Cartridge receiver quality problem tracing method
By collecting and preprocessing casing manufacturing data, constructing a traceability knowledge graph, and utilizing Bayesian network analysis, the problems of low efficiency and insufficient accuracy in tracing casing quality issues were solved, achieving efficient and accurate traceability analysis.
Patent Information
- Application Number
- CN202511479039.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-16
- Publication Date
- 2026-01-20
AI Technical Summary
Existing methods for tracing quality issues in casings rely on manual analysis, which is inefficient and susceptible to subjective factors. They also lack data integration and knowledge accumulation, resulting in insufficient accuracy in tracing the source.
Collect historical manufacturing data of the casing, preprocess and standardize it, construct a knowledge graph for tracing the source of quality problems, use Bayesian network analysis to analyze posterior probabilities, generate evidence chains and relationship chains, and form a traceability report.
It improved analysis efficiency, reduced the influence of subjective factors, broke down data silos, enabled rapid access to historical solutions, and improved the accuracy and precision of tracing the source.
Smart Images
Figure CN121365906A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to the technical field of an aero-engine, and discloses a method for tracing a quality problem of a case. BACKGROUND
[0002] The case of an aero-engine is a core load-bearing component, and the quality of the case directly affects the operation safety and reliability of the engine. The manufacturing process of the case involves complex machining processes and multi-link detection processes, and the quality problem may be caused by various factors such as design defects, performance fluctuations of raw materials, abnormal equipment parameters and operation errors.
[0003] The existing method for tracing the quality problem of the case has the following disadvantages: the method relies on manual analysis, is low in efficiency and is easily affected by subjective factors; quality data are stored in a scattered manner, a unified data integration mechanism is lacked, and a 'data island' is formed; a systematic knowledge accumulation and reuse mechanism is not established, and historical solutions cannot be quickly called when similar problems occur repeatedly; and data reliability and cause probability lack quantitative standards, resulting in insufficient tracing accuracy. SUMMARY
[0004] The application aims to provide a method for tracing a quality problem of a case, which can improve analysis efficiency, reduce the influence of subjective factors, break the data island, establish a systematic knowledge accumulation and reuse mechanism, quickly call historical solutions when similar problems occur repeatedly, and improve tracing accuracy.
[0005] In order to achieve the above technical effects, the application adopts the following technical scheme: A method for tracing a quality problem of a case, comprising: S1: collecting historical manufacturing data of the case, wherein the historical manufacturing data comprises case design parameter data in a design stage, machining process parameter data and machining equipment operation data in a machining stage, and case detection data and quality problem symptom description data in a detection stage; S2: preprocessing the historical manufacturing data to obtain a representation parameter of each type of data in the historical manufacturing data, then calculating a confidence degree of the representation parameter of each type of data by using a data confidence degree formula, and screening out the representation parameter with a confidence degree reaching a preset threshold to form a standardized data set; S3: constructing a case quality problem tracing knowledge graph according to the standardized data set, wherein the knowledge graph comprises a quality problem entity, an influence factor entity, and a relationship between the quality problem entity and the influence factor entity; S4: Based on the problem symptoms of the current quality problem to be traced in the casing, reasoning analysis is performed based on the casing quality problem tracing knowledge graph to obtain all the influencing factor entities that may cause the current quality problem in the casing, and the influencing factor entities are taken as the problem causes of the current quality problem in the casing; the posterior probability of each problem cause is obtained by analyzing the tracing model of Bayesian network, and the problem causes with posterior probabilities greater than a preset probability threshold are selected to form a problem cause candidate set, and the problem cause with the highest posterior probability in the problem cause candidate set is determined as the root cause; S5: For each problem cause in the candidate set of problem causes, retrieve the associated data in the standardized dataset, and construct an evidence chain and an influencing factor relationship chain about the current quality problem from problem symptoms to root cause; S6: Generate a traceability report based on the aforementioned chain of evidence and the chain of relationships between influencing factors.
[0006] Furthermore, the machining process parameters include cutting speed, feed rate, depth of cut, and machining equipment model; the casing inspection data includes three-dimensional dimension inspection data, material property inspection data, and surface quality inspection data.
[0007] Furthermore, the method for preprocessing the historical manufacturing data to obtain characterization parameters includes: based on The principle is to identify and remove outliers in the historical manufacturing data, and then fill in the missing values in the historical manufacturing data to obtain the historical manufacturing data after missing value filling. The historical manufacturing data from different sources, after missing values are filled in, is converted into a unified data format and unit to obtain standardized historical manufacturing data. The characterization parameters are obtained by extracting features from standardized historical manufacturing data using principal component analysis or wavelet transform.
[0008] Furthermore, the data confidence formula is as follows: ; in, The first in historical manufacturing data Confidence level of the representation parameters of class data; These are the weighting coefficients; The first in historical manufacturing data Valid data in the class data; The first in historical manufacturing data The total amount of data in the data category; The first in historical manufacturing data The percentage of valid data in this category; The first in historical manufacturing data The freshness of the data; is the number of the first quality problem, is the number of days from the data collection date of the data of the first quality problem to the current date; is the process cycle of the machine case.
[0009] Further, the inter-entity relationship includes a causal relationship, a correlation relationship and a time sequence relationship.
[0010] Further, the Bayesian network-based traceability model is constructed as: ; wherein, is the posterior probability of the problem cause under the evidence, is the likelihood probability of the evidence appearing when the problem cause occurs, is the prior probability of the problem cause, is the number of the first quality problem, is the cause of the generation of the first quality problem of the machine case, is the evidence of the quality problem of the machine case. is the number of the first quality problem, is the cause of the generation of the first quality problem of the machine case, is the evidence of the quality problem of the machine case. Further, in step S5, the method of constructing the evidence chain and the influence factor relationship chain from the problem symptom to the root cause of the current quality problem includes: for each problem cause in the candidate set of problem causes, searching and extracting the historical manufacturing data associated with each problem cause in the standardized data set;
[0011] according to the correlation relationship between the historical manufacturing data in the machine case quality problem traceability knowledge graph, constructing the correlation relationship between each problem cause in the candidate set and the corresponding historical manufacturing data, forming a link graph of the problem cause and the corresponding historical manufacturing data, i.e. relationship chain. The historical manufacturing data related to the problem cause are analyzed to construct the evidence chain including the problem symptom, the direct cause and the root cause, and the influence factor relationship chain including the equipment, the process and the personnel.
[0012] Further, in step S6, the traceability report includes the problem cause corresponding to the current quality problem to be traced of the machine case.
[0013] Compared with the prior art, the present application has the beneficial effects of:
[0014] Compared with the prior art, the present application has the beneficial effects of: The application reduces the proportion of manual intervention through data acquisition and preprocessing, avoids empiricism influence, improves the traceability analysis efficiency and reduces the influence of artificial subjective factors by analyzing the posterior probability of problem causes using a traceability model based on a Bayesian network; the data reliability is quantified using a data confidence formula, the interference of low-quality data on the traceability result is avoided, the data support is more scientific; the quantitative ordering of problem causes is realized through the Bayesian network posterior probability formula, the artificial subjective judgment error is reduced, and the traceability precision is improved; the data of the design stage, the processing stage and the detection stage of the machine case are integrated, the 'data island' is broken, the knowledge graph constructed on this basis realizes the reuse of historical experience, establishes a systematic knowledge accumulation and reuse mechanism, so that when the same problem occurs repeatedly, the historical solution can be quickly called, and the traceability efficiency is improved; the accuracy and interpretability of problem positioning are provided through the double verification of evidence chain and relationship chain. BRIEF DESCRIPTION OF DRAWINGS
[0015] Figure 1 A flowchart of the machine case quality problem traceability method in the embodiment is shown. DETAILED DESCRIPTION
[0016] The application will be further described in detail below in combination with the embodiments and drawings. However, it should not be understood that the above-mentioned subject matter of the application is limited to the following embodiments only, and any technology realized based on the content of the application belongs to the scope of the application.
[0017] Embodiment 1 Referring to Figure 1 , the embodiment provides a machine case quality problem traceability method, which comprises: S1: collecting historical manufacturing data of the machine case, wherein the historical manufacturing data comprises machine case design parameter data in the design stage, processing technology parameter data and processing equipment operation data in the processing stage, machine case detection data in the detection stage, and quality problem symptom description data.
[0018] Specifically, historical manufacturing data of the casing is collected through sensors, testing equipment, and manual input. This includes: design parameter data from the casing design phase, including 3D model parameters and material performance parameters; processing equipment operation data and processing technology parameter data from the processing phase, including equipment model, spindle speed, cutting fluid flow rate, and equipment maintenance records; and processing technology parameter data including cutting speed, feed rate, depth of cut, and processing equipment model. The testing phase includes casing inspection data and quality problem symptom description data; quality problem symptom description data includes crack location and deformation amount, etc.; casing inspection data includes 3D dimensional inspection data, material performance inspection data, and surface quality inspection data; the 3D dimensional inspection data includes casing dimensions, angles, and dimensional deviations; the material performance inspection data includes hardness and elongation; and the surface quality inspection data includes flatness and surface roughness, etc.
[0019] S2: Preprocess the historical manufacturing data to obtain the characterization parameters of each type of data in the historical manufacturing data. Then, use the data confidence formula to calculate the confidence of the characterization parameters of each type of data, and select the characterization parameters whose confidence reaches the preset threshold to form a standardized dataset.
[0020] Specifically, the historical manufacturing data is first preprocessed as follows: S21: Based on The principle is to identify and remove outliers from the historical manufacturing data. For the collected historical manufacturing data of the casing, the average value of each data category is first calculated. and standard deviation Then determine the average distance. Three standard deviations The range, i.e. [ -3 , +3 ], will not be [ -3 , +3 Data within the specified range is identified as outliers and removed. In practice, outliers in historical manufacturing data can also be identified and removed according to the 99.97% pass / fail principle typically required by manufacturing companies. This method effectively eliminates outliers in historical manufacturing data caused by measurement errors, equipment malfunctions, or human error, improving the accuracy and reliability of subsequent analysis.
[0021] Then, missing data from the collection and missing data after removal are filled in. For continuous data, mean imputation is used, and for discrete data, mode imputation is used to obtain historical manufacturing data with missing values filled in.
[0022] S22: Convert the data from different sources in the historical manufacturing data after missing values are filled into a unified data format and dimension to obtain standardized historical manufacturing data; S23: Use principal component analysis or wavelet transform to extract features from standardized historical manufacturing data to obtain the characterization parameters.
[0023] Then, the confidence level of the representation parameters for each type of data is calculated using the data confidence formula, and the representation parameters whose confidence levels reach a preset threshold are selected to form a standardized dataset; the data confidence formula is: ; in, The first in historical manufacturing data Confidence level of the representation parameters of class data. ; These are the weighting coefficients. Regarding the casing inspection data, A value of 0.7 was chosen to prioritize accuracy; this was based on the operating data of the processing equipment. The value is set to 0.6 to balance freshness; The first in historical manufacturing data Valid data in the class data; The first in historical manufacturing data The total amount of data in the data category; The first in historical manufacturing data The percentage of valid data in this category; The first in historical manufacturing data The freshness of the data; For the first The number of days from the current date to the data collection date for this type of data; The casing process cycle refers to the cycle including the casing design stage, the machining stage, and the inspection stage. =30 days. The preset threshold is set based on experience.
[0024] It should be noted that, taking the processing equipment operation data in historical manufacturing data as an example, the calculated confidence level refers to the confidence level of this type of data.
[0025] It should be noted that the weighting coefficients mentioned above... The determination of the weighting factor is based on business and industry. When historical manufacturing data of the casing is used as historical data, a high degree of accuracy is required, hence the weighting factor... The value is relatively large. For example, when assessing the confidence level of data related to computer box detection, the weighting coefficient... The value is 0.7.
[0026] S3: constructing a quality problem traceability knowledge graph of the machine case according to the standardized data set, the knowledge graph containing quality problem entities, influence factor entities, and relationships between the quality problem entities and the influence factor entities Specifically, the quality problem entity is a quality problem described in a quality problem symptom record, for example, a radial size out-of-tolerance. The influence factor entity is a factor that may cause the quality problem entity, for example, the influence factors that may cause the radial size out-of-tolerance include tool wear, insufficient material hardness, etc.
[0027] The relationship definition refers to determining the causal relationship, the association relationship, and the time sequence relationship between the quality problem entity and the attribute. The causal relationship is, for example, “tool wear → radial size out-of-tolerance”, the association relationship is, for example, “machining center A → operator B”, and the time sequence relationship is, for example, “after heat treatment → finish milling process”.
[0028] The causal relationship is determined through historical quality problem processing records and expert experience, and the time sequence relationship is established based on data collection time stamps.
[0029] According to the standardized data set, the quality problem entities and the relationships are constructed, that is, the quality problem traceability knowledge graph of the machine case is obtained. It should be noted that the quality problem traceability knowledge graph of the machine case can be constructed by using an existing traceability knowledge graph construction method.
[0030] S4: performing reasoning analysis based on the quality problem traceability knowledge graph of the machine case according to the problem symptom of the current quality problem of the machine case to be traced, obtaining all influence factor entities that may cause the current quality problem of the machine case, and taking the influence factor entities as problem causes of the current quality problem of the machine case; obtaining the posterior probability of each problem cause by using a traceability model analysis of a Bayesian network, selecting problem causes with posterior probabilities greater than a preset probability threshold to form a problem cause candidate set, and determining the problem cause with the maximum posterior probability in the problem cause candidate set as the root cause.
[0031] Specifically, first, the traceability model based on the Bayesian network is constructed as follows: ; wherein, is the posterior probability of the problem cause under the evidence, is the likelihood probability of the evidence appearing when the problem cause occurs, is the prior probability of the problem cause, is the nth quality problem, is the quality problem entity of the nth quality problem, is the influence factor entity of the nth quality problem, is the first influence factor entity of the nth quality problem, is the second influence factor entity of the nth quality problem, is the first quality problem, is the quality problem entity of the first quality problem, is the influence factor entity of the first quality problem. a cause of a quality problem, a piece of evidence of a quality problem of the cartridge. The evidence of the quality problem of the cartridge may be one of manufacturing data such as equipment operation data, if the data may cause a quality problem, the data is evidence of the occurrence of the quality problem, that is, the evidence of the quality problem. The evidence of the quality problem of the cartridge and the quality problem are one-to-one corresponding, one quality problem corresponds to one evidence of the quality problem of the cartridge .
[0032] Since the cartridge quality problem traceability knowledge graph has an association relationship between the problem cause and the evidence, the likelihood probability and the prior probability can be obtained by analyzing the historical manufacturing data collected in step S1.
[0033] Then, according to the problem symptom of the current quality problem of the cartridge to be traced, reasoning analysis is performed based on the cartridge quality problem traceability knowledge graph to obtain all influence factor entities that may cause the cartridge to have the current quality problem, and the influence factor entities are taken as the problem causes that cause the cartridge to have the current quality problem. The traceability model analysis of the Bayesian network is used to obtain the posterior probability of each problem cause, the problem causes with posterior probability greater than a preset probability threshold are selected to form a problem cause candidate set, and the problem cause with the maximum posterior probability in the problem cause candidate set is determined as the root cause. The preset probability threshold is set to 0 in this embodiment, and a person skilled in the art can adjust the preset probability threshold according to needs or experience.
[0034] S5: For each problem cause in the problem cause candidate set, retrieve the associated data in the standardized data set, and construct an evidence chain and an influence factor relationship chain from the problem symptom to the root cause about the current quality problem.
[0035] Specifically, first, for each problem cause in the problem cause candidate set, the associated historical manufacturing data such as equipment parameters are extracted by retrieving the standardized data set.
[0036] Second, according to the association relationship between the historical manufacturing data in the cartridge quality problem traceability knowledge graph, the association relationship between each problem cause in the candidate set and the corresponding historical manufacturing data is constructed to form a link graph of the problem cause and the corresponding historical manufacturing data, that is, a relationship chain.
[0037] Finally, the historical manufacturing data related to the cause of the problem is analyzed to build an evidence chain of "problem symptoms -> direct causes -> root causes" and a relationship chain of "equipment -> process -> personnel" related to the current quality problem. Among them, the direct cause is other influencing factors on the link from the problem symptom to the root cause.
[0038] S6: According to the evidence chain and the relationship chain of influencing factors, a traceability report is generated.
[0039] Specifically, a traceability report including the problem causes corresponding to the current quality problem to be traced of the machine case is generated, each problem cause in the traceability report is attached with a posterior probability value, and visual display is supported.
[0040] The embodiment adopts a data confidence formula to quantify data reliability, avoids interference of low-quality data on the traceability result, makes data support more scientific, realizes quantitative sorting of problem causes through a Bayesian network posterior probability formula, reduces artificial subjective judgment error, and improves traceability accuracy; through integration of data in the machine case design stage, the processing stage and the detection stage, the "data island" is broken, and the knowledge graph built on this basis realizes historical experience reuse, and improves traceability efficiency; through double verification of the evidence chain and the relationship chain, accuracy and explainability of problem positioning are provided.
[0041] The application reduces the proportion of artificial intervention through data acquisition and preprocessing, analyzes the posterior probability of problem causes by using a traceability model based on a Bayesian network, avoids empiricism influence, improves traceability analysis efficiency, and reduces influence of artificial subjective factors; through integration of data in the design, processing and detection whole process, and formation of a standardized data set, the data island in the prior art is broken; through construction of a machine case quality problem traceability knowledge graph, a systematic knowledge accumulation and reuse mechanism is established, so that when the same problem occurs repeatedly, historical solutions can be quickly called; low-quality data is removed through a data confidence formula to avoid interference with the traceability result, the posterior probability of problem causes is analyzed by using a traceability model based on a Bayesian network, and traceability accuracy is improved.
[0042] The application quantifies data reliability by using a confidence calculation formula, avoids interference of low-quality data on quality problem traceability. There is no concept of data reliability in the prior art, and data priority cannot be distinguished.
[0043] The application takes data screened through the confidence as input of the traceability model based on the Bayesian network, provides high-confidence data to the Bayesian network for reasoning, and improves traceability accuracy.
[0044] The application forms a complete closed loop through the cooperation of "confidence formula (high-quality data input) → knowledge graph (field knowledge storage) → Bayesian network (quantitative reasoning) → evidence chain (verifiable output)", meets the stringent requirements of the aviation field, and is not a simple patching of existing technologies.
[0045] Embodiment 2 This embodiment takes the traceability of the "flange surface flatness out-of-tolerance" quality problem of a certain type of engine casing as an example to illustrate the application, as follows: (1) Manufacturing data collection: Collect manufacturing data such as design drawings of the batch of casings, processing equipment operation data, casing three-coordinate detection data, equipment maintenance records, etc. The design drawings record design parameters such as flange surface flatness requirement ≤0.05mm, etc. The processing equipment operation data includes equipment model M7130, spindle speed 1800r / min, feed rate 0.1mm / r, etc. The casing three-coordinate detection data includes the actual flange surface flatness detection value of the casing, which is 0.08mm. The equipment maintenance record records that the processing equipment of model M7130 was calibrated for the last time 10 days before processing the casing, etc.
[0046] (2) Manufacturing data preprocessing: 2 abnormal values in the processing equipment operation data caused by instantaneous sensor failure are removed, and the effective data ratio is: =98 / 100=0.98; Calculate the confidence of the processing equipment operation data: Since the processing equipment operation data is collected 1 day before the current date, =1 day, =30 days, =0.7, the confidence of the processing equipment operation data is =0.7×0.98+0.3×(1 / 30)=0.686+0.01=0.696; Standardize the casing flange surface flatness detection data, etc.
[0047] (3) Knowledge graph retrieval: In the constructed casing quality problem traceability knowledge graph, match the "flatness out-of-tolerance" entity, and associate the possible causes of the "flatness out-of-tolerance" quality problem, including "equipment uncalibration" and "grinding wheel wear".
[0048] (4) Traceability analysis: analyze all problem causes using the traceability model, and obtain the posterior probability of each problem cause: =0.7, =0.3. At the same time, combined with the strong association relationship of "equipment uncalibration → flatness out-of-tolerance" in the knowledge graph, determine that the main cause candidate is equipment uncalibration.
[0049] (5) Evidence chain generation: The standardized data set is retrieved to find that the processing equipment of model M7130 has not been calibrated for nearly 10 days, and the flatness deviation of the first four products of the same batch of the machine case is all greater than or equal to 0.07 mm, forming an evidence chain of “equipment not calibrated→flatness out of tolerance” (the correlation data confidence is all greater than or equal to 0.6).
[0050] (6) Output result: The traceability report suggests that the equipment should be calibrated immediately and all the products of the same batch should be rechecked, and the report clearly marks “the probability of equipment not being calibrated as the reason is 70%”.
[0051] The above is only a preferred embodiment of the present application, and is not used to limit the present application, and any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims
1. A method of tracing a quality problem of a machine case, characterized by, The method comprises the following steps: S1: collecting historical manufacturing data of the cartridge, wherein the historical manufacturing data comprises cartridge design parameter data in a design stage, machining process parameter data and machining equipment operation data in a machining stage, cartridge detection data and quality problem symptom description data in a detection stage; S2: preprocessing the historical manufacturing data to obtain a representation parameter of each type of data in the historical manufacturing data, then calculating a confidence degree of the representation parameter of each type of data by using a data confidence degree formula, and screening out the representation parameter with a confidence degree reaching a preset threshold to form a standardized data set; S3: constructing a cartridge quality problem traceability knowledge graph according to the standardized data set, wherein the knowledge graph comprises quality problem entities, influence factor entities, and relationships between the quality problem entities and the influence factor entities; S4: performing reasoning analysis based on the cartridge quality problem traceability knowledge graph according to a problem symptom of a current quality problem of a cartridge to be traced, obtaining all influence factor entities that can cause the current quality problem of the cartridge, and taking the influence factor entities as problem causes of the current quality problem of the cartridge; obtaining a posterior probability of each problem cause by using a traceability model of a Bayesian network, selecting problem causes with a posterior probability greater than a preset probability threshold to form a problem cause candidate set, and determining a problem cause with the maximum posterior probability in the problem cause candidate set as a root cause; S5: for each problem cause in the problem cause candidate set, searching for associated data in the standardized data set to construct an evidence chain and an influence factor relationship chain from the problem symptom to the root cause of the current quality problem; S6: generating a traceability report according to the evidence chain and the influence factor relationship chain.
2. The châssis quality problem tracing method according to claim 1, characterized in that, The machining process parameters comprise cutting speed, feed rate, cutting depth and machining equipment model; and the cartridge detection data comprises three-dimensional size detection data, material performance detection data and surface quality detection data.
3. The châssis quality problem tracing method according to claim 1, characterized in that, The method for preprocessing the historical manufacturing data to obtain the representation parameter comprises: based on The abnormal values in the historical manufacturing data are identified and removed according to the principle, and then the historical manufacturing data is filled with missing values to obtain historical manufacturing data filled with missing values. converting data of different sources in the historical manufacturing data after filling in missing values into a unified data format and dimension to obtain standardized historical manufacturing data; performing feature extraction on the standardized historical manufacturing data by using a principal component analysis method or a wavelet transform method to obtain the representation parameter.
4. The châssis quality problem tracing method according to claim 3, characterized in that, The data confidence degree formula is: ； wherein, is a confidence of the representation parameter of the data of the is a weight coefficient; is valid data in the data of the is a total amount of data in the data of the is a valid data proportion of the data of the is a freshness of the data of the is a number of days from a data collection date of the data of the is a process cycle of the engine case. 5. The châssis quality problem tracing method according to claim 1, characterized in that, The relationships between the entities comprise causal relationships, association relationships and time sequence relationships.
6. The châssis quality problem tracing method according to claim 1, characterized in that, The traceability model based on the Bayesian network is constructed as: ； in, In the evidence The reason for the problem The posterior probability, Reason for the problem Evidence at the time of occurrence The likelihood probability of occurrence Reason for the problem The prior probability, For the first One quality issue, To generate the first casing The reason for the quality problem This serves as evidence of a quality issue with the casing.
7. The châssis quality problem traceability method according to claim 1, characterized in that, In step S5, the method for constructing the evidence chain and the influence factor relationship chain from the problem symptom to the root cause of the current quality problem comprises: for each problem cause in the problem cause candidate set, searching in the standardized data set to extract historical manufacturing data associated with each problem cause; constructing the association relationship between each problem cause in the candidate set and the corresponding historical manufacturing data according to the association relationship between the historical manufacturing data in the cartridge quality problem traceability knowledge graph to form a link graph of the problem cause and the corresponding historical manufacturing data, i.e., a relationship chain. The historical manufacturing data related to the problem cause is analyzed, and an evidence chain including a problem symptom, a direct cause, and a root cause and a relationship chain of influencing factors including equipment, process, and personnel are constructed.
8. The châssis quality problem traceability method according to claim 1, characterized in that, In step S6, the traceability report includes a problem cause corresponding to a current quality problem of the cartridge to be traced.