An industry chain quality analysis method based on big data

Abstract

The application discloses a kind of industry chain quality analysis methods based on big data, it is related to artificial intelligence technical field, the method is in the process of multi-vehicle mixed line production in automobile industry chain, based on order release cycle time, logistics arrival interval time and station actual operation cycle time Construction rhythm alignment dataset, and utilize cross correlation analysis to calculate rhythm phase shift Pm, combined with line edge work-in-process residence time and station beat adjustment frequency Construction rhythm disharmony effectiveness determination condition set, to convert the time mismatch relationship between upstream supply rhythm and downstream assembly rhythm into calculable, comparable quantitative index.Compared with prior art which only relies on result quality indicators such as pass rate and rework rate, the present application can identify the rhythm disharmony state of the industry chain in advance before the quality problem is explicit, significantly improving the foresight and sensitivity of the industry chain quality risk identification.

Description

Technical Field

[0001] This invention relates to the field of supply chain technology, specifically a supply chain quality analysis method based on big data. Background Technology

[0002] With the deep integration of new-generation information technology and manufacturing, big data-based supply chain analysis methods have been widely applied in industrial production management, supply chain collaboration, and quality control. Particularly in discrete manufacturing, supply chain quality analysis has evolved from single-enterprise internal quality management to a comprehensive quality analysis system covering multiple stages such as orders, logistics, manufacturing, and assembly. In the automotive manufacturing sector, due to the complex structure of complete vehicles and the large number of parts, the supply chain typically exhibits multi-level, multi-node, and highly coupled characteristics, and the operational status of the supply chain has a significant impact on the quality of the final product. In particular, under the mixed-model production model for automobiles, the coordination between the upstream parts delivery rhythm, logistics distribution rhythm, and downstream assembly rhythm has become a crucial factor affecting the stability and quality level of the supply chain.

[0003] Current supply chain quality analysis methods mostly focus on the statistical analysis of outcome-based quality data, such as non-conforming product rates, rework rates, and defect type distribution, or on retrospective analysis after quality problems occur. These methods typically treat quality problems as independent events, focusing primarily on the quality results of a single process or node, while rarely conducting a systematic analysis of the rhythmic relationships between various links in the supply chain. In multi-model mixed-line production scenarios, due to the significant differences in the demand rhythms for parts among different models, the order release rhythm, logistics arrival rhythm, and assembly station operation rhythm frequently change. Existing quality analysis methods struggle to identify the time mismatch between supply and demand rhythms in a timely manner, failing to incorporate this implicit factor of rhythm inconsistency into the quality analysis model.

[0004] Because each link in the supply chain operates relatively independently in terms of information systems and management objectives, rhythm data is typically scattered across order systems, logistics systems, and manufacturing execution systems. The lack of a unified rhythm correlation analysis mechanism makes it difficult to detect rhythm imbalances in the early stages. When a persistent mismatch occurs between the delivery rhythm of upstream components and the assembly rhythm of downstream components, although it may be mitigated in the short term through inventory buffers or manual adjustments, the accumulation of mismatch can easily lead to abnormal phenomena such as line stockpiling, frequent cycle time adjustments, temporary substitute materials, and operational stress, thereby increasing the risk of incorrect assembly, missing parts, rework, and quality fluctuations. These quality anomalies caused by rhythm mismatches are often delayed and insidious. By the time quality indicators significantly deteriorate, the problem has already spread throughout the supply chain, making it difficult to take timely and effective intervention measures, thus affecting the overall quality management level and operational stability of the supply chain. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a big data-based supply chain quality analysis method, which solves the problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a big data-based supply chain quality analysis method, comprising the following steps: S1. During the mixed production of multiple models in the automotive industry chain, production rhythm data and quality precursor data are continuously collected along each key node of the industry chain, and the production rhythm data and quality precursor data are transmitted to the industry chain quality analysis system. The production rhythm data is preprocessed to obtain the rhythm alignment dataset. S2. Based on the rhythm alignment dataset, complete the net supply rhythm dataset, calculate the rhythm phase offset Pm, and combine it with the quality precursor data to construct a set of rhythm detuning validity judgment conditions. S3. When the set of conditions for determining the effectiveness of rhythm detuning is determined to be in the effective state of rhythm detuning, the state of quality-related detuning is entered. Based on the quality precursor data and the rhythm phase offset Pm, the rhythm detuning effectiveness index Emis is calculated. The rhythm detuning effectiveness index Emis is compared with the preset rhythm detuning effectiveness threshold Eth to determine the current rhythm detuning state of the industrial chain. S4. When the state is determined to be detuned, the quality propagation observation parameters are collected, and the rhythm-driven quality propagation intensity value Ppro is calculated. The quality risk level of the industrial chain is determined based on the quality propagation intensity value Ppro, and rhythm control corresponding to the level determination result is executed.

[0007] Preferably, S1 includes S11; S11. The collection of production rhythm data and quality precursor data at each key node of the industrial chain adopts a hierarchical structure with multiple collection points to collect production rhythm data and quality precursor data in real time. After collection, the production rhythm data and quality precursor data are transmitted to the industrial chain quality analysis system in real time via a data bus. The acquisition point structure includes a first layer of rhythm source acquisition points and a second layer of quality precursor acquisition points. The first-layer rhythm source collection points are set up in the order management system, assembly manufacturing execution system and logistics management system to collect production rhythm data that reflects the production rhythm of the industrial chain; The second layer of quality precursor acquisition points are set in the assembly site line management module and the manufacturing execution system parameter recording module to collect quality precursor data that reflects the precursors of quality evolution; The production rhythm data includes order release cycle time, actual workstation operation cycle time Tcyc, and material arrival interval time sequence Tarr. The quality precursor data includes the work-in-process dwell time at the line edge (Tdwe) and the station cycle time adjustment frequency (Fadj).

[0008] Preferably, S1 further includes S12; S12. In the supply chain quality analysis system, rhythm alignment and derived sequence construction are performed on the production rhythm data to obtain the rhythm alignment dataset. The rhythm alignment and derived sequence construction uses a fixed sliding time window W of preset length to perform time alignment processing on the order release cycle time, the actual work station operation cycle time Tcyc, and the logistics arrival interval time, thereby unifying production rhythm data from different sources on the same time scale. Within each sliding time window W, window statistics are performed on the logistics arrival interval to obtain the net supply rhythm time Tsup, which reflects the logistics supply rhythm. The window statistics process includes at least the calculation of the window mean. Within each sliding time window W, the actual work cycle time Tcyc of the workstation is statistically processed to obtain the assembly demand rhythm time Tdem, which reflects the assembly demand rhythm. The window statistical processing includes the calculation of the window mean. The net supply rhythm time Tsup and the assembly demand rhythm time Tdem are written as derived fields as write items to form a rhythm-aligned dataset containing original production rhythm data and derived rhythm feature data.

[0009] Preferably, S2 includes S21; S21. By performing cross-correlation calculations on the net supply rhythm time Tsup and assembly demand rhythm time Tdem in the rhythm alignment dataset within a preset fixed sliding time window W, the rhythm phase offset Pm is obtained, and the degree of time mismatch between the net supply rhythm and the assembly demand rhythm is quantitatively analyzed. The rhythm phase offset Pm is calculated and output using the following algorithm formula; In the formula, This represents the time average of the net supply pace. This represents the average time-varying pace of assembly demand. Indicates the candidate lag. This represents the maximum allowed offset window, N represents the total number of sampling points within the current fixed sliding time window W, Tsup,k represents the net supply rhythm of the k-th sampling point, and Tdem,k+ This indicates that the assembly demand rhythm sequence is arranged according to candidate offsets. After translation at the k+th The assembly demand rhythm value corresponding to each collection point, and argmax is the lag that maximizes the cross-correlation value within the parentheses.

[0010] Preferably, S2 further includes S22; S22. Based on the rhythm phase offset Pm and the line edge work-in-process dwell time Tdwe and work station beat adjustment frequency time Fadj in the quality precursor data, in view of the actual conditions of rhythm detuning quality risk in the industrial chain, a rhythm detuning effectiveness judgment rule is adopted to construct a set of rhythm detuning effectiveness judgment conditions. The set of rhythm detuning effectiveness judgment conditions includes rhythm offset amplitude and persistence judgment condition C1, work station beat intervention judgment condition C2 and line edge accumulation evolution judgment condition C3. When at least two conditions in the set of conditions for determining the effectiveness of rhythm detuning are met, the current industrial chain is determined to be in a rhythm detuning effective state, and the rhythm detuning state flag Smis=1 is output. When fewer than two conditions in the set of conditions for determining the effectiveness of rhythm detuning are met, the current industrial chain is determined to be in a rhythm detuning ineffective state, and the rhythm detuning state flag Smis=0 is output. Among them, the determination rule for the effectiveness of rhythm detuning in the determination condition C1 of rhythm offset amplitude and continuity is: within three consecutive time windows, the rhythm phase offset Pm ≥ a·workstation actual operation cycle time Tcyc; where a represents the proportional threshold coefficient, which can be selected from 0.3 to 0.5 in engineering and is an empirically adjustable parameter. The rule for determining the effectiveness of rhythm detuning in workstation cycle intervention condition C2 is: workstation cycle adjustment frequency time Fadj≥Fbase+△F, where Fbase represents the average cycle adjustment frequency during the historical stable production period, and △F represents the allowable normal fluctuation redundancy, 1 time / hour–2 times / hour. The rhythmic detuning validity criterion for the evolution of line-edge accumulation, C3, is as follows: Tdwe avg (W) > Tdwe avg (W-1) and Tdwe avg (W)-Tdwel avg (W-1)≥β·Tcyc; Where β represents the growth threshold coefficient, with a value of 0.1-0.2, Tdwe avg (w) represents the average dwell time of work-in-process at the edge of the line within the W-th fixed sliding time window, Tdwe avg (W-1) represents the average dwell time of work-in-process at the edge of the line within the previous W-1 fixed sliding time window.

[0011] Preferably, S3 includes S31; S31. After outputting the rhythm detuning status indicator Smis=1, calculate the rhythm detuning effectiveness index Emis based on the quality precursor data and the rhythm phase offset Pm, and comprehensively analyze the superimposed impact of the degree of rhythm time misalignment in the industrial chain and the intensity of on-site rhythm intervention on the formation of quality risks. The rhythm detuning effectiveness index Emis is calculated and output using the following algorithm formula; In the formula, ln represents the natural logarithm function; This represents the average of the actual work cycle time Tcyc at the workstation.

[0012] Preferably, S3 further includes S32; S32. Calculate the mean value of the rhythm detuning effectiveness index Emis based on the historical stable production stage, and obtain the rhythm detuning effectiveness threshold Eth. The current production phase's rhythm misalignment effectiveness index, Emis, is compared with the rhythm misalignment effectiveness threshold, Eth, to deeply assess the current rhythm misalignment status of the industrial chain; the specific comparison is as follows: When the rhythm detuning effectiveness index Emis is greater than the rhythm detuning effectiveness threshold Eth, the current rhythm detuning state of the industrial chain is determined to be a rhythm detuning effective state, and the output rhythm detuning state flag Smis is kept at 1. When the rhythm detuning effectiveness index Emis is less than or equal to the rhythm detuning effectiveness threshold Eth, the current rhythm detuning state of the industrial chain is determined to be a tolerable rhythm detuning state, and the rhythm detuning state flag Smis=0 is output.

[0013] Preferably, S4 includes S41; S41. After maintaining the output rhythm detuning status flag Smis=1, start the rhythm detuning quality propagation path analysis. The rhythm detuning quality propagation path analysis collects and schedules quality propagation observation parameters, calculates the quality propagation intensity value Ppro based on the quality propagation observation parameters, classifies the quality propagation risk caused by rhythm detuning based on the quality propagation intensity value Ppro, and executes rhythm control corresponding to the level judgment result. The quality propagation observation parameters include the process rework time ratio Rtime, the process repeated scan interval △Tscan, and the process cycle time retraction amplitude △Tback; the collected quality propagation observation parameters are then transmitted to the supply chain quality analysis system through the manufacturing execution system interface.

[0014] Preferably, S4 further includes S42; S42. After initiating the quality propagation path analysis, calculate the rhythm-driven quality propagation intensity value Ppro based on the collected quality propagation observation parameters, wherein the quality propagation intensity value Ppro is calculated as follows: .

[0015] Preferably, S4 further includes S43; S43. During the historical stable production phase, calculate the corresponding drive mass propagation intensity value Ppro according to the same sliding time window as the current production phase, and select the upper limit of the low-risk interval as the first threshold P1 and the upper limit of the medium-risk interval as the second threshold P2 based on the statistical distribution of the drive mass propagation intensity value Ppro. In the current production stage, a level determination is made based on the calculated quality propagation intensity value Ppro and the first threshold P1 and the second threshold P2. The corresponding rhythm control based on the level determination results specifically includes: When the quality propagation intensity value Ppro ≤ the first threshold P1, it is determined to be a level 1 quality propagation risk. Only the rhythm detuning state and the quality propagation analysis results are recorded, and no rhythm control or quality intervention operations are performed. When the first threshold P1 < quality propagation intensity value Ppro ≤ the second threshold P2, it is determined to be a level 2 quality propagation risk, and the control strategy of rhythm constraint and quality control is implemented. The rhythm constraint and quality control strategy limits the frequency of rhythm adjustment at assembly stations, Fadj, to a preset maximum allowable adjustment frequency. The maximum allowable adjustment frequency, Flimit, is determined based on the average frequency of rhythm adjustment during historical stable production phases and is not higher than 1.5 times the average frequency of rhythm adjustment. At the same time, it mandates that the rework reason information be recorded for processes that require rework, in order to suppress the disorderly spread of rework behavior under rhythm out-of-sync conditions. When the quality propagation intensity value Ppro > the second threshold P2, it is determined to be a level 3 quality propagation risk, and an intervention strategy of mandatory rhythm stabilization and quality traceability is implemented. The intervention strategy of forced rhythm stabilization and quality traceability freezes the rhythm adjustment operation of the assembly station, so that the assembly station runs according to the standard rhythm; initiates quality responsibility path locking, and identifies the process, material batch and related operators that have quality abnormalities; and triggers the quality audit process.

[0016] This invention provides a supply chain quality analysis method based on big data. It has the following beneficial effects: (1) This method constructs a rhythm alignment dataset based on order release cycle time, logistics arrival interval time, and actual workstation cycle time during multi-model mixed-line production in the automotive industry chain. It then uses cross-correlation analysis to calculate the rhythm phase offset Pm, and combines this with the work-in-process dwell time at the line edge and the frequency of workstation rhythm adjustment to construct a set of conditions for determining the effectiveness of rhythm mismatch. This transforms the time mismatch between upstream supply rhythm and downstream assembly rhythm into a calculable and comparable quantitative indicator. Compared with existing technologies that rely solely on outcome-based quality indicators such as pass rate and rework rate, this invention can identify the rhythm mismatch state of the industry chain before quality problems become apparent, significantly improving the foresight and sensitivity of industry chain quality risk identification.

[0017] (2) This method, after determining that the industrial chain is in a state of rhythm mismatch, further collects quality propagation observation parameters such as the proportion of rework time in processes, the interval of repeated scanning in processes, and the magnitude of process cycle time regression, calculates the rhythm-driven quality propagation intensity value Ppro, and determines the level of quality propagation risk based on the first and second thresholds determined in the historical stable production stage. By classifying quality propagation risks into different levels and implementing corresponding rhythm control strategies, this invention can avoid excessive intervention in minor anomalies, and at the same time, take timely measures such as rhythm freezing, responsibility path locking, and quality auditing when quality propagation risks intensify, effectively suppressing the spread of quality problems caused by rhythm mismatch in the industrial chain, and improving the pertinence and execution efficiency of quality control.

[0018] (3) This method organically combines the front-end identification mechanism of supply chain rhythm misalignment with the rhythm-driven quality propagation analysis mechanism, constructing a complete closed-loop analysis and control process from "rhythm misalignment identification, misalignment effectiveness assessment, quality propagation intensity quantification, and hierarchical rhythm control". On the one hand, the rhythm phase offset Pm and the rhythm misalignment effectiveness index Emis are used to identify rhythm misalignment states with quality significance in advance, avoiding misjudging occasional fluctuations as quality risks; on the other hand, the quality propagation intensity value Ppro and its hierarchical control strategy are used to dynamically assess the degree of diffusion of quality risks under misalignment states and implement differentiated interventions. Through the above synergistic effect, this invention not only improves the accuracy and stability of supply chain quality analysis, but also enables hierarchical management and dynamic control of supply chain quality risks without adding new hardware, thereby improving the overall level of supply chain quality management and operational reliability. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the steps of a big data-based supply chain quality analysis method according to the present invention. Figure 2 A schematic diagram of a layered data collection structure that serves as a precursor to the rhythm and quality of the industrial chain. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1 This invention provides a supply chain quality analysis method based on big data. Please refer to [link / reference]. Figure 1 This includes the following steps: S1. During the mixed production of multiple models in the automotive industry chain, production rhythm data and quality precursor data are continuously collected along each key node of the industry chain, and the production rhythm data and quality precursor data are transmitted to the industry chain quality analysis system. The production rhythm data is preprocessed to obtain the rhythm alignment dataset. S2. Complete the net supply rhythm dataset based on the rhythm alignment dataset, calculate the rhythm phase offset Pm, and construct a set of rhythm detuning validity judgment conditions in combination with quality precursor data. S3. When the set of conditions for determining the effectiveness of rhythm detuning is determined to be in the effective state of rhythm detuning, the state of quality-related detuning is entered. Based on the quality precursor data and the rhythm phase offset Pm, the rhythm detuning effectiveness index Emis is calculated. The rhythm detuning effectiveness index Emis is compared with the preset rhythm detuning effectiveness threshold Eth to determine the current rhythm detuning state of the industrial chain. S4. When the state is determined to be detuned, the quality propagation observation parameters are collected, and the rhythm-driven quality propagation intensity value Ppro is calculated. The quality risk level of the industrial chain is determined based on the quality propagation intensity value Ppro, and rhythm control corresponding to the level determination result is executed.

[0022] In this embodiment, the method continuously collects production rhythm data and quality precursor data at key nodes in the multi-model mixed production line of automobiles through S1, and performs rhythm alignment processing on data from different sources at a unified time scale. The purpose is to eliminate rhythm misjudgment caused by differences in sampling cycles and recording times between the order system, logistics system, and manufacturing execution system. Without rhythm alignment, the rhythm data recorded by different systems are naturally misaligned on the time axis, easily misidentifying normal rhythm fluctuations as anomalies, leading to distorted subsequent analysis conclusions. By constructing a rhythm alignment dataset, the upstream supply rhythm and downstream assembly rhythm can be compared and analyzed within the same time window, providing a reliable basis for subsequent rhythm misalignment identification. In S2, the rhythm phase offset Pm is calculated by cross-correlation analysis of the net supply rhythm and assembly demand rhythm, and a set of rhythm misalignment validity judgment conditions is constructed by combining the work-in-process dwell time at the line edge and the frequency of workstation beat adjustments. The core purpose is to distinguish between "occasional fluctuations" and "structural rhythm misalignment with quality significance." In actual production, short-term logistics delays or temporary rhythm adjustments are normal phenomena. Treating these directly as quality risks can easily lead to frequent false alarms. By introducing criteria such as continuity, amplitude, and on-site intervention, a rhythm mismatch is only considered valid when it has already created continuous pressure on on-site operations. This ensures, from a physical perspective, that the identified object truly possesses the realistic conditions to evolve into a quality problem. In S3, after determining that a rhythm mismatch has entered a valid state, the rhythm mismatch effectiveness index Emis is further calculated and compared with the rhythm mismatch effectiveness threshold Eth. The purpose of this is to transform "whether there is mismatch" into a quantitative assessment of "how severe the mismatch is." By normalizing the rhythm phase offset to the rhythm and introducing a logarithmic compression form of the rhythm adjustment frequency, the incomparability issues between different production lines and different rhythm conditions are avoided. This makes Emis a dimensionless risk intensity indicator that can be used across scenarios, providing an objective basis for whether to enter the quality propagation analysis stage and avoiding excessive intervention in tolerable mismatch states. In S4, quality propagation observation parameters such as rework time ratio, repeated scan interval, and beat rewind amplitude are collected only after it is confirmed that rhythm mismatch has quality significance. The quality propagation intensity value Ppro is then calculated. The purpose of this is to determine whether rhythm mismatch has evolved from a "rhythm problem" to a "quality problem and has begun to spread." In actual production, if rhythm mismatch has not yet caused rework, repeated operations, or frequent beat rewinds, the quality risk is still in a locally controllable state. By classifying and judging Ppro and implementing differentiated rhythm control strategies, restrictive measures can be taken in the early stages of risk, and the rhythm can be frozen and responsibility traced in time when the risk intensifies, thereby effectively preventing the quality problem from spreading along the industrial chain.

[0023] Example 2 Please see Figure 1 and Figure 2 Specifically: S1 includes S11; S11. The collection of production rhythm data and quality precursor data at each key node of the industrial chain adopts a hierarchical structure with multiple collection points to collect production rhythm data and quality precursor data in real time. After collection, the production rhythm data and quality precursor data are transmitted to the industrial chain quality analysis system in real time via a data bus. The acquisition point structure includes a first layer of rhythm source acquisition points and a second layer of quality precursor acquisition points; The first-level rhythm source collection points are set up in the order management system, assembly manufacturing execution system, and logistics management system to collect production rhythm data that reflects the production rhythm of the industrial chain; The second layer of quality precursor collection points are set in the assembly site line-side management module and the manufacturing execution system parameter recording module to collect quality precursor data that reflects the precursors of quality evolution. Production rhythm data includes order release cycle time, actual workstation operation cycle time Tcyc, and material arrival interval time series Tarr; Precursor data for quality include the time spent on work-in-process at the line edge (Tdwe) and the frequency of station cycle time (Fadj). Specifically, the order release cycle time is obtained by releasing the order in the order management system or manufacturing execution system; the actual workstation operation cycle time Tcyc is obtained by recording the assembly workstation in the manufacturing execution system; and the material arrival interval time series Tarr is obtained by recording the logistics unit in the logistics management system. Tdwell calculates the dwell time of work-in-process at the assembly line by extracting the material entry time and requisition time at the assembly site line management module and then performing a difference calculation. The frequency and timing adjustment of the workstation cycle time (Fadjust) are obtained from the workstation cycle time parameter modification log in the manufacturing execution system parameter recording module.

[0024] S1 also includes S12; S12. In the supply chain quality analysis system, rhythm alignment and derived sequence construction are performed on the production rhythm data to obtain the rhythm alignment dataset. Rhythm alignment and derived sequence construction use a fixed sliding time window W of preset length to perform time alignment processing on order release cycle time, actual workstation operation cycle time Tcyc, and logistics arrival interval time, thereby unifying production rhythm data from different sources on the same time scale; Within each sliding time window W, window statistics are performed on the logistics arrival interval to obtain the net supply rhythm time Tsup, which reflects the logistics supply rhythm. The window statistics process includes at least the calculation of the window mean. Within each sliding time window W, the actual work cycle time Tcyc of the workstation is statistically processed to obtain the assembly demand rhythm time Tdem, which reflects the assembly demand rhythm. The window statistical processing includes the calculation of the window mean. The net supply rhythm time Tsup and assembly demand rhythm time Tdem are written as derived fields as write items, forming a rhythm-aligned dataset containing original production rhythm data and derived rhythm feature data.

[0025] In this embodiment, method S11 adopts a hierarchical acquisition structure of "first-layer rhythm source acquisition point + second-layer quality precursor acquisition point": the first layer directly obtains the order release cycle time, the actual work station operation cycle time Tcyc, and the material arrival interval time sequence Tarr from the order management system, assembly manufacturing execution system, and logistics management system, which are used to characterize the three types of rhythm sources: "how demand is initiated, how the production line consumes, and how supply arrives"; the second layer extracts the work-in-process dwell time Tdwell and the work station rhythm adjustment frequency time Fadjust from the line-side management module and the manufacturing execution system parameter recording module, which are used to capture the quality precursor signal that "rhythm mismatch has begun to pressure the site", and transmits it to the supply chain quality analysis system in real time through the data bus to avoid rhythm abnormalities being "discovered after the fact" due to system dispersion and reporting delay. For example, short-term fluctuations in logistics arrival rhythm are common when multiple vehicle models are mixed on the production line. If only Tarr is considered, occasional delays can easily be misjudged as risks. However, by introducing Tdwell (whether there is accumulation at the line edge) and Fadjust (whether the production line is forced to frequently change the takt time), it is possible to distinguish between "normal fluctuations" and "states that have affected the stability of the production line." Physically, this is equivalent to simultaneously observing whether "time misalignment has resulted in accumulation / intervention." S12 uses a fixed sliding time window W within the supply chain quality analysis system to align the order release cycle time, Tcyc, and Tarr. Within each time window, window statistics are performed on Tarr and Tcyc to form the net supply rhythm time Tsup and assembly demand rhythm time Tdem, which are then written as derived fields into the rhythm alignment dataset. The direct purpose of this setup is to eliminate "time axis misalignment" caused by different sampling frequencies and recording times in different systems. Otherwise, without alignment, it is possible that "supply actually arrives early, but the record falls in an adjacent time period" and is misjudged as a lag, or "instantaneous fluctuations in Tcyc caused by production line takt time switching" are amplified as long-term trends. By using a unified statistical standard within W, Tsup and Tdem are comparable on the same scale, enabling subsequent identification of supply and demand rhythm mismatches to be based on stable and reproducible data. This improves the consistency of data coverage, the reliability of analytical conclusions, and provides stronger noise resistance and real-time performance for subsequent rhythm mismatch determination.

[0026] Example 3 Please see Figure 1Specifically: S2 includes S21; S21. By performing cross-correlation calculations on the net supply rhythm time Tsup and assembly demand rhythm time Tdem in the rhythm alignment dataset within a preset fixed sliding time window W, the rhythm phase offset Pm is obtained, and the degree of time mismatch between the net supply rhythm and the assembly demand rhythm is quantitatively analyzed. The rhythm phase offset Pm is calculated and output using the following algorithm formula; In the formula, This represents the time average of the net supply pace. This represents the average time-varying pace of assembly demand. Indicates the candidate lag. This represents the maximum allowed offset window, N represents the total number of sampling points within the current fixed sliding time window W, Tsup,k represents the net supply rhythm of the k-th sampling point, and Tdem,k+ This indicates that the assembly demand rhythm sequence is arranged according to candidate offsets. After translation at the k+th For each collection point, the assembly demand rhythm value is argmax, which is the lag that maximizes the cross-correlation value within the parentheses. It should be noted that: the candidate lag τ specifically represents the time misalignment between supply and demand; Tsup represents when the parts will arrive and be usable; Tdem represents "when the assembly station needs them and how many are needed"; and τ represents the amount by which the assembly demand rhythm needs to be shifted on the time axis in order to align the "supply rhythm" with the "demand rhythm". The calculation basis of this formula comes from: the cross-correlation function in time series analysis, which is used in the field of signal processing to measure the similarity between two time series at different time offsets. The classic algorithm comes from: the cross-correlation analysis method widely used in statistical signal processing and industrial process control. This formula does not invent a completely new mathematical formula, but rather makes engineering scenario limitations and parameter replacements based on the above classic cross-correlation formula. Specific modifications include: explicitly limiting the classic abstract sequences xk and yk to: net supply rhythm time series Tsup and assembly demand rhythm time series Tdem. Will The physical meaning is limited to: the time misalignment between the supply rhythm of the industrial chain and the assembly demand rhythm. The calculation scope is limited to: a fixed sliding time window W and candidate lag intervals with engineering significance [- max , max Therefore, the innovation of this formula lies not in the mathematics itself, but in introducing the classic cross-correlation model into the industrial chain rhythm and quality analysis scenario, and giving it a clear engineering meaning; The physical meaning of Tsup×Tdema: Tsup alone only tells you "how fast the goods arrive"; Tdem alone only tells you "how busy the production line is"; only when the two are combined can you determine "whether the goods arrive at the right time". The physical meaning of "time shift + correlation": rhythmic mismatch is not an instantaneous error, but an overall time misalignment; shift can eliminate the impact of occasional jitter; correlation can capture the consistency of the overall trend. The rhythm phase offset Pm is the optimal alignment offset between the supply rhythm and the assembly demand rhythm on the time axis. It is the key mediating parameter that transforms rhythm mismatch from a "perceptual problem" into a "calculable engineering quantity". Dimensional consistency: Each item within parentheses is: Time - Time = Time, and the product of two sets of parentheses is: Time. 2 Multiplying by 1 / N yields a dimensionless proportionality coefficient, while Pm represents the maximum value summed within parentheses. The dimensionless result is time.

[0027] S2 also includes S22; S22. Based on the rhythm phase offset Pm and the work-in-process dwell time Tdwe and work station beat adjustment frequency time Fadj in the quality precursor data, in view of the actual conditions of rhythm mismatch quality risk in the industrial chain, a rhythm mismatch effectiveness judgment rule is adopted to construct a set of rhythm mismatch effectiveness judgment conditions. The set of rhythm mismatch effectiveness judgment conditions includes rhythm offset amplitude and persistence judgment condition C1, work station beat intervention judgment condition C2 and line edge accumulation evolution judgment condition C3. When at least two conditions in the set of conditions for determining the effectiveness of rhythm detuning are met, the current industrial chain is determined to be in a rhythm detuning effective state, and the rhythm detuning state flag Smis=1 is output. When fewer than two conditions in the set of conditions for determining the effectiveness of rhythm detuning are met, the current industrial chain is determined to be in a rhythm detuning ineffective state, and the rhythm detuning state flag Smis=0 is output. Among them, the determination rule for the effectiveness of rhythm detuning in the determination condition C1 of rhythm offset amplitude and continuity is: within three consecutive time windows, the rhythm phase offset Pm ≥ a·workstation actual operation cycle time Tcyc; where a represents the proportional threshold coefficient, which can be selected from 0.3 to 0.5 in engineering, and is an empirically adjustable parameter used to exclude "brief jitter" and "occasional delay", and only identify structural rhythm detuning. The rule for determining the effectiveness of rhythm mismatch in workstation cycle time intervention condition C2 is: workstation cycle time adjustment frequency Fadj ≥ Fbase + △F, where Fbase represents the average cycle time adjustment frequency during the historical stable production period, and △F represents the allowable normal fluctuation redundancy, 1 time / hour – 2 times / hour; to identify whether rhythm mismatch has interfered with the stability of on-site operations, more specifically, to explain that the production line has to frequently “manually or systematically correct the rhythm”, indicating that the rhythm problem has begun to put pressure on the on-site work. The rhythmic detuning validity criterion for the evolution of line-edge accumulation, C3, is as follows: Tdwe avg (W) > Tdwe avg (W-1) and Tdwe avg (W)-Tdwel avg (W-1)≥β·Tcyc; Where β represents the growth threshold coefficient, with a value of 0.1-0.2, Tdwe avg (w) represents the average dwell time of work-in-process at the edge of the line within the W-th fixed sliding time window, Tdwe avg (W-1) represents the average dwell time of work-in-process at the edge of the line within the previous W-1 fixed sliding time window; The judgment of whether rhythmic disharmony has caused physical accumulation and waiting is not an occasional accumulation, but an accumulation that is continuously aggravated.

[0028] In this embodiment, method S21 performs cross-correlation calculation on the net supply rhythm time Tsup and assembly demand rhythm time Tdem within a fixed sliding time window W and outputs the rhythm phase offset Pm. Its purpose is not to calculate the "instantaneous difference," but rather to use "time shift + correlation" to find the optimal alignment offset between the supply rhythm and demand rhythm on the time axis, thereby transforming the supply-demand mismatch from a subjective perception into a calculable time quantity. For example, when multiple vehicle models are mixed on the line, the demand rhythm at the same workstation will fluctuate periodically due to vehicle model switching. If only the difference between Tsup and Tdem at a certain moment is compared, it is easy to mistake the jitter caused by normal switching for detuning. By using candidate lag values… Within a certain range, the location with the highest correlation can filter out occasional jitter and identify "structural misalignment." Its physical meaning is equivalent to measuring whether the "rhythm of parts reaching usability" is ahead of or behind the "actual assembly consumption rhythm," and by how much. This provides a stable and interpretable mediator for subsequent risk identification. In S22, a set of judgment conditions for rhythm misalignment validity is further constructed based on Pm and quality precursor data, and a joint judgment rule of "at least two conditions being met" is used to output Smis. The core purpose is to avoid "treating all deviations as quality risks." Specifically, C1 uses the amplitude and persistence constraints of three consecutive time windows to exclude occasional delays; C2 uses whether Fadj exceeds the stable period baseline Fbase plus redundant ΔF to determine whether the site has been forced to frequently adjust the beat (this usually means that rhythm problems are beginning to pressure operational stability); C3 uses the continuous rise of Tdwe and the incremental threshold to determine whether there is a real accumulation evolution at the line edge. The direct effect of this setting is that false alarms will not be triggered when the supply and demand misalignment can still be buffered. Only when the misalignment has simultaneously manifested as a combination of symptoms such as "continuous offset + frequent intervention" or "continuous offset + accumulation and aggravation" can it be judged as an effective state of rhythm detuning and output Smis=1. This improves the accuracy and foresight of rhythm detuning identification, reduces invalid alarms and excessive intervention, and ensures that subsequent quality analysis and control are only carried out for detuning states that are truly of quality significance.

[0029] Example 4 Please see Figure 1 Specifically: S3 includes S31; S31. After outputting the rhythm detuning status indicator Smis=1, calculate the rhythm detuning effectiveness index Emis based on the quality precursor data and the rhythm phase offset Pm, and comprehensively analyze the superimposed impact of the degree of rhythm time misalignment in the industrial chain and the intensity of on-site rhythm intervention on the formation of quality risks. The rhythm detuning effectiveness index Emis is calculated and output using the following algorithm formula; In the formula, ln represents the natural logarithm function; This represents the average of the actual work cycle time Tcyc at the workstation. The engineering goal of this formula is to combine the "intensity of time misalignment" and the "intensity of on-site intervention" into a thresholdable risk intensity index after determining that Smis=1 (effective state of rhythm misalignment), so as to compare it with Eth and output the rhythm misalignment state of the industrial chain. The time misalignment is transformed into a dimensionless "relative misalignment"; the time misalignment amount is represented by Pm, and its unit is "seconds (or seconds converted from sampling points)"; directly using Pm is not conducive to cross-production line comparisons (different production lines have different cycle times); therefore, cycle time normalization is introduced: This indicates "the proportion of the average cycle time of a certain number of workstations equivalent to the misalignment," and is a common relative indicator construction in industrial engineering. The frequency of beat intervention is converted into "effective intervention intensity"; Fadj is a frequency quantity in "times / hour". Direct linear superposition will lead to masking time misalignment at extremely high frequencies; therefore, a logarithmic compression form is adopted: ln(1+Fadj) where "+1" ensures that the logarithmic term is 0 when Fadj=0, avoiding meaningless negative infinity or discontinuity. The combination method chosen is product (emphasizing "synergistic triggering"): This formula chooses to multiply the two parts rather than add them because: when the time misalignment is large but no on-site adjustment is required, it indicates that the system has a buffer capacity, and the risk may not be high; when on-site adjustments are frequent but the time misalignment is small, it may just be a strategy fluctuation or production schedule change, and the risk may not be high either; only when both deviate simultaneously is it a state of disharmony that "truly transforms into quality risk"; therefore, product is used to reflect the "synergistic enhancement" relationship, resulting in this formula; Dimensional consistency explanation: Pm is a time dimension. Measured in units of time. Since Fadj is a dimensionless number and the logarithmic function output ln(1+Fadj) is dimensionless, the rhythm detuning effectiveness index Emis is also dimensionless.

[0030] S3 also includes S32; S32. Calculate the mean value of the rhythm detuning effectiveness index Emis based on the historical stable production stage, and obtain the rhythm detuning effectiveness threshold Eth. The current production phase's rhythm misalignment effectiveness index, Emis, is compared with the rhythm misalignment effectiveness threshold, Eth, to deeply assess the current rhythm misalignment status of the industrial chain; the specific comparison is as follows: When the rhythm detuning effectiveness index Emis is greater than the rhythm detuning effectiveness threshold Eth, the current rhythm detuning state of the industrial chain is determined to be a rhythm detuning effective state, and the output rhythm detuning state flag Smis is kept at 1. When the rhythm detuning effectiveness index Emis is less than or equal to the rhythm detuning effectiveness threshold Eth, the current rhythm detuning state of the industrial chain is determined to be a tolerable rhythm detuning state, and the rhythm detuning state flag Smis=0 is output.

[0031] In this embodiment, after the rhythm misalignment status identifier Smis=1 is output in step S22, step S31 further calculates the rhythm misalignment effectiveness index Emis. The purpose is not to repeatedly determine "whether there is misalignment", but to quantify the confirmed misalignment status with quality significance into a continuous indicator of "how great is the risk intensity", which is convenient for subsequent threshold-based decision-making. Specifically, Pm reflects the time misalignment of supply and demand rhythm, but different production lines have different beat rates. If Pm (seconds) is used directly for comparison, it is easy to make the misjudgment that "small deviations in fast-beat production lines are also dangerous, while large deviations in slow-beat production lines are tolerable". Therefore, Pm is normalized to beat rate using the mean of Tcyc, so that its physical meaning becomes "how many workstation beat rates the misalignment is equivalent to". At the same time, Fadj is introduced and logarithmic compression of ln(1+Fadj) is used to avoid the linear bursting of the indicator by extreme high-frequency adjustments, and to ensure that the field law of "the more frequent the intervention, the greater the risk, but the marginal increase decreases" is reflected. Using product combinations instead of addition is equivalent to emphasizing that "large misalignments require frequent corrections to be dangerous," thus avoiding false alarm scenarios such as "large misalignments but sufficient system buffer" or "frequent adjustments but small misalignments indicating strategy fluctuations," making Emis closer to the actual quality risk formation mechanism. S32 uses the average Emis value from historical stable production phases to set the threshold Eth, and compares the current Emis with Eth to output the rhythm misalignment state. Its purpose is to ground the "risk intensity" in actionable engineering judgments, avoiding subjective triggering of control based solely on experience. For example, when model switching leads to increased short-term rhythm adjustments, if Emis does not exceed Eth, it is judged as a tolerable rhythm misalignment state and Smis=0 is output, avoiding excessive intervention caused by frequent subsequent quality propagation analysis. When supply-demand misalignment and on-site intervention simultaneously intensify, causing Emis to continuously exceed Eth, Smis=1 is maintained, ensuring that subsequent quality propagation analysis and tiered control are only carried out for truly high-risk misalignment periods. Through the above implementation, the comparability, noise resistance and decision consistency of rhythm misalignment assessment can be improved without adding new hardware, realizing a closed loop of "identification, quantification and judgment", reducing false alarms and missed alarms and enhancing the ability to proactively control quality risks.

[0032] Example 5 Please see Figure 1 Specifically: S4 includes S41; S41. After maintaining the output rhythm detuning status flag Smis=1, start the rhythm detuning quality propagation path analysis. The rhythm detuning quality propagation path analysis collects and schedules quality propagation observation parameters, calculates the quality propagation intensity value Ppro based on the quality propagation observation parameters, classifies the quality propagation risk caused by rhythm detuning based on the quality propagation intensity value Ppro, and executes rhythm control corresponding to the level judgment result. The quality propagation observation parameters include the process rework time ratio Rtime, the process repeated scan interval △Tscan, and the process cycle time back amplitude △Tback; then, the collected quality propagation observation parameters are transmitted to the supply chain quality analysis system through the manufacturing execution system interface. Among them, the rework time ratio Rtime is calculated by taking the rework time from the start and end times of the rework operation recorded in the manufacturing execution system and then comparing it with the normal operation time of the corresponding process. The process repetition scan interval △Tscan is obtained by calculating the difference between the consecutive scan timestamps generated by the same material or workpiece in the manufacturing execution system or automatic identification system. The process cycle time regression range △Tback is calculated by subtracting the cycle time adjustment value recorded in the manufacturing execution system from the standard cycle time value.

[0033] S4 also includes S42; S42. After initiating the quality propagation path analysis, calculate the rhythm-driven quality propagation intensity value Ppro based on the collected quality propagation observation parameters. The quality propagation intensity value Ppro is calculated as follows: ; The physical meaning of this formula lies in the "quality propagation intensity" resulting from the superposition of rework behavior, repetitive operation behavior, and beat reversal behavior, given that rhythmic misharmony has been confirmed to have quality significance. These three items were chosen because they correspond to quality repair, operational disturbance, and rhythm reversal, respectively, covering key aspects of the propagation mechanism; finally, the sum of squares and root extraction are used for: Nonnegativity: Risk intensity should not be negative; Sensitive to extreme anomalies: When a certain value is very large, it will be magnified by the square, which conforms to the industrial safety logic that "a single serious point can trigger high risk"; Explainability: Each contribution is clear and easy to trace; Explanation of dimensional consistency: The proportion of rework time in a process, Rtime, is dimensionless; the process repetition scan interval, △Tscan, and the process cycle back amplitude, △Tback, are time-dimensioned; the actual work cycle time, Tcyc, is also time-dimensioned. After squaring, summing, and taking the square root of this formula, the quality propagation intensity value, Ppro, is a dimensionless parameter.

[0034] S4 also includes S43; S43. During the historical stable production phase, calculate the corresponding drive mass propagation intensity value Ppro according to the same sliding time window as the current production phase, and select the upper limit of the low-risk interval as the first threshold P1 and the upper limit of the medium-risk interval as the second threshold P2 based on the statistical distribution of the drive mass propagation intensity value Ppro. In the current production stage, a level determination is made based on the calculated quality propagation intensity value Ppro and the first threshold P1 and the second threshold P2. The corresponding rhythm control based on the level determination results specifically includes: When the quality propagation intensity value Ppro ≤ the first threshold P1, it is judged as a level 1 quality propagation risk. Only the rhythm detuning state and the quality propagation analysis results are recorded, and no rhythm control or quality intervention operations are performed. When the first threshold P1 < quality propagation intensity value Ppro ≤ the second threshold P2, it is determined to be a level 2 quality propagation risk, and the control strategy of rhythm constraint and quality control is implemented. The rhythm constraint and quality control strategy limits the frequency of rhythm adjustment at assembly stations, Fadj, to a preset maximum allowable adjustment frequency. The maximum allowable adjustment frequency, Flimit, is determined based on the average rhythm adjustment frequency during historical stable production phases and is no higher than 1.5 times the average rhythm adjustment frequency. At the same time, it mandates that the rework reason information be recorded for processes that require rework, in order to suppress the disorderly spread of rework behavior under rhythm out-of-sync conditions. When the quality propagation intensity value Ppro > the second threshold P2, it is determined to be a level 3 quality propagation risk, and an intervention strategy of mandatory rhythm stabilization and quality traceability is implemented. The intervention strategy of forced rhythm stabilization and quality traceability freezes the rhythm adjustment operation of the assembly station to make the assembly station run according to the standard rhythm; initiates quality responsibility path locking to identify the process, material batch and related operators where quality abnormalities occur; and triggers the quality audit process to systematically investigate the quality transmission problem under the condition of rhythm disharmony.

[0035] In this embodiment, method S41 initiates quality propagation path analysis only after maintaining output Smis=1. The purpose is to further differentiate between "rhythm mismatch" and "the problem has begun to spread to quality behavior," avoiding premature intervention when there are only rhythm fluctuations but no quality disturbances. To this end, three continuously quantifiable observation parameters are collected: Rtime, △Tscan, and △Tback. Rtime directly reflects the proportion of rework time (quality repair intensity), △Tscan reflects the urgent error correction behavior of repeated scanning of the same material within a short period (operational disturbance intensity), and △Tback reflects the magnitude of the beat's regression from the adjusted value to the standard value (rhythm swing intensity). These parameters are all derived from MES / scan logs / beat parameter records, and can be obtained in real time without adding new hardware. They are then transmitted back to the supply chain quality analysis system via an interface, thus transforming the phenomenon of "the site being pressured by rhythm mismatch" into calculable data evidence. Step S42 further synthesizes the three into a dimensionless Ppro, whose physical meaning is equivalent to measuring the mass propagation intensity caused by the superposition of "rework-repetitive operation-rhythm relapse" after confirming that rhythm detuning has mass significance. The purpose of using the square root of the sum of squares is to ensure that the risk intensity is non-negative and to remain sensitive to extreme anomalies in a single item. For example, when there is a sudden increase in repeated scanning or a sudden increase in the beat relapse amplitude in a short period of time, even if other items have not increased significantly, Ppro can be raised in time to avoid missing the early diffusion signal. S43 establishes P1 and P2 through stable period statistics and implements tiered control, progressively advancing the handling strategy from "recording" to "constraining" and then to "loss prevention and traceability": When Ppro is lower than P1, only recording is performed to avoid false alarms and frequent disturbances; when Ppro is between P1 and P2, Fadj is limited to not exceeding Flimit (within 1.5 times the average value of the stable period), and the cause of rework is forcibly labeled to prevent the problem from spreading on-site through "frequent changes in rhythm + disorderly rework"; when Ppro is higher than P2, rhythm adjustment is frozen, and responsibility path locking and quality auditing are initiated to achieve rapid loss prevention and traceability closure. Through the above implementation, rhythm mismatch can be advanced from "whether it exists" to an executable chain of "whether it has spread, to what level it has spread, and what control actions to take," significantly improving the pertinence and timeliness of risk handling, while reducing production disturbances caused by ineffective intervention, achieving effective containment and traceability management of the spread of quality risks in the industrial chain.

[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A supply chain quality analysis method based on big data, characterized in that: Includes the following steps: S1. During the mixed production of multiple models in the automotive industry chain, production rhythm data and quality precursor data are continuously collected along each key node of the industry chain, and the production rhythm data and quality precursor data are transmitted to the industry chain quality analysis system. The production rhythm data is preprocessed to obtain the rhythm alignment dataset. S2. Based on the rhythm alignment dataset, complete the net supply rhythm dataset, calculate the rhythm phase offset Pm, and combine it with the quality precursor data to construct a set of rhythm detuning validity judgment conditions. S3. When the set of conditions for determining the effectiveness of rhythm detuning is determined to be in the effective state of rhythm detuning, the state of quality-related detuning is entered. Based on the quality precursor data and the rhythm phase offset Pm, the rhythm detuning effectiveness index Emis is calculated. The rhythm detuning effectiveness index Emis is compared with the preset rhythm detuning effectiveness threshold Eth to determine the current rhythm detuning state of the industrial chain. S4. When the state is determined to be detuned, the quality propagation observation parameters are collected, and the rhythm-driven quality propagation intensity value Ppro is calculated. The quality risk level of the industrial chain is determined based on the quality propagation intensity value Ppro, and rhythm control corresponding to the level determination result is executed.

2. The supply chain quality analysis method based on big data according to claim 1, characterized in that: S1 includes S11; S11. The collection of production rhythm data and quality precursor data at each key node of the industrial chain adopts a hierarchical structure with multiple collection points to collect production rhythm data and quality precursor data in real time. After collection, the production rhythm data and quality precursor data are transmitted to the industrial chain quality analysis system in real time via a data bus. The acquisition point structure includes a first layer of rhythm source acquisition points and a second layer of quality precursor acquisition points. The first-layer rhythm source collection points are set up in the order management system, assembly manufacturing execution system and logistics management system to collect production rhythm data that reflects the production rhythm of the industrial chain; The second layer of quality precursor acquisition points are set in the assembly site line-side management module and the manufacturing execution system parameter recording module to collect quality precursor data that reflects the precursors of quality evolution; The production rhythm data includes order release cycle time, actual workstation operation cycle time Tcyc, and material arrival interval time sequence Tarr. The quality precursor data includes the work-in-process dwell time at the line edge (Tdwe) and the station cycle time adjustment frequency (Fadj).

3. The supply chain quality analysis method based on big data according to claim 2, characterized in that: S1 further includes S12; S12. In the supply chain quality analysis system, rhythm alignment and derived sequence construction are performed on the production rhythm data to obtain the rhythm alignment dataset. The rhythm alignment and derived sequence construction uses a fixed sliding time window W of preset length to perform time alignment processing on the order release cycle time, the actual work station operation cycle time Tcyc, and the logistics arrival interval time, thereby unifying production rhythm data from different sources on the same time scale. Within each sliding time window W, window statistics are performed on the logistics arrival interval to obtain the net supply rhythm time Tsup, which reflects the logistics supply rhythm. The window statistics process includes at least the calculation of the window mean. Within each sliding time window W, the actual work cycle time Tcyc of the workstation is statistically processed to obtain the assembly demand rhythm time Tdem, which reflects the assembly demand rhythm. The window statistical processing includes the calculation of the window mean. The net supply rhythm time Tsup and the assembly demand rhythm time Tdem are written as derived fields as write items to form a rhythm-aligned dataset containing original production rhythm data and derived rhythm feature data.

4. The supply chain quality analysis method based on big data according to claim 3, characterized in that: S2 includes S21; S21. By performing cross-correlation calculations on the net supply rhythm time Tsup and assembly demand rhythm time Tdem in the rhythm alignment dataset within a preset fixed sliding time window W, the rhythm phase offset Pm is obtained, and the degree of time mismatch between the net supply rhythm and the assembly demand rhythm is quantitatively analyzed. The rhythm phase offset Pm is calculated and output using the following algorithm formula; In the formula, This represents the time average of the net supply pace. This represents the average time-varying pace of assembly demand. Indicates the candidate lag. This represents the maximum allowed offset window, N represents the total number of sampling points within the current fixed sliding time window W, Tsup,k represents the net supply rhythm of the k-th sampling point, and Tdem,k+ This indicates that the assembly demand rhythm sequence is arranged according to candidate offsets. After translation at the k+th The assembly demand rhythm value corresponding to each collection point, and argmax is the lag that maximizes the cross-correlation value within the parentheses.

5. The supply chain quality analysis method based on big data according to claim 4, characterized in that: S2 further includes S22; S22. Based on the rhythm phase offset Pm and the line edge work-in-process dwell time Tdwe and work station beat adjustment frequency time Fadj in the quality precursor data, in view of the actual conditions of rhythm detuning quality risk in the industrial chain, a rhythm detuning effectiveness judgment rule is adopted to construct a set of rhythm detuning effectiveness judgment conditions. The set of rhythm detuning effectiveness judgment conditions includes rhythm offset amplitude and persistence judgment condition C1, work station beat intervention judgment condition C2 and line edge accumulation evolution judgment condition C3. When at least two conditions in the set of conditions for determining the effectiveness of rhythm detuning are met, the current industrial chain is determined to be in a rhythm detuning effective state, and the rhythm detuning state flag Smis=1 is output. When fewer than two conditions in the set of conditions for determining the effectiveness of rhythm detuning are met, the current industrial chain is determined to be in a rhythm detuning ineffective state, and the rhythm detuning state flag Smis=0 is output. Among them, the determination rule for the effectiveness of rhythm detuning in the determination condition C1 of rhythm offset amplitude and continuity is: within three consecutive time windows, the rhythm phase offset Pm ≥ a·workstation actual operation cycle time Tcyc; where a represents the proportional threshold coefficient, which can be selected from 0.3 to 0.5 in engineering and is an empirically adjustable parameter. The rule for determining the effectiveness of rhythm detuning in workstation cycle intervention condition C2 is: workstation cycle adjustment frequency time Fadj≥Fbase+△F, where Fbase represents the average cycle adjustment frequency during the historical stable production period, and △F represents the allowable normal fluctuation redundancy, 1 time / hour–2 times / hour. The rule for determining the effectiveness of rhythmic detuning in the evolution criterion C3 of line-edge accumulation is: Tdwe avg (W) > Tdwe avg (W-1) and Tdwe avg (W)-Tdwel avg (W-1)≥β·Tcyc; Where β represents the growth threshold coefficient, with a value of 0.1-0.2, Tdwe avg (w) represents the average dwell time of work-in-process within the W-th fixed sliding time window, Tdwe avg (W-1) represents the average dwell time of the work-in-process at the edge of the line within the previous W-1 fixed sliding time window.

6. The supply chain quality analysis method based on big data according to claim 5, characterized in that: S3 includes S31; S31. After outputting the rhythm detuning status indicator Smis=1, calculate the rhythm detuning effectiveness index Emis based on the quality precursor data and the rhythm phase offset Pm, and comprehensively analyze the superimposed impact of the degree of rhythm time misalignment in the industrial chain and the intensity of on-site rhythm intervention on the formation of quality risks. The rhythm detuning effectiveness index Emis is calculated and output using the following algorithm formula; In the formula, ln represents the natural logarithm function; This represents the average of the actual work cycle time Tcyc at the workstation.

7. The supply chain quality analysis method based on big data according to claim 6, characterized in that: S3 further includes S32; S32. Calculate the mean value of the rhythm detuning effectiveness index Emis based on the historical stable production stage, and obtain the rhythm detuning effectiveness threshold Eth. The current production phase's rhythm misalignment effectiveness index, Emis, is compared with the rhythm misalignment effectiveness threshold, Eth, to deeply assess the current rhythm misalignment status of the industrial chain; the specific comparison is as follows: When the rhythm detuning effectiveness index Emis is greater than the rhythm detuning effectiveness threshold Eth, the current rhythm detuning state of the industrial chain is determined to be a rhythm detuning effective state, and the output rhythm detuning state flag Smis is kept at 1. When the rhythm detuning effectiveness index Emis is less than or equal to the rhythm detuning effectiveness threshold Eth, the current rhythm detuning state of the industrial chain is determined to be a tolerable rhythm detuning state, and the rhythm detuning state flag Smis=0 is output.

8. The supply chain quality analysis method based on big data according to claim 7, characterized in that: S4 includes S41; S41. After maintaining the output rhythm detuning status flag Smis=1, start the rhythm detuning quality propagation path analysis. The rhythm detuning quality propagation path analysis collects and schedules quality propagation observation parameters, calculates the quality propagation intensity value Ppro based on the quality propagation observation parameters, classifies the quality propagation risk caused by rhythm detuning based on the quality propagation intensity value Ppro, and executes rhythm control corresponding to the level judgment result. The quality propagation observation parameters include the process rework time ratio Rtime, the process repeated scan interval △Tscan, and the process cycle time retraction amplitude △Tback; the collected quality propagation observation parameters are then transmitted to the supply chain quality analysis system through the manufacturing execution system interface.

9. The supply chain quality analysis method based on big data according to claim 8, characterized in that: S4 also includes S42; S42. After initiating the quality propagation path analysis, calculate the rhythm-driven quality propagation intensity value Ppro based on the collected quality propagation observation parameters, wherein the quality propagation intensity value Ppro is calculated as follows: 。 10. A supply chain quality analysis method based on big data according to claim 9, characterized in that: S4 also includes S43; S43. During the historical stable production phase, calculate the corresponding drive mass propagation intensity value Ppro according to the same sliding time window as the current production phase, and select the upper limit of the low-risk interval as the first threshold P1 and the upper limit of the medium-risk interval as the second threshold P2 based on the statistical distribution of the drive mass propagation intensity value Ppro. In the current production stage, a level determination is made based on the calculated quality propagation intensity value Ppro and the first threshold P1 and the second threshold P2. The corresponding rhythm control based on the level determination results specifically includes: When the quality propagation intensity value Ppro ≤ the first threshold P1, it is determined to be a level 1 quality propagation risk. Only the rhythm detuning state and the quality propagation analysis results are recorded, and no rhythm control or quality intervention operations are performed. When the first threshold P1 < quality propagation intensity value Ppro ≤ the second threshold P2, it is determined to be a level 2 quality propagation risk, and the control strategy of rhythm constraint and quality control is implemented. The rhythm constraint and quality control strategy limits the frequency of rhythm adjustment at assembly stations, Fadj, to a preset maximum allowable adjustment frequency. The maximum allowable adjustment frequency, Flimit, is determined based on the average frequency of rhythm adjustment during historical stable production phases and is not higher than 1.5 times the average frequency of rhythm adjustment. At the same time, it mandates that the rework reason information be recorded for processes that require rework, in order to suppress the disorderly spread of rework behavior under rhythm out-of-sync conditions. When the quality propagation intensity value Ppro > the second threshold P2, it is determined to be a level 3 quality propagation risk, and an intervention strategy of mandatory rhythm stabilization and quality traceability is implemented. The intervention strategy of forced rhythm stabilization and quality traceability freezes the rhythm adjustment operation of the assembly station, so that the assembly station runs according to the standard rhythm; initiates quality responsibility path locking, and identifies the process, material batch and related operators that have quality abnormalities; and triggers the quality audit process.

Images