Method and system for detecting double-frequency quenching state of shaft metal part

Abstract

The invention relates to the technical field of non-destructive testing, and discloses a shaft metal piece double-frequency quenching state detection method and system, and the method comprises the steps: dividing a detection region according to geometric information and probe layout, synchronously collecting a first frequency detection signal and a second frequency detection signal to generate double-frequency original data, generating a sampling mark according to a sampling interval and waveform change, distinguishing a sampling stable detection area from an insufficient detection area, constructing a regional topological graph by combining the adjacency relation of the detection areas, and calculating a missing area registration factor by taking the sampling stable detection area as a reference under the constraint of the regional topological graph and the sampling mark; and substituting the sampling stable detection area into a preset quenching calibration relation to obtain a local hardening depth, adjusting the double-frequency original data of the insufficient detection area by using the missing area registration factor, calculating the hardening layer depth of the insufficient detection area, forming a hardening layer depth curve, and realizing nondestructive evaluation of consistent curve forms of different sections and step areas.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and more specifically, to a method and system for detecting the dual-frequency quenching state of shaft-type metal parts. Background Technology

[0002] After surface hardening, shaft-type metal parts typically require non-destructive evaluation of the hardened layer depth and distribution under limited detection space and signal processing computing power. Existing technologies mostly employ amplitude determination methods based on single-frequency or multi-frequency induction signals, transform domain characterization and correlation measurement methods, and time series smoothing and threshold screening methods to analyze the hardening state. These methods generally assume that the detection signal is fully sampled, the surface state is relatively stable after hardening, and the changes in shaft diameter and step structure are relatively gradual. Under the premise of high signal-to-noise ratio and slow change in response waveform, relatively stable hardened layer depth distribution results can be output.

[0003] In actual dual-frequency quenching testing, high-temperature residue and surface oxidation after quenching can cause local signal energy attenuation or even interruption, resulting in incomplete sampling and waveform distortion of the dual-frequency detection signal. At the same time, different shaft diameters and step regions can cause significant differences in electromagnetic field distribution and coupling conditions, making it difficult to make the response amplitudes of different sections and regions comparable. These unstable factors together weaken the ability of conventional judgment methods based on amplitude and waveform characteristics to stably characterize the morphology of the hardened layer depth distribution curve, making it difficult to keep the hardened layer depth distribution curves consistent for different sections and structural regions. Therefore, the technical problem to be solved is how to ensure the consistency of the hardened layer depth distribution curve morphology under the conditions of incomplete sampling and waveform distortion of the dual-frequency quenching detection signal, and under the condition that the response amplitudes of different shaft diameters and step regions are comparable.

[0004] In view of this, the present invention proposes a method and system for detecting the dual-frequency quenching state of shaft-type metal parts to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for detecting the dual-frequency quenching state of shaft-type metal parts.

[0006] To achieve the above objectives, the present invention provides the following technical solution: Firstly, a method for detecting the dual-frequency quenching state of shaft-type metal parts is provided, including: The detection zones are divided according to the geometric information of the shaft-type metal parts and the probe layout. During the quenching process, the first frequency detection signal and the second frequency detection signal are collected simultaneously in each detection zone to obtain dual-frequency raw data. For the dual-frequency raw data, sampling marks are generated by combining the sampling interval and waveform changes in each detection zone. Based on the sampling marks, the detection zone with stable sampling and the detection zone with insufficient sampling are divided. Based on the adjacency relationship of the detection areas and combined with the insufficient detection areas, a regional topology map is constructed. Under the constraints of the regional topology map and sampling markers, the registration factor of the missing area is calculated with reference to the dual-frequency original data in the sampling stable detection area. Under the constraints of the regional topology map and sampling markers, the sampling stable detection area in the dual-frequency raw data is substituted into the preset quenching calibration relationship to obtain the local hardening depth. The missing area registration factor is used to adjust the insufficient detection area in the dual-frequency raw data, and the hardening layer depth in the insufficient detection area is calculated in combination with the local hardening depth to obtain the hardening layer depth curve. The quenching state result is output based on the hardening layer depth curve.

[0007] In some embodiments, for dual-frequency raw data, sampling markers are generated by combining the sampling interval and waveform changes in each detection zone, including: Extract the sampling timestamps of the first frequency detection signal and the second frequency detection signal of each detection zone, and generate a sampling interval sequence according to the timestamp order; Locate the interval gap position index in the sampling interval sequence and write the interval gap position index into the detection area sampling interval table; Extract the envelope transition position index from the first frequency detection signal and the second frequency detection signal in each detection zone, and write the envelope transition position index into the waveform change table of the detection zone; A set of mutual verification anchor points is generated based on the sampling interval table and waveform change table of the detection area, and sampling markers are generated based on the set of mutual verification anchor points.

[0008] In some embodiments, a mutual verification anchor set is generated based on a detection area sampling interval table and a detection area waveform change table, and a sampling marker is generated based on the mutual verification anchor set, including: Map the gap position index to the gap time window, and write the gap time window into the detection area time window table; Outside the time window of the gap, retrieve the time proximity pairs between the first frequency envelope turning point index and the second frequency envelope turning point index, and record the time proximity pairs as candidate anchor point pairs; Perform chain-connection processing on candidate anchor point pairs in chronological order, and record the chain-connection processing results as an anchor point chain; Filter anchor chain segments that cover the time window table of the detection area from the anchor chain, and aggregate the anchor chain segments into a mutual verification anchor set; Based on the mutual verification anchor point set, write the sampling mark into the sampling interval table and waveform change table of the detection area, and output the sampling mark.

[0009] In some embodiments, a regional topology map is constructed based on the adjacency relationship of detection areas and in combination with insufficient detection areas, including: Based on the geometric information of shaft-type metal parts, axial adjacent pairs of the detection area are generated, and based on the probe layout, circumferential adjacent pairs of the detection area are generated. Perform duplicate elimination processing on axial and circumferential adjacent pairs in the detection area, and record the duplicate elimination processing results in the adjacency table of the detection area; Read the detection region partitioning table and extract the adjacent boundary pairs of the insufficient detection regions from the detection region adjacency table, and generate the missing region set based on the adjacent boundary pairs; Construct a regional topology graph using the detection area as nodes and the detection area adjacency list as the connection relationship, and write the missing area set into the node attributes of the regional topology graph; Based on the sampling tag table, write sampling tags into the node attributes of the regional topology map and output the regional topology map.

[0010] In some embodiments, reading the detection region partitioning table and extracting adjacent boundary pairs of insufficient detection regions from the detection region adjacency table, and generating a set of missing regions based on the adjacent boundary pairs, includes: Extract the adjacency pairs related to the insufficient detection area from the adjacency list of the detection area, and aggregate the adjacency pairs into a set of insufficient edges; Based on the insufficient edge set, perform connectivity aggregation processing on the insufficient detection region, and record the connectivity aggregation processing results as an aggregation cluster table; For each cluster in the cluster table, collect the insufficient detection areas and generate a missing area identifier for each cluster; Write the missing region identifier and its corresponding list of insufficient detection regions into the missing region table, then aggregate the missing region tables into a missing region set, output the missing region set and write it into the node attributes of the region topology graph.

[0011] In some embodiments, connectivity aggregation processing is performed on the insufficient detection region based on the insufficient edge set, and the results of the connectivity aggregation processing are recorded as an aggregation cluster table, including: Select any insufficient detection region from the insufficient edge set as the starting node, and write the starting node into the queue to be expanded; Take the head node from the queue to be expanded and search the adjacent insufficient detection area of ​​the head node, and write the search result into the connectivity record table; Add the unqueued insufficient detection areas in the connectivity record table to the queue to be expanded, and write access flags to the processed insufficient detection areas. When the queue to be expanded is empty, output the connectivity record table as an aggregate cluster and write the aggregate cluster to the aggregate cluster table.

[0012] In some embodiments, under the constraints of the region topology map and sampling markers, the registration factor for the missing region is calculated using the original dual-frequency data within the sampled stable detection area as a reference, including: Extract the sampling stable detection region adjacent to the insufficient detection region from the regional topology map, and record the sampling stable detection region as the boundary stable region; Within the boundary stable region, the first frequency detection signal segment and the second frequency detection signal segment are extracted according to the sampling mark, and the first frequency detection signal segment and the second frequency detection signal segment are written into the reference segment table; Establish a segment correspondence between the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table, and write the segment correspondence into the segment correspondence table; Generate an amplitude mapping table based on the segment correspondence table, and establish an association record between the amplitude mapping table and the boundary stable region in the regional topology map; The missing region registration factor is calculated based on the amplitude mapping table under the constraints of the regional topology map and sampling markers.

[0013] In some embodiments, a segment correspondence between a first frequency detection signal segment and a second frequency detection signal segment is established in a reference segment table, and the segment correspondence is written into the segment correspondence table, including: Read the reference segment table and extract the timestamp sequence of the first frequency detection signal segment and the timestamp sequence of the second frequency detection signal segment, and generate candidate segment pairs based on the timestamp proximity relationship; Extract the inflection point index of the first frequency detection signal segment and the inflection point index of the second frequency detection signal segment from the candidate segment pairs, and generate segment anchor point pairs based on the inflection point indexes; Based on the fragment anchor point pair, perform pairing convergence processing on the candidate fragment pairs, record the pairing convergence processing results as fragment pairings, and write the fragment pairings into the fragment correspondence table.

[0014] In some embodiments, the insufficient detection area in the dual-frequency raw data is adjusted using a missing region registration factor, and the hardened layer depth within the insufficient detection area is calculated in conjunction with the local hardening depth to obtain a hardened layer depth curve, including: Read the local hardening depth and write it into the depth anchor point table according to the spatial position of the detection area; Extract the stable sampling detection regions adjacent to the insufficient detection regions in the regional topology map, and write the adjacency relationship between the stable sampling detection regions and the insufficient detection regions into the boundary relationship table; Based on the boundary relationship table, the local hardening depth in the depth anchor point table is processed by boundary propagation, and the result of the boundary propagation is written into the boundary depth table. Write the dual-frequency raw data corresponding to the insufficient detection area, adjusted by the missing area registration factor, into the insufficient area input table, and establish an index association between the insufficient area input table and the boundary depth table; The depth of the hardened layer in the insufficient detection zone is calculated based on the input table of the insufficient zone and the boundary depth table, and the hardened layer depth curve is generated by combining them.

[0015] Secondly, a dual-frequency quenching state detection system for shaft-type metal parts is provided, which is used to implement the aforementioned dual-frequency quenching state detection method for shaft-type metal parts, including: The partition acquisition module is used to divide the detection area according to the geometric information of the shaft metal parts and the probe layout. During the quenching process, the first frequency detection signal and the second frequency detection signal are acquired simultaneously in each detection area to obtain dual-frequency raw data. Marking and Division Module: Used to generate sampling marks in each detection zone based on the comprehensive sampling interval and waveform changes of dual-frequency raw data, and to divide the stable sampling detection zone and the insufficient detection zone according to the sampling marks; Topology registration module: used to construct a regional topology map based on the adjacency relationship of the detection area and in combination with the insufficient detection area. Under the constraints of the regional topology map and sampling markers, the registration factor of the missing area is calculated with reference to the dual-frequency original data in the sampled stable detection area. The quenching depth calculation module is used to substitute the sampled stable detection area in the dual-frequency raw data into the preset quenching calibration relationship under the constraints of the region topology map and sampling markers to obtain the local hardening depth. It uses the missing area registration factor to adjust the insufficient detection area in the dual-frequency raw data, and calculates the hardened layer depth in the insufficient detection area in combination with the local hardening depth to obtain the hardened layer depth curve. Based on the hardened layer depth curve, it outputs the quenching state result.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention divides the detection area based on the geometric information of shaft-type metal parts and the probe layout, and simultaneously acquires the first frequency detection signal and the second frequency detection signal to form dual-frequency raw data. Sampling marks are generated in each detection area according to the sampling interval and waveform changes, dividing the area into a stable sampling detection area and an insufficient detection area. This allows incomplete sampling and waveform distortion of the dual-frequency quenching detection signal to be located and processed in a partitioned manner. A regional topology map is constructed by combining the adjacency relationship of the detection areas and the insufficient detection area. Under the constraints of the regional topology map and sampling marks, the registration factor for the missing area is calculated using the dual-frequency raw data in the stable sampling detection area as a reference, establishing comparable amplitude registration between different shaft diameters and step areas. Scale; Under the constraints of the regional topology map and sampling markers, the local hardening depth is obtained by substituting the sampling stable detection area into the preset quenching calibration relationship. The registration factor of the missing area is used to adjust the dual-frequency original data of the insufficient detection area. The hardening layer depth in the insufficient detection area is calculated by combining the local hardening depth, forming a hardening layer depth curve. By the joint constraints of the partition markers, registration factors and calibration relationship, the hardening layer depth distribution curve maintains the same shape in different cross sections and different structural areas. The quenching state result is output based on the hardening layer depth curve, realizing stable evaluation under the conditions of incomplete sampling of dual-frequency quenching detection signal and amplitude differences in different shaft diameters and step areas. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for detecting the dual-frequency quenching state of shaft-type metal parts according to the present invention. Figure 2This is a schematic diagram of a dual-frequency quenching state detection system for shaft-type metal parts according to the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. In the following detailed description, many specific details are set forth to provide a thorough understanding of the exemplary embodiments described. However, it will be apparent to those skilled in the art that the described embodiments may be practiced without some or all of these specific details. In other exemplary embodiments, well-known structures have not been described in detail to avoid unnecessarily obscuring the concepts of this disclosure. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention. Furthermore, the various aspects described in the embodiments may be combined arbitrarily without conflict.

[0019] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0020] Example 1

[0021] Figure 1 This disclosure illustrates a method for detecting the dual-frequency quenching state of shaft-type metal parts according to at least one embodiment, including: S10: The detection area is divided according to the geometric information of the shaft metal parts and the probe layout. During the quenching process, the first frequency detection signal and the second frequency detection signal are collected simultaneously in each detection area to obtain dual-frequency raw data. In this embodiment, the geometric information of shaft-type metal parts mainly refers to structural parameters such as shaft diameter, step position, step height, and length of each axial segment. The probe layout mainly refers to the installation position of the probe along the axial and circumferential directions, the gap between the probe and the shaft surface, and the probe coverage area. The detection area is first divided according to the geometric information and probe layout in order to limit the differences in electromagnetic field distribution caused by different shaft diameters and step areas to their respective detection areas, and to avoid mixing the response differences caused by structural changes into the same data segment, which would weaken the basis for subsequent amplitude comparability processing. Subsequently, the first frequency detection signal and the second frequency detection signal are simultaneously collected in each detection area during the quenching process to ensure that the responses of the two frequencies in the same detection area can be aligned on the time axis, which is convenient for subsequent construction of dual-frequency raw data with timestamps as indexes and connection with sampling marks.

[0022] For example, the axial direction is divided into three sections: D1 corresponds to a flat section detection area with a shaft diameter of 40 mm, D2 corresponds to a step transition detection area, and D3 corresponds to a flat section detection area with a shaft diameter of 30 mm. At the same time t0, the amplitude of the first frequency detection signal of D1 is 2.4 volts and the amplitude of the second frequency detection signal is 1.1 volts. At t1, 2.5 volts and 1.0 volts are collected, and at t2, 2.6 volts and 0.9 volts are collected. The dual-frequency amplitudes of the same detection area are combined in chronological order into a matrix 2.4 2.5 2.6, 1.1 1.0 0.9. This matrix is ​​used to represent the dual-frequency raw data of detection area D1 during the quenching process and serves as the input for subsequent sampling interval sequence and envelope transition position index extraction.

[0023] S20: For dual-frequency raw data, sampling marks are generated by combining the sampling interval and waveform changes in each detection zone, and the sampling marks are used to divide the stable sampling detection zone and the insufficient detection zone. For the dual-frequency raw data, sampling markers are generated by combining the sampling interval and waveform changes in each detection zone, including: Extract the sampling timestamps of the first frequency detection signal and the second frequency detection signal of each detection zone, and generate a sampling interval sequence according to the timestamp order; Locate the interval gap position index in the sampling interval sequence and write the interval gap position index into the detection area sampling interval table; Extract the envelope transition position index from the first frequency detection signal and the second frequency detection signal in each detection zone, and write the envelope transition position index into the waveform change table of the detection zone; A set of mutual verification anchor points is generated based on the sampling interval table and waveform change table of the detection area, and sampling markers are generated based on the set of mutual verification anchor points.

[0024] In this embodiment, the sampling interval sequence is used to characterize the spacing distribution of adjacent sampling points in the same detection area on the time axis. The interval gap position index is used to locate the segment where the sampling timestamp jumps and write it into the detection area sampling interval table. The envelope transition position index is used to locate the time position where the waveform shape of the first frequency detection signal and the second frequency detection signal changes and write it into the detection area waveform change table. The mutual verification anchor point set is used to cross-reference the gap time information in the detection area sampling interval table with the shape transition information in the detection area waveform change table and form a traceable anchor point link. First, the sampling interval sequence is used to determine whether there is a gap in the sampling. Then, the envelope transition position index is used to determine whether the waveform shape is distorted. Finally, sampling marks are generated on the basis that the two types of information can support each other. This is to avoid the deviation in the division of the stable sampling detection area and the insufficient detection area due to only relying on single time information or single shape information. The sampling marks can reflect both incomplete sampling and waveform distortion, thereby providing a verifiable basis for the detection area status for subsequent regional topology map constraints and missing area registration factor calculation.

[0025] For example, using the aforementioned D1 corresponding to the flat section detection area with a shaft diameter of 40 mm, D2 corresponding to the step transition detection area, and D3 corresponding to the flat section detection area with a shaft diameter of 30 mm, in D1, the timestamp sequence is t0, t1, t2, and the adjacent intervals are maintained at 0.02 seconds and 0.02 seconds. Therefore, the sampling interval sequence can be recorded as 0.02 seconds, 0.02 seconds, with the corresponding interval gap position index being an empty set. In D2, the timestamp sequence is t0, t1, t3, and the interval between t3 and t1 is 0.04 seconds. Therefore, the sampling interval sequence can be recorded as 0.02 seconds, 0.04 seconds, and the interval gap position index is recorded as {t1 to t3} and written into the detection area sampling interval table. Simultaneously, D1 and... The envelope transition position indices are extracted from the first and second frequency detection signals of D2, respectively. In D1, the envelope transition position index corresponds to {t1} and the transition times of the two frequencies are aligned. In D2, the envelope transition position index corresponds to {t1,t3} and the transition times of the two frequencies are not continuous. The gap time window {t1 to t3} of D2 is combined with the envelope transition position indices {t1,t3} to obtain the mutual verification anchor set {(t1,t1), (t3,t3)}, and the sampling marker is written accordingly. This causes D1 to enter the stable sampling detection region and D2 to enter the insufficient detection region. The mutual verification anchor set is used to characterize the traceable markers of the detection region on both the time and morphological links and serves as the input for subsequent processing.

[0026] A set of mutual verification anchor points is generated based on the sampling interval table and waveform change table of the detection area, and sampling markers are generated based on the set of mutual verification anchor points, including: Map the gap position index to the gap time window, and write the gap time window into the detection area time window table; Outside the time window of the gap, retrieve the time proximity pairs between the first frequency envelope turning point index and the second frequency envelope turning point index, and record the time proximity pairs as candidate anchor point pairs; Perform chain-connection processing on candidate anchor point pairs in chronological order, and record the chain-connection processing results as an anchor point chain; Filter anchor chain segments that cover the time window table of the detection area from the anchor chain, and aggregate the anchor chain segments into a mutual verification anchor set; Based on the mutual verification anchor point set, write the sampling mark into the sampling interval table and waveform change table of the detection area, and output the sampling mark.

[0027] In this embodiment, the gap time window is used to convert the time jump segment corresponding to the gap position index into a directly searchable time range and write it into the detection area time window table. The candidate anchor point pair is used to pair the first frequency envelope transition position index and the second frequency envelope transition position index according to their time proximity outside the gap time window to form dual-frequency comparable anchor point candidates. The anchor point chain is used to connect the candidate anchor point pairs into a continuous link in chronological order to check the continuity of the anchor points on the time axis. The anchor point chain segment is used to extract the continuous part that can cross the two sides of the gap time window from the anchor point chain and correspond to the detection area time window table. The mutual verification anchor point set is used to collect the anchor point chain segments and serve as the basis for writing the sampling mark. The processing order of first generating the gap time window, then searching for candidate anchor point pairs outside the window and forming the anchor point chain, and then filtering the anchor point chain segments that cover the gap boundary is to fix the correspondence between the sampling gap and the waveform transition on the same time link, so that the sampling mark can reflect both the gap boundary position and the transition connection state of the dual-frequency waveform at the gap boundary, thereby providing a traceable marking basis for the subsequent determination of the stable detection area and the insufficient detection area.

[0028] For example, using the sampling interval sequence of the aforementioned detection area D2 as 0.02 seconds and 0.04 seconds, and recording the interval gap position index as {t1 to t3}, {t1 to t3} is mapped to a gap time window and written into the detection area time window table. Within the time range outside the gap time window, the first frequency envelope transition position index {t1, t3} and the second frequency envelope transition position index {t1, t3} are retrieved and paired according to their temporal proximity to obtain candidate anchor point pairs {(t1, t1), (t3, t3)}. The candidate anchor point pairs are then sorted by time order. Anchor chain t1 to t3 is obtained by connecting the anchor chain. Anchor chain segments covering the gap time window boundary are selected from the anchor chain and aggregated into a mutual verification anchor set {(t1,t1), (t3,t3)}. Then, according to the mutual verification anchor set, gap markers are written for {t1 to t3} in the detection area sampling interval table and anchor markers are written for t1 and t3 in the detection area waveform change table. Sampling markers are output. The sampling markers are used to continue to reuse the partitioning results of D1 as the sampling stable detection area and D2 as the insufficient detection area in subsequent steps and to maintain the continuity of the instance link.

[0029] S30: Construct a regional topology map based on the adjacency relationship of the detection area and the insufficient detection area. Under the constraints of the regional topology map and sampling markers, calculate the registration factor of the missing area with reference to the dual-frequency original data in the stable sampling detection area. A regional topology map is constructed based on the adjacency relationships of the detection areas and in combination with insufficient detection areas, including: Based on the geometric information of shaft-type metal parts, axial adjacent pairs of the detection area are generated, and based on the probe layout, circumferential adjacent pairs of the detection area are generated. Perform duplicate elimination processing on axial and circumferential adjacent pairs in the detection area, and record the duplicate elimination processing results in the adjacency table of the detection area; Read the detection region partitioning table and extract the adjacent boundary pairs of the insufficient detection regions from the detection region adjacency table, and generate the missing region set based on the adjacent boundary pairs; Construct a regional topology graph using the detection area as nodes and the detection area adjacency list as the connection relationship, and write the missing area set into the node attributes of the regional topology graph; Based on the sampling tag table, write sampling tags into the node attributes of the regional topology map and output the regional topology map.

[0030] In this embodiment, the adjacency relationship of the detection area is used to characterize the adjacent connection structure of the detection area in the axial and circumferential directions. The axial adjacency pairs are determined by the geometric information of the shaft-like metal parts to determine the connection relationship of adjacent axial segments. The circumferential adjacency pairs are determined by the probe layout to determine the adjacency relationship of different circumferential detection areas at the same axial position. The purpose of collecting the axial and circumferential adjacency pairs to form the detection area adjacency table is to establish the boundary positioning of the subsequent insufficient detection area on the lookupable connection relationship, so that the insufficient detection area is no longer processed in a discrete point manner but organized in a spatially connected structure manner. The detection area partition table is read and the insufficient detection area is extracted. The adjacent boundary pairs in the detection area adjacency list are used to extract the boundary connections between the insufficient detection area and its adjacent stable sampling detection area, and use them as the basis for generating the missing area set. The missing area set is used to merge the insufficient detection areas that are connected in the detection area adjacency list into the missing area identifier and the list of the insufficient detection areas to which they belong. This allows the region topology map to describe the overall connection relationship of the detection area and carry the missing area set and sampling mark in the node attributes. This facilitates the subsequent calculation of the registration factor of the missing area under the constraints of the region topology map and sampling mark, and maintains the consistency of the structural expression of the step and axis diameter change areas.

[0031] For example, using the aforementioned D1 corresponding to the flat section detection area with a shaft diameter of 40 mm, D2 corresponding to the step transition detection area, and D3 corresponding to the flat section detection area with a shaft diameter of 30 mm, and using the aforementioned partitioning results, D1 is marked as the sampling stable detection area and D2 is marked as the insufficient detection area. The axial adjacency pairs are written as {(D1,D2), (D2,D3)} and the circumferential adjacency pairs are written as {(D1,D1), (D2,D2), (D3,D3)}, and after being collected, a detection area adjacency table {(D1,D2), (D2,D3)} is formed, and the detection area is partitioned. Extract the insufficient detection region D2 from the table and locate its adjacent boundary pairs {(D1,D2), (D2,D3)} in the detection region adjacency table. Generate the missing region set {(missing region H1,{D2})} based on the adjacent boundary pairs and write the missing region set into the node attributes of the regional topology map. Then, according to the sampling label table, write D1 into the sampling stable detection region label and D2 into the insufficient detection region label and output the regional topology map. The regional topology map is used in subsequent steps to continue to use D1 as a reference in the same example link and calculate the missing region registration factor for the missing region H1.

[0032] Read the detection region partitioning table and extract the adjacent boundary pairs of the insufficient detection regions from the detection region adjacency table. Generate a set of missing regions based on the adjacent boundary pairs, including: Extract the adjacency pairs related to the insufficient detection area from the adjacency list of the detection area, and aggregate the adjacency pairs into a set of insufficient edges; Based on the insufficient edge set, perform connectivity aggregation processing on the insufficient detection region, and record the connectivity aggregation processing results as an aggregation cluster table; For each cluster in the cluster table, collect the insufficient detection areas and generate a missing area identifier for each cluster; Write the missing region identifier and its corresponding list of insufficient detection regions into the missing region table, then aggregate the missing region tables into a missing region set, output the missing region set and write it into the node attributes of the region topology graph.

[0033] Based on the insufficient edge set, connectivity aggregation processing is performed on the insufficient detection region, and the results of the connectivity aggregation processing are recorded in an aggregation cluster table, including: Select any insufficient detection region from the insufficient edge set as the starting node, and write the starting node into the queue to be expanded; Take the head node from the queue to be expanded and search the adjacent insufficient detection area of ​​the head node, and write the search result into the connectivity record table; Add the unqueued insufficient detection areas in the connectivity record table to the queue to be expanded, and write access flags to the processed insufficient detection areas. When the queue to be expanded is empty, output the connectivity record table as an aggregate cluster and write the aggregate cluster to the aggregate cluster table.

[0034] In this embodiment, the insufficient edge set is used to collect the adjacency pairs related to the insufficient detection area in the detection area adjacency table, thereby fixing the connectivity relationship between insufficient detection areas into a searchable data structure. The connectivity aggregation process is used to merge interconnected insufficient detection areas into the same aggregation cluster based on the insufficient edge set and record it in the aggregation cluster table. The missing area identifier is used to establish a referenceable region number for each aggregation cluster and form a missing area table with the list of its corresponding insufficient detection areas. The missing area set is used to collect the missing area table and write it into the node attributes of the regional topology graph so that the missing area can be used as a whole object in the calculation of the missing area registration factor. The processing order of first extracting the insufficient edge set, then performing connectivity aggregation, then generating the missing area identifier and collecting it into the missing area set is to avoid the insufficient detection areas being mixed or split when they are distributed in multiple segments in space, so that the missing area set can stably represent the connectivity boundary of the insufficient detection area and maintain consistency with the connection relationship of the regional topology graph.

[0035] For example, using the aforementioned adjacency list {(D1,D2), (D2,D3)} and the result that D2 is an insufficient detection region in the detection region partitioning table, we extract the adjacency pairs related to the insufficient detection region D2 from the detection region adjacency list to obtain the insufficient edge set {(D1,D2), (D2,D3)}. We select the starting node D2 from the insufficient edge set and write it into the queue to be expanded. When retrieving the head node D2 from the queue and searching for adjacent insufficient detection regions, we only retain the adjacent nodes within the insufficient detection region set. Therefore, the adjacent insufficient detection regions are an empty set, and the search results are written into the connectivity record table {D2}. After the extended queue is empty, the connectivity record table is output as an aggregation cluster and written into the aggregation cluster table {cluster1:{D2}}. The aggregation cluster 1 in the aggregation cluster table is aggregated to form the list of insufficient detection regions {D2} and a missing region identifier H1 is generated. The missing region identifier H1 and the list of insufficient detection regions {D2} are written into the missing region table and aggregated into a missing region set {(H1,{D2})} and written into the node attributes of the regional topology graph. The missing region set continues to be used as the calculation object of the missing region registration factor in subsequent steps and remains consistent with the example link where D1 is a sampled stable detection region.

[0036] Under the constraints of the regional topology map and sampling markers, the registration factor for the missing region is calculated using the original dual-frequency data within the stable sampling detection area as a reference, including: Extract the sampling stable detection region adjacent to the insufficient detection region from the regional topology map, and record the sampling stable detection region as the boundary stable region; Within the boundary stable region, the first frequency detection signal segment and the second frequency detection signal segment are extracted according to the sampling mark, and the first frequency detection signal segment and the second frequency detection signal segment are written into the reference segment table; Establish a segment correspondence between the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table, and write the segment correspondence into the segment correspondence table; Generate an amplitude mapping table based on the segment correspondence table, and establish an association record between the amplitude mapping table and the boundary stable region in the regional topology map; The missing region registration factor is calculated based on the amplitude mapping table under the constraints of the regional topology map and sampling markers.

[0037] In this embodiment, the boundary stabilization region is used to locate the adjacent boundary of the insufficient detection region in the regional topology map, which can serve as a reference sampling stabilization detection region. This limits the source of reference data to the detection region where the sampling marks are continuous and the waveform has no gaps. The reference segment table is used to record the first frequency detection signal segment and the second frequency detection signal segment extracted according to the sampling marks in the boundary stabilization region as a reusable segment set in timestamp order. The segment correspondence table is used to form a segment correspondence between the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table according to the same time position, so that the two frequency signals have a comparable relationship within the same reference segment. The amplitude mapping table is used to organize the dual-frequency amplitude correspondence relationship in the segment correspondence table into a mapping relationship and establish an association record with the boundary stabilization region, so that the mapping relationship can be referenced in the boundary propagation path of the regional topology map. The missing area registration factor is used to convert the amplitude scale offset of the insufficient detection region under the structural regional differences into an applicable adjustment parameter according to the amplitude mapping table, so as to maintain the comparability of response amplitudes between different shaft diameters and step regions and reduce the interference of incomplete sampling on subsequent depth estimation when adjusting the dual-frequency raw data of the insufficient detection region.

[0038] For example, using the aforementioned regional topology map where D2 is an insufficient detection region and the missing region is identified as H1, and D1 is a stable sampling detection region adjacent to D2, we extract the stable sampling detection region adjacent to the insufficient detection region D2 in the regional topology map and record the boundary stable region as {D1}. Within the boundary stable region D1, we extract reference segments according to the sampling markers and write them into a reference segment table. The reference segment table can be written as {(D1, first frequency: {2.4, 2.5, 2.6}, second frequency: {1.1, 1.0, 0.9})}, and keep the timestamps as t0, t1, t2. We then establish a segment correspondence table for the reference segment table and write it into a segment correspondence table. The segment correspondence table can be written as {(t0, 2... The amplitude mapping table is generated based on the segment correspondence table and associated with the boundary stable region {D1}. The amplitude mapping table can be written as {(first frequency 2.4, second frequency 1.1), (first frequency 2.5, second frequency 1.0), (first frequency 2.6, second frequency 0.9)}. Then, under the constraints of the regional topology map and sampling markers, the missing region registration factor corresponding to the missing region H1 is calculated according to the amplitude mapping table and the missing region registration factor is obtained. The missing region registration factor is used in subsequent steps to adjust the dual-frequency raw data of the insufficient detection region D2 and keep it consistent with the aforementioned example link.

[0039] Establish a segment correspondence between the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table, and write the segment correspondence into the segment correspondence table, including: Read the reference segment table and extract the timestamp sequence of the first frequency detection signal segment and the timestamp sequence of the second frequency detection signal segment, and generate candidate segment pairs based on the timestamp proximity relationship; Extract the inflection point index of the first frequency detection signal segment and the inflection point index of the second frequency detection signal segment from the candidate segment pairs, and generate segment anchor point pairs based on the inflection point indexes; Based on the fragment anchor point pair, perform pairing convergence processing on the candidate fragment pairs, record the pairing convergence processing results as fragment pairings, and write the fragment pairings into the fragment correspondence table.

[0040] In this embodiment, candidate segment pairs are used to form a pairable candidate set by the proximity of the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table according to the timestamp proximity. The inflection point index is used to characterize the shape turning position of the waveform inside the segment on the time axis and serve as an internal constraint for the alignment between segments. The segment anchor point pair is used to establish a correspondence between the inflection point index of the first frequency detection signal segment and the inflection point index of the second frequency detection signal segment to limit the anchor point position of the segment pairing. The pairing convergence process is used to filter out candidates that do not meet the segment anchor point pair constraints from the candidate segment pairs and retain candidates whose anchor point positions can be compared. The segment pair is used to record the correspondence between the first frequency detection signal segment and the second frequency detection signal segment after being filtered by the constraints and write it into the segment correspondence table. The processing order of first limiting the candidate range by the proximity of timestamps, then generating segment anchor point pairs by inflection point indexes, and then performing pairing convergence process is to avoid misalignment of the internal shape of the segment due to pairing only by timestamps, so that the segment correspondence table can stably carry the correspondence relationship of the dual frequency signals in the reference segment and provide a traceable data foundation for the subsequent amplitude mapping table generation.

[0041] For example, using the aforementioned reference segment table {(D1, first frequency: {2.4, 2.5, 2.6}, second frequency: {1.1, 1.0, 0.9})} with timestamp sequences t0, t1, t2, the reference segment table is read, and the timestamp sequences of the first frequency detection signal segment {t0, t1, t2} and the second frequency detection signal segment {t0, t1, t2} are extracted. Candidate segment pairs {(first frequency segment t0 to t2, second frequency segment t0 to t2)} are generated based on the timestamp proximity. The first frequency segment is then extracted from the candidate segment pairs. The inflection point index of the first frequency detection signal segment and the inflection point index of the second frequency detection signal segment are used to generate segment anchor point pairs. The segment anchor point pairs can be written as {(t1,t1)} and used as alignment constraints. Based on the segment anchor point pairs, the candidate segment pairs are paired and converged to obtain the segment correspondence {(t0,2.4,1.1), (t1,2.5,1.0), (t2,2.6,0.9)}. The segment correspondence table is used in subsequent steps to generate the amplitude mapping table and maintains continuity with the example links of the aforementioned boundary stable region D1 and insufficient detection region D2.

[0042] S40: Under the constraints of the regional topology map and sampling markers, the sampling stable detection area in the dual-frequency raw data is substituted into the preset quenching calibration relationship to obtain the local hardening depth. The insufficient detection area in the dual-frequency raw data is adjusted by the missing area registration factor, and the hardening layer depth in the insufficient detection area is calculated by combining the local hardening depth to obtain the hardening layer depth curve. The quenching state result is output based on the hardening layer depth curve.

[0043] The insufficient detection area in the dual-frequency raw data is adjusted using the missing region registration factor, and the hardened layer depth within the insufficient detection area is calculated in conjunction with the local hardening depth to obtain the hardened layer depth curve, including: Read the local hardening depth and write it into the depth anchor point table according to the spatial position of the detection area; Extract the stable sampling detection regions adjacent to the insufficient detection regions in the regional topology map, and write the adjacency relationship between the stable sampling detection regions and the insufficient detection regions into the boundary relationship table; Based on the boundary relationship table, the local hardening depth in the depth anchor point table is processed by boundary propagation, and the result of the boundary propagation is written into the boundary depth table. Write the dual-frequency raw data corresponding to the insufficient detection area, adjusted by the missing area registration factor, into the insufficient area input table, and establish an index association between the insufficient area input table and the boundary depth table; The depth of the hardened layer in the insufficient detection zone is calculated based on the input table of the insufficient zone and the boundary depth table, and the hardened layer depth curve is generated by combining them.

[0044] In this embodiment, a preset quenching calibration relationship is used to convert the dual-frequency raw data within the stable sampling detection area into a local hardening depth that can be used for depth expression. This local hardening depth then becomes a reference anchor point for estimating the depth of the insufficient detection area. A depth anchor point table records the local hardening depth according to the spatial location of the detection area and maintains a correspondence with the node positions on the region topology map. A boundary relationship table records the adjacency relationship between the insufficient detection area and its adjacent stable sampling detection areas to fix the boundary reference source. A boundary depth table organizes the local hardening depths in the depth anchor table along the boundary relationship table into a boundary depth record that can be referenced by the insufficient detection area. An insufficient area input table records the insufficient... The dual-frequency raw data corresponding to the detection area, adjusted by the registration factor of the missing area, is indexed and associated with the boundary depth table. The processing order is to first obtain the local hardening depth from the sampled and stable detection area, then form a depth anchor point table and a boundary relationship table, generate a boundary depth table, and then introduce the insufficient area input table. This is to ensure that even in the case of incomplete sampling and waveform distortion, the depth calculation path of the insufficient detection area is still constrained by the boundary reference in the regional topology map. This allows the dual-frequency raw data of the insufficient detection area to be combined with the boundary depth record on the same spatial link after the registration factor of the missing area is adjusted, thereby obtaining a continuously organized hardened layer depth curve and using it to output the quenching state result.

[0045] For example, using the aforementioned boundary stable region as {D1}, insufficient detection region as D2, and missing region as H1, the dual-frequency original data matrix of D1 (2.4, 2.5, 2.6, 1.1, 1.0, 0.9) is substituted into the preset quenching calibration relationship to obtain the local hardening depth and written into the depth anchor table. The depth anchor table can be written as {(D1, 1.6 mm)}, maintaining the spatial position as the axial flat segment region. The sampling stable detection region adjacent to D2 in the region topology map is extracted as D1, and the adjacency relationship is written into the boundary relationship table. The boundary relationship table can be written as {(D2, D1)}. Based on the boundary relationship table, the depth anchor table is organized to obtain the boundary depth table, which can be written as {(D2 boundary corresponds to D1, 1.6 mm)}. 6 mm)}, write the dual-frequency raw data corresponding to D2, adjusted by the missing region registration factor, into the insufficient region input table and establish an index association with the boundary depth table. The insufficient region input table can be written as {(D2, first frequency: {2.2, 2.3}, second frequency: {1.0, 0.95})} and establish an association index with {(D2 boundary corresponds to D1, 1.6 mm)}. Calculate the hardened layer depth of D2 based on the insufficient region input table and the boundary depth table, and generate a hardened layer depth curve by combining it with the 1.6 mm of D1 according to its spatial position. The hardened layer depth curve can be written as {(D1, 1.6 mm), (D2, 1.4 mm)} and continue to retain the extension position of D3 to support the output quenching state result.

[0046] In this embodiment, sampling incompleteness and waveform distortion are located at the detection area granularity using sampling markers, and the detection area is divided into a stable sampling detection area and an insufficient detection area based on this. Then, the adjacent boundary relationship between the insufficient detection area and the stable sampling detection area is solidified into a referable structure using a regional topology map. This allows the registration factor of the missing area to scale the original dual-frequency data of the insufficient detection area based on the dual-frequency correspondence of the reference segment in the stable boundary area. This maintains comparability constraints even when there are differences in response amplitude in different shaft diameters and step areas. Furthermore, the depth anchor point table and boundary depth table are used to transfer the local hardening depth along the boundary relationship to the depth estimation process of the insufficient detection area. This ensures that the hardened layer depth curve maintains morphological consistency when generated across cross sections and across structural areas. As a result, the quenching state results still have traceable partitioning basis and continuous depth distribution expression in detection scenarios with local signal attenuation or interruption.

[0047] Example 2

[0048] Please see Figure 2 As shown, based on the same inventive concept, this embodiment discloses a dual-frequency quenching state detection system for shaft-type metal parts. For details not covered in this embodiment, please refer to the relevant sections of Embodiment 1. The system includes: The partition acquisition module is used to divide the detection area according to the geometric information of the shaft metal parts and the probe layout. During the quenching process, the first frequency detection signal and the second frequency detection signal are acquired simultaneously in each detection area to obtain dual-frequency raw data. Marking and Division Module: Used to generate sampling marks in each detection zone based on the comprehensive sampling interval and waveform changes of dual-frequency raw data, and to divide the stable sampling detection zone and the insufficient detection zone according to the sampling marks; Topology registration module: used to construct a regional topology map based on the adjacency relationship of the detection area and in combination with the insufficient detection area. Under the constraints of the regional topology map and sampling markers, the registration factor of the missing area is calculated with reference to the dual-frequency original data in the sampled stable detection area. The quenching depth calculation module is used to substitute the sampled stable detection area in the dual-frequency raw data into the preset quenching calibration relationship under the constraints of the region topology map and sampling markers to obtain the local hardening depth. It uses the missing area registration factor to adjust the insufficient detection area in the dual-frequency raw data, and calculates the hardened layer depth in the insufficient detection area in combination with the local hardening depth to obtain the hardened layer depth curve. Based on the hardened layer depth curve, it outputs the quenching state result.

[0049] The detailed description above, in conjunction with the accompanying drawings, describes examples but does not represent all examples that can be implemented or fall within the scope of the claims. The terms “example” and “exemplary” are used in this specification to mean “serving as an example, instance or illustration” and do not mean “superior to or better than other examples”.

[0050] Throughout this specification, the phrase "an embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with that embodiment is included in at least one embodiment of the invention. Therefore, the use of these phrases may refer to more than one embodiment. Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0051] It should also be noted that these embodiments may be described as processes depicted as flowcharts, structural diagrams, or block diagrams. Although a flowchart may describe the operations as sequential processes, many of these operations can be performed in parallel or concurrently, and the order of these operations may be rearranged.

Claims

1. A method for detecting the dual-frequency quenching state of shaft-type metal parts, characterized in that, include: The detection zones are divided according to the geometric information of the shaft-type metal parts and the probe layout. During the quenching process, the first frequency detection signal and the second frequency detection signal are collected simultaneously in each detection zone to obtain dual-frequency raw data. For the dual-frequency raw data, sampling marks are generated by combining the sampling interval and waveform changes in each detection zone. Based on the sampling marks, the detection zone with stable sampling and the detection zone with insufficient sampling are divided. Based on the adjacency relationship of the detection areas and combined with the insufficient detection areas, a regional topology map is constructed. Under the constraints of the regional topology map and sampling markers, the registration factor of the missing area is calculated with reference to the dual-frequency original data in the sampling stable detection area. Under the constraints of the regional topology map and sampling markers, the sampling stable detection area in the dual-frequency raw data is substituted into the preset quenching calibration relationship to obtain the local hardening depth. The missing area registration factor is used to adjust the insufficient detection area in the dual-frequency raw data, and the hardening layer depth in the insufficient detection area is calculated in combination with the local hardening depth to obtain the hardening layer depth curve. The quenching state result is output based on the hardening layer depth curve.

2. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 1, characterized in that, For the dual-frequency raw data, sampling markers are generated by combining the sampling interval and waveform changes in each detection zone, including: Extract the sampling timestamps of the first frequency detection signal and the second frequency detection signal of each detection zone, and generate a sampling interval sequence according to the timestamp order; Locate the interval gap position index in the sampling interval sequence and write the interval gap position index into the detection area sampling interval table; Extract the envelope transition position index from the first frequency detection signal and the second frequency detection signal in each detection zone, and write the envelope transition position index into the waveform change table of the detection zone; A set of mutual verification anchor points is generated based on the sampling interval table and waveform change table of the detection area, and sampling markers are generated based on the set of mutual verification anchor points.

3. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 2, characterized in that, A set of mutual verification anchor points is generated based on the sampling interval table and waveform change table of the detection area, and sampling markers are generated based on the set of mutual verification anchor points, including: Map the gap position index to the gap time window, and write the gap time window into the detection area time window table; Outside the time window of the gap, retrieve the time proximity pairs between the first frequency envelope turning point index and the second frequency envelope turning point index, and record the time proximity pairs as candidate anchor point pairs; Perform chain-connection processing on candidate anchor point pairs in chronological order, and record the chain-connection processing results as an anchor point chain; Filter anchor chain segments that cover the time window table of the detection area from the anchor chain, and aggregate the anchor chain segments into a mutual verification anchor set; Based on the mutual verification anchor point set, write the sampling mark into the sampling interval table and waveform change table of the detection area, and output the sampling mark.

4. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 2, characterized in that, A regional topology map is constructed based on the adjacency relationships of the detection areas and in combination with insufficient detection areas, including: Based on the geometric information of shaft-type metal parts, axial adjacent pairs of the detection area are generated, and based on the probe layout, circumferential adjacent pairs of the detection area are generated. Perform duplicate elimination processing on axial and circumferential adjacent pairs in the detection area, and record the duplicate elimination processing results in the adjacency table of the detection area; Read the detection region partitioning table and extract the adjacent boundary pairs of the insufficient detection regions from the detection region adjacency table, and generate the missing region set based on the adjacent boundary pairs; Construct a regional topology graph using the detection area as nodes and the detection area adjacency list as the connection relationship, and write the missing area set into the node attributes of the regional topology graph; Based on the sampling tag table, write sampling tags into the node attributes of the regional topology map and output the regional topology map.

5. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 4, characterized in that, Read the detection region partitioning table and extract the adjacent boundary pairs of the insufficient detection regions from the detection region adjacency table. Generate a set of missing regions based on the adjacent boundary pairs, including: Extract the adjacency pairs related to the insufficient detection area from the adjacency list of the detection area, and aggregate the adjacency pairs into a set of insufficient edges; Based on the insufficient edge set, perform connectivity aggregation processing on the insufficient detection region, and record the connectivity aggregation processing results as an aggregation cluster table; For each cluster in the cluster table, collect the insufficient detection areas and generate a missing area identifier for each cluster; Write the missing region identifier and its corresponding list of insufficient detection regions into the missing region table, then aggregate the missing region tables into a missing region set, output the missing region set and write it into the node attributes of the region topology graph.

6. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 5, characterized in that, Based on the insufficient edge set, connectivity aggregation processing is performed on the insufficient detection region, and the results of the connectivity aggregation processing are recorded in an aggregation cluster table, including: Select any insufficient detection region from the insufficient edge set as the starting node, and write the starting node into the queue to be expanded; Take the head node from the queue to be expanded and search the adjacent insufficient detection area of ​​the head node, and write the search result into the connectivity record table; Add the unqueued insufficient detection areas in the connectivity record table to the queue to be expanded, and write access flags to the processed insufficient detection areas. When the queue to be expanded is empty, output the connectivity record table as an aggregate cluster and write the aggregate cluster to the aggregate cluster table.

7. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 4, characterized in that, Under the constraints of the regional topology map and sampling markers, the registration factor for the missing region is calculated using the original dual-frequency data within the stable sampling detection area as a reference, including: Extract the sampling stable detection region adjacent to the insufficient detection region from the regional topology map, and record the sampling stable detection region as the boundary stable region; Within the boundary stable region, the first frequency detection signal segment and the second frequency detection signal segment are extracted according to the sampling mark, and the first frequency detection signal segment and the second frequency detection signal segment are written into the reference segment table; Establish a segment correspondence between the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table, and write the segment correspondence into the segment correspondence table; Generate an amplitude mapping table based on the segment correspondence table, and establish an association record between the amplitude mapping table and the boundary stable region in the regional topology map; The missing region registration factor is calculated based on the amplitude mapping table under the constraints of the regional topology map and sampling markers.

8. The method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 7, characterized in that, Establish a segment correspondence between the first frequency detection signal segment and the second frequency detection signal segment in the reference segment table, and write the segment correspondence into the segment correspondence table, including: Read the reference segment table and extract the timestamp sequence of the first frequency detection signal segment and the timestamp sequence of the second frequency detection signal segment, and generate candidate segment pairs based on the timestamp proximity relationship; Extract the inflection point index of the first frequency detection signal segment and the inflection point index of the second frequency detection signal segment from the candidate segment pairs, and generate segment anchor point pairs based on the inflection point indexes; Based on the fragment anchor point pair, perform pairing convergence processing on the candidate fragment pairs, record the pairing convergence processing results as fragment pairings, and write the fragment pairings into the fragment correspondence table.

9. A method for detecting the dual-frequency quenching state of shaft-type metal parts according to claim 8, characterized in that, The insufficient detection area in the dual-frequency raw data is adjusted using the missing region registration factor, and the hardened layer depth within the insufficient detection area is calculated in conjunction with the local hardening depth to obtain the hardened layer depth curve, including: Read the local hardening depth and write it into the depth anchor point table according to the spatial position of the detection area; Extract the stable sampling detection regions adjacent to the insufficient detection regions in the regional topology map, and write the adjacency relationship between the stable sampling detection regions and the insufficient detection regions into the boundary relationship table; Based on the boundary relationship table, the local hardening depth in the depth anchor point table is processed by boundary propagation, and the result of the boundary propagation is written into the boundary depth table. Write the dual-frequency raw data corresponding to the insufficient detection area, adjusted by the missing area registration factor, into the insufficient area input table, and establish an index association between the insufficient area input table and the boundary depth table; The depth of the hardened layer in the insufficient detection zone is calculated based on the input table of the insufficient zone and the boundary depth table, and the hardened layer depth curve is generated by combining them.

10. A dual-frequency quenching state detection system for shaft-type metal parts, used to implement the dual-frequency quenching state detection method for shaft-type metal parts according to any one of claims 1-9, characterized in that, include: The partition acquisition module is used to divide the detection area according to the geometric information of the shaft metal parts and the probe layout. During the quenching process, the first frequency detection signal and the second frequency detection signal are acquired simultaneously in each detection area to obtain dual-frequency raw data. Marking and Division Module: Used to generate sampling marks in each detection zone based on the comprehensive sampling interval and waveform changes of dual-frequency raw data, and to divide the stable sampling detection zone and the insufficient detection zone according to the sampling marks; Topology registration module: used to construct a regional topology map based on the adjacency relationship of the detection area and in combination with the insufficient detection area. Under the constraints of the regional topology map and sampling markers, the registration factor of the missing area is calculated with reference to the dual-frequency original data in the sampled stable detection area. The quenching depth calculation module is used to substitute the sampled stable detection area in the dual-frequency raw data into the preset quenching calibration relationship under the constraints of the region topology map and sampling markers to obtain the local hardening depth. It uses the missing area registration factor to adjust the insufficient detection area in the dual-frequency raw data, and calculates the hardened layer depth in the insufficient detection area in combination with the local hardening depth to obtain the hardened layer depth curve. Based on the hardened layer depth curve, it outputs the quenching state result.

Images