipment acquisition confidence level; when color or reflection intensity is missing, the template weight of the corresponding feature channel is set to zero.

[0039] The point cloud quality audit module first checks the file summary, coordinate units, number of points, spatial range, and scan frame integrity. Then, based on the external parameters of the scanning equipment, it converts each scan frame to a unified engineering coordinate system. Adjacent scan frames are registered from point to plane through overlapping areas. After registration, the local point density, normal consistency, scan incidence direction, multi-frame registration residual, and equipment acquisition confidence level of each point are recorded. These five indicators are normalized to 0 to 1 and summed according to the fixed weights in the acceptance rule template to form the measurement confidence weight.

[0040] Points with a measurement confidence weight below 0.50 are marked as non-metric points; points with a measurement confidence weight between 0.50 and 0.75 are marked as low-weight points; and points with a measurement confidence weight above 0.75 are marked as valid metric points. Non-metric points are retained in the original point cloud and 3D display but are not used for reference axis fitting, machining boundary fitting, or virtual caliper contact calculations.

[0041] The engineering reference coordinate construction module uses effective measurement points to perform weighted principal component analysis to obtain the initial axial direction; candidate sections are established every 2 mm along the initial axial direction, and robust ellipse fitting is performed on each candidate section to extract the section center; cubic B-spline fitting is used to fit the center of continuous sections to form the engineering reference axis, and a local normal section is established using a parallel moving frame to ensure that the circumferential zero-degree direction of adjacent normal sections is transferred according to the minimum rotation angle.

[0042] The structural identification module generates an identification cross section every 0.5 mm along the engineering reference axis. Each identification cross section records the radial upper quantile envelope, radial lower quantile envelope, outer layer point ratio, normal dispersion, curvature distribution, cross section coverage, and the material category probability of conductor, insulation layer, semiconducting layer, and sheath. The system calculates the characteristic changes of adjacent cross sections and forms a boundary response sequence, excluding isolated responses with a continuous length of less than 2 mm.

[0043] The system filters boundary combinations according to the axial topological order of conductor, insulating layer, semiconductive layer, and sheath. Each processing boundary is divided into 48 circumferential units at 7.5 degrees circumferentially. The location of the maximum boundary response is recorded in each circumferential unit, thus forming a spatial boundary that varies with the circumferential angle. The system calculates the average axial position, maximum position, minimum position, circumferential range, and overall tilt angle based on the spatial boundary.

[0044] The dimensional measurement module determines the measurement center based on the back-off distance specified in the peeling process template, and selects continuous measurement sections within a 2 mm range before and after the measurement center. Within each measurement section, two sets of parallel virtual plates are established, one horizontally and one vertically. The virtual plates move from the outside of the section inwards. When the number of contact points is not less than 12, the axial span of the contact points is not less than 3 mm, the circumferential span of the contact points is not less than 15 degrees, and the average measurement confidence weight of the contact points is not less than 0.75, a stable contact zone is formed and movement stops.

[0045] The system uses the normal distance between the stopping positions of two virtual card plates as the diameter in the corresponding direction, the median of the diameters of continuously measured cross-sections as the output diameter, and the ratio of the difference between the transverse and longitudinal diameters to their average value as the ellipticity. The core length, insulation break location, semiconductive layer break location, and transition slope are all calculated along the arc length coordinate of the engineering reference axis.

[0046] During surface defect identification, the system unfolds the cable surface into a two-dimensional grid consisting of axial arc length and circumferential angle. A local reference surface is fitted with a neighborhood of 5 mm x 15 degrees, and the height residual, curvature abrupt change, and normal variation are calculated. Areas with continuously negative height residuals forming directional grooves are recorded as scratches; areas with continuously positive height residuals are recorded as protrusions; and thin-layer areas with color or reflectivity inconsistent with the insulation template and height close to the reference surface are recorded as semiconductive layer residues.

[0047] The acceptance judgment module reads the scanning calibration error, multi-frame registration error, measurement section positioning deviation, and continuous section repeatability measurement deviation, and uses the square and square root method to form the combined uncertainty of the measurement item. A measurement value and its uncertainty interval are recorded as qualified when they are completely within the acceptance limits; a measurement value's uncertainty interval intersects with the acceptance limits when it is verified; and a measurement value and its uncertainty interval are recorded as unqualified when they are completely outside the acceptance limits.

[0048] The report evidence chain module saves a summary of the original point cloud file, point cloud region index, section number, virtual card stop position, rule entry identifier, algorithm version, template version, and manual correction differences for each measurement field. When the structural boundary or measurement section is manually modified, the system retains the results before the modification and recalculates the affected dimensions, acceptance status, and report fields.

[0049] After the basic software processing is complete, an acceptance report is directly generated, including a 3D overview, structural partitions, measurement sections, dimension tables, defect tables, measurement uncertainties, rule-based judgments, and manual correction records. When the extended data interface is closed, the task does not invoke the micro-loading device, load sensor, full-field displacement acquisition device, load transfer operator, or accessory contact pressure solution.

[0050] Specific Implementation Example 2: Hardware Extension Implementation Example.

[0051] This extended embodiment uses the point cloud model, structural region, processing boundary, and acceptance results generated by the basic software embodiment as input, and forms additional verification data through independent equipment and independent algorithm services. The openable ring-shaped support frame, loading ring, micro-load execution unit, load sensor, full-field displacement acquisition device, load transfer operator, and contact pressure analysis are not essential components of the basic software embodiment.

[0052] A digital acceptance system for cable joints includes an openable ring-shaped support frame, a first loading ring, a second loading ring, a micro-load actuator, a load sensor, a flexible contact head, a full-field displacement acquisition device, an axial positioning device, a circumferential reference component, a temperature sensor, an inclination sensor, a controller, and an industrial computer. The two loading rings are spaced apart along the cable axis, and each loading ring has the same number of micro-load channels circumferentially.

[0053] The controller applies trial loading to a stable insulation section of the same cable. The formal test load is determined based on the linearity of the response when the load changes exponentially and the residual displacement after complete unloading. The formal test uses an orthogonal coding matrix to simultaneously control each micro-load actuator. Load sensors record the actual load, and a full-field displacement acquisition device synchronously records the cable surface displacement.

[0054] The industrial computer converts the channel load applied by the flexible contact head into a surface element load, forms a surface load matrix from the loads of each encoded state, forms a displacement response matrix from the radial displacements of each encoded state, identifies the cable load transfer operator through the local support space basis function, displacement reconstruction residual, operator reciprocity, spatial smoothing and semi-positive definite constraints, and subtracts the system flexibility operator calibrated by the detection equipment on a high-stiffness standard cylinder.

[0055] The system identifies the reference load transfer operator and the processing area load transfer operator in the stable insulation section and the processing area, respectively. A spatial mapping matrix is ​​established based on the axial arc length, circumferential angle and surface unit area to map the reference load transfer operator to the processing area coordinates, obtain the transfer anomaly operator, and extract local compliance, load propagation radius, load propagation direction, loading and unloading hysteresis and residual displacement from the transfer anomaly operator.

[0056] The system reads the inner surface data, functional area data, and load transfer operator of the specified accessory, calculates the initial gap between the cable processing surface and the inner surface of the accessory when the accessory is in different installation postures, and then solves the gap vector and contact pressure vector after assembly based on the measured load transfer operator of the cable processing area and the accessory load transfer operator to form an installation posture window that satisfies the pressure range of the accessory's functional area, contact coverage, and no-pressure breakage conditions.

[0057] When the installation attitude window is empty, the system distinguishes between local protrusions and local increases in flexibility based on the geometric surface and transfer anomaly characteristics. Regions with increased flexibility are designated as prohibited removal areas, and the repair depth distribution is calculated under constraints of remaining insulation thickness, maximum material removal depth, and surface slope. After repair, the system re-acquires surface data and micro-load response of the repair area, re-identifies the load transfer operator, and re-solves the accessory contact state.

[0058] The openable annular support frame consists of a first half-ring and a second half-ring. The first and second half-rings are connected on one side by a hinge and on the other side by a locking mechanism. When the locking mechanism is closed, the first and second half-rings form a circular structure. The circular structure is equipped with three unloaded positioning rollers, spaced 120 degrees apart circumferentially and in contact with the outer surface of the cable. The radial position of the positioning rollers is adjusted by a lead screw and locked with a scale, thereby adjusting the center of the support frame to be near the center of the cable.

[0059] The first and second loading rings are fixed within an openable annular support frame, with a 12 mm interval between them along the cable axis. Each loading ring has 16 micro-load actuators circumferentially arranged, resulting in a total of 32 loading channels. The circumferential angle between adjacent channels is 22.5 degrees. The corresponding channels of the first and second loading rings are aligned circumferentially to form paired axial loading positions.

[0060] Each micro-load actuator unit includes a closed-loop linear actuator, a linear guide, a tension / compression load sensor, a flexible contact head, and a mechanical travel limiter. The contact surface of the flexible contact head is an arc surface with an axial width of 6 mm and a circumferential width of 8 mm. The radius of the arc is selected according to the nominal outer diameter of the cable under test, and the contact head material is silicone rubber with a Shore A hardness of 50. A spherical self-aligning seat is set between the contact head and the load sensor, so that the contact head automatically conforms to the local surface within a range of no more than 2 degrees.

[0061] The full-field displacement acquisition device includes two 5-megapixel industrial cameras, a binocular synchronous controller, a narrowband illuminator, and a projection speckle generator. The two industrial cameras are positioned at an angle of 25 to 40 degrees, with a working distance of 350 to 500 millimeters and a lens focal length of 35 millimeters. The projection speckle generator projects random black and white textures onto the cable surface without contacting the surface, and the industrial cameras acquire the loading process at a frequency of 100 frames per second. Four laser displacement sensors are arranged at 0, 90, 180, and 270 degrees circumferentially to verify the absolute values ​​of the binocular displacement results.

[0062] The axial positioning device includes a linear guide, a slider, a handwheel screw, and an absolute position encoder. An openable annular support frame is fixed to the slider. The position encoder has a resolution of 0.05 mm. The circumferential reference component uses a 2 mm wide reflective strip, adhered to a stable surface that is not subject to loading. The centerline of the reflective strip is defined as the 0-degree circumferential direction. The temperature sensor is a surface-mount platinum resistance thermometer mounted on the cable surface, and the tilt sensor is fixed to the support frame.

[0063] The controller employs a real-time industrial controller, featuring 32 load acquisition channels, 32 actuator control channels, a camera trigger channel, a laser displacement acquisition channel, a temperature acquisition channel, a tilt angle acquisition channel, and an emergency stop input channel. An industrial computer connects to the controller via gigabit Ethernet, with a control cycle of 1 millisecond. Displacement images, actual loads, and status timestamps all utilize the same controller clock.

[0064] Each test is assigned a unique task number. The task record includes the cable manufacturer, cable model, voltage rating, conductor cross-sectional area, nominal insulation outer diameter, factory measured insulation outer diameter, accessory manufacturer, accessory model, accessory batch, construction location, operator identification, rule template version, testing equipment number, and testing time.

[0065] The geometric data file records the original point cloud coordinates, point cloud normal vector, point cloud reflection intensity, sampling time, scanning device number, coordinate units, and point cloud quality indicators; the micro-load data file records the target load, actual load, actuator displacement, load rise start time, stabilization start time, stabilization end time, and unloading end time for each loading channel; the displacement data file records the three-dimensional displacement, radial displacement, axial displacement, circumferential displacement, and measurement confidence level of each surface unit.

[0066] The accessory template consists of the three-dimensional surface of the accessory's inner cavity, functional area labels, lower and upper limits of contact pressure for each functional area, minimum contact coverage, upper limit of the no-pressure breakage angle, accessory load transfer operator, calibration temperature, and template version. The inner surface of the accessory's inner cavity is divided into a guide zone, stress control zone, insulation crimping zone, sealing zone, axial limiting zone, and tolerance zone.

[0067] The three-dimensional surface of the accessory's inner cavity is obtained by converting the accessory's design model; the accessory's load transfer operator is obtained by calibrating the partitioned force-measuring mandrel. Radial loading blocks and displacement sensors corresponding to the accessory template surface units are set on the outer periphery of the partitioned force-measuring mandrel. Loads are applied at 20 degrees Celsius, 30 degrees Celsius, and 40 degrees Celsius respectively to form a correspondence matrix between the accessory's surface load and the accessory's inner cavity radial displacement. A template index is established using the accessory model, batch, and temperature range.

[0068] Each micro-load actuator performs six levels of loading and unloading (0.5 N, 1 N, 2 N, 4 N, 6 N, and 8 N) on a standard force gauge, recording the actuator control input and load sensor readings. Piecewise linear interpolation is used to establish a mapping between the control input and the actual load. The relative load error of the i-th loading channel is calculated using the following formula:

[0069]

[0070] In the formula, Indicates the first The relative load error of each loading channel; Indicates the first The target load for each loading channel is expressed in Newtons. Indicates the first The actual load of each loading channel is expressed in Newtons. Indicates the channel number to be loaded; This indicates absolute value operations. Loading states with a relative load error greater than 0.01 are not included in the operator identification.

[0071] The system flexibility is calibrated using a high-rigidity steel standard cylinder with the same diameter as the cable under test and a radial compressive deformation of less than 1 micrometer. After the load-bearing frame is installed on the standard cylinder, a complete orthogonal coded loading process is performed to obtain the system flexibility operator formed by the testing equipment, flexible contact head, and connection structure. The cable net load transfer operator is calculated according to the following formula:

[0072]

[0073] In the formula, This represents the net load transfer operator of the cable after deducting the deformation of the testing equipment itself. This represents the total load transfer operator formed by the testing equipment and the cable under test. This represents the compliance operator of the detection system calibrated on a high-stiffness standard cylinder. The three matrices have the same number of rows and columns; the matrix rows correspond to displacement response surface elements, and the matrix columns correspond to surface load elements.

[0074] The binocular industrial camera undergoes intrinsic and extrinsic parameter calibration using a 200mm x 150mm planar calibration plate, followed by displacement calibration using a three-axis micro-displacement platform with a resolution of 1 micrometer. Four laser displacement sensors are calibrated for zero point and scale factor using gauge blocks and the micro-displacement platform. If the difference between the binocular displacement and the laser displacement in the four reference directions exceeds 8 micrometers, the current acquisition cycle is discarded and recalibrated.

[0075] The system determines the formal test load on a stable insulation section of the same cable. A stable insulation section is an insulation section located outside the joint machining area, with no cutting marks on the surface and an axial length of not less than 60 mm. The controller first applies a trial load of 0.5 Newtons, holds it for 0.5 seconds, and records the displacement vector. Then, a second trial load of 1 Newton is applied, and the displacement vector is recorded. The load is then unloaded, and after waiting 3 seconds, the residual displacement is recorded.

[0076] The response nonlinearity index corresponding to the test load F is calculated according to the following formula:

[0077]

[0078] In the formula, Indicates test load The corresponding nonlinear response index; This indicates the first test load, in Newtons. This indicates the second test load, in Newtons; Indicates load The radial displacement vector of all measurement units; Indicates load The radial displacement vector of all measurement units; Represents the Euclidean norm; This indicates a positive number to prevent the denominator from being zero; in this embodiment, it is taken as 0.5 micrometers.

[0079] The residual displacement ratio after complete unloading is calculated according to the following formula:

[0080]

[0081] In the formula, Indicates the residual displacement rate; This represents the residual displacement vector after complete unloading and waiting for 3 seconds. This represents the peak displacement vector in the current test load level; The meaning and value of are consistent with the aforementioned formula.

[0082] When the response nonlinearity index is less than 0.03 and the residual displacement rate is less than 0.01, the test load is increased by 0.5 Newtons and the test is repeated; when either index reaches the corresponding threshold for the first time, the current load is recorded as the nonlinear boundary load. The formal test load is determined according to the following formula:

[0083]

[0084] In the formula, This indicates the formal test load, expressed in Newtons. This represents the safety factor, which is 0.40 in this embodiment; The nonlinear boundary load, expressed in Newtons, represents the nonlinear response index or residual displacement rate that first reaches the threshold.

[0085] The basic contact load is set to 0.30 times the formal test load, and the coded load variation amplitude is set to 0.20 times the formal test load. Therefore, the basic contact load is greater than the coded load variation amplitude, and the flexible contact head still maintains compressive contact under negative coding conditions. When the actual load exceeds 1.05 times the formal test load, the actuator reaches the mechanical stroke limit, or the residual displacement rate after unloading exceeds 0.01, the controller releases the load from all loading channels within 50 milliseconds.

[0086] The system performs a no-load 3D scan of the stable insulation section and the processing area, with a point cloud spacing of no more than 0.15 mm. An industrial computer extracts a point cloud cross-section every 2 mm along the cable axis, obtains the cross-section center using robust ellipse fitting, and then uses cubic B-spline fitting to fit the center trajectory.

[0087] At each arc length position of the center trajectory, a tangential unit vector is established, and two mutually orthogonal radial unit vectors perpendicular to the tangential unit vector are established using a parallel moving frame. Cable surface points are represented by the following formula:

[0088]

[0089] In the formula, Indicates the position of axial arc length and circumferential angle The corresponding three-dimensional coordinates of the cable surface points; Indicates the position of arc length The corresponding center trajectory coordinates; This represents the arc length along the center trajectory, in millimeters. This represents the radial distance from a point on the cable surface to the center trajectory, in millimeters. and Let these represent the first and second radial unit vectors of the local cross-section, respectively. It represents the circumferential angle with the centerline of the circumferential reference component as the zero-degree direction; and They represent angles respectively. The cosine and sine values.

[0090] The processing area is divided into surface units of 3 mm axially and 7.5 degrees circumferentially. The stable insulation section uses the same circumferential division and 3 mm axial division. Each surface unit records the center coordinates, actual surface area, radial unit normal vector, average radius, point density, effective image pixel count, displacement measurement confidence level, and loading response sequence. Within 15 mm before and after the semiconductive layer fracture, the axial unit length is subdivided into 1.5 mm units.

[0091] The system comprises 32 micro-payload execution units, employing a 32-order Hadamard matrix as the orthogonal encoding matrix. The encoding matrix is ​​represented as follows:

[0092]

[0093] In the formula, Represents an orthogonal coding matrix; Indicates the first The microload execution unit in the first The encoded value in each encoded state, the encoded value is or ; This indicates the total number of micro-load execution units, in this embodiment... ; In this embodiment, the total number of encoded states is represented. ; Subscript The matrix represents having lines and List.

[0094] The target load of the i-th micro-load execution unit in the k-th encoded state is determined by the following formula:

[0095]

[0096] In the formula, Indicates the first The microload execution unit in the first The target load in each encoded state is expressed in Newtons. This represents the basic contact load, expressed in Newtons. This represents the magnitude of the coded load variation, in Newtons. This represents the corresponding encoded value; Indicates the channel number to be loaded; Indicates the encoding status number.

[0097] Orthogonal coding matrices satisfy the following equation:

[0098]

[0099] In the formula, This represents the summation operation; Indicates the summation index; Indicates the total number of encoded states; Indicates the first The first channel in the The encoded value in each encoded state; Indicates the first The first channel in the The encoded value in each encoded state; Indicates another loading channel number; This indicates that the two numbers are the same; This indicates that the two numbers are different. This relationship ensures that the inner product of the encoded sequences of different loaded channels is zero.

[0100] The load rise time for each coded state is 0.20 seconds, the stabilization time is 0.40 seconds, and the state transition time is 0.10 seconds. The controller averages the actual load and surface displacement during the last 0.20 seconds of the stabilization interval. After 32 formal coded states, 6 repeating states are added, with the repeating states being the 1st, 6th, 11th, 16th, 21st, and 26th coded columns, respectively. If the root mean square of the displacement difference between repeating states is greater than 0.05 times the peak displacement of the formal state, the current position data is discarded.

[0101] The overall translation and rotation of the cable and support frame during loading are compensated for by rigid reference markers. Four reference markers are set on the support frame, and four reference markers are set in the unloaded stable region of the cable. The industrial computer calculates the rigid body transformation based on the three-dimensional coordinates of the reference markers in the initial state and the k-th coded state.

[0102]

[0103] In the formula, Indicates the first The optimal rotation matrix for each encoded state; Indicates the first The optimal translation vector for each encoded state; and Let them represent the rotation matrix and translation vector to be determined, respectively. This represents the parameter that minimizes the objective function; Indicates the reference marker number; In this embodiment, the total number of reference markers is indicated. ; Indicates the first The coordinates of each reference marker in the initial state; Indicates the first The reference marker is at the _ The coordinates of each encoded state; This represents the square of the Euclidean norm.

[0104] The radial net displacement of the j-th surface element in the k-th encoded state is calculated according to the following formula:

[0105]

[0106] In the formula, Indicates the first The surface unit at the first The radial net displacement under each coded state, in millimeters; Indicates the first The radial unit normal vector of each surface element; This represents the transpose of the radial unit normal vector; Indicates the first The coordinates of each surface unit in the initial state; Indicates the first The surface unit at the first The coordinates of each encoded state; and The meaning is consistent with the aforementioned formula.

[0107] If the root mean square of the three-dimensional residuals of the eight reference markers after rigid body compensation is greater than 3 micrometers, the current encoded state will not be included in the displacement response matrix. If there are fewer than 32 valid encoded states, the system will re-execute the entire encoding cycle.

[0108] The flexible contact head forms a finite area load on the cable surface. The system lays a pressure-sensitive membrane on the surface of an elastic standard cylinder, applying 1 Newton, 2 Newtons, 4 Newtons, and 6 Newtons sequentially to each channel to obtain the pressure distribution in the contact area, and converts the actual load of the actuator unit into an equivalent load for the surface unit. The surface load vector in the k-th coded state is calculated according to the following formula:

[0109]

[0110] In the formula, Indicates the first The equivalent load vector borne by all surface units under each encoded state; The load distribution matrix represents the transformation of the micro-load actuator load into the surface element load; Indicates the first The actual load vector measured by all load sensors under each coded state; This represents the encoded state number. The sum of the elements in each column of the load distribution matrix is ​​1.

[0111] The surface load matrix and displacement response matrix for all encoded states are respectively composed according to the following formulas:

[0112]

[0113] In the formula, Represents the surface load matrix; to They represent the first to the second. A surface load vector of coded states; square brackets indicate column concatenation.

[0114]

[0115] In the formula, Represents the displacement response matrix; to They represent the first to the second. The radial net displacement vector of all surface units in each encoded state; This indicates the total number of valid encoded states.

[0116] Since the number of surface elements is greater than the number of loading channels, the system uses cubic B-spline basis functions with local support along the axial and circumferential directions to establish the basis matrix, and the load transfer operator is expressed as:

[0117]

[0118] In the formula, The load transfer operator represents the relationship between surface element load and surface element displacement; This represents the basis matrix formed by the basis functions of the local support space; Represents the coefficient matrix of a low-dimensional operator; This represents the transpose of the basis matrix. The number of basis functions is 64 in this embodiment.

[0119] The coefficient matrix of the low-dimensional operator is solved by the following formula:

[0120]

[0121] In the formula, This represents the obtained low-dimensional operator coefficient matrix; Representation matrix It is a positive semi-definite matrix; This represents the confidence matrix for displacement measurements. This indicates element-wise multiplication; Denotes the Frobenius norm; Represents the reciprocity constraint coefficient; Indicates the spatial smoothing constraint coefficient; This represents the discrete difference matrix between adjacent spatial basis functions; This represents the transpose of a discrete difference matrix. and The method determined by five-fold cross-validation of the repeated coding data of the stable insulation segment minimizes the reconstruction error of the repeated state displacement that is not involved in the solution.

[0122] The measured load transfer operator, after deducting the compliance of the detection system, is obtained according to the following formula:

[0123]

[0124] In the formula, This represents the measured load transfer operator after deducting the compliance of the detection system. Represents the basis matrix; This represents the obtained low-dimensional operator coefficient matrix; This represents the compliance operator of the detection system. The hat symbol indicates that the matrix was obtained from measured data.

[0125] The low-dimensional operator solution employs an interior-point semi-positive definite optimization algorithm with a maximum of 300 iterations, and the relative change in the objective function is less than [value missing]. The calculation stops when the displacement reconstruction residuals no longer decrease. When negative eigenvalues ​​appear in the results, those with absolute values ​​less than the largest eigenvalue are selected. Negative eigenvalues ​​that are times the threshold are truncated to zero; if there are negative eigenvalues ​​whose absolute value is not less than the threshold, the current operator determines that the physical constraints are not met and re-acquires data.

[0126] The measured load transfer operator for the stable insulation section is denoted as the reference load transfer operator, and the measured load transfer operator for the processed region is denoted as the load transfer operator to be measured. A bilinear interpolation mapping matrix is ​​constructed based on the axial arc length, circumferential angle, and surface unit area of ​​the two regions. The mapped reference load transfer operator is calculated according to the following formula:

[0127]

[0128] In the formula, This represents the reference load transfer operator for transformation to the coordinate domain of the machining area; The reference load transfer operator represents the stable insulation section; This represents the spatial mapping matrix from the surface units of the stable insulation section to the surface units of the processed area; This represents the transpose of the space mapping matrix.

[0129] The transitive exception operator is calculated according to the following formula:

[0130]

[0131] In the formula, This indicates the change in load transfer in the processing area relative to the stable insulation section of the same cable; The measured load transfer operator represents the processing area; This represents the reference load transfer operator after mapping; Indicates the difference.

[0132] The local compliance of the j-th surface element is obtained according to the following formula:

[0133]

[0134] In the formula, Indicates the first The local compliance of a surface unit, expressed in millimeters per Newton; The first term of the measured load transfer operator Line 1 Column elements; Indicates the surface unit number.

[0135] The load propagation radius of the j-th surface element is calculated according to the following formula:

[0136]

[0137] In the formula, Indicates the first The average load propagation distance of a surface unit after being loaded, in millimeters; Indicates the total number of surface units in the detection area; Indicates the response surface element number; Indicates the first The surface unit and the first The surface geodesic distance between surface units, in millimeters; Indicates the first When the surface unit bears a unit load, the first Radial displacement of each surface unit; This indicates a positive number that prevents the denominator from being zero.

[0138] The load / unload hysteresis index is calculated using the following formula:

[0139]

[0140] In the formula, Indicates the first Load and unload hysteresis index of each surface unit; Indicates the loading phase at the 1st stage. Load levels Radial displacement below; This indicates the radial displacement during the unloading phase under the same load level. In this embodiment, the total number of load classes is indicated. ; Indicates the load class number; This indicates a positive number that prevents the denominator from being zero.

[0141] The system constructs a feature vector from local compliance, load propagation radius, axial propagation second moment, circumferential propagation second moment, loading / unloading hysteresis index, and residual displacement rate, and calculates the anomaly degree based on the characteristic mean and covariance of the stable insulation section.

[0142]

[0143] In the formula, Indicates the first The degree of anomaly of the surface unit of each processing area relative to the stable insulation section of the same cable; Indicates the first Feature vectors of surface units; This represents the characteristic mean vector of a stable insulation segment; Represents the characteristic covariance matrix of the stable insulation section; The inverse matrix of the covariance matrix; superscript The square root represents the vector transpose; it makes the outlier scale consistent with the standard deviation.

[0144] The anomalous region consists of circumferentially and axially adjacent high-anomaly surface units, with the high-anomaly threshold set at the 95th percentile of the anomalous degree of the stable insulation section. When the axial length of the anomalous region is less than 3 mm and the circumferential width is less than 7.5 degrees, the system requires repeated testing at the same location; if the two results are inconsistent, no structural anomaly record is generated.

[0145] When the local surface radius is higher than that of the adjacent stable surface and the local flexibility does not increase, the abnormal area is recorded as a geometric protrusion; when the surface radius is within the range of the adjacent stable surface and the local flexibility increases but the hysteresis index does not increase, the abnormal area is recorded as a decrease in remaining thickness; when the local flexibility increases, the propagation radius decreases and the hysteresis index increases, the abnormal area is recorded as a structural continuity anomaly.

[0146]

[0147] In the formula, The displacement response matrix represents each encoded state, in millimeters; The surface load matrix represents each encoded state, in Newtons; This represents the load transfer operator, with units of millimeters per Newton; load transfer operator The Line 1 Column elements represent the first When the surface unit bears a unit load, the first The radial displacement generated by each surface element. The load transfer operator is obtained from the surface load matrix and displacement response matrix under displacement reconstruction residual, reciprocity, spatial smoothing and positive semidefinite constraints.

[0148] The inner surface of the accessory cavity and the processed surface of the cable are resampled under the same surface element topology. The mounting attitude vector of the accessory relative to the cable is defined by the following formula:

[0149]

[0150] In the formula, This represents the mounting orientation vector of the accessory relative to the cable. Indicates the position of the accessory along the cable axis, in millimeters; This indicates the angle of rotation of the attachment relative to the circumferential reference direction, in degrees; and These represent the offsets of the accessory axis along the first and second radial directions, respectively, in millimeters; and These represent the inclination angles of the accessory axis about the first and second radial directions, respectively, in degrees; superscript This represents the transpose of a vector.

[0151] The system searches for the installation posture within the range specified in the attached template. In this embodiment, the axial position search range is 5 mm before and after the nominal position, the circumferential angle search range is 0 degrees to 357.5 degrees, and the coarse search step size is 2.5 degrees. The search ranges for the two radial offsets are both -1.5 mm to +1.5 mm, with a step size of 0.25 mm, and the search ranges for the two tilt angles are both -1 degree to +1 degree, with a step size of 0.25 degrees. After obtaining the posture that meets the basic contact conditions through the coarse search, a bisection refinement is performed centered on the coarse search posture until the circumferential angle resolution reaches 0.25 degrees and the axial position resolution reaches 0.1 mm.

[0152] Under the coordinate transformation corresponding to the installation attitude vector, the system calculates the signed initial gap vector between the inner surface of the accessory cavity and the processed surface of the cable. A positive initial gap indicates that the two surfaces are separated, while a negative initial gap indicates that there is a geometric interference between the two surfaces. The contact state after assembly satisfies:

[0153]

[0154] In the formula, This represents the gap vector after assembly deformation, in millimeters. Indicates installation posture The initial gap vector is given below, in millimeters; The measured load transfer operator represents the cable processing area; Indicates the load transfer operator for attachments; This represents the contact load vector, measured in Newtons.

[0155] The gap vector and the contact load vector simultaneously satisfy:

[0156]

[0157] In the formula, Represents a vector whose all elements are zero; This indicates that the assembly gap of each surface unit is not less than zero; This indicates that the contact load of each surface element is not less than zero; Indicates the first The assembly clearance of each surface unit; Indicates the first Contact load of each surface unit; This indicates that the contact load is zero when a positive gap exists in the same unit, and the gap is zero after assembly when a positive contact load exists.

[0158]

[0159] In the formula, This represents the contact pressure vector, in megapascals (MPA). This represents an area diagonal matrix with the area of ​​each surface unit as its diagonal element, in square millimeters. The inverse matrix representing the area diagonal matrix; This represents the contact load vector, in Newtons. The first element of the contact pressure vector... The element equals the first element. The contact load of a surface unit divided by the area of ​​that surface unit.

[0160] The system uses projective gradient iteration to solve for the contact state:

[0161]

[0162] In the formula, represents the contact load vector of the i-th iteration; represents the contact load vector of the i-th iteration; represents the gap vector corresponding to the i-th iteration; represents the iteration step size, which is 0.8 times the reciprocal of the largest eigenvalue of the joint load transfer operator; represents the larger value between zero and the value in parentheses for each element; represents the iteration number. The maximum number of iterations is 2000, and the complementary residual is less than 1. Stop when the time comes.

[0163] The effective contact coverage of the Z-zone of the accessory function area is calculated according to the following formula:

[0164]

[0165] In the formula, Indicates the attachment function area Effective contact coverage; This represents the set of surface units corresponding to the functional area; Indicates the first The area of ​​each surface unit, in square millimeters; This represents an indicator function, which takes the value 1 if the condition inside the parentheses is true and 0 if it is false. Indicates the first Contact pressure of each surface unit, in megapascals; Indicates functional area The lower limit of contact pressure; Indicates functional area The upper limit of contact pressure.

[0166] The system constructs a circumferential connectivity graph for surface units with pressure below the lower limit of the functional area pressure. When the angular span of the circumferential connectivity region is greater than the upper limit of the angle of the pressureless break zone in the accessory template, it is recorded as a pressureless break zone. The installation posture window that meets the conditions is represented by the following formula:

[0167]

[0168] In the formula, Indicates the installation attitude window; Represents an installation attitude vector; Indicates the functional area under this attitude. Effective contact coverage; Indicates functional area Minimum contact coverage; This represents the maximum contact pressure among all surface units in this orientation; This indicates the upper limit of the contact pressure recorded in the attached template; This indicates the area of ​​the pressure-free gap within the critical functional area; The vertical line represents the upper limit of the pressureless gap area recorded in the attached template; the vertical line represents the set of postures that satisfy all the subsequent conditions.

[0169] When the installation attitude window includes the nominal axial position and 0-degree circumferential position of the accessory, the acceptance record is direct assembly; when the installation attitude window does not include the 0-degree circumferential position but includes a continuous circumferential angle range, the system projects alignment marks on the surface of the cable and accessory, and writes the circumferential angle range and axial position range into the acceptance record; when the installation attitude window is empty, the system enters rework calculation.

[0170] Before rework calculations, the system determines permissible rework areas and prohibited removal areas based on geometric surface and load transfer anomaly results. Surface elements with local compliance greater than 1.15 times the mapped reference compliance, and where further increasing the initial gap would expand the low-pressure area, are designated as prohibited removal areas; surface elements with local surface radii higher than the neighboring fitted surface, local compliance within the reference range, and contact pressure higher than the upper limit of the attachment pressure are designated as permissible rework areas.

[0171] The material removal depth of the allowable rework area constitutes the rework depth vector:

[0172]

[0173] In the formula, Represents the rework depth vector; Indicates the first The material removal depth of each repair surface unit, in millimeters; Indicates the total number of surface elements participating in the rework solution; superscript This represents the transpose of a vector.

[0174] After the material is removed, the cable surface moves towards the center. The initial gap after repair is calculated using the following formula:

[0175]

[0176] In the formula, This represents the initial gap vector after repair; This represents the initial gap vector before rework; This represents the rework depth vector; the apostrophe indicates the amount after rework. The rework depth corresponding to the area that cannot be removed is fixed at zero.

[0177] The optimal repair depth is determined by the following formula:

[0178]

[0179] In the formula, Represents the optimal rework depth vector; Indicates the material's volume weight after removal; Indicates the first The area of ​​each repaired surface unit; Indicates the first Material removal depth of each repair surface unit; Indicates the continuity weight of the reworked surface; This represents the discrete gradient matrix between adjacent repair surface units; This represents the sum of squares of the changes in adjacent repair depths; Indicates the weight of the contact state deviation; This represents the normalized weighted sum of the low-pressure area, overpressure area, and unpressured fault length obtained by solving using the initial gap after repair. The three weights are taken as the reciprocals of the normalized values ​​that make the three terms have the same initial magnitude.

[0180] The depth of repair must meet the following constraints:

[0181]

[0182] In the formula, Indicates the first Material removal depth of each repair surface unit; This indicates the maximum allowable removal depth determined jointly by the attached construction rules and the remaining insulation thickness; the left side of the inequality restricts the material removal depth to be non-negative, while the right side restricts the material removal depth to not exceed the maximum allowable value.

[0183]

[0184] In the formula, Indicates the first day before repair The remaining insulation thickness corresponding to each surface unit; This indicates the material removal depth of the surface unit; This indicates the minimum remaining insulation thickness recorded in the attached construction rules; this inequality limits the remaining insulation thickness after repair to be no less than the minimum value.

[0185] The rework optimization uses a sequential quadratic programming algorithm with a maximum of 200 iterations, and the relative change in the objective function is less than... The process stops when all constraints are met. The system divides the rework depth distribution into areas that must be removed, gradual transition areas, and areas that cannot be removed, and projects the boundary and depth classification marks onto the actual cable surface using a projector.

[0186] After rework, a new no-load point cloud was acquired to confirm that the absolute difference between the actual material removal depth and the rework depth distribution was no greater than 0.10 mm. Subsequently, orthogonal coded micro-load detection was re-executed at the same location to re-identify the load transfer operator and the contact status of the attachments. When the installation attitude window is not empty after rework, the coverage of each functional area reaches the attachment template threshold, and there are no pressure-free broken strips exceeding the threshold, the system generates a re-inspection pass record.

[0187] The industrial computer's software includes a task and template management module, a point cloud data access module, a point cloud quality review module, an engineering reference coordinate construction module, a structure identification module, a dimension measurement module, an acceptance judgment module, a report evidence chain module, and an extended data interface module. Point cloud quality review, structure identification, dimension measurement, acceptance judgment, and report output constitute the basic software processing chain, while the extended data interface module separately receives micro-load, displacement response, and accessory contact analysis data.

[0188] The data processing sequence of the basic software is fixed as follows: point cloud file integrity check, multi-frame coordinate unification, measurement confidence weight calculation, engineering reference axis establishment, section sequence generation, structural area identification, processing boundary determination, virtual caliper measurement, defect quantification, measurement uncertainty calculation, acceptance rule judgment, and report evidence chain generation. When the extended data interface is enabled, load transfer operators and attachment contact status verification are added after the basic software forms the acceptance results.

[0189] Point clouds, original images, load data, and displacement data are stored as read-only objects. Operator matrices, attachment templates, attitude search parameters, and rework parameters all record version numbers. When surface partitions, circumferential reference directions, or attachment attitude ranges are manually modified, the system saves the data before modification, the data after modification, the operator's identification, the modification time, and the reason for modification; the original results are not overwritten.

[0190] The evidence chain module calculates SHA-256 digest values ​​for the original point cloud, original image, load file, displacement file, reference load transfer operator, processing area load transfer operator, attachment template, installation attitude window, rework depth vector, and re-verification results, and writes the digest value of the previous data object into the metadata of the next data object according to the task processing order, thereby establishing a chain verification relationship within the task.

[0191] On-site operation procedures:

[0192] The first step is for the operator to scan the identification codes of the cable and accessories. The system then reads the cable structure template, stripping process template, and acceptance rule template, establishes a task number, and checks the template version.

[0193] The second step involves importing multi-frame point clouds of the stable segment and processing area, as well as the calibration parameters of the scanning equipment. Then, it performs checks on file integrity, coordinate units, point density, scan coverage, and multi-frame overlap, and converts each scan frame to a unified engineering coordinate system.

[0194] The third step is to calculate the measurement confidence weights of each point, establish an engineering reference axis in the form of a space curve and a non-twisted curve coordinate system, generate a cross-section sequence along the engineering reference axis and form a cable surface unfolded mesh.

[0195] The fourth step involves identifying the structural regions and processing boundaries based on the radial envelope, outer layer point ratio, normal variation, curvature distribution, cross-sectional coverage, and material type probability. Horizontal and longitudinal virtual caliper measurements are then performed at the specified retraction positions to quantify surface defects.

[0196] The fifth step is to calculate the uncertainties introduced by scanning calibration, multi-frame registration, cross-section positioning, and repeated measurements, and map the measurement items and defect items to the acceptance rule template to form a qualified, verified, or unqualified status.

[0197] The sixth step involves writing the original point cloud, point cloud version, algorithm version, template version, measurement section, virtual card position, manual calibration record, 3D screenshot, and acceptance field into the evidence chain and generating an acceptance report.

[0198] Step 7: When hardware extension is enabled in the project configuration, the system receives micro-load detection data through the extended data interface, calculates the load transfer operator, contact load and contact pressure, and appends the extended verification results to the acceptance report; when hardware extension is not enabled, the task ends directly after the basic software acceptance is completed.

[0199] Specific Implementation Example 3: Hardware Extension Application Example.

[0200] This embodiment uses a cross-linked polyethylene cable with a rated voltage of 110 kV and an insulation outer diameter of 86 mm, along with a corresponding prefabricated intermediate joint. The length of the stable insulation section is 60 mm, and the length of the processed area is 120 mm. The system is equipped with two loading rings, each with 16 micro-load actuators, and the axial distance between the two loading rings is 12 mm.

[0201] The test load on the stable insulation section started at 0.5 Newtons and increased by 0.5 Newtons each time. When the test load reached 6 Newtons, the nonlinear response index reached 0.031 and the residual displacement rate was 0.008. Therefore, 6 Newtons was recorded as the nonlinear boundary load. The formal test load was taken as 2.4 Newtons, the foundation contact load was taken as 0.72 Newtons, and the coded load change amplitude was taken as 0.48 Newtons.

[0202] The system performed orthogonal encoding detection in both the stable insulation section and the processing area. In the processing area, the local compliance of surface units in the circumferential direction (37.5° to 67.5°) and axial direction (42 mm to 57 mm) was higher than the mapped reference compliance, the load propagation radius decreased, and the loading / unloading hysteresis index did not exceed the 95th percentile of the stable section. In the unloaded point cloud, the radius of this region was not higher than the neighboring fitted surface; therefore, the system recorded this region as a region with remaining thickness reduction and marked it as a region to be prohibited from removal.

[0203] Contact calculations for the accessories showed that when installed at 0 degrees circumferentially, the compressive displacement in this area increased, resulting in a 28-degree low-pressure break in the stress control zone on the opposite side. After performing a circumferential attitude search, the system obtained a continuous installation angle window from 146.25 degrees to 158.50 degrees. Within this window, the effective contact coverage of the stress control zone reached the accessory template threshold, and the maximum contact pressure did not exceed the upper limit of the accessory template. The system projected alignment marks on the cable and accessory surfaces and recorded the installation angle window and axial position window in the acceptance record.

[0204] Specific Implementation Example 4: Hardware Expansion Repair Case

[0205] This embodiment uses a cross-linked polyethylene cable with a rated voltage of 35 kV and an insulation outer diameter of 58 mm, along with corresponding cold-shrink termination accessories. The no-load point cloud analysis shows a localized protrusion with a maximum height of 0.32 mm existing in the circumferential direction (210° to 235°) and the axial direction (31 mm to 46 mm). The difference between the local flexibility in this area and the reference flexibility of the stable section of the same cable is less than 0.05 times the reference flexibility, and the load / unload hysteresis index is within the stable section range.

[0206] Contact calculations showed that localized protrusions formed contact areas exceeding the pressure limit, creating low-pressure break zones in adjacent circumferential regions. The system classified these localized protrusions into permissible rework areas, optimizing the rework process to achieve a maximum material removal depth of 0.24 mm. Gradual transition zones of 8 mm in length were set at both axial ends, while areas where removal was prohibited maintained at zero rework depth.

[0207] After on-site repair, a rescan was performed. The actual maximum material removal depth was 0.22 mm, with a maximum difference of 0.06 mm from the repair depth distribution. Micro-load testing and accessory contact calculations were then re-executed after repair. No increase in local flexibility was observed, the effective contact coverage of the stress control zone and sealing zone reached the accessory template threshold, the low-pressure breakage disappeared, and the system generated a re-inspection pass record.

[0208] If the actual load is not within ±1% of the target load, the displacement measurement confidence level is less than 0.90, the rigid body compensation residual is greater than 3 micrometers, or the number of effective coding states is less than 32, the current detection position data will not be included in the operator identification.

[0209] The difference between the two repeated identifications of the reference operator of the stable insulation section is calculated by matrix Frobenius norm normalization. When the difference is greater than 0.08, the system checks the bearing frame locking, the cleaning of the flexible contact head, environmental vibration and temperature changes. After the check is completed, the stable insulation section is re-inspected.

[0210] When the surface temperature difference between the stable insulation section and the processing area exceeds 2 degrees Celsius, the system suspends in-cable reference differential; it resumes data acquisition once the temperature difference between the two areas reaches a state where it is no greater than 2 degrees Celsius. If the calibration temperature range of the accessory template does not include the field surface temperature, the system will not output accessory contact acceptance results.

[0211] If the actual load on any loading channel exceeds 1.05 times the formal test load, any actuator triggers the travel limit, the camera synchronization signal is lost, or the residual displacement rate of unloading exceeds 0.01, the controller will perform full channel unloading and save the data from 500 milliseconds before the fault to 500 milliseconds after the fault.

[0212] When the processing area is recorded as having structural continuity anomalies, the system does not generate a material removal plan, and the acceptance record is marked as remaking the joint; when the local flexibility increases but there is an installation posture window that meets the requirements of the attachment template, the system only outputs the installation posture window and does not generate a grinding command for the area with increased flexibility.

[0213] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0214] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A digital acceptance system for cable joints, characterized in that, The system includes a task and template management module, a point cloud data access module, a point cloud quality review module, an engineering reference coordinate construction module, a structure identification module, a dimension measurement module, an acceptance judgment module, and a report evidence chain module. The task and template management module stores scanning equipment calibration parameters, cable structure templates, stripping process templates, and acceptance rule templates. The point cloud data access module receives multiple frames of raw point clouds and acquisition parameters from cable joints. The point cloud quality review module converts the multiple frames of raw point clouds to a unified engineering coordinate system and determines the measurement confidence weight of each point based on point density, normal consistency, scanning incident direction, multi-frame registration residuals, and equipment acquisition confidence. The engineering reference coordinate construction module establishes an engineering reference axis based on the weighted point cloud and establishes a torsion-free curve coordinate system along the engineering reference axis, consisting of axial arc length, cross-sectional direction angle, and radial position. The structure identification module generates a cross-sectional sequence along the engineering reference axis, and determines the structural regions and processing boundaries of the conductor, insulation layer, semiconducting layer, and sheath based on the radial envelope of the cross-section, the proportion of outer points, normal variation, cross-sectional coverage, material category probability, and axial continuity. The dimension measurement module establishes a measurement cross-section at the specified retraction position corresponding to the processing boundary, constructs mutually parallel virtual cards with stable contact band constraints, and calculates the axial length, transverse diameter, longitudinal diameter, ellipticity, slope, and defect size based on the normal distance between the virtual cards. The acceptance judgment module maps the measurement results, measurement uncertainty, and acceptance rule template to form a qualified, reviewed, or unqualified status. The report evidence chain module establishes an associated record of the original point cloud, point cloud version, algorithm version, template version, measurement cross-section, manual correction record, 3D screenshot, and acceptance field.

2. The digital acceptance system for cable joints according to claim 1, characterized in that, The point cloud quality review module converts the points in each scan frame to the engineering coordinate system according to the calibration parameters of the scanning equipment. It performs rigid registration and local non-rigid correction through the overlapping area of ​​adjacent scan frames, normalizes the point density, normal consistency, scanning incident direction, registration residual, and equipment acquisition confidence into quality components, and calculates the measurement confidence weight according to the five weights in the acceptance rule template. Points with measurement confidence weights lower than the effective threshold retain their original index and are marked as non-metric points. Non-metric points do not participate in the processing boundary fitting and virtual caliper contact calculation.

3. The digital acceptance system for cable joints according to claim 1, characterized in that, The engineering reference coordinate construction module uses weighted principal component analysis to obtain the axial direction, extracts the cross-section center along the axial direction, fits the center trajectory of a cubic B-spline with the cross-section center as the control point, and establishes a parallel moving frame along the center trajectory of the cubic B-spline. The parallel moving frame includes a tangential unit vector of the center trajectory and a first radial unit vector and a second radial unit vector located within the local cross-section. The first radial unit vector and the second radial unit vector are transferred between adjacent cross-sections according to the minimum rotation angle.

4. The digital acceptance system for cable joints according to claim 1, characterized in that, The structure recognition module forms a cross-section sequence according to the axial step size. For each cross-section, it calculates the radial upper quantile envelope, radial lower quantile envelope, outer layer point ratio, normal dispersion, curvature distribution, cross-section coverage, and material category probability, and forms the processing boundary response based on the characteristic difference between adjacent cross-sections. The structure recognition module performs continuous interval filtering on the processing boundary response according to the structural order of conductor, insulating layer, semiconducting layer, and sheath arranged along the axial direction, and outputs the axial range of each structural region and the axial position of the processing boundary at different circumferential angles.

5. The digital acceptance system for cable joints according to claim 1, characterized in that, The dimension measurement module establishes a transverse virtual caliper and a longitudinal virtual caliper within a series of measurement sections at a specified retraction position. Each virtual caliper includes a first virtual caliper plate and a second virtual caliper plate that move toward each other along the same normal direction. The first virtual caliper plate and the second virtual caliper plate stop at a position where the number of contact points, axial span, circumferential span, and measurement confidence weight satisfy the stable contact zone threshold. The normal distance between the two stop positions is determined as the measurement diameter.

6. The digital acceptance system for cable joints according to claim 1, characterized in that, The structure recognition module unfolds the cable surface into a two-dimensional surface grid according to the axial arc length and circumferential angle. It calculates the height residual, curvature abrupt change and normal change based on the local fitted surface, and determines the three-dimensional regions of semiconductive layer residue, scratches, protrusions and depressions by combining color features or reflection intensity features. The size measurement module records the axial length, circumferential angle range, maximum height difference, maximum depth and surface area of ​​each region.

7. The digital acceptance system for cable joints according to claim 1, characterized in that, The acceptance judgment module generates uncertainty intervals for each measurement item based on scanning calibration error, multi-frame registration error, measurement section positioning deviation, and continuous section repeated measurement deviation. It then compares the measurement item and its uncertainty interval with the lower limit, upper limit, and verification zone in the acceptance rule template. The report evidence chain module saves the point cloud region index, section number, virtual card position, rule entry identifier, algorithm parameter summary, and differences before and after manual correction for each measurement item.

8. A digital acceptance method for cable joints, characterized in that, The process includes the following steps: acquiring multi-frame original point clouds of the cable joint, scanning equipment calibration parameters, cable structure template, stripping process template, and acceptance rule template; converting the multi-frame original point clouds to a unified engineering coordinate system; determining measurement confidence weights based on point density, normal consistency, scanning incident direction, multi-frame registration residuals, and equipment acquisition confidence; establishing an engineering reference axis based on the weighted point cloud; establishing a torsion-free curve coordinate system along the engineering reference axis and generating a cross-section sequence; determining the structural region and processing boundary based on the cross-section radial envelope, outer layer point ratio, normal variation, cross-section coverage, material type probability, and axial continuity; establishing a measurement cross-section at the specified retraction position corresponding to the processing boundary; obtaining dimensional measurement results using transverse and longitudinal virtual calipers with stable contact band constraints; obtaining defect measurement results based on the height residual, curvature abrupt change, and normal variation of the unfolded surface; forming an acceptance status based on the dimensional measurement results, defect measurement results, measurement uncertainty, and acceptance rule template; and writing the original point cloud, algorithm version, template version, measurement cross-section, manual correction record, and acceptance fields into the acceptance report and evidence chain.

9. The digital acceptance method for cable joints according to claim 8, characterized in that, When determining the processing boundary, the radial upper quantile envelope difference, outer layer point ratio difference, normal dispersion difference, curvature distribution difference, and material category probability difference of the continuous cross section are normalized and weighted. The continuous boundary interval is screened according to the structural arrangement order defined by the cable structure template, and the position with the largest boundary response within the continuous boundary interval is determined as the processing boundary. When performing virtual caliper measurement, points with measurement confidence weights lower than the effective threshold are excluded from contact calculation. The median diameter of multiple consecutive measurement cross sections is used as the measurement diameter, and the diameter dispersion is used as the repeatability measurement deviation.

10. The digital acceptance method for cable joints according to claim 8, characterized in that, After completing the point cloud-based acceptance processing, the surface load matrix output by the external micro-loading detection device is received. and displacement response matrix ,according to The load transfer operator is derived under displacement reconstruction residuals, reciprocity, spatial smoothing, and positive semidefinite constraints. According to the load transfer operator of the cable processing area Attachment load transfer operator Initial gap vector and contact load vector Establish and satisfy , and ; a diagonal matrix formed by the areas of surface units Calculate the contact pressure vector , the contact pressure vector The corresponding extended review results are written into the acceptance record.

Images