A method for precise marking and positioning of an eddy current testing of a pipe and related equipment

Abstract

The embodiment of the application provides a kind of pipe inner eddy current detection precision mark positioning method and related equipment, to solve the technical problems that traditional mileage positioning method is not accurate in corresponding inside and outside position due to the deformation of flexible transmission component when detecting internal defects of complex curved pipe.The method comprises: establishing a pipe space path model;Synchronously collect probe travel parameters and eddy current detection signals;Compensate the positioning error caused by the deformation of flexible component through curvature feature matching;Convert the defect coordinates to the external reference coordinate system and project radially to the outer surface.This method extracts the curvature feature sequence of flexible component and performs sliding window matching with the space path model, combines coordinate transformation and radial projection, solves the cumulative error problem caused by the deformation of flexible component in the curved section of pipe, improves the accuracy and reliability of external positioning of internal defects, and can be applied to the field of pipe detection and maintenance of industrial equipment such as heat exchanger and condenser.

Description

Technical Field

[0001] This application relates to the fields of non-destructive testing and industrial pipeline maintenance technology, and in particular to a method and related equipment for accurate marking and positioning of eddy current detection inside pipe fittings. Background Technology

[0002] Eddy current testing is a non-destructive testing technique for conductive materials, widely used in industrial equipment such as heat exchangers, condensers, and steam generators for surface or near-surface defect detection in pipe fittings. In practice, the eddy current testing probe is typically connected to the front end of a flexible transmission component such as a flexible cable or pusher, and is pushed into the pipe fitting by an external drive mechanism. When the probe passes through cracks, corrosion, or thinning areas in the pipe wall, it picks up abnormal electromagnetic induction signals, thereby identifying internal defects.

[0003] In some related technologies, in order to locate the location of internal defects on the outside of pipe fittings for subsequent repair welding, cutting, or pipe replacement, operators usually use a length estimation method based on mileage. The system records the insertion length or retraction length of the flexible transmission component when the defect is found. Then, based on this length data, the operator uses measuring tools such as a tape measure to measure the same distance along the outer surface of the pipe fitting and marks it manually at that location.

[0004] However, in real-world industrial scenarios, the objects to be inspected often include numerous bends, U-shaped tubes, or serpentine tubes. When the flexible transmission component travels inside these tubes with complex spatial curves, it is constrained by local frictional resistance in the bends and the geometry of the tube walls. This stress state causes the flexible transmission component to undergo localized compressive buckling (compression) or tensile elongation within the tube. As the number of bends the probe passes through increases, this nonlinear physical deformation accumulates, leading to a discrepancy between the externally recorded "cable insertion length" and the "true position" of the probe within the tube. The actual location of internal defects cannot be accurately mapped to the outer surface of the tube, easily causing unnecessary losses in normal tube sections and potentially missing genuine defects, reducing equipment maintenance efficiency and increasing overall maintenance costs. Summary of the Invention

[0005] This application provides a method and related equipment for precise marking and positioning of eddy current detection inside pipe fittings, which can improve the maintenance efficiency of the equipment and reduce the overall maintenance cost.

[0006] Firstly, this application provides a precise marking and positioning method for eddy current detection inside pipe fittings. The method includes: extracting coordinate feature points of the inlet reference and external reference reference of the pipe fitting to be inspected to calculate a coordinate transformation matrix; analyzing the geometric data of the pipe fitting to be inspected to generate a spatial path model containing three-dimensional spatial coordinates and curvature features; simultaneously acquiring the change in travel arc length, the bending morphology parameters of the flexible transmission component, and the amplitude of the eddy current detection signal picked up by the detection probe as the probe advances or retracts inside the pipe fitting via a flexible transmission component; using the spatial path model as a trajectory set, and taking the change in travel arc length as the initial arc length reference, extracting the curvature feature sequence formed during the advancement or retraction process from the bending morphology parameters; and processing the curvature feature sequence... A sliding window matching process is performed with the curvature characteristics of each curved segment in the spatial path model to compensate for the deviation between the arc length and the actual position caused by the deformation of the flexible transmission component. The matched three-dimensional coordinate points are used as the actual detection positions of the detection probe. Based on the amplitude change of the eddy current detection signal and the actual detection position, the starting and ending three-dimensional coordinates of the internal defect space interval are determined. The coordinate transformation matrix is ​​called to transform the starting and ending three-dimensional coordinates of the internal defect space interval to the coordinate system of the external reference datum. The transformed starting and ending three-dimensional coordinates are projected radially along the outer surface of the pipe to be inspected onto the outer surface contour of the pipe to be inspected to generate the coordinates of the outer surface marking interval of the pipe to be inspected for guiding external marking.

[0007] This embodiment extracts the curvature feature sequence during the movement of the flexible transmission component and performs sliding window matching with a pre-generated spatial path model to establish a spatial position mapping mechanism based on morphological features. This compensates for arc length deviations caused by compressive buckling or tensile elongation. Furthermore, by combining coordinate transformation and radial projection, the start and end coordinates of internal defects are directly converted into marked intervals on the outer surface contour of the pipe fitting. This reduces the cumulative error caused by relying on single mileage data for external positioning and improves the reliability of marking internal defects in the external physical space.

[0008] In some embodiments of the first aspect, the curvature feature sequence is matched with the curvature features of each curved segment in the spatial path model using a sliding window to compensate for the deviation between the arc length and the actual position caused by the deformation of the flexible transmission component. The matched three-dimensional coordinate point is used as the actual detection position of the detection probe. Specifically, this includes: constructing a sliding window with the change in the traveling arc length as the independent variable, and extracting the local curvature feature subsequence to be matched within the sliding window; introducing a time warping algorithm to calculate the cumulative distance matrix between the local curvature feature subsequence and the reference curvature sequence in the corresponding search interval of the spatial path model; finding the optimal warping path based on the cumulative distance matrix to eliminate the nonlinear spatial scale scaling caused by the buckling under pressure or the elongation under tension during the advancement or retraction of the flexible transmission component; aligning the feature points in the local curvature feature subsequence to the corresponding points of the reference curvature sequence according to the mapping relationship of the optimal warping path, and using the three-dimensional coordinate point of the aligned reference curvature sequence as the actual detection position of the detection probe.

[0009] This embodiment introduces a time warping algorithm to find the optimal warping path, and performs nonlinear spatial alignment between the local curvature feature subsequence and the reference curvature sequence. This reduces the interference of sequence stretching or compression caused by local forces on the matching process, and maps the feature points to the corresponding positions of the reference sequence, thereby improving the objectivity of the actual detection position calculation of the probe.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, before constructing the sliding window with the change in travel arc length as the independent variable, the method further includes: obtaining the axial force parameters and travel speed of the flexible transmission component in the current travel state; determining whether the flexible transmission component is in a compressive buckling state or a tensile elongation state based on the axial force parameters, and calculating the deformation compensation coefficient in combination with the travel speed; dynamically and adaptively adjusting the window length and sliding step size of the sliding window using the deformation compensation coefficient, increasing the window length when in a compressive buckling state, and decreasing the window length when in a tensile elongation state.

[0011] This embodiment calculates the deformation compensation coefficient by acquiring axial force parameters and travel speed, and dynamically adjusts the length and step size of the sliding window accordingly. The window length is increased during compressive buckling and decreased during tensile elongation, ensuring that the window parameters match the real-time physical deformation state of the flexible component. This improves the stability of the curvature feature sequence extraction and matching process under different stress states.

[0012] In some embodiments of the first aspect, the determination of the starting and ending three-dimensional coordinates of the internal defect space interval based on the amplitude change of the eddy current detection signal and the actual detection position includes: extracting the local spatial curvature corresponding to the actual detection position in the spatial path model; establishing a correlation mapping model between the probe lift-off distance and the local spatial curvature, and calculating the lift-off compensation gain coefficient of the current actual detection position based on the correlation mapping model; using the lift-off compensation gain coefficient to dynamically amplify or attenuate the amplitude of the synchronously acquired eddy current detection signal to eliminate the signal baseline drift caused by probe eccentricity in the pipe bending section; extracting the envelope of the compensated eddy current detection signal, and truncating the defect feature band based on a preset adaptive signal-to-noise ratio threshold, and using the actual detection positions corresponding to the starting point of the rising edge and the ending point of the falling edge of the defect feature band as the starting and ending three-dimensional coordinates of the internal defect space interval, respectively.

[0013] This embodiment establishes a correlation mapping model between probe lift-off distance and local spatial curvature, and uses the calculated lift-off compensation gain coefficient to dynamically amplify or attenuate the amplitude of the eddy current detection signal. This compensation mechanism counteracts the interference of spatial geometric changes on electromagnetic induction characteristics during signal processing. Combined with envelope extraction and adaptive signal-to-noise ratio thresholding to truncate the defect band, it reduces the misjudgment rate caused by geometric eccentricity and improves the reliability of defect spatial interval boundary determination.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, establishing a correlation mapping model between the probe lift-off distance and the local spatial curvature specifically includes: obtaining the outer diameter of the probe and the inner diameter of the pipe to be inspected; simplifying the probe into a rigid cylinder, and performing geometric interference simulation between the rigid cylinder and the inner wall of the pipe to be inspected at various local spatial curvatures in the spatial path model; calculating the theoretical lift-off distance between the probe surface coil and the inner wall of the pipe to be inspected based on the radial offset vector between the central axis of the rigid cylinder and the center line of the spatial path model obtained from the simulation, so as to generate the correlation mapping model.

[0015] In this embodiment, the detection probe is simplified as a rigid cylinder, and geometric interference simulations are performed between the probe and the inner wall of the pipe at various local spatial curvatures. By calculating the radial offset vector between the central axis of the rigid cylinder and the center line of the path, the degree of spatial attitude deviation of the probe under different degrees of curvature is quantified, providing a geometric basis for the calculation of the theoretical lift-off distance, thereby providing objective data support for signal gain compensation.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, the process of projecting the transformed starting point three-dimensional coordinates and ending point three-dimensional coordinates radially along the outer surface of the pipe fitting to the outer surface contour of the pipe fitting to generate pipe fitting outer surface marking interval coordinates for guiding external marking specifically includes: calculating the normal vectors of the positions of the starting point three-dimensional coordinates and the ending point three-dimensional coordinates based on the spatial path model in the coordinate system of the external reference datum; extending rays outward along their respective normal vectors, starting from the starting point three-dimensional coordinates and the ending point three-dimensional coordinates; analyzing the outer surface three-dimensional mesh model of the pipe fitting to be inspected and calculating the spatial intersection points of the rays and the outer surface three-dimensional mesh model; using the obtained spatial intersection points as outer surface marking points, and using the outer surface geodesic line segments formed by each outer surface marking point as pipe fitting outer surface marking interval coordinates.

[0017] This embodiment calculates the normal vectors of the starting and ending coordinates based on the spatial path model, and extends rays outward along the normal vectors. By analyzing the three-dimensional mesh model of the outer surface of the pipe to be inspected, spatial intersections are calculated. The abstract three-dimensional coordinates inside the pipe are mapped onto the three-dimensional solid outline of the outer surface of the pipe along the real physical normal, generating geodesic line segments composed of outer surface marker points, providing direct surface spatial guidance for external physical marking operations.

[0018] In some embodiments of the first aspect, before calculating the spatial intersection of the ray and the three-dimensional mesh model of the outer surface, the method further includes: obtaining the actual wall thickness distribution data of the pipe segment corresponding to the starting point three-dimensional coordinates and the ending point three-dimensional coordinates of the pipe to be inspected; determining whether there are eccentric thickening or thinning features in the actual wall thickness distribution data; if so, performing vector deflection correction on the extension angle of the normal vector according to the eccentric direction and eccentricity corresponding to the eccentric thickening or thinning feature to compensate for the radial projection refraction error caused by the uneven pipe wall thickness, and then using the corrected normal vector to extend the ray to calculate the spatial intersection.

[0019] This embodiment introduces actual wall thickness distribution data. When there are eccentric thickening or thinning features, vector deflection correction is performed based on the extension angle of the normal vector according to the eccentricity direction and eccentricity amount. This compensates for the geometric projection deviation caused by non-uniform pipe wall thickness, reduces the interference of pipe wall structure variation on the internal and external space mapping process, and improves the objectivity of the coordinates of the outer surface marked interval.

[0020] Secondly, embodiments of this application provide a precise marking and positioning device for eddy current detection inside pipe fittings. The device includes one or more processors and a memory. The memory is coupled to the one or more processors and is used to store computer program code, which includes computer instructions. The one or more processors call the computer instructions to cause the device to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a pipe fitting eddy current detection precision marking and positioning device, cause the pipe fitting eddy current detection precision marking and positioning device to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a pipe fitting eddy current detection precision marking and positioning device, cause the pipe fitting eddy current detection precision marking and positioning device to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the precise marking and positioning device for eddy current detection inside pipes provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. This application compensates for the arc length deviation caused by deformation by extracting the curvature feature sequence of flexible components and matching it with the spatial path model through a sliding window; by combining coordinate transformation and radial projection, the coordinates of internal defects are converted into the marking interval of the outer surface, which reduces the cumulative error of single mileage positioning and improves the reliability of the external marking.

[0026] 2. This application establishes a correlation mapping model between probe lift-off distance and local spatial curvature, calculates the lift-off compensation gain coefficient to dynamically compensate the amplitude of eddy current detection signal, offsets the interference of spatial geometric shape changes on electromagnetic induction characteristics, reduces the misjudgment rate caused by probe eccentricity, and improves the reliability of defect boundary determination.

[0027] 3. This application calculates the normal vectors of the defect's start and end coordinates and extends them outwards. It then uses actual wall thickness distribution data to correct the vector deflection angle of the normal vector extension, compensating for the radial projection refraction error caused by uneven pipe wall thickness. This accurately maps the internal coordinates to the external entity contour, providing high-precision guidance for external marking. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a precise marking and positioning method for eddy current detection inside pipe fittings in an embodiment of this application.

[0029] Figure 2 This is a schematic diagram of the physical device structure of a precise marking and positioning device for eddy current detection inside pipe fittings in the embodiments of this application. Detailed Implementation

[0030] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0031] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0032] For ease of understanding, the method provided in this implementation is described in process below. Please refer to [link / reference]. Figure 1 This is a flowchart illustrating a precise marking and positioning method for eddy current detection inside pipe fittings in an embodiment of this application.

[0033] S101. Extract the coordinate feature points of the inlet datum and the external reference datum of the pipe fitting to be inspected in order to calculate the coordinate transformation matrix.

[0034] The entry reference refers to the spatial positioning reference of the end face where the detection probe enters the pipe fitting under inspection. It is used to determine the origin and initial axial direction of the probe's starting position in the pipe fitting's local coordinate system. Typical implementations include a combination of the center point of the entry flange and the keyway direction vector, or a combination of the center of the entry section and the entry tangential unit vector. The external reference reference refers to a coordinate reference system located outside the pipe fitting that can be directly measured or identified in physical space. It is used to provide a unified spatial orientation reference for the final external marking operation. Typical implementations include a system of positioning pin holes on the tooling fixture, or pre-set visual markers on the outer wall (such as high-precision QR code targets). Coordinate feature points refer to a set of sampling points that can establish a stable correspondence between the entry local coordinate system and the external reference coordinate system. It must contain at least three non-collinear points to fully constrain the six-degree-of-freedom transformation. The larger the point set, the higher the degree of overdeterminacy of the registration equation system and the better the robustness of the matrix solution. The coordinate transformation matrix is ​​a linear transformation operator that maps the three-dimensional coordinates in the local coordinate system of the pipe fitting to the external reference coordinate system. It is usually represented by a 4×4 homogeneous transformation matrix, which contains a 3×3 rotation component R and a 3×1 translation component t. The matrix can be obtained by least squares point pair registration, three-point orientation determination method or ICP (Iterative Closest Point) algorithm.

[0035] This step is performed before each inspection operation begins. Its purpose is to establish a coordinate mapping link that runs through the entire process, allowing all 3D coordinates obtained internally within the pipe fitting to be uniformly converted to the external physical space. Specifically, at least two paths are available: Path 1: Acquire the coordinates of the center of the inlet flange end face and the keyway direction vector, while simultaneously selecting the coordinates of three or more locating pin holes on the tooling base. Substitute these two sets of points into the least squares registration equations to solve for the rotation matrix R and the translation vector t. Path 2: Using the center of the inlet section as the local origin, extract the coordinates of preset marker points on the outer wall using a visual recognition system. Determine the rotation component using the three-point orientation method, and then calibrate the translation component using a known reference distance. For both paths, the residuals at each registration point should be calculated after matrix solving. Typically, the root mean square residual of the registration should not exceed 0.5 to 2 mm to ensure that the contribution error of the coordinate mapping link to the final marking accuracy is within a controllable range.

[0036] During engineering implementation, thermal expansion of pipe fittings or multiple clamping operations can cause attitude shifts, rendering the original matrix invalid. To address this, the system supports rapid re-acquisition of feature points and re-solution before each inspection, ensuring the transformation matrix always corresponds to the actual geometric state of the current clamping, thus preventing batch-to-batch attitude deviations from being transmitted to the final marking results. Furthermore, when a pre-calibrated tooling coordinate system with known accuracy or external positioning equipment such as a laser tracker is available, its output can be directly used as an equivalent source of coordinate correspondence, eliminating the need for feature point registration each time.

[0037] S102. Analyze the geometric data of the pipe fitting to be inspected and generate a spatial path model containing three-dimensional spatial coordinates and curvature features.

[0038] Geometric data refers to the original input information describing the three-dimensional shape of the pipe fitting under inspection, used to drive the generation of the spatial path model. The sources are at least two categories: one is directly reading CAD or 3D design files (such as STEP or IGES formats) to parse the pipe fitting's centerline geometry; the other is extracting the centerline coordinate sequence through reverse engineering using 3D laser scanning, industrial CT reconstruction, or pipe bending machine processing parameters. The spatial path model refers to a three-dimensional digital trajectory representation with the pipe fitting's internal centerline as its framework. Its core data structure should at least include the following fields: three-dimensional coordinates of discrete points on the centerline. Corresponding arc length parameters Tangential unit vector Principal normal vector binormal vector Local curvature With torque And the start and end marks of the bends; it can be expanded as needed to a "centerline + cross-sectional dimension model" including the inner and outer diameters of the cross-section, or a complete geometric model with an outer surface mesh. Curvature characteristics refer to the quantitative value of the local curvature of the path, in units of 1 / mm or rad / mm, which can be calculated by the finite difference method of tangent vectors of adjacent discrete points, the three-point circle fitting method, or the second derivative method of spline curves.

[0039] This step is completed before the formal start of the inspection process. The generated spatial path model will serve as the reference matrix for subsequent position matching and coordinate transformation. Specifically, when the geometric data comes from a CAD file, the system analyzes the pipe centerline and performs discrete sampling at intervals of 1 to 10 mm along the arc length. At each sampling point, the local curvature and torsion are calculated using the second derivative of cubic spline fitting. When the geometric data comes from reverse engineering of a physical object, the segmented bending radius and bending angle reported by the pipe bending machine can be converted into equivalent curvature. Using the measurement point as an anchor point, the intermediate segment is completed through interpolation, thus forming an ordered discrete point sequence. In terms of model field design, the introduction of the torsion field is significant because curvature amplitude alone cannot distinguish between two bends with different bending planes but similar curvature magnitudes. The combined use of curvature and torsion can significantly reduce ambiguity in candidate positions in subsequent matching steps, which is particularly noticeable in improving positioning accuracy in spatial pipe fittings with multiple adjacent similar bends.

[0040] The spatial path model can be expanded into a complete geometric model with an outer surface mesh, depending on the scenario requirements, to directly support subsequent intersection operations of the outer surface normal rays. In scenarios with dense pipe components such as heat exchanger tube bundles or boiler coils, the absolute coordinates of structural beacons such as support plates, clamps, welds, and tube sheets can be stored as additional fields in the model. These structural beacons form identifiable feature responses in the detection signal and can be used as absolute coordinate calibration anchor points during probe travel, providing a basis for phased correction of position estimation.

[0041] S103. When the detection probe is pushed or pulled back inside the pipe to be inspected via the flexible transmission component, the change in the travel arc length, the bending morphology parameters of the flexible transmission component, and the amplitude of the eddy current detection signal picked up by the detection probe are collected simultaneously.

[0042] The change in travel arc length refers to the cumulative increase in the path length of the detection probe traveling along the inside of the pipe from the inlet reference, used as an initial arc length estimate for the nominal position of the probe. It can be obtained by pulse counting conversion from a rotary encoder coaxial with the push-pull mechanism, rolling count of the odometer wheel on the surface of the flexible component, or output from a servo screw displacement sensor. The typical sampling resolution ranges from 0.1 to 2 mm. The flexible transmission component refers to the flexible carrier that connects the detection probe to the external drive mechanism, is responsible for transmitting push-pull force inside the curved pipe and guiding the probe along the pipe path. Specific forms include, but are not limited to, push-pull flexible shafts, guide sheaths, chain-link flexible rods, and composite probes. Their common feature is that local compression or tension occurs when the curved section is subjected to force, resulting in a cumulative deviation between the arc length value measured at the end and the actual position of the probe. Bending morphology parameters refer to the sensor outputs reflecting the actual bending distribution along the path of the flexible transmission component in its current propulsion state. Unlike arc length counts, which only reflect end displacement, these parameters can be acquired through at least two methods: First, by distributing fiber optic shape sensors (based on multi-point FBG strain sensing or Rayleigh scattering OFDR) along the flexible component, directly outputting a three-dimensional curvature field distributed along the sensing arc length; second, by deploying multiple miniature IMUs or gimbal angle encoders in series along the flexible component, reconstructing the local bending angle sequence along the path through the accumulation of joint angles at each node. Synchronous acquisition means that the three signals (arc length change, bending morphology parameters, and eddy current detection signal amplitude) use a unified hardware clock source, are triggered by the same acquisition controller, and are stamped with the same timestamp, ensuring strict alignment of the three data streams on the time axis.

[0043] This step is performed continuously throughout the entire stroke of the probe as it advances or retracts within the tube. The three raw data streams collected form the common basis for subsequent position determination and defect localization. Specifically, on the arc length acquisition side, the encoder or odometer wheel continuously outputs pulse counts at a resolution of 0.1 to 2 mm and converts them into arc length increments. On the shape acquisition side, the fiber optic shape sensor or IMU array continuously outputs the curvature or attitude values ​​of each sensing node at a frequency of 50 to 500 Hz, updating in real time as the flexible component moves. On the eddy current signal side, the detection coil picks up the induced electromotive force amplitude sequence at a sampling frequency of 10 to 500 kHz. After the three signals are timestamped, they form a synchronous data frame indexed by a unified time axis, allowing simultaneous querying of the current arc length mileage, current component shape status, and current eddy current response at any given time.

[0044] The reason for using "synchronization" as a core design constraint is that any temporal misalignment of the three signals will lead to incorrect binding of morphological parameters at different spatial locations with defect signals at different locations, causing systematic deviations in location solving and defect boundary extraction. The morphological parameters are acquired through distributed acquisition along the entire process rather than single-point observation at the end because the local stress and bending state of the flexible component within the curved tube cannot be reflected by the end signals; only a complete morphological distribution along the entire process can provide effective input for subsequent curvature sequence matching. In some embodiments, auxiliary quantities such as push-pull motor current, frictional resistance, or pullback tension can also be recorded synchronously. These quantities can be used for deformation state identification and subsequent adaptive adjustment of window parameters, and can also provide alternative data sources when the morphological sensor malfunctions or its accuracy degrades.

[0045] S104. Using the spatial path model as the trajectory set, and taking the change in the traveling arc length as the initial arc length reference, extract the curvature feature sequence formed in the bending morphology parameters during the advancement or retraction process.

[0046] The initial arc length reference refers to the cumulative arc length mileage value provided by the change in the traveling arc length. It serves only as the initial candidate search starting point for the probe in the spatial path model, not the final determined actual detection position. The deviation between it and the probe's actual position originates from the local compression or stretching of the flexible transmission component in the curved section. This deviation accumulates with the propulsion distance and exhibits nonlinearity under different stress states, requiring correction through subsequent matching steps before it can be used for coordinate mapping. The curvature feature sequence refers to the ordered discrete curvature value set extracted from the bending morphology parameters, with arc length as the independent variable, in the form of...

[0047]

[0048] in The component arc length position corresponding to the j-th sampling point. This is the local curvature value at that location; this sequence describes the actual bending distribution of the flexible transmission component along its own arc length direction during the current journey, and is the core feature carrier for similarity comparison with the pre-stored benchmark curvature sequence in the spatial path model.

[0049] After the arc length change and morphological parameters are synchronously acquired, this step transforms the original morphological quantities into curvature feature representations that can be directly used for sequence matching. Specifically, when the morphological acquisition method is a distributed fiber optic shape sensor, the local curvature values ​​output by the sensor along the sensing unit position naturally form a sequence indexed by the sensor arc length coordinates, which can be used directly after resampling and alignment. When the acquisition method is a serial IMU or joint encoder, the joint angle data must first be converted into local tangent vectors through the coordinate difference between adjacent nodes, and then the equivalent curvature is calculated by the ratio of the angle between adjacent tangent vectors to the inter-segment arc length. Finally, the data are arranged in arc length order to form a discrete sequence.

[0050] Regardless of the sensing method used, smoothing should be applied after the sequence is constructed—using methods such as moving average, Savitzky-Golay filtering, or cubic spline smoothing—to suppress local fluctuations caused by sensor noise and avoid false mismatches caused by high-frequency disturbances in subsequent similarity calculations. From a positioning logic perspective, the initial arc length benchmark is only used to narrow down the candidate search range; the system does not rely on its absolute accuracy. Transforming the arc length mileage accumulation error problem into a curvature sequence feature matching problem is the core mechanism that distinguishes this scheme from traditional pure mileage positioning, giving the position estimation results structural tolerance to the slippage, compression, and stretching of flexible components. Furthermore, in scenarios where pipes exhibit spatial torsion (such as spiral coils), a torsion sequence can be added to the curvature sequence to form a curvature-torsion joint feature, which can distinguish pipe segments with similar curvature amplitudes but different bending planes, further improving the discriminative power of the matching steps in complex three-dimensional pipes.

[0051] S105. The curvature feature sequence is matched with the curvature features of each curved segment in the spatial path model using a sliding window to compensate for the deviation between the arc length and the actual position caused by the deformation of the flexible transmission component. The three-dimensional coordinate points determined by the matching are used as the actual detection positions of the detection probe.

[0052] The sliding window refers to a local search unit within the arc-length domain of the spatial path model, centered on an initial arc-length reference and advancing by a fixed length L. The typical range for the window length L is 20 to 150 mm, and the typical range for the sliding step size is 1 to 20 mm. Specific values ​​are determined comprehensively based on the pipe bending density, the resolution of the shape sensor, and the allowable computational cost. The reference curvature sequence refers to the theoretical curvature discrete sequence pre-stored in the spatial path model, corresponding to the arc-length positions of each bending segment. This sequence serves as a reference standard for similarity comparison with the measured local curvature feature subsequence. The actual detection position refers to the estimated position of the probe coil in the three-dimensional coordinate system of the spatial path model after sliding window matching and correction. This position can be given as a single three-dimensional coordinate point or as a short coordinate interval with the highest confidence level, to accommodate different positioning accuracy requirements.

[0053] This step is the core of the entire positioning link. It replaces pure arc length mileage estimation with a morphological feature alignment mechanism, decoupling the local slippage error of the flexible transmission component in curved sections from the position estimation. Specifically, at each sampling time, the system locates the center of the search interval based on the initial arc length reference. Within a certain range before and after this center (typically ±50mm of the reference position, dynamically adjusted according to the maximum expected deformation), a search interval is constructed. A sliding window is constructed within the search interval, and candidate segments of the reference curvature sequence are extracted step by step. The distance metric is calculated with the measured local curvature feature subsequence. The distance metric can be the sum of weighted absolute differences, Pearson correlation coefficient loss, or DTW cumulative path distance. DTW finds the optimal regular path by establishing a cumulative distance matrix and applying a constrained bandwidth (typically 5% to 30% of the window length), which can maintain robust alignment under nonlinear stretching conditions of the sequence.

[0054] The 3D coordinates corresponding to the candidate segment with the smallest distance are the actual detection position of the probe. When more robust output is required, the shortest position interval with the highest confidence can also be given. When the matching distance exceeds the historical mean plus 1 to 3 times the standard deviation at a certain moment, the system judges it as a low-confidence match and triggers an alarm. The matching method is not limited to DTW. In other embodiments, constrained correlation matching, hidden Markov model path inference, particle filter state estimation, or regression localization based on a trained model can also be used.

[0055] The common feature of the above methods is that they all use spatial path models and morphological parameters to jointly constrain the probe position, rather than relying solely on mileage data. When the positions of structures such as pipe sheets, support plates, and clamps form identifiable characteristic responses in eddy current signals or morphological data during the detection process, the time corresponding to this response has known absolute coordinates prior. These coordinates can be used as observation update points to perform phased recalibration or filtering fusion of the current matching position, so that the positioning error in long-distance or multi-bend pipe fittings does not monotonically accumulate with the travel distance.

[0056] In some preferred embodiments, the above steps specifically include:

[0057] A sliding window with the change in travel arc length as the independent variable is constructed, and the local curvature feature subsequence to be matched is extracted within the sliding window; a time warping algorithm is introduced to calculate the cumulative distance matrix between the local curvature feature subsequence and the benchmark curvature sequence in the corresponding search interval in the spatial path model;

[0058] The optimal regularization path is found based on the cumulative distance matrix to eliminate the nonlinear spatial scale scaling caused by the buckling under pressure or the elongation under tension of the flexible transmission component during the advancement or retraction process; according to the mapping relationship of the optimal regularization path, the feature points in the local curvature feature subsequence are aligned to the corresponding points of the reference curvature sequence, and the three-dimensional coordinate points of the aligned reference curvature sequence are used as the actual detection positions of the detection probe.

[0059] The local curvature feature subsequence refers to a short subset of segments extracted from the measured curvature feature sequence within the current sliding window, with the arc length change as the independent variable. This subset is denoted as... Its length *m* corresponds to the arc length range covered by the current window, serving as the query sequence for this matching iteration. The corresponding search interval of the baseline curvature sequence refers to the candidate reference segment extracted from the spatial path model according to the search interval, denoted as... The values ​​of m and n may differ due to compression or stretching of the flexible component. The cumulative distance matrix refers to the m×n two-dimensional matrix D in the DTW algorithm, which records the cumulative matching cost of Q and R point by point and is obtained through recursive filling. The optimal regularized path is the monotonic path in D that connects the lower left corner D(1,1) to the upper right corner D(m,n) and minimizes the total cumulative cost; each step (i*, j*) on the path establishes a... and The optimal point correspondence relationship, the three-dimensional coordinates of the reference sequence corresponding to the middle segment of the path are the estimate of the actual detection position of the probe.

[0060] This preferred embodiment provides a complete DTW implementation path for the matching sub-step of S105, enabling the sliding window matching process to handle non-uniform deformation of flexible components. Specifically, the query sequence Q and model candidate segment R are extracted within the current sliding window, and then the cumulative distance matrix is ​​filled element by element according to the following recursive relationship: in For local distance measurement between points, the absolute difference can be used. Or weighted squared difference; matrix boundary conditions set , , To prevent misalignment due to excessive offset from local minima, a Sakoe-Chiba constraint bandwidth w (typically 5% to 30% of the window length) is applied, stipulating that the path must not deviate from the diagonal by more than w steps. The optimal regularized path is obtained by backtracking from D(m, n), and the reference sequence 3D coordinates corresponding to the middle segment of the path are used as the estimate of the actual detection position of the probe. When an interval format is required for output, the coordinates corresponding to the start and end points of the path can be used to form the boundary of the position interval. When the final cumulative distance D(m, n) exceeds the historical mean plus k times the standard deviation (typically k is 1 to 3), it is judged as a low-confidence match, and a second scan is recommended.

[0061] The DTW scheme is preferred in engineering scenarios due to its mature algorithm, interpretable parameters, and lack of pre-training data. In other embodiments, the same nonlinear position correction function can also be achieved by using constrained correlation sequence alignment, sequence alignment neural network, hidden Markov model or particle filter. All of the above methods are based on the joint constraint of spatial path model and morphological parameters.

[0062] In some other embodiments of this application, before the step of constructing a sliding window with the change in travel arc length as the independent variable, the method further includes:

[0063] Obtain the axial force parameters and travel speed of the flexible transmission component in its current traveling state;

[0064] The axial force parameters are used to determine whether the flexible transmission component is in a state of compressive buckling or tensile elongation, and the deformation compensation coefficient is calculated in combination with the travel speed.

[0065] The window length and sliding step size of the sliding window are dynamically and adaptively adjusted using the deformation compensation coefficient. The window length is increased when the window is in a compressive buckling state and decreased when the window is in a tensile elongation state.

[0066] Among them, the axial force parameter refers to the quantified value of the compressive or tensile force acting on the flexible transmission component along its own axial direction, used to determine the deformation state of the component at the current moment of travel. There are two typical acquisition methods: First, the real-time drive current of the push-pull motor is converted into an equivalent axial thrust or pull force according to the rated torque coefficient of the motor and the efficiency of the transmission mechanism; Second, a range-matched tension and compression sensor is connected in series in the transmission link to directly output the axial force amplitude, with a sampling resolution typically better than 1N. Travel speed refers to the instantaneous advance or retraction rate of the arc length at the end of the flexible transmission component, in mm / s, and is obtained by quotienting the increment of adjacent sampling points of the arc length encoder with respect to the time difference. The deformation compensation coefficient is a dimensionless correction factor determined jointly based on the current stress state and travel speed. It is used to drive the adaptive adjustment of the sliding window length and step size. Under compression, the coefficient is greater than 1 (typically 1.10 to 1.80 in the extended range), and less than 1 under tension (typically 0.60 to 0.95 in the contraction range). When there is no obvious deformation, it is taken as 1.00. Its value can be determined by a piecewise linear function or by an offline calibration lookup table indexed by the stress amplitude and speed.

[0067] This sub-step is executed before constructing the sliding window. Its purpose is to inject the current physical deformation level of the flexible component into the window parameter decision in advance, so that the window length is adapted to the current arc length scaling degree when matching starts, thereby reducing curvature sequence truncation or interval overlap caused by the mismatch between the window width and the actual coverage arc length.

[0068] Specifically, the system continuously reads axial force parameters and travel speed at each acquisition trigger moment: when the axial force exceeds the preset compression threshold F1, it is determined to be in a state of compressive buckling, and the window length is adjusted accordingly. Increase, of which The reference window length is defined as α (typically ranging from 20 to 150 mm), and α is the expansion coefficient obtained by interpolating the current force amplitude and travel speed (typically ranging from 0.10 to 0.80). When the axial force is lower than the preset tensile threshold F2, it is determined to be in a tensile elongation state, and the window length is adjusted accordingly. The shrinkage coefficient β is reduced (typically ranging from 0.05 to 0.40). The sliding step size is adjusted proportionally in the same direction to maintain a stable ratio between the window length and the step size. Including the travel speed in the calculation of the deformation compensation coefficient is because the actual compression accumulated at low speed and high speed under the same force amplitude is significantly different—at low speed, the frictional jamming effect between the component and the pipe wall is more pronounced, and local accumulation is more severe—if the degree of deformation is judged solely by the force amplitude, the compression will be continuously underestimated in the low-speed segment, resulting in a shorter window and insufficient coverage of the matching interval.

[0069] Compression buckling and tension elongation are two typical working conditions. In actual engineering, atypical deformation states such as low-speed stagnation, stick-slip alternation, local jamming, and reverse pullback hysteresis may also occur. The system can identify the above states by monitoring the frequency of speed fluctuations and the amplitude of force value changes in real time. When the deformation compensation coefficient exceeds the design range, it will output a marker indicating a decrease in positioning reliability, prompting the operator to perform a second scan on the abnormal section. This allows the positioning uncertainty caused by deformation to enter the human-machine collaborative management process rather than being completely digested by the algorithm.

[0070] S106. Based on the amplitude change of the eddy current detection signal and the actual detection location, determine the starting three-dimensional coordinates and the ending three-dimensional coordinates of the internal defect space interval.

[0071] The amplitude change of the eddy current detection signal refers to the signal response characteristics caused by local anomalies in the electromagnetic induction characteristics of the wall surface when the detection probe coil travels inside the pipe fitting. This application uses amplitude as the basic feature for defect judgment. When it is necessary to improve the ability to distinguish different defect mechanisms, it can be extended to combined features such as phase shift, complex impedance plane trajectory area, or envelope integral energy. Among them, the amplitude response reflects the absolute amplitude change of the induced electromotive force of the coil, while the phase response can distinguish different defect formation mechanisms such as wall thickness reduction and surface cracks. The internal defect spatial interval refers to the spatial segment describing the defect distribution range in the local three-dimensional coordinate system of the pipe fitting. According to the defect morphology, it can be represented as point defects (interval length on the order of single detection spatial resolution), short segment defects (length on the order of millimeters to centimeters), and continuous corrosion zones or crack propagation zones (a collection of multiple discrete sub-intervals), all unified by the three-dimensional coordinates of the starting point. Three-dimensional coordinates of the endpoint Characterization.

[0072] After S105 completes the actual detection position determination of the probe, this step uses the synchronously aligned eddy current signal sequence and the position time axis as dual inputs to perform a complete extraction process from the original signal to the coordinates of the defect interval with spatial attributes. Specifically, the system first performs preprocessing on the original eddy current signal: applying bandpass or lowpass filtering to suppress high-frequency acquisition noise and power frequency interference, and then obtaining the absolute value of the analytical signal as the amplitude envelope through Hilbert transform, or calculating an approximate envelope using the moving root mean square. Subsequently, it enters the threshold determination stage, which can use two methods: First, a fixed threshold is set based on the calibration baseline of the standard test block or blank tube, using the mean amplitude of the calibration segment plus k times the standard deviation (k is typically 2 to 4) as the cutoff threshold, which is suitable for batch testing scenarios where the pipe material and probe specifications are stable; Second, the local root mean square of a signal segment several millimeters near the current position is used as the dynamic background noise floor, multiplied by a preset signal-to-noise ratio gain (typically 2 to 6 times) to obtain an adaptive threshold, which is suitable for scenarios where the signal baseline continuously drifts with the bending segment. After extracting the defect feature bands, the sampling times at the start and end of the rising edge are located on the position-time axis and their corresponding actual detection positions are found, and output as the three-dimensional coordinates of the start point. Three-dimensional coordinates of the endpoint .

[0073] The defect boundary is described using three-dimensional coordinates rather than a single mileage value because in spatially curved pipe fittings, the same mileage increment corresponds to different three-dimensional spatial spans at different bends. A single mileage value is difficult to accurately indicate the distribution range of defects in physical space. However, three-dimensional start and end coordinates can directly serve subsequent coordinate transformations and external surface projections, and support the accurate visualization and overlay of defect distribution on the three-dimensional digital model of the pipe fitting. When structural beacons such as tube sheets, support plates, clamps, or welds form identifiable characteristic responses in the eddy current signal during the inspection process, the system identifies and eliminates these responses based on pre-stored structural templates, marking the corresponding moment as an absolute coordinate calibration event rather than a defect boundary. Defect judgment is only performed on the residual signal segments filtered by the structural template, reducing misjudgments caused by structural beacons while allowing the beacon response to serve the dual purpose of position calibration and signal classification, enhancing the collaborative reliability of positioning and defect judgment in the inspection of long-stroke multi-bend pipe fittings.

[0074] In some preferred embodiments, the above steps specifically include:

[0075] Extract the local spatial curvature corresponding to the actual detection position in the spatial path model; establish a correlation mapping model between the probe lift-off distance and the local spatial curvature, and calculate the lift-off compensation gain coefficient of the current actual detection position based on the correlation mapping model; use the lift-off compensation gain coefficient to dynamically amplify or attenuate the amplitude of the synchronously acquired eddy current detection signal to eliminate the signal baseline drift caused by probe eccentricity in the bent section of the pipe fitting; extract the envelope of the compensated eddy current detection signal, and extract the defect feature band based on a preset adaptive signal-to-noise ratio threshold. Use the actual detection positions corresponding to the starting point of the rising edge and the ending point of the falling edge of the defect feature band as the starting three-dimensional coordinates and the ending three-dimensional coordinates of the internal defect spatial interval, respectively.

[0076] The lift-off distance refers to the normal distance between the induction coil of the detection probe and the inner wall of the inspected pipe fitting, measured in mm. In straight pipe sections, the probe is approximately coaxial with the pipe wall, and the lift-off distance is close to the nominal gap value. However, in curved sections, due to the lateral offset of the probe relative to the center of curvature, it locally approaches the inner arc side of the pipe wall, resulting in a decrease in the lift-off distance on the inner arc side and an increase on the outer arc side, forming an asymmetric signal baseline drift. The correlation mapping model refers to a function mapping with local spatial curvature k as input and theoretical lift-off distance δ as output. Its implementation forms include, but are not limited to: analytical geometric models constructed based on geometric interference simulation of rigid cylinders; empirical lookup table models formed by fitting measured calibration data of different pipe diameters and probe specifications; and multivariate mapping models further using the pipe fitting material conductivity, wall thickness, nominal diameter, and excitation frequency as joint input variables. Different forms involve a trade-off between computational cost and accuracy, and the appropriate model can be selected based on site conditions and calibration resources. The lift-off compensation gain coefficient G is a dimensionless correction factor output by the correlation mapping model. Its physical relationship with the lift-off distance is determined based on the monotonically decreasing law of induced electromotive force as the lift-off distance increases: when the lift-off distance is too large, G>1 performs signal amplification compensation; when the lift-off distance is too small, G<1 performs attenuation compensation. The average lift-off distance measured in the straight pipe section is used as the benchmark. The reference point is used as the baseline. The adaptive signal-to-noise ratio threshold is a defect truncation threshold dynamically determined based on the local background noise floor of the compensated signal. It is obtained by multiplying the local root mean square noise floor of the signal segment several millimeters before and after the current position by a preset signal-to-noise ratio gain (typically 2 to 6 times), and automatically follows and adjusts as the signal baseline fluctuates.

[0077] After determining the actual detection location of each sampling point, this sub-step uses the curvature information in the spatial path model to drive point-by-point signal compensation, and then extracts the defect boundary coordinates from the compensated signal. Specifically, for each actual detection location p, the system reads the corresponding local spatial curvature k from the spatial path model, substitutes it into the correlation mapping model to obtain the theoretical lift-off distance δ(k), and then... The compensation gain coefficient is calculated, where the exponent m depends on the excitation frequency and coil geometry, typically ranging from 1.0 to 2.5. During the calibration phase, it is determined by fitting measured data from a standard test block with a known lift-off distance. If an empirical lookup table is used, the value of G is obtained directly from the pre-stored (k,G) point interpolation without explicit calculation of δ. G(k) is multiplied by the signal amplitude at the current position to complete the compensation point by point, eliminating the baseline slope effect caused by the eccentricity of the curved section and restoring the induction amplitude level to an approximately equivalent straight pipe section.

[0078] After compensation, the system extracts the envelope of the amplitude sequence and applies an adaptive signal-to-noise ratio threshold, extracts the defect feature band, and outputs the actual detection positions bound to the rising edge start point and falling edge end point of the band as the three-dimensional coordinates of the start and end points of the defect spatial interval, respectively. A gain correction mechanism driven by local spatial curvature is chosen instead of adaptive filtering based on historical baseline drift because the latter relies on the assumption of signal stationarity. In multi-bend, continuously variable curvature sections, the baseline change rate easily exceeds the filtering tracking capability, resulting in undercompensation or overcompensation in the curvature jump neighborhood. In contrast, spatial curvature-based mapping compensation directly drives the correction amount based on physical causes, providing immediate response to curvature abrupt changes. Furthermore, by expanding the model input variables, its adaptability to different pipe fitting specifications and excitation conditions can be further improved, providing an adjustable accuracy margin for subsequent expansion to multiple pipe fitting models and multi-frequency band excitation.

[0079] In other embodiments of this application, the above-mentioned establishment of the correlation mapping model between the detection probe lift-off distance and the local spatial curvature specifically includes:

[0080] Obtain the outer diameter of the detection probe and the inner diameter of the pipe fitting to be inspected;

[0081] The detection probe is simplified into a rigid cylinder, and the geometric interference between the rigid cylinder and the inner wall of the pipe to be inspected is simulated at various local spatial curvatures in the spatial path model.

[0082] Based on the radial offset vector between the central axis of the rigid cylinder and the center line of the spatial path model obtained from the simulation, the theoretical lift-off distance between the detection probe surface coil and the inner wall of the pipe is calculated to generate an associated mapping model.

[0083] Among them, the outer diameter of the detection probe This refers to the outer diameter of the largest cross-section of the probe body, and the inner diameter of the pipe fitting being inspected. This refers to the nominal inner diameter of the through hole in the inner wall of the pipe fitting. Half the difference between the two determines the nominal radial clearance of the probe in a straight pipe section. To accurately estimate the lift-off distance at the coil, the length of the probe's front guide head also needs to be taken into account. (Typical 10 to 50 mm) and the axial offset distance between the center of the induction coil and the probe head. The reason is that when the probe experiences lateral displacement in a curved section, the axial offset of the coil determines the relative position between the coil cross-section and the geometric apex of the bend, thus affecting the accuracy of radial displacement estimation at that cross-section. Ignoring this... This will introduce systematic bias.

[0084] The rigid cylinder simplification refers to approximating the probe body and its front-end transmission components of a certain length as a finite-length rigid cylinder, sacrificing the fine morphology of the flexible end segments in exchange for the analytical solvability of geometric interference simulation. This simplification has good accuracy when the rigid section of the probe is long and the ratio of the bending radius of the pipe to the inner diameter of the pipe is greater than 3:1. When the ratio of the bending radius to the inner diameter is less than 2:1, a multi-section rigid-flexible coupling chain model should be used instead.

[0085] Geometric interference simulation refers to determining the lateral offset equilibrium state of the cylinder axis relative to the pipe centerline at a given local spatial curvature k pipe section, with the rigid cylinder outer wall not penetrating the pipe wall as a hard constraint. The contact assumption is that the outer wall of one side of the cylinder forms a line contact with the pipe wall on the curved inner arc side, and the direction of the contact point is determined by the direction of the curvature center; penetration is not allowed. The radial offset vector Δr is the lateral displacement vector of the center of the rigid cylinder axis at the corresponding section of the coil relative to the centerline point of that section in the spatial path model, pointing towards the curvature center. The theoretical lift-off distance is calculated by the following formula: δ = c - |Δr|, which reaches a minimum value close to zero on the inner arc side and a maximum value of approximately 2c on the outer arc side.

[0086] This sub-step is executed offline during the parameter configuration phase before the detection operation. It simulates each curvature sampling point in the spatial path model one by one, forming a (k,δ) discrete mapping table indexed by curvature k, which can be queried point by point during runtime without consuming online computing resources. Specifically, the system reads... , , and Four geometric parameters, for each curvature value in the model. The lateral offset of the rigid cylinder axis at the coil section is calculated based on the geometric constraints of the single-sided wall attachment. ,get: All Point pairs are arranged in ascending order of curvature to form an ordered discrete table for linear interpolation lookup, thus constituting the associative mapping model of this embodiment. The applicable scope of the model should be clearly defined in the calibration report; typically, the bending radius should be no less than 1.5 times the nominal diameter of the pipe, and the probe outer diameter fill rate should be... The model accuracy is optimal when the value is between 0.85 and 0.98.

[0087] From the perspective of scheme integrity, when the probe tip has a long flexible tail section, or the ratio of the bending radius to the pipe inner diameter is less than 3:1, the accuracy of the rigid cylinder simplification decreases. In such scenarios, the probe and its front-end transmission section can be divided into several rigid segments and flexible hinged sections. A multi-segment rigid-flexible coupling chain model can be used to solve the chain link equilibrium configuration at each bending curvature to obtain a more realistic lift-off distance estimate. Alternatively, an empirical lookup table can be directly established using physical calibration data covering different curvatures, pipe diameters, and probe specifications as an equivalent substitute or comparative verification of the above analytical model. In any implementation, the core technical point of the correlation mapping model lies in estimating the coupling geometric relationship between the probe and the inner wall of the pipe fitting based on the current local spatial state, and correcting the detection signal accordingly, without relying on the statistical assumptions of the signal historical baseline. This characteristic enables the model to independently correct each spatial position in multi-bend and variable curvature pipe fittings, without introducing cross-segment interference due to accumulated historical deviations.

[0088] S107. Call the coordinate transformation matrix to transform the starting three-dimensional coordinates and ending three-dimensional coordinates of the internal defect space interval to the coordinate system of the external reference datum.

[0089] The local coordinate system of the pipe fitting refers to a right-handed rectangular coordinate system established with the center of the inlet reference section as the origin, the unit vector of the inlet axis as the positive Z-axis, and the XY plane determined by the principal orientation features in the inlet section plane (such as the keyway direction or clamping alignment axis). The actual detection position of the probe and the start and end coordinates of the defects obtained in steps S104 to S106 are described in this coordinate system, which is the unified expression framework for the internal detection results. Coordinate transformation refers to calling the 4×4 homogeneous transformation matrix T pre-solved in S101 to transform the three-dimensional coordinate points in the local coordinate system of the pipe fitting. Perform the following homogeneous linear transformation:

[0090]

[0091] in Given homogeneous column vectors, R is a 3×3 rotation matrix, and t is a 3×1 translation vector, output... The first three components are the three-dimensional coordinates of the point in the external reference coordinate system.

[0092] S106 determines the starting coordinates of the internal defect space interval. coordinates of the endpoint Subsequently, this step completes the coordinate system transition with a single matrix multiplication, transferring the internal abstract description of the defect to an externally executable marking operation framework. This is a necessary prerequisite for subsequent external surface projection mapping. Specifically, the system performs the following steps respectively: and Perform the above homogeneous coordinate transformation and output the transformed coordinates. and In the same batch of inspections, if the clamping posture of the pipe remains unchanged throughout the inspection process, the matrix solved by S101 can be directly reused in this step without recalculation. If the posture shift is caused by re-clamping, thermal deformation, or secondary positioning, the system performs rapid re-acquisition of feature points and updates the matrix before the current inspection to ensure that the transformation result always corresponds to the current physical clamping state, thus isolating the posture deviation between batches from the mark output link.

[0093] In the extended design of the coordinate link, this step realizes a single-level direct transformation from the local coordinate system of the tube fitting to the external reference coordinate system. When facing complex scenarios such as heat exchanger tube bundles or tube sheet installation structures, a tube sheet end face coordinate system can be further introduced on the basis of the external reference coordinate system to form a multi-level transformation link of local coordinate system → external reference coordinate system → tube sheet end face coordinate system. Finally, the defect coordinates are resolved into a structured output of tube sheet hole position number plus axial depth. All transformations at each level adopt the same form of 4×4 homogeneous matrix multiplication to maintain the consistency of mathematical expression. Moreover, the matrix at any level can be updated independently without affecting other levels, which has good engineering scalability and maintainability.

[0094] S108. Project the converted starting point three-dimensional coordinates and ending point three-dimensional coordinates onto the outer surface contour of the pipe fitting along the radial direction of the outer surface of the pipe fitting to generate the pipe fitting outer surface marking interval coordinates for guiding external marking.

[0095] The outer surface marking interval coordinates refer to the structured coordinate representation of the executable physical marking range on the outer wall of the pipe fitting, corresponding to the position of the internal defect, under an external reference coordinate system. Its output includes at least two formats: Format 1, a structured description composed of the three-dimensional coordinates of the starting and ending marking points on the outer wall, and the geodesic line segment connecting the two points, which can directly drive laser pointers, inkjet printers, or AR overlay rendering engines; Format 2, a triplet description composed of the circumferential angle of the outer wall, the corresponding axial position, and the axial length of the interval, suitable for rotation alignment tools and manual marking operations. Both formats include a marking tolerance band parameter along the axial direction, typically ±2 to 10 mm wide, to cover the uncertain range of coordinate mapping accuracy, providing an executable margin for external physical marking operations. The outer surface profile refers to the three-dimensional geometric description of the outer surface of the pipe fitting to be inspected, stored in the form of a three-dimensional mesh model or a parametric surface model, for subsequent geometric operations such as normal ray intersection or section projection.

[0096] After S107 completes the coordinate transformation and Now positioned in the external reference coordinate system, this step uses these two points as source points to perform a geometric mapping from the internal 3D coordinates to the visible area of ​​the outer wall, generating the final output directly serving the marking operation. Specifically, the system extracts from the spatial path model... and For the cross-sectional normal at the centerline of the corresponding pipe section, extend the rays along the normal and intersect with the three-dimensional mesh model of the outer surface to obtain two marker points on the outer wall; combine the geodesic line segment on the outer surface connecting the two marker points with the tolerance zone parameters to form the final marked interval coordinate output.

[0097] The marked interval is expressed in the form of three-dimensional coordinates plus geodesic segments, rather than just outputting a single mileage value. This is because in spatially curved pipe fittings, the same mileage increment corresponds to different external surface spatial areas and cross-sectional orientations in different curved sections. A single mileage cannot uniquely determine the circumferential orientation of the outer wall marking position. The three-dimensional coordinates plus geodesic segments can be accurately superimposed and visualized on the digital model, and can also directly drive the automated execution end, covering various operation modes from manual assistance to fully automatic execution. In dense installation scenarios such as heat exchanger tube bundles or boiler coils where the outer wall cannot be directly accessed, the system can automatically switch the output mode according to the pre-loaded assembly topology information, translating the defect coordinates into a substitute description of the tube sheet end hole position number plus axial depth, or generating robotic arm positioning instructions and AR guidance information for the maintenance execution end. This makes the practicality of the marked interval coordinates not limited by the installation configuration, and achieves adaptive coverage of various field conditions.

[0098] In some preferred embodiments, the above steps specifically include:

[0099] In the coordinate system of the external reference datum, the normal vectors of the starting point's three-dimensional coordinates and the ending point's three-dimensional coordinates are calculated based on the spatial path model; with the starting point's three-dimensional coordinates and the ending point's three-dimensional coordinates as the starting points, rays are extended outward along their respective normal vectors; the three-dimensional mesh model of the outer surface of the pipe fitting to be inspected is analyzed, and the spatial intersection points of the rays and the three-dimensional mesh model of the outer surface are calculated; the obtained spatial intersection points are used as outer surface marker points, and the geodesic line segments formed by the outer surface marker points are used as the coordinates of the outer surface marker interval of the pipe fitting.

[0100] In the Frenet coordinate system of the spatial path model, the tangent vectors of each centerline point Principal normal vector with the binormal vector To form a complete local orthogonal basis, any outward normal direction within the cross-sectional plane can be expressed by the following equation:

[0101]

[0102] For circular tubes with no need to distinguish the direction of eccentricity, take... The default outward normal is along the direction of the principal normal vector. When mapping the defect orientation in a specific quadrant, the θ value is determined based on S106 signal characteristic analysis or measured wall thickness information. For pipe fittings with spatial torsion (such as spiral coils), the path model must accurately record the cumulative torsion along the path. This ensures that the direction of the secondary normal vector is updated correctly with the torque, and avoids the accumulation of cross-sectional attitude errors being passed on to the calculation of the outer normal.

[0103] After the coordinate transformation is complete, this step performs geometric ray intersection calculations from the internal coordinates to the external wall physical coordinates in the external reference coordinate system. Specifically, the system reads... and For the Frenet coordinate frame corresponding to the centerline position, determine the outward normal unit vector according to the above formula. and Each ray extends outward along its corresponding normal direction from its starting point. The ray-patch intersection algorithm of the three-dimensional mesh on the outer surface (typically implemented as the Möller–Trumbore algorithm) is called to obtain the spatial intersection points of the two rays with the outer surface, namely the starting and ending marker points of the outer wall. Then, the geodesic segments on the outer surface are calculated using these two points as endpoints. Finally, the coordinates of the marked interval are output by combining the parameters with the tolerance zone parameters.

[0104] The radial normal of the cross-section is used as the ray direction, rather than the shortest distance projection or the nearest point projection. This is because for pipe sections with wall eccentricity or local curvature changes, the direction of the shortest distance projection will drift due to local geometric inhomogeneities on the outer surface, causing the correspondence between the internal and external coordinates to deviate from the physical radial direction of the cross-section. The radial normal of the cross-section, geometrically, strictly maintains the direct alignment between the "cross-section where the internal defect is located" and the "cross-section where the external wall marker point is located," ensuring that the external wall marker point accurately corresponds to the actual location of the internal defect in the circumferential direction, providing a geometrically rigorous guarantee for precise orientation marking. In other embodiments, the external surface mapping can also be achieved by intersecting the cross-section normal plane and expanding along the geodesic of the outer surface, thus covering the technical path of using cross-section projection without relying on the normal ray. The common logic of both implementation methods is to constrain the correspondence between internal and external coordinates based on the physical cross-section relationship, rather than relying on a distance minimization criterion without direction constraints.

[0105] In other embodiments of this application, prior to the step of calculating the spatial intersection of the ray and the three-dimensional mesh model of the outer surface, the method further includes:

[0106] Obtain the actual wall thickness distribution data of the pipe segment corresponding to the starting point three-dimensional coordinates and the ending point three-dimensional coordinates of the pipe fitting to be inspected; determine whether there are eccentric thickening or thinning characteristics in the actual wall thickness distribution data; if so, perform vector deflection correction on the extension angle of the normal vector according to the eccentric direction and eccentricity corresponding to the eccentric thickening or thinning characteristics to compensate for the radial projection refraction error caused by the uneven pipe wall thickness, and then use the corrected normal vector extension ray to calculate the spatial intersection point.

[0107] Eccentricity (The difference between the maximum and minimum wall thickness, in mm) is quantitatively characterized; the typical threshold for judgment is set as follows. Exceeding the nominal wall thickness A significant eccentricity is identified when the deviation is between 5% and 20%, triggering vector deflection correction. Vector deflection correction refers to... and Apply an angle deflection to the initial normal vector This ensures the corrected ray direction points towards the actual geometric centroid of the outer wall, compensating for radial projection refraction errors caused by the non-uniform pipe wall; the reference calculation relationship for the deflection angle is...

[0108]

[0109] The corrected ray direction is rotated from the original normal vector around the tangent direction of the cross section orthogonal to the eccentric direction. Angle obtained.

[0110] Finding the intersection of the ray and the 3D mesh model of the outer surface is the core geometric operation for outer surface projection. When the pipe wall has a non-uniform wall thickness, using the nominal section normal as the ray direction will cause the ray system to be systematically biased towards the thinner wall side, resulting in the outer wall markers deviating circumferentially from directly above the actual defect. Therefore, this step first determines whether there are significant eccentric features before performing the intersection calculation to decide whether to initiate vector deflection correction. Specifically, the system reads... and Based on the actual wall thickness distribution data of the corresponding pipe section, calculate the difference between the maximum and minimum wall thickness along the circumferential direction of the cross-section. ;like If the deviation exceeds a preset threshold (typically 5% to 20% of the nominal wall thickness), then the eccentricity direction angle is extracted. Calculate the deflection angle using the above formula, along with the eccentricity. Perform a corresponding rotation on the initial normal vector to obtain the corrected ray direction vector, and then use the corrected vector to enter the intersection process; if If the threshold is not exceeded, the ray is extended directly in the nominal normal direction to maintain compatibility with the conventional normal projection path.

[0111] The design of wall thickness eccentricity correction as a separate pre-step for ray intersection, rather than incorporating it into the lift-off compensation model of S106, stems from the following technical considerations: Lift-off compensation addresses the geometric influence in the signal amplitude domain, with its correction amount acting on the eddy current signal value in the form of a gain coefficient; wall thickness eccentricity correction addresses the projection direction error in the spatial geometric domain, with its correction amount acting on the ray orientation in the form of an angular deflection. Both are closed at independent data stream nodes, allowing for separate verification, closure, or replacement, making the overall error source clear and traceable, and facilitating precise location and quantification of the independent contribution of each error during engineering debugging. Furthermore, after outputting the outer wall marking coordinates, the system can automatically generate a marking confidence level based on the magnitude of the wall thickness distribution deviation and the assembly obstruction situation. When the deviation is significant or the obstruction risk is high, a recommended verification strategy is provided, guiding operators to perform secondary confirmation on abnormal pipe sections. This incorporates the residual uncertainty caused by uneven wall thickness into the human-machine collaborative management process, rather than relying entirely on the algorithm to handle it, thus improving automation while retaining necessary manual judgment interfaces.

[0112] The following describes the precise marking and positioning device for eddy current detection inside pipes in the embodiments of this invention from the perspective of hardware processing. Please refer to [link / reference needed]. Figure 2 This is a schematic diagram of a physical device structure of a pipe fitting eddy current detection and precise marking positioning device in the embodiments of this application.

[0113] It should be noted that, Figure 2 The structure of the eddy current detection and precise marking positioning device inside the pipe shown is merely an example and should not impose any limitation on the functions and scope of use of the embodiments of the present invention.

[0114] like Figure 2 As shown, the eddy current detection and precise marking positioning device for pipe fittings includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 202 or a program loaded from storage section 208 into random access memory (RAM) 203, such as performing the methods described in the above embodiments. The RAM 203 also stores various programs and data required for system operation. The CPU 201, ROM 202, and RAM 203 are interconnected via bus 204. An input / output (I / O) interface 205 is also connected to bus 204.

[0115] The following components are connected to I / O interface 205: input section 206 including audio input devices, push-button switches, etc.; output section 207 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 208 including a hard disk, etc.; and communication section 209 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 209 performs communication processing via a network such as the Internet. Drive 210 is also connected to I / O interface 205 as needed. Removable media 211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 210 as needed so that computer programs read from them can be installed into storage section 208 as needed.

[0116] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 209, and / or installed from removable medium 211. When the computer program is executed by central processing unit (CPU) 201, it performs the various functions defined in the present invention.

[0117] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0118] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0119] Specifically, the pipe fitting eddy current detection and precise marking positioning device of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements the pipe fitting eddy current detection and precise marking positioning method provided in the above embodiment.

[0120] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the pipe fitting eddy current detection precise marking and positioning device described in the above embodiments; or it may exist independently and not assembled into the pipe fitting eddy current detection precise marking and positioning device. The storage medium carries one or more computer programs, which, when executed by a processor of the pipe fitting eddy current detection precise marking and positioning device, cause the pipe fitting eddy current detection precise marking and positioning device to implement the pipe fitting eddy current detection precise marking and positioning method provided in the above embodiments.

[0121] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0122] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0123] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for precise marking and positioning of eddy current detection inside pipe fittings, characterized in that, The method includes: Extract the coordinate feature points of the inlet datum and the external reference datum of the pipe fitting to be inspected in order to calculate the coordinate transformation matrix; The geometric data of the pipe fitting to be inspected is analyzed to generate a spatial path model containing three-dimensional spatial coordinates and curvature features; As the detection probe is advanced or retracted inside the tube under test via the flexible transmission component, the change in travel arc length, the bending morphology parameters of the flexible transmission component, and the amplitude of the eddy current detection signal picked up by the detection probe are collected simultaneously. Using the spatial path model as a trajectory set and the change in travel arc length as an initial arc length reference, the curvature feature sequence formed during the advance or retraction process in the bending morphology parameters is extracted. The curvature feature sequence is matched with the curvature features of each curved segment in the spatial path model using a sliding window to compensate for the deviation between the arc length and the actual position caused by the deformation of the flexible transmission component. The three-dimensional coordinate points determined by the matching are used as the actual detection positions of the detection probe. Based on the amplitude change of the eddy current detection signal and the actual detection location, the starting three-dimensional coordinates and the ending three-dimensional coordinates of the internal defect space interval are determined. The coordinate transformation matrix is ​​invoked to transform the three-dimensional coordinates of the starting point and the three-dimensional coordinates of the ending point of the internal defect space interval to the coordinate system of the external reference datum. The converted three-dimensional coordinates of the starting point and the three-dimensional coordinates of the ending point are projected onto the outer surface contour of the pipe fitting along the radial direction of the outer surface of the pipe fitting to generate the outer surface marking interval coordinates of the pipe fitting for guiding external marking.

2. The method according to claim 1, characterized in that, The step of performing sliding window matching between the curvature feature sequence and the curvature features of each curved segment in the spatial path model to compensate for the deviation between the arc length and the actual position caused by the deformation of the flexible transmission component, and using the matched three-dimensional coordinate points as the actual detection positions of the detection probe, specifically includes: Construct a sliding window with the change in the travel arc length as the independent variable, and extract the local curvature feature subsequence to be matched within the sliding window; A time warping algorithm is introduced to calculate the cumulative distance matrix between the local curvature feature subsequence and the benchmark curvature sequence within the corresponding search interval in the spatial path model; The optimal regular path is found based on the cumulative distance matrix to eliminate the nonlinear spatial scale scaling caused by the buckling or elongation of the flexible transmission component during the advancement or retraction process. Based on the mapping relationship of the optimal regularized path, the feature points in the local curvature feature subsequence are aligned to the corresponding points of the reference curvature sequence, and the three-dimensional coordinate point of the aligned reference curvature sequence is used as the actual detection position of the detection probe.

3. The method according to claim 2, characterized in that, Prior to the step of constructing a sliding window with the change in travel arc length as the independent variable, the method further includes: Obtain the axial force parameters and travel speed of the flexible transmission component in its current traveling state; The axial force parameters are used to determine whether the flexible transmission component is in a state of compressive buckling or tensile elongation, and the deformation compensation coefficient is calculated in conjunction with the travel speed. The window length and sliding step size of the sliding window are dynamically and adaptively adjusted using the deformation compensation coefficient. When the window is in a compressive buckling state, the window length is increased, and when the window is in a tensile elongation state, the window length is decreased.

4. The method according to claim 1, characterized in that, The step of determining the starting and ending three-dimensional coordinates of the internal defect space interval based on the amplitude change of the eddy current detection signal and the actual detection position specifically includes: Extract the local spatial curvature corresponding to the actual detection location in the spatial path model; Establish a correlation mapping model between the lift-off distance of the detection probe and the local spatial curvature, and calculate the lift-off compensation gain coefficient at the current actual detection position based on the correlation mapping model; The amplitude of the synchronously acquired eddy current detection signal is dynamically amplified or attenuated using the lift-off compensation gain coefficient to eliminate signal baseline drift caused by probe eccentricity in the bent section of the pipe fitting. The envelope of the compensated eddy current detection signal is extracted, and the defect feature band is truncated based on a preset adaptive signal-to-noise ratio threshold. The actual detection positions corresponding to the starting point of the rising edge and the ending point of the falling edge of the defect feature band are respectively used as the starting three-dimensional coordinates and the ending three-dimensional coordinates of the internal defect space interval.

5. The method according to claim 4, characterized in that, The establishment of the correlation mapping model between the detection probe lift-off distance and the local spatial curvature specifically includes: Obtain the outer diameter of the detection probe and the inner diameter of the pipe fitting to be inspected; The detection probe is simplified as a rigid cylinder, and the geometric interference simulation between the rigid cylinder and the inner wall of the pipe to be inspected is performed at each local spatial curvature of the spatial path model. Based on the radial offset vector between the central axis of the rigid cylinder and the center line of the spatial path model obtained from the simulation, the theoretical lift-off distance between the surface coil of the detection probe and the inner wall of the pipe is calculated to generate the correlation mapping model.

6. The method according to claim 1, characterized in that, The process of projecting the converted starting point three-dimensional coordinates and ending point three-dimensional coordinates radially along the outer surface of the pipe fitting to the outer surface contour of the pipe fitting, generating pipe fitting outer surface marking interval coordinates for guiding external marking, specifically includes: In the coordinate system of the external reference datum, the normal vectors of the positions of the starting point's three-dimensional coordinates and the ending point's three-dimensional coordinates are calculated based on the spatial path model; Starting from the three-dimensional coordinates of the starting point and the three-dimensional coordinates of the ending point, rays are extended outward along their respective normal vectors. Analyze the three-dimensional mesh model of the outer surface of the pipe fitting to be inspected, and calculate the spatial intersection point between the ray and the three-dimensional mesh model of the outer surface. The obtained spatial intersection points are used as outer surface marker points, and the outer surface geodesic line segment formed by each outer surface marker point is used as the coordinate of the outer surface marker interval of the pipe fitting.

7. The method according to claim 6, characterized in that, Prior to the step of calculating the spatial intersection point of the ray and the three-dimensional mesh model of the outer surface, the method further includes: Obtain the actual wall thickness distribution data of the pipe segment corresponding to the three-dimensional coordinates of the starting point and the three-dimensional coordinates of the ending point of the pipe fitting to be inspected; Determine whether the actual wall thickness distribution data exhibits eccentric thickening or thinning characteristics; If present, the extension angle of the normal vector is vector deflected and corrected according to the eccentric direction and eccentricity corresponding to the eccentric thickening or thinning feature to compensate for the radial projection refraction error caused by uneven pipe wall thickness. Then, the corrected normal vector is used to extend the ray to calculate the spatial intersection point.

8. A precise marking and positioning device for eddy current detection inside pipe fittings, characterized in that, The precise marking and positioning device for eddy current detection inside pipe fittings includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the precise marking and positioning device for eddy current detection inside pipe fittings to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the eddy current detection precision marking and positioning device inside the pipe fitting, the eddy current detection precision marking and positioning device inside the pipe fitting performs the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the pipe fitting eddy current detection precision marking and positioning device, the pipe fitting eddy current detection precision marking and positioning device performs the method as described in any one of claims 1-7.

Images