Search

CN-121615049-B - Quality anomaly prediction and tracing method based on riveting process

CN121615049BCN 121615049 BCN121615049 BCN 121615049BCN-121615049-B

Abstract

The invention discloses a quality anomaly prediction and tracing method based on a riveting process, in particular to the field of data processing and quality prediction and tracing of press riveting equipment, which comprises the steps of acquiring force data, displacement data and time data acquired by the press riveting equipment in one riveting process, and constructing a riveting process curve corresponding to the riveting process based on the acquired data; binding the riveting process curve with the equipment identifier, the technological parameter version, the die life, the material batch and the time stamp corresponding to the riveting, and generating a riveting event corresponding to the riveting. The riveting process curve is systematically bound with the equipment state, the service life of the die, the technological parameter version, the material batch and other contextual information, and mutation monitoring, risk prediction and source backtracking analysis are performed on fingerprint evolution of the continuous riveting process.

Inventors

  • WANG YUNQING

Assignees

  • 一浦莱斯精密技术(深圳)有限公司

Dates

Publication Date
20260508
Application Date
20260129

Claims (8)

  1. 1. A quality anomaly prediction and tracing method based on a riveting process is characterized by comprising the following steps: S1, acquiring force data, displacement data and time data acquired by riveting equipment in a one-time riveting process, and constructing a riveting process curve corresponding to the one-time riveting based on the acquired data; s2, executing process decomposition calculation on a riveting process curve in a riveting event, decomposing the riveting process curve into material fingerprints and equipment fingerprints, wherein the material fingerprints are used for representing material forming characteristics in the riveting process, and the equipment fingerprints are used for representing structural response and driving characteristics of riveting equipment; S3, additionally writing the currently generated riveting event into an event sequence according to the time sequence of occurrence of riveting, generating an event summary based on the summary value of the previous riveting event and the content of the current riveting event in the event sequence, and recording the event summary in the current riveting event to form an event chain with an association check relation between adjacent riveting events; S4, based on the process fingerprints of each riveting event in the event chain, similarity calculation is respectively carried out on the material fingerprints and the equipment fingerprints, abrupt indexes taking the material fingerprints and the equipment fingerprints as index keys are constructed, and continuity monitoring is carried out on the process fingerprints of adjacent riveting events in the event chain by utilizing the abrupt indexes so as to determine whether structural abrupt changes of the process fingerprints exist.
  2. 2. The quality anomaly prediction and tracing method based on the riveting process according to claim 1, wherein the quality anomaly prediction and tracing method based on the riveting process is characterized in that: Further comprising S5: S5-1, when the mutation index detects that the structural mutation of the process fingerprints exists in the event chain or the offset between the process fingerprints is larger than a preset offset threshold value, determining that the corresponding riveting event is a target riveting event; The method comprises the steps of selecting a preset number of historical riveting events taking a target riveting event as an end point in time sequence in an event chain to form a historical event set, respectively reading material fingerprints and equipment fingerprints of each historical riveting event in the historical event set, sequentially performing item-by-item differential operation on the material fingerprints of each historical riveting event by taking the material fingerprints and the equipment fingerprints corresponding to the target riveting event as reference fingerprints to obtain a material differential sequence arranged in time sequence, and sequentially performing item-by-item differential operation on the equipment fingerprints of each historical riveting event to obtain an equipment differential sequence arranged in time sequence; S5-2, calculating the change increment between two adjacent items according to the time sequence of the material differential sequence to form a material change increment sequence, calculating the average value of the absolute value of the material change increment sequence to be used as a material offset amplitude and calculating the difference value between the first item and the last item of the material change increment sequence to be used as a material offset trend, calculating the change increment between two adjacent items according to the time sequence of the equipment differential sequence to form an equipment change increment sequence, calculating the average value of the absolute value of the equipment change increment sequence to be used as an equipment offset amplitude and calculating the difference value between the first item and the last item of the equipment change increment sequence to be used as an equipment offset trend, and forming an offset evolution quantity set by the material offset amplitude, the material offset trend, the equipment offset amplitude and the equipment offset trend together; S5-3, comparing the material offset amplitude in the offset evolution quantity set with a first amplitude threshold value, and generating a material amplitude abnormality mark when the first amplitude threshold value is exceeded; Comparing the device offset amplitude with a second amplitude threshold, and generating a device amplitude abnormality flag when the second amplitude threshold is exceeded; comparing the equipment offset trend with a second trend threshold, and generating equipment trend abnormality marks when the second trend threshold is exceeded; Executing combination judgment on the material amplitude abnormal mark, the material trend abnormal mark, the equipment amplitude abnormal mark and the equipment trend abnormal mark, when a preset combination condition is met, generating a quality abnormal prediction risk of a subsequent riveting process as a trigger state, and determining a target riveting event as a prediction starting point event; S5-4, selecting retrospective riveting events one by one according to time reverse sequence in an event chain by taking the prediction starting point event as a starting point, repeatedly executing fingerprint differential sequence construction of S5-1 and offset evolution quantity calculation of S5-2 for each retrospective riveting event, and regenerating a corresponding prediction risk according to a prediction risk generation rule of S5-3; And when the backtracking riveting event with the predicted risk being the trigger state appears for the first time, determining the backtracking riveting event as a source riveting event, and reading the equipment identifier, the technological parameter version, the die life and the material batch bound in the source riveting event as a backtracking result corresponding to the quality abnormity predicted risk and outputting the backtracking result.
  3. 3. The quality anomaly prediction and tracing method based on riveting process according to claim 2, wherein the quality anomaly prediction and tracing method based on riveting process is characterized in that: The step S1 includes: S1-1, in a one-time riveting process, synchronous sampling is carried out on force data, displacement data and time data according to a unified sampling clock, corresponding time indexes are respectively added to the acquired force data and displacement data, and a force sampling sequence and a displacement sampling sequence which are in one-to-one correspondence with each other according to the time indexes are formed; S1-2, performing threshold starting judgment on the force sampling sequence to determine a riveting contact starting point, and performing slope change judgment on the displacement sampling sequence to determine a forming end point; s1-3, performing one-to-one pairing on a force sampling sequence and a displacement sampling sequence in the effective sampling section according to time indexes to form a force-displacement data pair sequence driven by the time indexes, and taking the force-displacement data pair sequence as a riveting process curve corresponding to riveting; s1-4, uniformly binding the riveting process curve with the equipment identifier, the technological parameter version, the die life, the material batch and the time stamp corresponding to the riveting, and generating a riveting event corresponding to the riveting.
  4. 4. A quality anomaly prediction and tracing method based on a riveting process according to claim 3, wherein: the step S2 includes: s2-1, acquiring a corresponding riveting process curve in a riveting event, determining a curve section between a riveting contact starting point and a forming end point determined in the S1-2 as a loading section, and determining a curve section corresponding to a curve section after the forming end point until a force value is reduced to a preset unloading threshold value as a return section; S2-2, rearranging the force-displacement data pair sequence in the return section according to a time index reverse order, and performing symbol mapping on the displacement value after reverse order in a mode consistent with the displacement change direction of the loading section to obtain a return mapping sequence aligned with the displacement direction of the loading section; s2-3, aligning the force-displacement data pair sequence in the loading section and the return mapping sequence point by point according to the time index, averaging the corresponding force values under the same time index to form an equipment projection curve, and determining the equipment projection curve as an equipment fingerprint of the riveting event; S2-4, taking the equipment fingerprint as a reference, performing force value differential operation on the riveting process curve in the loading section point by point according to time indexes to form a material residual curve, and determining the material residual curve as the material fingerprint of the riveting event; s2-5, taking the material fingerprint and the equipment fingerprint as process fingerprints of riveting events.
  5. 5. The method for predicting and tracing quality anomalies based on riveting processes according to claim 4, wherein the method comprises the steps of: the step S3 includes: s3-1, writing a currently generated riveting event into the tail of the event sequence according to the time sequence of occurrence of riveting, and recording the event sequence number of the currently generated riveting event in the event sequence for the current riveting event; s3-2, sequentially performing abstract extraction on equipment identifiers, process parameter versions, mold lives, material batches, time stamps and riveting process curves bound in the current riveting event, and splicing according to a preset sequence to generate a current content abstract; S3-3, if the current riveting event is the first event, performing summary operation by taking the current content summary as a basic input string to generate a basic event summary, otherwise, reading the event summary recorded in the previous riveting event in the event sequence as a preamble summary, splicing the preamble summary and the current content summary to form a basic input string, and performing summary operation on the basic input string to generate a basic event summary.
  6. 6. The method for predicting and tracing quality anomalies based on riveting processes according to claim 5, wherein the method comprises the steps of: in S3, further comprising: S3-4, determining whether a preset level condition is met according to an event sequence number of a current riveting event, if so, backtracking and selecting a historical riveting event which is spaced with the event sequence number by a preset step length in an event sequence, reading an event abstract of the event abstract as a level precursor abstract, splicing the level precursor abstract and the basic event abstract to form a level input string, and executing abstract operation on the level input string to generate a level event abstract, otherwise, directly taking the basic event abstract as the level event abstract; S3-5, selecting a preset number of historical riveting events taking the current riveting event as an end point from an event sequence, sequentially reading the current content abstracts of the historical riveting events, and splicing the current content abstracts in time sequence to form a section input string; S3-6, writing the current event abstract and the preamble abstract into current riveting events, then selecting preset number of riveting events including the current riveting events in an event sequence, repeatedly executing calculation from S3-3 to S3-5 on each selected riveting event to generate a recalculation event abstract, judging that an event chain is invalid at the riveting event and marked as a chain abnormal event if any recalculation event abstract is inconsistent with the event abstract recorded in the corresponding riveting event, and otherwise, maintaining the event chain to be effective.
  7. 7. The method for predicting and tracing quality anomalies based on riveting processes according to claim 6, wherein the method comprises the steps of: In S4, it includes: S4-1, sequentially selecting two adjacent riveting events in an event chain according to time sequence, respectively reading a material fingerprint and an equipment fingerprint of a previous riveting event and a material fingerprint and an equipment fingerprint of a subsequent riveting event, differencing the material fingerprint of the previous riveting event and the material fingerprint of the subsequent riveting event item by item according to corresponding positions to obtain a material difference sequence, and calculating an average value of absolute values of the material difference sequence to be used as a material similarity value; s4-2, using the similarity record as an index entry, writing an event sequence number, a material similarity value and a device similarity value of adjacent riveting event pairs into a mutation index, enabling the mutation index to be searched according to a time sequence corresponding to the event sequence number, and carrying out positioning search according to the material similarity value and the device similarity value.
  8. 8. The method for predicting and tracing quality anomalies based on riveting processes according to claim 7, wherein the method comprises the steps of: In S4, further comprising: S4-3, traversing similarity records in the mutation indexes sequentially along the sequence numbers of the events, comparing the material similarity value with a first preset mutation threshold value, comparing the equipment similarity value with a second preset mutation threshold value, judging that the process fingerprints corresponding to adjacent riveting events have structural mutation if the material similarity value is larger than the first preset mutation threshold value or the equipment similarity value is larger than the second preset mutation threshold value, and determining the latter riveting event as the mutation event, otherwise, judging that the process fingerprints corresponding to the adjacent riveting events are continuous and maintaining the mutation indexes unchanged.

Description

Quality anomaly prediction and tracing method based on riveting process Technical Field The invention relates to the technical field of data processing and quality prediction traceability of riveting equipment, in particular to a quality anomaly prediction and traceability method based on a riveting process. Background The intelligent riveting equipment comprises C-shaped riveting equipment and four-column riveting equipment, wherein the whole equipment comprises a frame, a guide structure, a power executing structure, a pressure head, a die structure, a fixture structure and a control and detection structure, the C-shaped riveting equipment forms an operation space through the C-shaped frame with a single side opening, the four-column riveting equipment forms a closed guide frame through four upright posts so as to ensure the parallelism and rigidity of the movement of a sliding block, the power executing structure generally adopts a hydraulic cylinder or a servo electric cylinder to drive the sliding block and the pressure head to do axial reciprocating movement along the guide structure under the driving of a control system, the pressure head and the die directly act on a rivet and a connected member, the fixture structure is used for realizing the positioning, supporting and limiting of a workpiece, and the control and detection structure takes a programmable controller or an industrial personal computer as a core and is matched with a pressure sensor and a displacement sensor and a time sampling unit so as to collect and control the force, displacement and beat in the riveting process; When the equipment runs, the continuous stages of workpiece clamping and positioning, quick downward approaching of a pressure head, contact loading, plastic forming of materials, final pressure shaping, return unloading and the like are generally carried out, a force-displacement-time response curve with a specific form is formed in the process, and the prior art judges whether the single riveting process is qualified or not according to peak force, final displacement or whether the final displacement reaches the result indexes such as a set stroke or not; With the continuous operation of the equipment for a long time, the progressive wear of the die and the pressure head, the continuous change of the friction state of the guide pair and the temperature rise of the hydraulic system, the objective difference of the yield characteristics of rivets and connected materials in different batches exists, meanwhile, the process parameters can be subjected to multiple version switching in the process of dispatching or changing the shape, under the superposition of the multiple factors, even if the end point index of the single riveting process still meets the set threshold value, the middle-front section form, the yield inflection point position and the slope of the forming stage of the force-displacement curve tend to have systematic deviation, and further the hidden quality problems of insufficient connection strength, reduced fatigue life and the like are induced in the subsequent assembly stress or service working condition; Therefore, the key problem of the prior art is that in the continuous riveting process of the intelligent riveting equipment, a tracing mechanism capable of systematically binding a single riveting process curve with equipment state, die life, process parameter version, material batch and other multi-source information and supporting calculation and analysis is lacked, so that the source of quality abnormality is difficult to position. Disclosure of Invention In order to overcome the above-mentioned drawbacks of the prior art, an embodiment of the present invention provides a quality anomaly prediction and tracing method based on a riveting process, which is to bind a riveting process curve with context information such as a device state, a die life, a process parameter version, a material batch, etc., and perform mutation monitoring, risk prediction and source backtracking analysis on fingerprint evolution of a continuous riveting process, so as to solve the problems presented in the above-mentioned background art. In order to achieve the purpose, the invention provides the following technical scheme that the quality anomaly prediction and tracing method based on the riveting process comprises the following steps: S1, acquiring force data, displacement data and time data acquired by riveting equipment in a one-time riveting process, and constructing a riveting process curve corresponding to the one-time riveting based on the acquired data; s2, executing process decomposition calculation on a riveting process curve in a riveting event, decomposing the riveting process curve into material fingerprints and equipment fingerprints, wherein the material fingerprints are used for representing material forming characteristics in the riveting process, and the equipment fingerprints are used for representing stru