Parallel digital signal synchronization implementation method for long-distance pipeline magnetic flux leakage internal detection

Abstract

This invention discloses a method for parallel digital signal synchronization in the internal detection of magnetic flux leakage in long-distance pipelines, relating to the field of pipeline nondestructive testing technology, and is used to solve the problem of multi-channel synchronization distortion. This invention achieves synchronous organization and collaborative interpretation of multi-channel magnetic flux leakage signals, reference measurement signals, and operating status signals in the internal detection of magnetic flux leakage in long-distance pipelines by constructing a processing chain that includes a unified measurement frame sequence, candidate event segment extraction, cross-channel mapping, state compensation, and defect response distribution reconstruction. This enables the differentiation between real defect responses and false anomalies caused by weld disturbances, lift-off changes, velocity fluctuations, and local magnetization anomalies, reducing misjudgments and missed judgments, and improving the stability and reliability of defect boundary positioning, circumferential coverage determination, axial extension identification, and detection results under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of pipeline nondestructive testing technology, and more specifically, to a method for synchronous implementation of parallel digital signals in the internal detection of magnetic flux leakage in long-distance pipelines. Background Technology

[0002] In oil and gas transportation, refined oil transportation, and chemical media transportation, long-distance pipeline defect detection methods based on internal magnetic flux leakage (MFL) detection have been widely used. These methods typically involve an internal detection device running along the pipeline's interior to continuously magnetize the pipe wall. Multiple circumferentially deployed detection channels simultaneously collect axial and radial MFL signals, as well as operational status information such as mileage, speed, and attitude, to identify and locate corrosion thinning, localized metal loss, and related abnormal defects. Existing technologies focus on improving MFL signal acquisition density, detection coverage, and defect identification accuracy. They organize multi-source detection data through parallel acquisition links, digital processing units, and mileage correlation mechanisms to achieve continuous signal recording and defect interpretation during long-distance pipeline inspection.

[0003] In parallel digital signal synchronous processing, there is an inherent deviation between the discrete sampling process of multi-channel magnetic flux leakage signals and the physical process of the internal detection device running continuously in the pipeline. When the detection device passes through welds, bends, bottom liquid accumulation areas, and local corrosion areas, it is often accompanied by speed fluctuations, probe lift-off changes, and local magnetization state disturbances. The response time, peak-valley boundaries, and amplitude patterns of different detection channels are prone to deviation. Existing solutions mostly rely on unified clocks, fixed triggers, or post-processing splicing methods for alignment, which makes it difficult to ensure consistent correspondence of multi-channel signals in the time and spatial dimensions. This synchronization inconsistency makes it difficult to form a continuous, unified, and stable defect response distribution in different detection channels for the same defect, thereby affecting the accuracy of defect location determination, circumferential range identification, and axial extension analysis, and weakening the reliability of magnetic flux leakage internal detection results in long-distance pipelines. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the following solution is proposed to solve the problem of multi-channel synchronization distortion in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for synchronizing parallel digital signals in the internal detection of magnetic flux leakage in long-distance pipelines includes the following steps:

[0007] During the internal detection process of magnetic flux leakage, multi-channel magnetic flux leakage signals, reference measurement signals, and operating status signals are collected to generate a unified measurement frame sequence.

[0008] Candidate event segments representing the defect passage process are extracted based on a unified measurement frame sequence;

[0009] Perform cross-channel mapping based on the temporal relationship, spatial adjacency relationship, and response correspondence relationship of candidate event fragments to generate a cluster of synchronous events;

[0010] By combining reference measurement signals and operating status signals, state compensation is performed on the synchronization event cluster to generate a compensation synchronization signal sequence;

[0011] The defect response distribution is reconstructed based on the compensated synchronization signal sequence, and the leakage magnetic field detection results are output.

[0012] Furthermore, the multi-channel leakage magnetic signals include axial leakage magnetic signals and radial leakage magnetic signals distributed along the circumference, the reference measurement signals include lift-off characterization signals and magnetization state characterization signals, and the operating status signals include mileage signals, speed signals, and attitude signals.

[0013] Furthermore, generating a unified measurement frame sequence includes:

[0014] The frame index is determined based on the operating status signal. The multi-channel leakage magnetic signal, reference measurement signal and operating status signal under the same frame index are written into the same measurement frame. The channel identifier, component identifier and status identifier are associated in the measurement frame to form a unified measurement frame sequence arranged continuously according to the frame index.

[0015] Furthermore, the extraction of candidate event fragments representing the defect passage process includes:

[0016] Baseline recovery and trend segmentation are performed on the multi-channel leakage magnetic signals in the unified measurement frame sequence to determine the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary, and candidate event segments of the corresponding defect passage process are formed based on the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary.

[0017] Furthermore, performing cross-channel mapping based on the temporal relationships of candidate event fragments includes:

[0018] Candidate event segments in each channel are sequentially sorted according to the frame index arrangement order and mileage advancement direction in the unified measurement frame sequence.

[0019] For any candidate event segment, search for candidate event segments in other channels within the corresponding frame index interval and mileage continuity interval, determine the segment combination with consistent event sequence and continuous mileage change direction, and generate a candidate mapping set.

[0020] Furthermore, performing cross-channel mapping based on the spatial adjacency of candidate event fragments includes:

[0021] Based on the circumferential distribution order corresponding to the leakage magnetic field signal, a channel adjacency table is established. According to the circumferential adjacency continuity between channels corresponding to each candidate event segment in the candidate mapping set and the propagation continuity along the pipeline running direction, an adjacency consistency check is performed on the candidate mapping set. Segment combinations that are discretely distributed in the circumferential direction, cross non-adjacent channels, or whose propagation order is inconsistent with the circumferential adjacency relationship are eliminated, and an adjacency mapping set is generated.

[0022] Furthermore, performing cross-channel mapping and generating synchronization event clusters based on the response correspondence of candidate event fragments includes:

[0023] Compare the corresponding order of the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary of each candidate event segment in the adjacency mapping set, and verify the correspondence between the axial leakage magnetic signal and the radial leakage magnetic signal in different channels in terms of polarity change, peak-valley continuity and continuous interval overlap.

[0024] Candidate event fragments that satisfy the response correspondence are grouped into synchronous event clusters, and a unified event identifier and channel association information are associated with the synchronous event clusters.

[0025] Furthermore, performing state compensation on the synchronization event cluster by combining reference measurement signals and operating status signals includes:

[0026] Based on the trends of reference measurement signal changes, speed signal changes, mileage advancement continuity, and attitude change continuity within the frame index interval corresponding to the synchronization event cluster, identify the speed change state, lift-off change state, and local magnetization disturbance state.

[0027] Based on the correspondence between velocity change state, lift-off change state, and local magnetization disturbance state and synchronization event cluster, compensation state identifier and compensation interval identifier are associated with synchronization event cluster.

[0028] Furthermore, state compensation is performed on the synchronization event cluster to generate a compensated synchronization signal sequence, including:

[0029] The corresponding compensation rules are invoked for each cluster of synchronous events associated with different compensation status identifiers;

[0030] Under speed change conditions, position corrections are performed on the abnormal entry boundary, peak-valley transition boundary, and abnormal exit boundary based on the mileage advance offset between adjacent frame indices.

[0031] Under the lift-off change state, the leakage magnetic response amplitude in the synchronous event cluster is corrected according to the change range of the reference measurement signal;

[0032] Under local magnetization disturbance, boundary pruning, missing completion or anomaly removal are performed on the abnormal response segment according to the continuity relationship of the synchronous event cluster in adjacent channels.

[0033] After completing position correction, amplitude correction, and boundary correction, the synchronization event clusters are written back to the unified measurement frame sequence according to the unified event identifier to generate the compensation synchronization signal sequence.

[0034] Furthermore, the reconstruction of the defect response distribution based on the compensated synchronization signal sequence and the output of the leakage magnetic field detection results include:

[0035] The synchronization event clusters in the compensation synchronization signal sequence are expanded along the circumferential distribution direction and the pipeline extension direction according to the unified event identifier to construct the defect response distribution of the corresponding defect passage process;

[0036] The defect location, circumferential range, and axial extension characteristics are determined based on the boundary turning points, peak and valley continuation trajectories, and channel coverage in the defect response distribution. The defect level is generated based on the abnormal persistence morphology and component response relationship of the defect response distribution, and the leakage magnetic field internal detection results are output.

[0037] The technical effects and advantages of the parallel digital signal synchronization method for internal detection of magnetic flux leakage in long-distance pipelines according to the present invention are as follows:

[0038] This invention achieves synchronous organization and collaborative interpretation of multi-channel magnetic flux leakage signals, reference measurement signals, and operating status signals during the internal detection of magnetic flux leakage in long-distance pipelines by constructing a closed-loop processing chain that includes a unified measurement frame sequence, candidate event fragment extraction, cross-channel mapping, state compensation, and defect response distribution reconstruction. It can effectively distinguish between real defect responses and false anomalies caused by weld disturbances, lift-off changes, velocity fluctuations, and local magnetization anomalies, significantly reducing misjudgments and missed judgments.

[0039] Based on this, a cluster of synchronous events is generated according to the temporal relationship, spatial adjacency relationship and response correspondence. Combined with the lift-off characterization signal, magnetization state characterization signal, mileage signal, speed signal and attitude signal to perform state compensation, making the defect boundary location, circumferential coverage determination and axial extension identification more accurate and stable. This improves the synchronous consistency of leakage magnetic field detection results, the reliability of defect identification and the engineering applicability under complex working conditions. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating the parallel digital signal synchronization method for internal detection of leakage magnetic flux in long-distance pipelines according to the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] In order to achieve the above objectives, Figure 1 A schematic diagram of the parallel digital signal synchronization implementation method for internal leakage magnetic field detection in long-distance pipelines according to the present invention is given, which specifically includes the following steps;

[0043] During the internal detection process of magnetic flux leakage, multi-channel magnetic flux leakage signals, reference measurement signals, and operating status signals are collected to generate a unified measurement frame sequence.

[0044] Candidate event segments representing the defect passage process are extracted based on a unified measurement frame sequence;

[0045] Perform cross-channel mapping based on the temporal relationship, spatial adjacency relationship, and response correspondence relationship of candidate event fragments to generate a cluster of synchronous events;

[0046] By combining reference measurement signals and operating status signals, state compensation is performed on the synchronization event cluster to generate a compensation synchronization signal sequence;

[0047] The defect response distribution is reconstructed based on the compensated synchronization signal sequence, and the leakage magnetic field detection results are output.

[0048] Step 1: Acquire multi-channel leakage magnetic field signals, reference measurement signals, and operating status signals during the leakage magnetic field detection process to generate a unified measurement frame sequence. The specific implementation is as follows:

[0049] In this embodiment, when the long-distance pipeline magnetic flux leakage detection device travels along the pipeline axis, the magnetization component forms a continuous magnetization state on the pipe wall of the current detection section. Multiple detection channels are uniformly arranged circumferentially behind the magnetized area. Each detection channel corresponds to a unique channel identifier and synchronously outputs axial magnetic flux leakage signals and radial magnetic flux leakage signals. The axial magnetic flux leakage signal is used to characterize the magnetic field disturbance change in the defect area along the pipeline extension direction, and the radial magnetic flux leakage signal is used to characterize the magnetic flux leakage change in the defect area perpendicular to the pipe wall direction.

[0050] Simultaneously acquire lift-off characterization signals and magnetization state characterization signals at corresponding positions in each detection channel or near the magnetization component, so that the subsequent synchronization process can identify measurement offsets caused by changes in the relative distance between the probe and the tube wall and local magnetization instability.

[0051] The lift-off characterization signal is used to reflect the increase, decrease and continuity of the distance between the detection channel and the inner wall of the pipeline. The magnetization characterization signal is used to reflect the changes in magnetization strength and continuity in the current detection section. At the same time, the internal detection device also acquires mileage signal, speed signal and attitude signal. The mileage signal is used to indicate the advancing position of the internal detection device along the pipeline extension direction. The speed signal is used to indicate the changes in the running speed of the internal detection device in adjacent advancing sections. The attitude signal is used to indicate the changes in the pitch, yaw or roll state of the internal detection device in the current detection section.

[0052] After the aforementioned axial leakage magnetic field signal, radial leakage magnetic field signal, lift-off characterization signal, magnetization state characterization signal, odometer signal, velocity signal, and attitude signal enter their respective data acquisition links, they undergo signal conditioning, digital sampling, and time-sequence buffering in sequence to form raw digital sampling results that can be used for unified framing. The velocity signal can be directly output by an independent velocity measurement unit or calculated based on the odometer signal changes and sampling time intervals at adjacent sampling times according to the time sequence. However, regardless of the acquisition method used, it is uniformly used as a velocity signal in subsequent processing.

[0053] When the speed signal is converted from the mileage signal, the speed signal value corresponding to the current sampling time can be expressed as: ,in, This represents the velocity signal value corresponding to the current sampling time. This represents the mileage signal value corresponding to the current sampling time. This represents the mileage signal value corresponding to the previous sampling time. Indicates the current sampling time. This indicates the previous sampling time.

[0054] When generating a unified measurement frame sequence, this embodiment uses the odometer signal as the primary basis for determining the frame index, and the velocity signal and attitude signal as auxiliary basis for frame index verification and state marking, specifically including:

[0055] First, the mileage signals are divided into continuous advancement intervals according to their increasing order. Each continuous advancement interval corresponds to a frame index. The continuous advancement interval is determined by the consistency of the advancement direction and the continuity of the advancement rhythm between two adjacent valid mileage signals. When two adjacent valid mileage signals change along the same advancement direction and there is no reverse jump, long-term stagnation, or interruption exceeding the preset buffer waiting time, the mileage segment is determined as the same continuous advancement interval. When there is mileage rollback, mileage remains unchanged and continues to exceed the preset stagnation confirmation time, or the buffer interval between the current sampling time and the previous valid sampling time exceeds the preset framing waiting time, the previous continuous advancement interval is terminated and a new continuous advancement interval is generated.

[0056] The preset pause confirmation duration is defined as the time length used to distinguish between short-term jitter pauses and true pause states, and is preset as an integer multiple of the device's normal sampling cycle;

[0057] The preset framing wait time is defined as the maximum time allowed for the reference measurement signal and the running status signal to arrive together at the current frame index.

[0058] When the odometer signal enters a new propulsion zone, a new frame index is generated and assigned to the axial leakage magnetic signal, radial leakage magnetic signal, lift-off characterization signal, magnetization state characterization signal, odometer signal, velocity signal and attitude signal collected in the current propulsion zone.

[0059] If the speed signal display device significantly accelerates, decelerates, or stops within the current propulsion interval, the frame index order determined by the mileage signal remains unchanged. Only the status flag reflecting the speed change is written into the data corresponding to that frame index, so as to avoid splitting the same propulsion interval into multiple frame indices simply because of changes in the sampling cycle.

[0060] If the attitude signal display device experiences a sudden attitude change, the frame index determined by the odometer signal will not be changed. Instead, the attitude change result will be written into the status identifier for use in status compensation to identify the operating state corresponding to the current frame index. The status identifier includes at least the odometer status identifier, the speed status identifier, and the attitude status identifier. The odometer status identifier is used to characterize the propulsion position range corresponding to the current frame index, the speed status identifier is used to characterize the change trend of the speed signal in the current frame index relative to the previous frame index, and the attitude status identifier is used to characterize the change of the attitude signal in the current frame index.

[0061] Mileage status identifiers are generated based on the start and end points of the continuous advancement interval corresponding to the current frame index, and are used to record the mileage segment to which the frame index belongs. Speed ​​status identifiers are generated based on the direction of change of the speed signal in the current frame index relative to the speed signal in the previous frame index. When the speed continues to increase, it is marked as an acceleration state; when the speed continues to decrease, it is marked as a deceleration state; when the speed remains stable within a preset fluctuation tolerance range, it is marked as a stable state; when the speed drops to near zero and exceeds a preset pause confirmation time, it is marked as a stagnant state. Attitude status identifiers are generated based on the direction and magnitude of change of the attitude signal in the current frame index relative to the attitude signal in the previous frame index. When the attitude change is within a preset stable range, it is marked as a stable attitude state; when any component of pitch, yaw, or roll continuously crosses a preset attitude change limit, it is marked as an abrupt attitude change state.

[0062] Furthermore, the validity flag is written when the sampled value is digitized and within the allowable range of the acquisition link; the missing flag is written when the corresponding detection channel under the same frame index does not obtain a valid sampled value; and the continuation flag is written when the reference measurement signal or running status signal under the same frame index does not arrive in time and is filled with the most recent valid signal, and the continuation source frame index is recorded at the same time.

[0063] Through the above processing, each frame index corresponds to a unique propulsion position basis, and the velocity state and attitude state corresponding to that propulsion position are retained synchronously, thereby ensuring that the extraction of subsequent candidate event segments is not based on isolated sampling points, but on unified data units with propulsion position meaning and running state meaning.

[0064] After the same frame index is determined, the axial leakage magnetic field signal, radial leakage magnetic field signal, lift-off characterization signal, magnetization state characterization signal, and corresponding odometer, velocity, and attitude signals output by each detection channel under the same frame index are written into the same measurement frame to form a unified measurement frame. These unified measurement frames are then arranged sequentially according to their frame indices to form a unified measurement frame sequence. Each unified measurement frame is equipped with a channel identifier, component identifier, and status identifier, specifically including:

[0065] Channel identifiers are used to distinguish detection channels at different circumferential positions. Component identifiers are used to distinguish axial leakage magnetic signals, radial leakage magnetic signals, lift-off characterization signals, and magnetization state characterization signals. Status identifiers are used to record the mileage status identifier, velocity status identifier, and attitude status identifier corresponding to the frame index. In addition to the above identifiers, the digital sample value of the corresponding signal and its validity flag are also written. For detection channels that have not obtained valid sample values ​​under the same frame index, the position of the detection channel in the unified measurement frame is not deleted, but a missing flag is written at the corresponding position to maintain the consistency of channel arrangement between different frame indices.

[0066] If the reference measurement signal or operating status signal under the same frame index arrives in the buffer area later than the leakage magnetic signal, the current frame index is filled with the most recent valid lift-off characterization signal, magnetization status characterization signal, mileage signal, speed signal or attitude signal from the previous moment, and a continuation mark is written in the status identifier to ensure that all types of signals under the current frame index still maintain the same frame association relationship.

[0067] The unified measurement frame sequence formed by the above writing rules contains, at each frame index, axial leakage magnetic signal, radial leakage magnetic signal, lift-off characterization signal, magnetization state characterization signal, odometer signal, velocity signal, and attitude signal, as well as their corresponding channel identifier, component identifier, and status identifier.

[0068] Step two involves extracting candidate event segments representing the defect passage process based on the unified measurement frame sequence. The specific implementation is as follows:

[0069] In this embodiment, based on the unified measurement frame sequence formed in the aforementioned embodiments, candidate event segments characterizing the defect passage process are extracted for each detection channel. In specific processing, the axial and radial magnetic flux leakage signals of the same detection channel are read in the order of the frame index using the channel identifier and component identifier as indexes to form a continuous magnetic flux leakage sequence of the current detection channel.

[0070] During the reading process, only the sampled values ​​with normal validity markers and not occupied by missing markers are included in the continuous magnetic flux leakage sequence of the current detection channel. At the same time, the mileage signal, velocity signal, attitude signal, lift-off characterization signal and magnetization state characterization signal corresponding to the continuous magnetic flux leakage sequence are retained. Since defects in magnetic flux leakage detection will manifest as different peak values, valley values ​​or peak-valley combinations on different magnetic flux leakage components, and lift-off changes will weaken the magnetic flux leakage response, and high-speed operation will cause waveform drag and distortion, this embodiment first performs baseline restoration and then performs trend segmentation to avoid misjudging background drift, lift-off attenuation or velocity disturbance as defect events.

[0071] It should be noted that when adjacent frame indices in a continuous magnetic leakage sequence remain continuous and the corresponding mileage signal monotonically advances, this segment is defined as the same continuous processing segment. When a missing marker, frame index jump, mileage signal rollback, or attitude change occurs and continuously crosses the preset boundary confirmation length, the current continuous processing segment ends and a new continuous processing segment begins. The preset boundary confirmation length is defined as the minimum number of consecutive frames used to confirm the abnormal entry boundary, peak-valley transition boundary, and abnormal exit boundary. The background fluctuation range is jointly determined by the upper and lower boundaries of the natural fluctuation of the sampled values ​​of adjacent frames in the background sample segment. Only when the current sampled value continuously crosses the upper or lower boundary and continuously reaches the preset boundary confirmation length is it considered a valid boundary crossing. Subsequent consecutive frames uniformly refer to the frame segment that continuously reaches the preset boundary confirmation length.

[0072] During baseline recovery, the continuous magnetic flux leakage sequence of the current detection channel is divided into continuous processing segments according to the frame index order. Within each continuous processing segment, frames with smooth changes in the lift-off characterization signal, continuous magnetization state characterization signal, smooth changes in velocity signal, and no obvious abrupt changes in axial and radial magnetic flux leakage signals are selected as baseline candidate frames.

[0073] The sampled values ​​corresponding to the baseline candidate frames are arranged according to their numerical values. The middle and continuous segment is taken as the baseline reference segment. The median of the baseline reference segment is then used as the baseline reference value for the continuous processing segment. The sampled values ​​of each frame in the current continuous processing segment are offset and corrected relative to the baseline reference value to obtain the axial and radial magnetic flux leakage signals after baseline recovery. This process separates the defect from the magnetic flux leakage response through the slow drift background before and after the defect. This allows the boundary judgment to target the local anomaly caused by the defect, rather than the overall rise, overall fall, or slow drift caused by magnetization instability or lift-off changes. The quantitative analysis of magnetic flux leakage detection depends on the identification and processing of the magnetic flux leakage signal. Different components will show different peak and valley shapes. Moreover, the increase in lift-off distance will reduce the magnetic flux leakage signal. High-speed operation will cause motion-induced effects and increase the difficulty of identification. Therefore, baseline recovery is performed first, and signal processing is carried out in combination with the constraints of the operating state.

[0074] Specifically, the axial and radial magnetic flux leakage signal values ​​after baseline recovery are expressed as follows: , ,in, This represents the axial leakage magnetic field value after baseline recovery corresponding to the current frame index. This represents the original sampled value of the axial leakage magnetic field signal corresponding to the current frame index. This indicates the axial baseline reference value corresponding to the currently continuously processed segment; This represents the radial leakage magnetic field value after baseline recovery corresponding to the current frame index. This represents the original sampled value of the radial leakage magnetic field signal corresponding to the current frame index. This indicates the radial baseline reference value corresponding to the current continuous processing segment.

[0075] After baseline restoration is completed, trend segmentation is performed on the axial and radial magnetic flux leakage signals of the current detection channel to determine the anomaly entry boundary, peak-valley transition boundary, and anomaly exit boundary, specifically including:

[0076] The changes in sampled values ​​after baseline recovery are examined frame by frame along the increasing direction of the frame index to first determine the background fluctuation range. The background fluctuation range is formed by selecting the offset-corrected sampled values ​​corresponding to the baseline candidate frames to form a background sample segment, and then determined according to the natural fluctuation amplitude between adjacent frames in the background sample segment.

[0077] If, at the beginning of a certain frame, the axial or radial magnetic flux leakage signal continuously exceeds the range of background fluctuations, and for several consecutive frames thereafter maintains a trend of increasing or decreasing in the same direction, while the corresponding mileage signal continues to advance continuously, then that frame is recorded as an abnormal entry into the boundary.

[0078] If, after entering the abnormal boundary, the sampled value changes from continuously increasing to continuously decreasing, or from continuously decreasing to continuously increasing, and this reversal continues in several subsequent consecutive frames, then the index of the first frame where the reversal occurs is recorded as the peak-valley transition boundary. If, after entering the peak-valley transition boundary, the axial and radial magnetic flux leakage signals return to the background fluctuation range and no longer exhibit the same type of boundary-crossing change in several subsequent consecutive frames, then the index of the first frame that returns to the background fluctuation range and remains stable is recorded as the abnormal exit boundary.

[0079] To avoid false detections caused by weld disturbances, sudden attitude changes, or transient noise, this embodiment requires that the determination of abnormal entry boundaries, peak-valley transition boundaries, and abnormal exit boundaries must simultaneously reference velocity signals, attitude signals, lift-off characterization signals, and magnetization state characterization signals.

[0080] When the velocity signal changes drastically but the leakage magnetic signal only shows a one-sided tail, it is preferentially identified as a velocity disturbance section instead of directly generating a boundary.

[0081] When the lift-off characterization signal changes significantly and the overall amplitude of the leakage magnetic signal decreases synchronously but does not form a complete peak-valley transition, it is preferentially retained as a lift-off disturbance record rather than directly judged as a complete defect event.

[0082] When the magnetization state characterization signal shows local magnetization discontinuity, the axial leakage magnetic signal and the radial leakage magnetic signal must have a mutually corroborating trend change near the boundary in order to determine the effective boundary.

[0083] It should be noted that different defect geometric dimensions will produce different waveform characteristics, and the key to defect identification is to separate the abnormal response with the physical meaning of the defect from the complex background. The above boundary determination method can clearly pinpoint the start, transition and end of the abnormality to the continuous frame index.

[0084] After obtaining the abnormal entry boundary, peak-valley transition boundary, and abnormal exit boundary, a candidate event segment is formed by consecutive frame segments with the same detection channel, the same component identifier, and the boundary order satisfying that the abnormal entry boundary comes first, the peak-valley transition boundary is in the middle, and the abnormal exit boundary comes last. The candidate event segment is written as an event entry that can be called by subsequent cross-channel mapping.

[0085] Each candidate event fragment should record at least the following:

[0086] The corresponding channel identifier, component identifier, start frame index, transition frame index, end frame index, corresponding mileage interval, polarity change mode within the boundary segment, peak and valley continuation direction, lift-off characterization signal change state, magnetization characterization signal change state, and attitude signal change state.

[0087] If multiple separate and clearly restored abnormal segments appear successively in adjacent frame index intervals in the same detection channel, multiple candidate event segments are generated respectively.

[0088] If the axial and radial magnetic flux leakage signals partially overlap in the same channel but their boundaries are not completely consistent, the start frame index, turning frame index, and end frame index of each of the two types of candidate event segments are retained, and a channel association mark is added to the event entry. This allows for joint comparison of the multi-component responses of the same defect passage process when mapping temporal relationships, spatial adjacency relationships, and response correspondence relationships.

[0089] When the candidate event segments corresponding to the axial and radial magnetic flux leakage signals under the same channel identifier overlap in the mileage interval, the offset between the start frame index and the end frame index does not exceed the preset component alignment window, and the peak-valley transition boundaries of both fall within the same continuous advancement interval, write the same channel association mark for these two types of candidate event segments.

[0090] When two components partially overlap but remain within the same continuous advance interval, their independent start frame index, turning frame index, and end frame index are retained, and the partial overlap is recorded in the same-channel association marker. When their corresponding mileage intervals are separate and the sources of disturbance are different, no same-channel association marker is written. The preset component alignment window is defined as the maximum frame index offset range within which different component candidate event segments under the same detection channel are allowed to retain same-channel association relationships.

[0091] For example, when there is localized corrosion and depression on the inner wall of a long-distance pipeline, the radial magnetic flux leakage signal often shows a significant trough before a peak, while the axial magnetic flux leakage signal may show a single peak or an offset peak. In this embodiment, as long as the two types of magnetic flux leakage signals are continuous in the corresponding mileage intervals within the same channel, the boundary sequence is established, and they can be distinguished from the disturbance records of the lift-off characterization signal and the magnetization state characterization signal, candidate event fragments are formed respectively and their boundary information is retained, without performing cross-channel merging in advance in this step.

[0092] Step 3: Perform cross-channel mapping based on the temporal relationship, spatial adjacency relationship, and response correspondence relationship of candidate event fragments to generate a synchronization event cluster. The specific implementation is as follows:

[0093] In this embodiment, based on the candidate event segments already obtained in the previous embodiment, cross-channel mapping is first performed according to the temporal relationship of the candidate event segments to form a candidate mapping set. Specifically, all candidate event segments are sorted according to the frame index arrangement order and mileage advancement direction in the unified measurement frame sequence.

[0094] The sorting is based on the starting frame index as the first order, the mileage signal corresponding to the starting frame index as the second order, and when the starting frame indices of two candidate event segments are the same, the circumferential arrangement order of the channel identifiers is used as the third order, thus obtaining a globally ordered queue of candidate event segments.

[0095] From the globally ordered candidate event segment queue, select any candidate event segment that has not been marked as mapped as the current reference segment. Read the start frame index, turning frame index, end frame index and corresponding mileage interval of the current reference segment. Search for candidate event segments in other detection channels that have overlapping frame index intervals or overlapping mileage intervals with it.

[0096] During the search, a first candidate window is established centered on the start frame index to the end frame index of the current reference segment, and a second candidate window is established centered on the mileage interval corresponding to the current reference segment. Only candidate event segments that fall into the first candidate window or the second candidate window at the same time are entered into the event sequence check.

[0097] When multiple candidate event segments in the same detection channel simultaneously meet the entry conditions, the candidate event segment with the largest overlap length with the current reference segment mileage interval is selected first. If the overlap lengths are the same, the candidate event segment with the smallest mileage difference between the peak-valley transition boundary and the current reference segment peak-valley transition boundary is selected first. If they are still the same, the candidate event segment with a channel adjacency relationship closer to the current reference segment is selected first.

[0098] For each candidate event fragment found in the search, check whether the order of the events is consistent. The check method is as follows:

[0099] Compare whether the order of the abnormal entry boundary, the order of the peak-valley transition boundary, and the order of the abnormal exit boundary of the current benchmark segment and the searched candidate event segment are aligned in the same direction, and at the same time compare whether the corresponding mileage signals of the two segments change continuously along the same direction of advancement.

[0100] If the boundary sequence remains consistent and the mileage change direction is continuous, the candidate event segment is added to the segment combination corresponding to the current baseline segment; if the boundary sequence is reversed, the mileage interval jumps in the opposite direction, or although there is short-term frame index overlap, they are not continuous in the mileage interval, they are not included in the same segment combination.

[0101] After completing the full-channel search of the current reference segment, the current reference segment and segments that meet the above conditions are combined and recorded as a candidate mapping set. The reason for adopting this processing method is that the leakage magnetic field detection signal is affected by the running speed, while the defect actually corresponds to the same physical passing event during the pipeline advancement process. Therefore, by constraining the order of events with the frame index order and the mileage advancement direction, it is possible to avoid incorrectly merging different defects or non-defect disturbances into the same event simply because the sampling time is close.

[0102] After obtaining the candidate mapping set, cross-channel mapping is further performed based on the spatial adjacency relationship of the candidate event fragments to filter out fragment combinations that do not conform to the circumferential distribution continuity and generate an adjacency mapping set. In specific implementation, a channel adjacency relationship table is established based on the actual installation order of each detection channel along the circumferential direction.

[0103] For the detection channels arranged in a closed circle, the channel adjacency table not only records the adjacency relationship between each channel and its left and right adjacent channels, but also records the closed adjacency relationship between the first and last channels of the circle, so as to ensure that the circumferential unfolding is consistent with the actual circumferential structure.

[0104] Perform adjacency consistency checks on each candidate mapping set one by one:

[0105] First, read the channel identifiers corresponding to all candidate event fragments in the candidate mapping set, and restore their arrangement positions on the circumference according to the channel adjacency table. Then, check whether these channel identifiers form a continuous circumferential adjacency chain, and whether the candidate event fragments corresponding to these channels show a continuous propagation relationship in the mileage direction.

[0106] Circumferential adjacency continuity means that any two adjacent channel identifiers in the candidate mapping set have a direct adjacency relationship in the channel adjacency table;

[0107] The propagation continuity along the pipeline running direction refers to the overlap, connection, or smooth offset of candidate event segments in adjacent detection channels over a mileage interval, rather than a sudden long-distance mileage break under circumferential adjacent conditions;

[0108] If a candidate mapping set contains channel identifiers that cross non-adjacent channels, are discretely distributed on the circumference, or are adjacent on the circumference but whose corresponding candidate event segments are reversed or propagated in the mileage direction, then the candidate mapping set is deemed not to meet the spatial adjacency requirements and is removed. The remaining candidate mapping set is denoted as the adjacency mapping set.

[0109] For example, when a localized corrosion defect is located in the common coverage area of ​​two circumferentially adjacent detection channels, the radial magnetic flux leakage signal will often appear continuously in the adjacent detection channels, rather than appearing in one channel and then crossing multiple unresponsive channels before reappearing in distant channels. Conversely, if an anomaly only appears in a single channel or appears discretely in every channel, it is more likely to be local noise, probe jitter, or discontinuous disturbance, and should not be directly identified as a cross-channel response of the same defect.

[0110] After the adjacency mapping set is formed, cross-channel mapping is performed and a synchronization event cluster is generated according to the response correspondence of the candidate event segments. Specifically, for each adjacency mapping set, the corresponding order of the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary of each candidate event segment is compared first.

[0111] If each candidate event fragment satisfies the following relationships on the adjacent circumferential channels: the abnormal entry boundary appears consecutively according to the mileage progression, the peak-valley transition boundary is located in the corresponding mileage segment, and the abnormal exit boundary falls back within the similar mileage segment, then it enters the response correspondence verification.

[0112] Response correspondence verification includes three aspects:

[0113] First, compare whether the polarity changes of axial and radial magnetic flux leakage signals in different channels have corresponding similar patterns. For example, whether the radial magnetic flux leakage signals all show a continuous change of first valley and then peak or first peak and then valley, and whether the axial magnetic flux leakage signals all show a main peak in the same direction or a main peak accompanied by a shifted tail.

[0114] Secondly, compare whether the peak and trough continuity is consistent, that is, whether the peak and trough change trend in the previous channel can continue in the next adjacent channel, rather than being suddenly interrupted or reversed.

[0115] Third, compare the continuous interval overlap relationship, that is, the mileage intervals of the candidate event segments corresponding to each channel should have continuous overlap or smooth shift, and should not be completely separated;

[0116] Only candidate event fragments that simultaneously satisfy the following criteria are merged into the same synchronization event cluster: boundary order correspondence, polarity change correspondence, peak-valley continuity correspondence, and continuous interval overlap.

[0117] After merging, a unified event identifier is assigned to the synchronous event cluster, and channel association information is recorded. The unified event identifier is used to indicate that the synchronous event cluster is a unified expression of the same defect passing event in multiple detection channels. The channel association information includes at least the set of channel identifiers contained in the synchronous event cluster, the start frame index, transition frame index, end frame index and corresponding component identifier for each channel.

[0118] The unified event identifier can be generated incrementally according to the order in which candidate event fragments first enter the merging process, or it can be generated by combining the starting frame index and the center channel identifier, as long as it is unique and traceable in the same detection process.

[0119] For example, for a metal loss defect covering three adjacent detection channels, the radial leakage magnetic signal may exhibit a continuous bipolar peak-valley structure in all three channels, while the axial leakage magnetic signal is strongest in the middle channel and gradually weakens in the two side channels. As long as the boundaries of the three correspond, the polarity continues, and the mileage intervals overlap, they can be merged into a synchronous event cluster and will not be mistakenly divided into three independent defect events.

[0120] Step four: Combine the reference measurement signal and the operating status signal to perform state compensation on the synchronization event cluster, generating a compensated synchronization signal sequence. The specific implementation is as follows:

[0121] In this embodiment, based on the synchronization event clusters already generated in the previous embodiments, the synchronization event clusters are further compensated by combining reference measurement signals and operating status signals. Specifically, the unified event identifier is used as an index to read the start frame index, transition frame index and end frame index covered by the synchronization event cluster one by one, and the lift-off characterization signal, magnetization state characterization signal, odometer signal, velocity signal and attitude signal within the frame index interval are read synchronously.

[0122] First, change trend sequences are established for the lift-off characterization signal, magnetization state characterization signal, and velocity signal, respectively. The change trend sequence is not a simple comparison of individual sample values, but rather an observation of whether the signal continuously rises, continuously falls, fluctuates back and forth, or remains stable within the current synchronization event cluster interval according to the frame index order. Then, the main sources of disturbance to the current synchronization event cluster are determined by combining the mileage advancement continuity and attitude change continuity.

[0123] If the speed signal continuously increases, continuously decreases, or increases first and then decreases within the coverage area of ​​the synchronous event cluster, and the corresponding mileage advancement rhythm changes significantly compared to the adjacent synchronous event cluster, it is identified as a speed change state.

[0124] If the lift-off characterization signal continuously deviates from the previous stable range within the coverage area of ​​the synchronous event cluster, and the axial leakage magnetic signal and the radial leakage magnetic signal simultaneously show an overall amplitude contraction or an overall amplitude increase, then it is identified as a lift-off change state.

[0125] If the magnetization state characterization signal exhibits continuous fluctuations, abrupt interruptions, or magnetization continuity imbalances between adjacent detection channels within the coverage area of ​​the synchronous event cluster, and the attitude signal shows deflection or rolling changes, it is identified as a local magnetization disturbance state.

[0126] For a cluster of synchronous events that simultaneously satisfies two or more state characteristics, the main compensation state is determined in the order that the local magnetization disturbance state takes precedence over the lift-off change state, and the lift-off change state takes precedence over the velocity change state. At the same time, other states are recorded as accompanying states. After identification is completed, a compensation state identifier and a compensation interval identifier are written for each cluster of synchronous events. The compensation state identifier is used to record the main compensation state and accompanying state corresponding to the cluster of synchronous events, and the compensation interval identifier is used to record the specific frame index interval for which state compensation needs to be performed.

[0127] When a synchronization event cluster has both a primary compensation state and a secondary state, the compensation rule corresponding to the primary compensation state is executed first, followed by the compensation rule corresponding to the secondary state. The compensation interval identifier corresponding to the primary compensation state covers all the start frame index to the end frame index of the synchronization event cluster, while the compensation interval identifier corresponding to the secondary state only covers the local frame index segment where it actually occurs. This avoids the situation where step four only has state priority and no compensation execution order.

[0128] The reason for the above settings is that the results of magnetic flux leakage detection are significantly affected by the detection speed, lift-off distance and magnetization state. High-speed movement will cause magnetic field distortion, and increased lift-off will reduce detection sensitivity and weaken the defect signal. Therefore, only by clearly identifying the disturbance state of the section where the synchronous event cluster is located can subsequent compensation have a clear object and boundary.

[0129] After associating the compensation status identifier and the compensation interval identifier, the corresponding compensation rules are invoked for different synchronization event clusters to generate a compensation synchronization signal sequence, specifically including:

[0130] For a cluster of synchronous events in a state of velocity change, the position is corrected for the start frame index, transition frame index, and end frame index, using the odometer signal as the primary reference. The specific method is as follows:

[0131] First, use the stable interval before and after the current synchronization event cluster as the position reference segment. Then check whether the mileage advancement offset between adjacent frame indices is uniform. If it is found that the frame index advancement is still continuous in a certain local interval but the mileage advancement is suddenly compressed or stretched, then according to the continuity of mileage advancement, the abnormal entry boundary is pushed forward or moved backward, the peak-valley transition boundary is adjusted to the position where the axial leakage magnetic signal and the radial leakage magnetic signal both turn, and the abnormal exit boundary is adjusted to the position where the mileage advancement returns to stability and the leakage magnetic response falls back.

[0132] For a cluster of synchronous events in a state of lift-off change, the amplitude of the leakage magnetic response is corrected according to the change range of the lift-off characterization signal. During the correction, instead of directly amplifying or reducing all frames by a uniform ratio, a reference amplitude range under the same detection channel and the same component identifier is first extracted from the adjacent stable intervals. Then, the leakage magnetic response affected by lift-off within the current synchronous event cluster is reverted frame by frame according to the principle that if the lift-off increases, it will be compensated and restored, and if the lift-off decreases, it will suppress over-amplification. This makes the peak value, valley value and peak-valley interval of the axial leakage magnetic signal and the radial leakage magnetic signal return to a state that is continuously connected with the adjacent stable intervals.

[0133] For a cluster of synchronous events in a state of local magnetization disturbance, boundary pruning, missing completion, or anomaly removal are performed based on the continuity relationship of the synchronous event cluster in adjacent channels.

[0134] If the boundary of the synchronous event cluster of a certain detection channel is significantly longer than that of the adjacent detection channels, and the magnetization state characterization signal shows that there is a local magnetization collapse in this section, then the boundary is pruned for the part that exceeds the common coverage area of ​​the adjacent channels.

[0135] If a detection channel has a short-term missing information in the middle when all adjacent channels respond continuously, then missing information will be filled in according to the continuous interval corresponding to the same unified event identifier in the adjacent channels.

[0136] If an isolated peak or valley value appears in a detection channel that is completely inconsistent with the adjacent channels, it will be rejected as an anomaly after confirming that it has no common support with the lift-off characterization signal, velocity signal, and attitude signal.

[0137] After the position correction, amplitude correction, and boundary correction are completed, the corrected synchronization event clusters are written back to the unified measurement frame sequence according to the unified event identifier to form a compensated synchronization signal sequence, which specifically includes:

[0138] During the write-back, the original frame index structure is used as the skeleton. The overall arrangement order of the unified measurement frame sequence is not changed. Only the start frame index, turning frame index, end frame index, axial leakage magnetic signal sample value, radial leakage magnetic signal sample value, and corresponding compensation status and compensation interval identifiers corresponding to the current unified event identifier are updated.

[0139] Write a boundary clipping marker to the frame index position that was removed due to boundary clipping.

[0140] For frame index positions added due to missing padding, write padding markers;

[0141] For local responses that are deleted due to abnormal removal, a removal flag is written to ensure that the source and method of compensation can be traced during subsequent reconstruction of defect response distribution.

[0142] For example, when a unified event identifier covers three circumferentially adjacent detection channels, and the radial leakage magnetic signal peak of the middle detection channel shrinks significantly due to short-term lift-up, while the peak and valley positions shift slightly backward relative to the preceding and following channels due to a slight increase in speed, the radial leakage magnetic signal amplitude of the middle detection channel is first restored to the interval that is continuously connected with the preceding and following detection channels based on the lift-up characterization signal. Then, the turning boundary of the channel is finely adjusted forward based on the mileage advancement offset. Finally, the three detection channels re-form a continuous synchronous event cluster under the same unified event identifier, and are written back to the unified measurement frame sequence as part of the compensation synchronization signal sequence.

[0143] The compensated synchronization signal sequence obtained through the above processing not only retains the original cross-channel correspondence of the synchronization event cluster, but also makes targeted corrections to the main distortions caused by velocity change state, lift-off change state and local magnetization disturbance state. This provides a stable and traceable data foundation for reconstructing the defect response distribution based on the compensated synchronization signal sequence and outputting leakage magnetic field detection results.

[0144] For example, during a periodic magnetic flux leakage internal inspection of a long-distance pipeline, when the internal inspection device is advanced along the pipeline axis to a local corrosion thinning section adjacent to a circumferential weld, the middle inspection channel at the corresponding circumferential position of that section and its two adjacent inspection channels simultaneously collect continuous abnormal responses.

[0145] The radial leakage magnetic field signal shows a continuous change from valley to peak in all three adjacent detection channels. The axial leakage magnetic field signal shows a main peak response in the middle detection channel and a continuous response with gradually decreasing amplitude in the two adjacent detection channels. At the same time, the lift-off characterization signal corresponding to the middle detection channel shows a short-term rise, and the velocity signal shows a trend of rising first and then stabilizing in this section.

[0146] Based on the unified measurement frame sequence generated in step one, the axial leakage magnetic signal, radial leakage magnetic signal, lift-off characterization signal, magnetization state characterization signal, odometer signal, velocity signal and attitude signal of this section are first written into the unified measurement frame corresponding to the continuous frame index. Then, according to step two, baseline recovery and trend segmentation are performed on the three adjacent detection channels respectively to determine their respective abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary, and form corresponding candidate event segments.

[0147] Because this section is adjacent to the circumferential weld, although some non-adjacent detection channels showed short-term abnormal fluctuations, these abnormal fluctuations did not simultaneously satisfy the complete boundary sequence and continuous mileage progression relationship. Therefore, they were only retained as weld disturbance records and did not form effective candidate event fragments.

[0148] Based on this, according to step three, cross-channel mapping is performed on the above candidate event segments. First, according to the frame index arrangement order and mileage advancement direction, the candidate event segments in the three adjacent detection channels are grouped into the same candidate mapping set. Then, according to the channel adjacency relationship table, it is confirmed that the three detection channels form a continuous adjacency chain in the circumferential direction. Finally, according to the consistency of the polarity change of the radial leakage magnetic signal from valley to peak, the continuity relationship of the main peak of the axial leakage magnetic signal, and the overlapping relationship of the continuous interval, the candidate event segments in the three detection channels are merged into the same synchronous event cluster and assigned a unified event identifier. Since the intermediate detection channel has a composite disturbance of short-term rise of the lift-off characterization signal and first rise and then stabilization of the velocity signal in this section, according to step four, the synchronous event cluster is identified as a compensation state with the lift-off change state as the main state and the velocity change state as the accompanying state. The amplitude of the radial leakage magnetic signal of the intermediate detection channel is corrected, the position of its peak-valley transition boundary is corrected, and the corrected synchronous event cluster is written back to the unified measurement frame sequence according to the unified event identifier to form a compensation synchronous signal sequence.

[0149] After compensation, the three adjacent detection channels maintain continuous boundary correspondence, peak-valley continuity, and mileage interval overlap under the same unified event identifier. Thus, in step five, they can stably expand along the circumferential distribution direction and the pipeline extension direction, forming a defect response distribution corresponding to the local corrosion thinning event near the circumferential weld, and further outputting the defect location, circumferential range, axial extension characteristics, and defect level.

[0150] For example, during an internal magnetic flux leakage detection of a long-distance gas transmission pipeline, when the internal detection device passed through a liquid accumulation-prone area at the bottom of the pipeline, two adjacent detection channels in the circumferential region at the bottom of the pipeline and the adjacent transition detection channel successively collected continuous abnormal responses. The radial magnetic flux leakage signal showed a long-lasting valley-to-peak response in the two adjacent detection channels at the bottom, and a smaller amplitude but consistent boundary sequence response in the transition detection channel. The axial magnetic flux leakage signal formed a main peak in the central detection channel at the bottom and a continuous attenuation peak in the detection channels on both sides.

[0151] Meanwhile, the velocity signal in this section remained stable, the lift-off characterization signal did not show any significant abrupt change, and the magnetization state characterization signal remained continuous throughout the entire abnormal section. Thus, the dominant influence of velocity disturbance, lift-off disturbance, and local magnetization disturbance on the abnormal response can be eliminated. Based on the processing results of steps one to three, multiple candidate event segments in this section can be grouped into the same synchronous event cluster according to temporal relationship, spatial adjacency relationship, and response correspondence relationship.

[0152] Since the channel identifiers covered by this cluster of synchronous events are continuously distributed in the circumferential area at the bottom of the pipeline, and the corresponding mileage intervals have a continuous overlapping relationship, it can be determined that the abnormal response originates from the same bottom corrosion event, rather than multiple discrete noise sources.

[0153] Step 5: Reconstruct the defect response distribution based on the compensation synchronization signal sequence and output the leakage magnetic field detection results. The specific implementation is as follows:

[0154] After the compensation synchronization signal sequence has been formed, the synchronization event clusters in the compensation synchronization signal sequence are first collected according to the unified event identifier. In specific processing, the unified event identifier is used as the main index to extract all synchronization event clusters with the same unified event identifier from the compensation synchronization signal sequence, and then expand them bidirectionally according to the circumferential arrangement order corresponding to the channel identifier and the mileage order corresponding to the starting frame index.

[0155] When unfolding along the circumferential distribution direction, the axial and radial magnetic flux leakage signals corresponding to each channel identifier are read sequentially to restore the response coverage of the same defect at different circumferential positions.

[0156] As the pipeline extends, the mileage signals corresponding to the start frame index, turning frame index, and end frame index are read sequentially. The leakage magnetic response under the same unified event identifier is arranged in the order of mileage progression to form a continuous response band during the defect passage process.

[0157] At each circumferential position and each mileage position, the axial leakage magnetic signal, radial leakage magnetic signal, and their corresponding component identifiers, channel identifiers, and boundary information are simultaneously retained, thereby constructing the defect response distribution of the corresponding defect passage process. The defect response distribution is not the result of waveform splicing of a single channel, but the result of joint expansion of the same unified event identifier in the circumferential and axial directions. Therefore, it can uniformly express the continuous coverage relationship of the defect in different detection channels and the extension relationship in the mileage direction, serving as the direct basis for subsequent extraction of defect location, circumferential range, and axial extension features.

[0158] Multi-component magnetic flux leakage signals are used for defect quantification and defect profile reconstruction, while retaining information on magnetic field disturbances in different directions.

[0159] After constructing the defect response distribution, the defect location, circumferential range, and axial extension characteristics are further determined based on the boundary turning points, peak and valley continuation trajectories, and channel coverage in the defect response distribution. Specifically, the defect location is determined in the following ways:

[0160] First, search for the central response segment that is covered by all synchronous event clusters in the defect response distribution. Then, take the mileage position where the axial leakage magnetic signal and radial leakage magnetic signal in the central response segment maintain a continuous response as the segment where the defect is located.

[0161] If the central response segment spans multiple consecutive mileage positions, the midpoint between the mileage corresponding to the start frame index and the mileage corresponding to the end frame index is used as the defect location marker. The circumferential range is determined by the channel coverage range, that is, by counting the set of channel identifiers that have consecutive valid responses under the same unified event identifier, and determining the coverage interval between the leftmost consecutive channel identifier and the rightmost consecutive channel identifier as the defect circumferential range.

[0162] If the channel covers the beginning and end of the circle and is closed and continuous, then the beginning and end closure determination is performed according to the channel adjacency table before outputting the circumferential range. Axial extension features are determined by the boundary turning points and peak-valley continuation trajectories, specifically:

[0163] Compare the mileage span of each channel between the start frame index and the end frame index, extract the continuous mileage segment with the most sufficient overlap as the main extension segment, and then take the continuous length of the peak and valley continuation trajectory within the main extension segment as the axial extension feature.

[0164] If the peak-valley continuation trajectory of a certain channel appears only locally and does not continuously correspond to adjacent channels, then this local anomaly is not counted separately as an axial extension feature. For example, when a synchronous event cluster corresponding to a unified event identifier covers three circumferentially adjacent detection channels, and the radial leakage magnetic signals in the three detection channels all show continuous bipolar responses within similar mileage sections, while the axial leakage magnetic signals are strongest in the middle detection channel and gradually weaken in the two side detection channels, then the central mileage section corresponding to the middle detection channel can be identified as the defect location, the continuous coverage area of ​​the three circumferentially adjacent detection channels can be identified as the circumferential range, and the continuous mileage section where the bipolar response persists can be identified as an axial extension feature.

[0165] After extracting the defect location, circumferential range, and axial extension features, the defect level is further generated based on the abnormal persistence morphology and component response relationship of the defect response distribution, and the leakage magnetic field detection results are output. Specifically, the abnormal persistence morphology in the defect response distribution is first classified:

[0166] If the axial and radial magnetic flux leakage signals remain continuous, have clear boundaries, and have stable peak and valley continuation trajectories within the main extension section, they are recorded as a continuous, stable, abnormal, and persistent morphology.

[0167] If there is a local interruption in the main extension section but it can still maintain cross-channel continuity after the aforementioned compensation, it is recorded as a continuous repair abnormal continuous form.

[0168] If unstable responses remain only in individual channels or individual local sections, they are recorded as local discrete anomaly persistence mode. Then, the component response relationship between the axial leakage magnetic signal and the radial leakage magnetic signal is compared. If the two types of components maintain good correspondence in boundary position, turning sequence and continuation direction, the reliability of the defect response is considered to be high.

[0169] If the two types of components correspond only in some segments, the reliability of the defect response is considered to be moderate; if there is a long-term lack of correspondence between the two types of components, the reliability of the defect response is considered to be low.

[0170] Finally, the abnormal persistence pattern and component response relationship are combined to form the defect level:

[0171] For defect response distributions with continuous, stable, and anomalous persistent patterns and complete component response relationships, a higher defect level is generated.

[0172] For defect response distributions with continuously repaired abnormal and persistent morphologies and basically complete component response relationships, intermediate defect levels are generated.

[0173] For defect response distributions with persistent local discrete anomalies and incomplete component response relationships, a lower defect level is generated.

[0174] After generating the defect level, the unified event identifier, defect location, circumferential range, axial extension characteristics, defect level, and the corresponding channel identifier set and mileage interval will be output as the internal detection result of magnetic flux leakage.

[0175] It should be noted that the defect level is generated according to a preset level table, which includes at least a first defect level, a second defect level, and a third defect level. The first defect level is generated when the defect response distribution corresponds to a continuous and stable abnormal pattern and the component response relationship is complete. The second defect level is generated when the defect response distribution corresponds to a continuous and repairable abnormal pattern and the component response relationship is basically complete. The third defect level is generated when the defect response distribution corresponds to a local discrete abnormal pattern and the component response relationship is incomplete. The internal detection results of magnetic flux leakage include at least a unified event identifier, defect location, circumferential range, axial extension characteristics, defect level, corresponding channel identifier set, starting mileage position, and ending mileage position.

[0176] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0177] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0178] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0179] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0180] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for synchronous implementation of parallel digital signals in internal leakage magnetic flux detection of long-distance pipelines, characterized in that: Includes the following steps: During the internal detection process of magnetic flux leakage, multi-channel magnetic flux leakage signals, reference measurement signals, and operating status signals are collected to generate a unified measurement frame sequence. Candidate event segments representing the defect passage process are extracted based on a unified measurement frame sequence; Perform cross-channel mapping based on the temporal relationship, spatial adjacency relationship, and response correspondence relationship of candidate event fragments to generate a cluster of synchronous events; By combining reference measurement signals and operating status signals, state compensation is performed on the synchronization event cluster to generate a compensation synchronization signal sequence; The defect response distribution is reconstructed based on the compensated synchronization signal sequence, and the leakage magnetic field detection results are output.

2. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 1, characterized in that: The multi-channel magnetic flux leakage signals include axial magnetic flux leakage signals and radial magnetic flux leakage signals distributed along the circumference. The reference measurement signals include lift-off characterization signals and magnetization state characterization signals. The operating status signals include mileage signals, speed signals and attitude signals.

3. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 1, characterized in that: Generating a unified measurement frame sequence includes: The frame index is determined based on the operating status signal. The multi-channel leakage magnetic signal, reference measurement signal and operating status signal under the same frame index are written into the same measurement frame. The channel identifier, component identifier and status identifier are associated in the measurement frame to form a unified measurement frame sequence arranged continuously according to the frame index.

4. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 1, characterized in that: Candidate event fragments for characterizing the defect process include: Baseline recovery and trend segmentation are performed on the multi-channel leakage magnetic signals in the unified measurement frame sequence to determine the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary, and candidate event segments of the corresponding defect passage process are formed based on the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary.

5. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 4, characterized in that: Performing cross-channel mapping based on the temporal relationship of candidate event fragments includes: Candidate event segments in each channel are sequentially sorted according to the frame index arrangement order and mileage advancement direction in the unified measurement frame sequence. For any candidate event segment, search for candidate event segments in other channels within the corresponding frame index interval and mileage continuity interval, determine the segment combination with consistent event sequence and continuous mileage change direction, and generate a candidate mapping set.

6. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 5, characterized in that: Performing cross-channel mapping based on the spatial adjacency of candidate event fragments includes: Based on the circumferential distribution order corresponding to the leakage magnetic field signal, a channel adjacency table is established. According to the circumferential adjacency continuity between channels corresponding to each candidate event segment in the candidate mapping set and the propagation continuity along the pipeline running direction, an adjacency consistency check is performed on the candidate mapping set. Segment combinations that are discretely distributed in the circumferential direction, cross non-adjacent channels, or whose propagation order is inconsistent with the circumferential adjacency relationship are eliminated, and an adjacency mapping set is generated.

7. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 6, characterized in that: Performing cross-channel mapping and generating synchronization event clusters based on the response correspondence of candidate event fragments includes: Compare the corresponding order of the abnormal entry boundary, peak-valley transition boundary and abnormal exit boundary of each candidate event segment in the adjacency mapping set, and verify the correspondence between the axial leakage magnetic signal and the radial leakage magnetic signal in different channels in terms of polarity change, peak-valley continuity and continuous interval overlap. Candidate event fragments that satisfy the response correspondence are grouped into synchronous event clusters, and a unified event identifier and channel association information are associated with the synchronous event clusters.

8. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 7, characterized in that: Combining reference measurement signals and operating status signals to perform state compensation on the synchronization event cluster includes: Based on the trends of reference measurement signal changes, speed signal changes, mileage advancement continuity, and attitude change continuity within the frame index interval corresponding to the synchronization event cluster, identify the speed change state, lift-off change state, and local magnetization disturbance state. Based on the correspondence between velocity change state, lift-off change state, and local magnetization disturbance state and synchronization event cluster, compensation state identifier and compensation interval identifier are associated with synchronization event cluster.

9. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 8, characterized in that: Perform state compensation on the synchronization event cluster to generate a compensated synchronization signal sequence, including: The corresponding compensation rules are invoked for each cluster of synchronous events associated with different compensation status identifiers; Under speed change conditions, position corrections are performed on the abnormal entry boundary, peak-valley transition boundary, and abnormal exit boundary based on the mileage advance offset between adjacent frame indices. Under the lift-off change state, the leakage magnetic response amplitude in the synchronous event cluster is corrected according to the change range of the reference measurement signal; Under local magnetization disturbance, boundary pruning, missing completion or anomaly removal are performed on the abnormal response segment according to the continuity relationship of the synchronous event cluster in adjacent channels. After completing position correction, amplitude correction, and boundary correction, the synchronization event clusters are written back to the unified measurement frame sequence according to the unified event identifier to generate the compensation synchronization signal sequence.

10. The method for synchronous implementation of parallel digital signals in the internal detection of leakage magnetic flux in long-distance pipelines according to claim 9, characterized in that: The defect response distribution is reconstructed based on the compensated synchronization signal sequence, and the internal leakage magnetic field detection results are output, including: The synchronization event clusters in the compensation synchronization signal sequence are expanded along the circumferential distribution direction and the pipeline extension direction according to the unified event identifier to construct the defect response distribution of the corresponding defect passage process; The defect location, circumferential range, and axial extension characteristics are determined based on the boundary turning points, peak and valley continuation trajectories, and channel coverage in the defect response distribution. The defect level is generated based on the abnormal persistence morphology and component response relationship of the defect response distribution, and the leakage magnetic field internal detection results are output.

Images