A product quality life cycle management method for a projection lamp for a vehicle

Abstract

The application discloses a product quality life cycle management method of a vehicle projection lamp, and belongs to the technical field of quality management, and specifically comprises the following steps: disassembling and recycling a scrapped vehicle projection lamp to generate a production recycling database; generating a reliability coefficient curve of each component under each supplier by statistically analyzing failure duration data; analyzing co-occurrence data of multiple component failures in the same whole machine, and calculating failure correlation coefficients between components of different suppliers; combining the reliability coefficient and the failure correlation coefficient, automatically traversing and evaluating all possible supplier combinations, constructing a system survival probability model, and integrating and calculating the expected reliability of each combination to generate several supply chain combination schemes; screening a target scheme by comprehensively considering cost parameters and expected reliability to guide production, and establishing a traceable traceability file for the product. The application realizes dynamic closed-loop optimization of the supply chain based on inversion of recycling data, and improves the scientificity and accuracy of product quality and supply chain management.

Description

Technical Field

[0001] This invention relates to the field of quality management technology, specifically to a product quality lifecycle management method for automotive projection lights. Background Technology

[0002] As complex automotive electronic components integrating optics, electronics, and structural elements, automotive projection lights directly impact user experience and brand reputation in terms of quality and reliability. Traditional product quality management methods, which focus primarily on design verification, production sampling, and after-sales troubleshooting, have significant limitations.

[0003] Currently, supply chain management and optimization in the industry mainly rely on supplier onboarding audits, incoming quality control (IQC), and subjective judgment based on historical procurement experience. When selecting component suppliers, decisions often focus on the procurement cost, delivery time, or the factory pass rate based on a limited sample. However, this static and isolated evaluation method cannot reveal deeper quality risks: First, components from different suppliers have different long-term lifespans and degradation patterns in actual automotive environments, which are difficult to accurately assess based on short-term testing alone; second, and more importantly, when components from different suppliers work together in the overall system, they may produce a "1+1>2" negative effect due to electrical matching, thermal characteristics, or mechanical stress coupling, i.e., triggering correlated failures. This risk is seriously overlooked in the traditional component-based management model.

[0004] On the other hand, in traditional quality systems, product end-of-life recycling is often seen as the end point of management, and the value of the key failure data attached to it has not been fully explored. Recycled faulty parts essentially record the performance evolution of components under real loads, environments, and a complete service life, making them a high-value data source for assessing long-term supplier quality consistency and identifying potential design flaws and systemic failure mechanisms. However, in current practices, this data is either lost due to a lack of systematic collection or is only used for macro-level failure rate statistics, failing to be linked to specific component production batches and supply sources through effective traceability mechanisms. Furthermore, it is not deeply utilized to build quantifiable and predictable evaluation models, making it difficult to provide forward-looking guidance for subsequent product design improvements and precise supply chain decisions. Summary of the Invention

[0005] The purpose of this invention is to provide a product quality lifecycle management method for automotive projection lamps, solving the following technical problems:

[0006] Existing quality management methods do not consider assessing the associated failure risks caused by different combinations of components from different suppliers in real-world use, making it difficult to achieve dynamic optimization of the supply chain and closed-loop improvement of product reliability.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A product quality lifecycle management method for automotive projection lights includes the following steps:

[0009] S1, read the whole machine identification and disassemble the components of the scrapped vehicle projection lamp. During the disassembly process, determine the failure of the components and record the cumulative usage time of the failed components. Integrate the whole machine identification, component identification, corresponding supplier identification and usage time of the failed components into recycling data entries and store them in the recycling database.

[0010] S2, classify the recycled data entries in the recycling database according to component identifier and supplier identifier, and generate the reliability coefficient of each component under each supplier based on the usage time data of the failed components in each group after classification.

[0011] S3, extract recycling data entries of multiple component failures under the same whole machine identifier from the recycling database, and calculate the failure correlation coefficient of components from different suppliers under the same whole machine based on the reliability coefficient;

[0012] S4. Based on the reliability coefficient and the failure correlation coefficient, all suppliers are combined to generate several supply combination schemes and the overall reliability coefficient corresponding to each supply combination scheme is calculated.

[0013] S5. Obtain the cost parameters of each supplier combination scheme, and calculate the priority score of each supplier combination scheme based on the overall reliability coefficient, and mark the supplier combination scheme with the highest priority score as the target scheme.

[0014] S6. Based on the target solution, generate production guidance information, and after the automotive projection lamp is assembled, generate a product traceability file that is associated with the whole machine identifier and the target solution.

[0015] As a further aspect of the present invention: the specific process for generating the reliability coefficients of each component under each supplier in step S2 is as follows:

[0016] The recycling database is used to filter out recycling data entries whose component status is determined to be failed. The filtered entries are then categorized and grouped according to component identifier and supplier identifier. For each categorized group, the cumulative usage time data and the total number of failed components are extracted from all failed component records within the group. All extracted cumulative usage times are sorted by numerical value, and the sorted usage time sequence is divided into several consecutive time intervals.

[0017] The number of failed components within each time interval is counted, and the proportion of the number of failed components to the total number of failed components is calculated. The end time of each time interval is correlated with its corresponding cumulative failure ratio, and a curve of the cumulative failure ratio increasing with the usage time is plotted. The curve is then labeled as the reliability coefficient of the component under this supplier.

[0018] As a further aspect of the present invention: the specific calculation process of the failure correlation coefficient in S3 is as follows:

[0019] Entries with multiple component failures under the same complete machine are screened from the recycling database. The component identifier and supplier identifier of the failed component are extracted and aggregated into observed failure combinations. The frequency of occurrence of each observed failure combination is counted, and combinations with a frequency exceeding a threshold are identified as associated combinations.

[0020] Select a preset usage time point, and read the cumulative failure ratio at that preset usage time point from the reliability coefficient of each component identifier under its corresponding supplier identifier in the associated combination; multiply the cumulative failure ratios of all components to obtain the theoretical independent failure combined ratio;

[0021] The total number of complete machines with supplier identifiers recorded using this association combination is counted, as well as the total number of complete machines that have not exceeded the preset usage time point in terms of the cumulative usage time of all components; the ratio of the total number of complete machines that have not exceeded the time limit to the total number of complete machines is calculated to obtain the actual observed failure joint ratio.

[0022] Calculate the difference between the actual observed failure joint ratio and the theoretical independent failure joint ratio, and define the difference as the failure correlation coefficient of the correlation combination at the preset usage time point.

[0023] As a further aspect of the present invention: the specific process for generating several supply combination schemes in S4 is as follows:

[0024] Select any component identifier as the target component identifier, and obtain any supplier identifier corresponding to the target component identifier as the current benchmark supplier; obtain a set of candidate supplier identifiers for the remaining component identifiers, and combine the current benchmark supplier with a supplier identifier selected from each of the remaining sets to form a combination to be evaluated;

[0025] Obtain the reliability coefficients of each component in the supplier identification combination to be evaluated under the corresponding supplier and the failure correlation coefficients of the supplier identification combination to be evaluated under the same whole machine, and calculate the expected reliability of the whole lamp of the supplier identification combination to be evaluated.

[0026] Enumerate all supplier identification combinations to be evaluated, calculate the expected reliability of the whole lamp for each supplier identification combination to be evaluated, and mark the supplier identification combination with the highest expected reliability of the whole lamp as the best cooperative supplier combination scheme of the current benchmark supplier.

[0027] Iterate through each candidate supplier identifier of the target component identifier, and use them as the current benchmark supplier in turn. Repeat the above process to determine all the generated best cooperative supplier combinations as several supply combination schemes.

[0028] As a further aspect of the present invention: the specific process for generating the expected reliability of the entire lamp is as follows:

[0029] A set of candidate usage duration points arranged in chronological order is pre-defined. For each candidate usage duration point, the corresponding cumulative failure ratio is read from the reliability coefficient corresponding to each component identifier in the combination to be evaluated. The component survival ratio is obtained by subtracting the cumulative failure ratio from one.

[0030] Obtain the failure correlation coefficient of the combination to be evaluated at the candidate usage time point; subtract the failure correlation coefficient from one to obtain the coupling adjustment coefficient; multiply the component survival rate by the coupling adjustment coefficient to obtain the corrected survival rate;

[0031] Multiply the corrected survival rates of all components in the assembly to be evaluated at the candidate usage time points to obtain the overall lamp system survival rate; iterate through all candidate usage time points to obtain a sequence of overall lamp system survival rates;

[0032] Calculate the duration difference between each pair of adjacent candidate usage duration points in the sequence, and calculate the average survival rate of the two whole lamp systems corresponding to the pair of adjacent candidate usage duration points; multiply the duration difference by the average value to obtain the area of ​​the sub-interval corresponding to each pair of adjacent points; sum all the sub-interval areas, and label the summation result as the expected reliability coefficient of the whole lamp for the supplier identifier combination to be evaluated.

[0033] As a further aspect of the present invention, it also includes a pre-set set of candidate usage duration points arranged in chronological order, wherein the value of the largest candidate usage duration point is less than the end point of the shortest time range covered by the reliability coefficients corresponding to all component identifiers in the combination to be evaluated; when any candidate usage duration point does not correspond to the failure correlation coefficient of the combination to be evaluated, the failure correlation coefficient used to calculate the coupling adjustment coefficient is set to zero.

[0034] As a further aspect of the present invention: in step S5, the specific process for generating the target solution is as follows:

[0035] The cost parameters and the overall reliability coefficient are standardized numerically to obtain the corresponding standard cost score and standard reliability score. Based on the preset weight allocation, the standard reliability score and the standard cost score are weighted and summed to obtain the priority score of each supplier combination scheme. The priority scores of all supplier combination schemes are compared, and the scheme with the highest priority score is designated as the target scheme.

[0036] As a further aspect of the present invention: S6 further includes, when the vehicle projection lamp is scrapped and recycled, reading the product traceability file according to its whole machine identification, adding a new recycling record to the recycling database according to the supplier combination information in the product traceability, and triggering the update of the reliability parameters and failure correlation relationship.

[0037] The beneficial effects of this invention are:

[0038] 1) This invention constructs a component-level failure database by recycling dismantling records and failure data of scrapped products. Based on this database, it statistically analyzes the failure duration of various components from each supplier, generating reliability coefficient curves that characterize the failure patterns throughout the entire lifecycle. Simultaneously, by analyzing records of failures of multiple components within the same complete machine, it statistically analyzes their co-occurrence patterns and frequencies, calculating failure correlation coefficients that quantify the risk of collaborative failures between components from different suppliers. This method transforms the traditionally difficult-to-detect and even more difficult-to-quantify "compatibility" issues or implicitly coupled failure risks between components into explicit numerical indicators. This upgrades the supplier evaluation system from relying on static and isolated dimensions such as single-piece sampling pass rates and supply prices to a dynamic and comprehensive data-driven model encompassing long-term actual service life and system-wide interactive risks, providing supply chain management with empirically based in-depth insights and decision support.

[0039] 2) This invention achieves evidence-based supply chain portfolio optimization by establishing a quantitative link between product end-of-life recycling data and front-end supply chain decisions. The method first uses component failure durations recorded in the recycling database to generate reliability curves characterizing the long-term failure patterns of components from various suppliers. Then, by statistically analyzing the frequency and patterns of collaborative failures of multiple components within the same complete unit, a failure correlation coefficient is calculated to quantify the coupling risk between components. Based on this, the algorithm automatically iterates through all supplier combinations, constructs a system-wide survival probability model, and integrates the expected reliability to match each core component supplier with a collaborative supply combination that maximizes its expected reliability. Finally, by standardizing and weighting the expected reliability and cost parameters of each combination, a quantified priority score is generated, thereby objectively selecting the target supplier solution that achieves the optimal balance between reliability and economy. This method replaces the traditional manual matching mode that relies on experience with data- and model-driven automated decision-making.

[0040] 3) This invention establishes a unique product traceability file for each product manufactured according to the target solution, binding it with supplier portfolio information, thus constructing a data traceability link connecting the production start point and the recycling end point. When a product is scrapped and recycled, by identifying its whole machine identifier and reading the corresponding file, the failure data and usage duration of the recycled components can be accurately classified under their original supplier portfolio and fed back to the recycling database. This mechanism forms a complete data closed loop from "historical recycling data analysis → generating optimization solutions → guiding new product production → new product use and recycling → data feedback and model update". Based on this, supplier reliability curves, component failure correlation coefficients, and optimal supply chain portfolio solutions can all be periodically and dynamically recalculated and updated as real market recycling data continues to be incorporated, thereby enabling the entire quality management system to have the ability to self-iterate and continuously optimize, ensuring that quality decisions always evolve in sync with the product's performance in actual use and the latest failure modes, achieving a true full lifecycle management closed loop from design, production, use to recycling feedback. Attached Figure Description

[0041] The invention will now be further described with reference to the accompanying drawings.

[0042] Figure 1 This is a schematic diagram of a product quality lifecycle management method for automotive projection lamps according to the present invention. Detailed Implementation

[0043] 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.

[0044] Please see Figure 1 As shown, this invention is a product quality lifecycle management method for automotive projection lamps, comprising the following steps:

[0045] S1, read the whole machine identification and disassemble the components of the scrapped vehicle projection lamp. During the disassembly process, determine the failure of the components and record the cumulative usage time of the failed components. Integrate the whole machine identification, component identification, corresponding supplier identification and usage time of the failed components into recycling data entries and store them in the recycling database.

[0046] First, on a dedicated recycling line, operators or automated equipment scan the unique machine identification code, such as a one-dimensional barcode or QR code, on the casing of the scrapped vehicle's projection lamp. This identification code is assigned when the vehicle leaves the factory and remains in use throughout the product's entire lifecycle.

[0047] Subsequently, the projection lamp was disassembled in a standardized manner. The disassembly process followed a predetermined workflow, sequentially separating core components such as the outer casing, heat dissipation module, optical lens, LED light source module, and driver circuit board. Failure assessment was performed simultaneously with each component disassembly. Assessment methods included visual inspection (e.g., lens cracks, casing corrosion), basic functional power-on testing (e.g., whether the LED emits light, whether the driver circuit has output), and, if necessary, more precise instrument testing (e.g., measuring LED light decay, capacitor capacitance). Once a component was determined to be in a failed state, its component identification (e.g., "LED Module - Type A") and its corresponding supplier identification (e.g., "Supplier A") were recorded. Simultaneously, by querying the vehicle's vehicle network history data or maintenance records, the cumulative usage time of the failed component from the vehicle's first registration to its scrapping and recycling (e.g., "5 years and 2 months" or "1520 days") was obtained and recorded.

[0048] Finally, the four key information fields—machine ID, component ID, supplier ID, and cumulative usage time—are integrated to form a structured recycling data entry and stored in the central recycling database. For example, a record might be: "Machine ID: VIN123456, Component: Drive Circuit-V2, Supplier: Supplier E, Failure Time: 4.8 years".

[0049] S2, classify the recycled data entries in the recycling database according to component identifier and supplier identifier, and generate the reliability coefficient of each component under each supplier based on the usage time data of the failed components in each group after classification.

[0050] The database undergoes data processing. First, data cleaning is performed to filter out entries with a clearly defined "failed" status. Then, entries are grouped using the component identifier and supplier identifier as a joint key. For example, all failure records with "component identifier 'LED Module-A' and supplier identifier 'Supplier A'" are grouped into one group; records with "component identifier 'LED Module-A' but supplier identifier 'Supplier B'" are grouped into another group.

[0051] For each categorized group, the cumulative usage time data for all items within that group is extracted to form a usage time dataset. The usage time values ​​in this dataset are then sorted in ascending order. To analyze the failure distribution, this ordered usage time sequence is divided into several consecutive intervals (e.g., every 0.5 years or every 1 year). The number of failed components within each usage time interval is counted, and the proportion of this number to the total number of failed components in that group is calculated; this is the cumulative failure rate for that interval.

[0052] Next, points were plotted on a coordinate system with the end time of each time interval as the x-axis and the corresponding cumulative failure rate as the y-axis. Connecting these points formed a curve showing how the cumulative failure rate increases with usage time. This curve visually illustrates the statistical pattern of cumulative failures of a specific component supplied by a particular vendor over time. Ultimately, this complete curve itself is defined as the reliability coefficient of the component from that specific vendor. It is no longer a single number, but rather contains rich information about the entire process from early failure to wear-out failure.

[0053] S3, extract recycling data entries of multiple component failures under the same whole machine identifier from the recycling database, and calculate the failure correlation coefficient of components from different suppliers under the same whole machine based on the reliability coefficient;

[0054] The database was filtered to identify sets of data entries where two or more components failed under the same unit identifier. These entries revealed instances of "co-failure" of components in a real-world operating environment. For example, records from all units that simultaneously contained failures of both "Supplier A's LED module" and "Supplier E's driver circuit" were filtered out.

[0055] For each frequently co-occurring specific component and supplier combination (referred to as the observed failure combination) selected by the system, the frequency of its occurrence in all recovery records is counted. Combinations whose frequency exceeds a preset empirical threshold are identified as associated combinations that require further analysis.

[0056] To calculate the degree of correlation, a representative preset usage period (e.g., "4 years") is selected. For each component in this correlation combination (e.g., "Supplier A's LED module"), the cumulative failure rate at the "4-year" point is read from its reliability coefficient curve generated in S2 (assuming readings of 30% and 15% respectively). Assuming that the failures of these two components are independent, their respective cumulative failure rates are multiplied (30% × 15% = 4.5%) to obtain the theoretical independent failure joint rate. This means that if the two components are completely unrelated, theoretically only 4.5% of the products will fail after 4 years of use.

[0057] Next, practical verification was conducted: the total number of complete units using this specific supplier combination (i.e., using "LED modules from supplier A" and "driver circuits from supplier E") was counted from the recycling database. Then, the number of complete units in which both components failed within a cumulative usage period of less than 4 years was counted. The ratio of the latter to the former was calculated to obtain the actual observed joint failure rate (assuming that the statistics showed this rate to be as high as 18%).

[0058] Finally, the difference between the actual observed combined failure rate and the theoretical combined independent failure rate was calculated (18% - 4.5% = 13.5%). This positive difference of 13.5% is quantified as the failure correlation coefficient of this associated combination (LED module from supplier A & driver circuit from supplier E) at a 4-year service point. A significant positive value strongly indicates the risk of incompatibility or mutual exacerbation of failure between the two components.

[0059] S4. Based on the reliability coefficient and the failure correlation coefficient, all suppliers are combined to generate several supply combination schemes and the overall reliability coefficient corresponding to each supply combination scheme is calculated.

[0060] With the goal of optimizing the overall lamp lifespan, the system intelligently matches the best suppliers for the remaining components to each potential "core component" supplier. For example, if the "LED module" is selected as the core component, there are two suppliers, A and B.

[0061] First, assuming supplier A's LED module is used as the current benchmark, all alternative supplier combinations for "optical lens" and "driving circuit" will be automatically traversed (e.g., there are two suppliers, C and D, for lenses, and two suppliers, E and F, for circuits), forming multiple supplier identification combinations to be evaluated, such as {A,C,E}, {A,C,F}, {A,D,E}, and {A,D,F}.

[0062] For each combination to be evaluated (taking {A,C,E} as an example), based on the reliability coefficient curves of A, C, and E generated in S2, and the failure correlation coefficients of all relevant pairs (such as {A,E}, {C,E}, etc., if they exist) calculated in S3, a model is constructed and integrated to calculate the survival probability of the entire lighting system over time. This model is then used to quantify and predict the expected reliability of the entire lighting system under this combination (e.g., a value expressed as "mean time between failures" or "the number of years to reach a certain low failure probability").

[0063] After completing the calculations for all combinations to be evaluated, the combination that maximizes the expected reliability of the entire lamp (let's say {A,D,F}) will be selected for the current benchmark supplier A, and this combination will be marked as the best cooperative supplier combination for supplier A.

[0064] Switching the benchmark to LED modules from supplier B, the above traversal, calculation, and screening process is repeated to obtain the optimal cooperation scheme for supplier B (e.g., {B,C,E}). Ultimately, all these optimal supplier combinations (e.g., {A,D,F} for A and {B,C,E} for B) constitute several supply combination schemes for the decision-making level to choose from. Each scheme is accompanied by a calculated overall reliability coefficient (i.e., the predicted expected reliability of the entire lamp).

[0065] S5. Obtain the cost parameters of each supplier combination scheme, and calculate the priority score of each supplier combination scheme based on the overall reliability coefficient, and mark the supplier combination scheme with the highest priority score as the target scheme.

[0066] After obtaining several supply combination options and their corresponding overall reliability coefficients (representing quality levels), the system will retrieve the cost parameters of all component procurement involved in each option from the Enterprise Resource Planning (ERP) or procurement system, and summarize them into the total cost of the option.

[0067] Next, multi-objective decision-making is carried out. First, the data in the "overall reliability coefficient" and "total cost" columns are standardized to eliminate the influence of dimensions, so that they are converted into standard reliability scores and standard cost scores in the range of 0 to 1 (generally, the higher the reliability, the higher the score, and the lower the cost, the higher the score).

[0068] Then, based on the company's strategy (whether quality or cost is prioritized), appropriate weights are assigned to reliability and cost (e.g., 70% for reliability and 30% for cost). For each option, its standard reliability score and standard cost score are weighted and summed according to their respective weights to calculate the priority score for that option.

[0069] Finally, the priority scores of all options are compared, and the supply combination with the highest priority score is designated as the final target option to guide production. This process achieves an objective and quantifiable optimal balance between quality and cost.

[0070] S6. Based on the target solution, generate production guidance information, and after the automotive projection lamp is assembled, generate a product traceability file that is associated with the whole machine identifier and the target solution.

[0071] Once the target solution is determined (e.g., solution {A,D,F}), the system will automatically generate detailed production guidance information, including a bill of materials (specifying that LED modules from supplier A, optical lenses from supplier D, and driver circuits from supplier F must be purchased), assembly process requirements, etc., and send it to the Manufacturing Execution System (MES) to guide the production line assembly.

[0072] Each automotive projection light manufactured according to this target solution is assigned a unique unit identifier after assembly. The system creates a product traceability file, strongly associating this unit identifier with the target solution adopted (i.e., the supplier combination {A,D,F}), and persistently storing it. This file is crucial for future quality traceability and data analysis, ensuring the integrity of the data chain from production to recycling. When the light is eventually scrapped and recycled, its identifier can be read to immediately identify which suppliers' components it was assembled from, allowing for accurate attribution of failure data for the next round of model optimization, thus forming a closed-loop management system.

[0073] In a preferred embodiment of the present invention, the specific process of generating the reliability coefficients of each component under each supplier in step S2 is as follows:

[0074] The recycling database is used to filter out recycling data entries whose component status is determined to be failed. The filtered entries are then categorized and grouped according to component identifier and supplier identifier. For each categorized group, the cumulative usage time data and the total number of failed components are extracted from all failed component records within the group. All extracted cumulative usage times are sorted by numerical value, and the sorted usage time sequence is divided into several consecutive time intervals.

[0075] The number of failed components within each time interval is counted, and the proportion of the number of failed components to the total number of failed components is calculated. The end time of each time interval is correlated with its corresponding cumulative failure ratio, and a curve of the cumulative failure ratio increasing with the usage time is plotted. The curve is then labeled as the reliability coefficient of the component under this supplier.

[0076] First, all recycling data entries whose component status is determined to be failed are automatically filtered from the recycling database. For example, all LED light source module records marked "failed" are filtered out. Next, the filtered entries are categorized and grouped according to specific component and supplier identifiers. For example, all failure records with the component identifier "LED Module - Model L1" and the supplier identifier "Supplier A" are grouped into the same independent analysis group. Then, for this group, the cumulative usage time data recorded in each failure component record within the group is extracted. For example, a series of usage time values ​​such as "3 years", "4.5 years", and "5.2 years" are extracted. Finally, all extracted usage time values ​​are sorted in ascending order. An ordered time sequence is formed; then the ordered sequence is divided into several consecutive time intervals, such as "0-1 year", "1-2 years", etc., with a one-year interval; then the number of failed components falling into each time interval is counted, and the proportion of this number to the total number of failed components in that group is calculated; then the end time of each time interval is correlated with its calculated cumulative failure ratio, for example, the end time of "2 years" is correlated with the corresponding cumulative failure ratio; finally, based on all these correlation points, a smooth curve is drawn showing how the cumulative failure ratio changes with the increase of usage time, and this complete curve is officially labeled as the reliability coefficient of the component under that specific supplier.

[0077] Abandoning traditional, one-sided evaluation methods that rely on a single pass rate or average lifespan, this approach directly constructs a probability distribution model reflecting the entire process of a component group from its initial use to its final failure through statistical analysis of the lifespan of failed components in a large number of real scrapped products. The advantage is that the generated reliability coefficient curves can intuitively reveal the failure risk characteristics of components at different stages of use, such as whether they are prone to early failure or experience concentrated wear after stable operation. This provides an objective, comprehensive, and traceable data foundation for accurately assessing the true quality level of suppliers. This step lays a crucial empirical data foundation for subsequent accurate calculations of failure correlations between components and the scientific optimization of the supply chain portfolio. It ensures that the entire quality management decision-making process is based on real, dynamic product lifecycle performance from the outset, rather than abstract theoretical indicators or isolated sampling results.

[0078] In another preferred embodiment of the present invention, the specific calculation process of the failure correlation coefficient in step S3 is as follows:

[0079] Entries with multiple component failures under the same complete machine are screened from the recycling database. The component identifier and supplier identifier of the failed component are extracted and aggregated into observed failure combinations. The frequency of occurrence of each observed failure combination is counted, and combinations with a frequency exceeding a threshold are identified as associated combinations.

[0080] Select a preset usage time point, and read the cumulative failure ratio at that preset usage time point from the reliability coefficient of each component identifier under its corresponding supplier identifier in the associated combination; multiply the cumulative failure ratios of all components to obtain the theoretical independent failure combined ratio;

[0081] The total number of complete machines with supplier identifiers recorded using this association combination is counted, as well as the total number of complete machines that have not exceeded the preset usage time point in terms of the cumulative usage time of all components; the ratio of the total number of complete machines that have not exceeded the time limit to the total number of complete machines is calculated to obtain the actual observed failure joint ratio.

[0082] Calculate the difference between the actual observed failure joint ratio and the theoretical independent failure joint ratio, and define the difference as the failure correlation coefficient of the correlation combination at the preset usage time point.

[0083] First, the system automatically filters out all data entries from the recycling database that record the failure of two or more components under the same whole unit identifier. For example, it retrieves all records of projection lamps where both the "LED module" and the "driver circuit" have failed. Next, it extracts the specific component identifier and its corresponding supplier identifier for each failed component from these entries, and aggregates these identifiers from the same whole unit into an observed failure combination. For example, it aggregates the identifiers of "LED module from supplier A" and "driver circuit from supplier E" that failed in a lamp into a combination {A,E}. Subsequently, the system counts the frequency of each observed failure combination like {A,E} in all filtered entries, and sets a frequency threshold based on historical data analysis and engineering experience. Combinations that occur at a frequency significantly higher than random levels and exceed the threshold are identified as related combinations that require further attention. Then, for quantitative analysis, a preset usage time point with engineering evaluation significance is selected, such as a key time point "4 years" after the product warranty period. For the confirmed associated combination, the system will accurately read the cumulative failure ratio at the "4-year" time point from the reliability coefficient curve generated in S2 for each component in the combination (such as the LED module of supplier A). Under the assumption that the failures of these components are completely independent, the cumulative failure ratios of each component are multiplied to obtain a theoretical independent failure joint ratio. After that, the system moves to actual data verification: it counts the total number of complete machines in the database that use the precise supplier identification recorded for this associated combination, that is, the number of projection lamps that use the combination of "LED module of supplier A" and "driver circuit of supplier E"; and further counts the number of complete machines in which all components in the associated combination have failed before the cumulative usage time of all components has not exceeded "4 years"; by calculating the ratio of the latter to the former, an actual observed failure joint ratio is obtained. Finally, the system calculates the difference between the actual observed combined failure ratio and the theoretical combined independent failure ratio, and formally labels this difference, which can be positive or negative, as the failure correlation coefficient of the associated combination at the preset usage time of "4 years".

[0084] Understandably, by using mathematical modeling and comparing actual data, we can proactively identify and quantify the hidden risks of component failures exacerbated by design compatibility, electrical mismatch, or physical stress coupling in real-world usage environments. The core benefit lies in transforming the previously difficult-to-define "component matching problem," which relied solely on post-failure analysis or engineer experience, into an objective, calculable, and comparable quantitative indicator. This failure correlation coefficient clearly shows whether the risk of shared failure among specific supplier combinations is higher or lower than the statistically expected risk of their individual failures, thus providing direct evidence for judging the merits of the combination. This step plays a crucial role in achieving the ultimate goal of scientifically optimizing the supply chain combination. It ensures that subsequent supplier combination selection is no longer a simple aggregation of cost or individual component quality, but rather based on a deep understanding and prevention of systemic risks, thereby improving the overall reliability and durability of the product from the source.

[0085] In another preferred embodiment of the present invention, the specific process of generating several supply combination schemes in step S4 is as follows:

[0086] Select any component identifier as the target component identifier, and obtain any supplier identifier corresponding to the target component identifier as the current benchmark supplier; obtain a set of candidate supplier identifiers for the remaining component identifiers, and combine the current benchmark supplier with a supplier identifier selected from each of the remaining sets to form a combination to be evaluated;

[0087] Obtain the reliability coefficients of each component in the supplier identification combination to be evaluated under the corresponding supplier and the failure correlation coefficients of the supplier identification combination to be evaluated under the same whole machine, and calculate the expected reliability of the whole lamp of the supplier identification combination to be evaluated.

[0088] Enumerate all supplier identification combinations to be evaluated, calculate the expected reliability of the whole lamp for each supplier identification combination to be evaluated, and mark the supplier identification combination with the highest expected reliability of the whole lamp as the best cooperative supplier combination scheme of the current benchmark supplier.

[0089] Iterate through each candidate supplier identifier of the target component identifier, and use them as the current benchmark supplier in turn. Repeat the above process to determine all the generated best cooperative supplier combinations as several supply combination schemes.

[0090] First, select one component from all components of the automotive projection lamp as the target component identifier, for example, select the core light-emitting component "LED module"; then, obtain all candidate supplier identifiers corresponding to the target component identifier, such as supplier A and supplier B, and designate one of them, such as supplier A, as the current benchmark supplier; then, obtain the candidate supplier identifier sets for other components besides the LED module, such as "optical lens" and "driving circuit", for example, suppliers C and D for the lens, and suppliers E and F for the circuit; then, combine the current benchmark supplier A with an identifier selected from the lens supplier set (such as C) and an identifier selected from the circuit supplier set (such as E) to form a supplier identifier combination {A,C,E} to be evaluated. Based on the combination to be evaluated, the system retrieves the reliability coefficient curves generated in S2 for each component (LED of A, lens of C, circuit of E) and searches for the failure correlation coefficients that may have been calculated in S3 (e.g., there may be a correlation between LED of A and circuit of E). Using a mathematical model that integrates the independent failure probabilities of each component and the impact of associated risks, the system calculates the expected reliability of the entire lamp under this specific combination—a quantitative value that comprehensively characterizes its overall durability. Then, the system uses an enumeration method to iterate through all other possible supplier choices for the lens and circuit, constructing other combinations to be evaluated, such as {A,C,F}, {A,D,E}, {A,D,F}, etc., and repeats the above calculation process to obtain the expected reliability of the entire lamp for each combination. After completing all calculations, the system compares the expected reliability of all combinations to be evaluated and officially designates the combination with the highest value, such as {A,D,F}, as the best cooperative supplier combination for the current benchmark supplier A. Finally, the system switches the benchmark supplier to the next alternative supplier for the target component (LED module), namely supplier B, and completely repeats the entire process of specifying the benchmark, constructing combinations, calculating reliability, and enumerating and filtering to obtain the best cooperation solution for supplier B, which may be {B,C,E}. By traversing each alternative supplier for the target component, the system ultimately identifies all the generated best cooperation solutions, such as {A,D,F} for A and {B,C,E} for B, as several supply combination solutions available for final decision-making.

[0091] This approach elevates supplier selection from isolated evaluations of individual components to a global optimization of the overall system's compatibility, thus addressing the inherent challenge in traditional procurement where "simple assembly of excellent components does not necessarily result in an excellent product." Its advantage lies in its data- and model-driven approach, systematically exploring and evaluating every possible supplier combination configuration. It considers not only the quality history of each component but, more importantly, incorporates empirically proven collaborative failure risks between components, transforming the prediction of the entire lamp's lifespan from a rough estimate to a precise extrapolation. This process transforms the complex systems engineering matching problem into a quantifiable, comparable, and automated optimization task. For the ultimate goal of the solution, this step is a crucial juncture, bridging the micro-risk insights revealed by the preliminary basic analysis (reliability coefficients and failure correlation coefficients) and transforming them into a macro-optimal configuration scheme that directly guides production and procurement. This ensures that the final selected target solution is a rational choice that maximizes overall performance based on a deep understanding and mitigation of internal system risks, thereby providing a fundamental guarantee for manufacturing products with longer lifespans and higher reliability.

[0092] In another preferred embodiment of the present invention, the specific process for generating the expected reliability of the entire lamp is as follows:

[0093] A set of candidate usage duration points arranged in chronological order is pre-defined. For each candidate usage duration point, the corresponding cumulative failure ratio is read from the reliability coefficient corresponding to each component identifier in the combination to be evaluated. The component survival ratio is obtained by subtracting the cumulative failure ratio from one.

[0094] Obtain the failure correlation coefficient of the combination to be evaluated at the candidate usage time point; subtract the failure correlation coefficient from one to obtain the coupling adjustment coefficient; multiply the component survival rate by the coupling adjustment coefficient to obtain the corrected survival rate;

[0095] Multiply the corrected survival rates of all components in the assembly to be evaluated at the candidate usage time points to obtain the overall lamp system survival rate; iterate through all candidate usage time points to obtain a sequence of overall lamp system survival rates;

[0096] Calculate the duration difference between each pair of adjacent candidate usage duration points in the sequence, and calculate the average survival rate of the two whole lamp systems corresponding to the pair of adjacent candidate usage duration points; multiply the duration difference by the average value to obtain the area of ​​the sub-interval corresponding to each pair of adjacent points; sum all the sub-interval areas, and label the summation result as the expected reliability coefficient of the whole lamp for the supplier identifier combination to be evaluated.

[0097] First, the system pre-sets a set of candidate usage durations evenly arranged in chronological order. For example, starting from the product's initial use, it sets a series of time points such as "0.5 years, 1 year, 1.5 years, ..., 6 years" at six-month intervals. For each candidate usage duration, such as "4 years," the system precisely reads the cumulative probability (i.e., cumulative failure ratio) of each component identifier (i.e., LED module from supplier A, optical lens from supplier C, and drive circuit from supplier E) at the specific time point "4 years" from the reliability coefficient curves corresponding to each component identifier in the supplier identifier combination to be evaluated (e.g., combination {A,C,E}). Then, by subtracting each cumulative failure ratio from the given number "one," the system obtains the probability that each component is still alive at that time point, i.e., the component survival ratio. The system then retrieves the pre-calculated failure correlation coefficient for the evaluated combination at the same "4-year" candidate point. This coefficient may reflect, for example, the additional risk of synergistic failure between the LED module from supplier A and the driver circuit from supplier E due to current matching issues. Similarly, subtracting this failure correlation coefficient from the number "one" yields a coupling adjustment coefficient, typically less than or equal to one. Subsequently, the survival rate of each component at the "4-year" point is multiplied by its corresponding coupling adjustment coefficient to obtain a corrected survival rate that more closely approximates the actual coupled operating state of the system. This corrected survival rate is typically lower than its individual survival rate because it considers the associated risks. Then, at the "4-year" point, the corrected survival rates of all components (LED modules, optical lenses, driver circuits) in the evaluated combination are multiplied together. The product represents the probability that the entire lighting system will still function normally at "4 years" after considering both the individual reliability of each component and the associated risks between them—that is, the overall lighting system survival rate at that point. The system iterates through all preset candidate usage duration points, repeating the above-described reading, calculation, and multiplication process for each point to obtain a complete sequence consisting of each time point and its corresponding overall lamp system survival rate. Finally, the system performs numerical integration on this sequence to estimate the average lifespan of the entire lamp: specifically, it calculates the duration difference between each pair of adjacent candidate usage duration points in the sequence and calculates the average of the two overall lamp system survival rates corresponding to these adjacent points; it multiplies the duration difference of each sub-interval with its corresponding average survival rate to obtain the contribution area of ​​that sub-interval to the overall lifespan estimation; and it sums the areas of all sub-intervals. This final sum is then designated as the expected reliability of the lamp for the supplier identification combination being evaluated. Essentially, it is a mathematical expectation of the average mean time between failures (MTBF) of the entire lamp after considering the risk of component coupling failures.

[0098] An integral method that iterates through all time points and calculates the area under the curve yields a robust comprehensive indicator—the expected lifespan of the entire system. This indicator more comprehensively represents the overall product lifespan than simply observing the survival probability at a single time point. This process is the actuarial cornerstone for the scientific optimization of supply chain combinations. It provides an objective, unified, and comparable quantitative benchmark, allowing us to accurately determine which supplier combinations, such as {A,C,E} and {A,D,F}, deliver more durable and reliable products. This ensures that subsequent comprehensive decisions based on cost and reliability are grounded in sound systems engineering calculations, rather than vague experience or guesswork, thus fundamentally guaranteeing that the selected optimal supply chain combination can deliver the expected high-quality products.

[0099] In another preferred embodiment of the present invention, a set of candidate usage duration points arranged in chronological order is further included, wherein the value of the largest candidate usage duration point is less than the end point of the shortest time range covered by the reliability coefficients corresponding to all component identifiers in the combination to be evaluated; when any candidate usage duration point does not have a corresponding failure correlation coefficient for the combination to be evaluated, the failure correlation coefficient used to calculate the coupling adjustment coefficient is set to zero.

[0100] Setting the maximum candidate usage duration to not exceed the shortest coverage endpoint of all component reliability data is to strictly limit the entire evaluation to a range supported by actual observational data. This avoids unfounded extrapolations or guesses for periods exceeding the shortest effective observational lifetime of components, ensuring that every cumulative failure ratio value read from the reliability coefficient curve originates from real recovery statistics, thus making all subsequent calculations based on credible empirical evidence. When certain candidate usage durations lack corresponding failure correlation coefficients, they are set to zero. This essentially provides a clear and reasonable default rule. The underlying logic is to temporarily adopt the conservative neutral assumption of "no additional correlation effect" when there is a lack of specific statistical evidence of a correlation risk. This ensures that the calculation process will not be interrupted due to sparse data for some combinations or failure to reach the statistical significance threshold, guaranteeing the continuity and robustness of the system when handling various real-world data situations.

[0101] In another preferred embodiment of the present invention, the specific process of generating the target solution in step S5 is as follows:

[0102] The cost parameters and the overall reliability coefficient are standardized numerically to obtain the corresponding standard cost score and standard reliability score. Based on the preset weight allocation, the standard reliability score and the standard cost score are weighted and summed to obtain the priority score of each supplier combination scheme. The priority scores of all supplier combination schemes are compared, and the scheme with the highest priority score is designated as the target scheme.

[0103] In another preferred embodiment of the present invention, step S6 further includes reading the product traceability file based on the whole machine identification when the vehicle projection lamp is scrapped and recycled, adding a new recycling record to the recycling database based on the supplier combination information in the product traceability, and triggering the update of the reliability parameters and failure correlation relationship.

[0104] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A product quality lifecycle management method for automotive projection lamps, characterized in that, Includes the following steps: S1, read the whole machine identification and disassemble the components of the scrapped vehicle projection lamp. During the disassembly process, determine the failure of the components and record the cumulative usage time of the failed components. Integrate the whole machine identification, component identification, corresponding supplier identification and usage time of the failed components into recycling data entries and store them in the recycling database. S2, classify the recycled data entries in the recycling database according to component identifier and supplier identifier, and generate the reliability coefficient curve of each component under each supplier based on the usage time data of the failed components in each group after classification. S3, extract recycling data entries of multiple component failures under the same whole machine identifier from the recycling database, and calculate the failure correlation coefficient of components from different suppliers under the same whole machine based on the reliability coefficient curve; The specific calculation process for the failure correlation coefficient is as follows: Entries with multiple component failures under the same complete machine are screened from the recycling database. The component identifier and supplier identifier of the failed component are extracted and aggregated into observed failure combinations. The frequency of occurrence of each observed failure combination is counted, and combinations with a frequency exceeding a threshold are identified as associated combinations. Select a preset usage time point, and read the cumulative failure ratio at that preset usage time point from the reliability coefficient curve of each component identifier in the associated combination under its corresponding supplier identifier; multiply the cumulative failure ratios of all components to obtain the theoretical independent failure combined ratio; The total number of complete machines with supplier identifiers recorded using this association combination is counted, as well as the total number of complete machines that have not exceeded the preset usage time point in terms of the cumulative usage time of all components. Calculate the ratio of the total number of units that did not exceed the timeout to the total number of units to obtain the actual observed joint failure ratio; Calculate the difference between the actual observed failure joint ratio and the theoretical independent failure joint ratio, and define the difference as the failure correlation coefficient of the correlation combination at the preset usage time point; S4. Based on the reliability coefficient curve and the failure correlation coefficient, all suppliers are combined to generate several supply combination schemes and the expected reliability of the whole lamp corresponding to each supply combination scheme is calculated; wherein, the expected reliability of the whole lamp is essentially a mathematical expectation of the mean time between failures of the whole lamp after considering the risk of component coupling failure. S5. Obtain the cost parameters of each supplier combination scheme, and calculate the priority score of each supplier combination scheme based on the expected reliability of the whole lamp, and mark the supplier combination scheme with the highest priority score as the target scheme. S6. Based on the target solution, generate production guidance information, and after the automotive projection lamp is assembled, generate a product traceability file that is associated with the whole machine identifier and the target solution.

2. The product quality lifecycle management method for an automotive projection lamp according to claim 1, characterized in that, In step S2, the specific process of generating the reliability coefficient curves of each component under each supplier is as follows: The recycling database is used to filter out recycling data entries whose component status is determined to be invalid, and the filtered entries are classified and grouped according to component identifier and supplier identifier; For each category group, extract the cumulative usage time data and the total number of failed parts from all records of failed parts within the group. Sort all extracted cumulative usage times according to their numerical values ​​and divide the sorted usage time sequence into several consecutive time intervals. The number of failed components within each time interval is counted, and the proportion of the number of failed components to the total number of failed components is calculated. The end time of each time interval is correlated with its corresponding cumulative failure ratio, and a curve of the cumulative failure ratio increasing with the usage time is plotted. The curve is then labeled as the reliability coefficient curve of the component under this supplier.

3. The product quality lifecycle management method for an automotive projection lamp according to claim 1, characterized in that, In step S4, the specific process for generating several supply combination schemes is as follows: Select any component identifier as the target component identifier, and obtain any supplier identifier corresponding to the target component identifier as the current benchmark supplier; Obtain a set of alternative supplier identifiers for the remaining component identifiers, and combine the current benchmark supplier with a supplier identifier selected from each of the remaining sets to form a combination to be evaluated; Obtain the reliability coefficient curves of each component in the supplier identification combination to be evaluated under the corresponding supplier and the failure correlation coefficient of the supplier identification combination to be evaluated under the same whole machine, and calculate the expected reliability of the whole lamp of the supplier identification combination to be evaluated. Enumerate all supplier identification combinations to be evaluated, calculate the expected reliability of the whole lamp for each supplier identification combination to be evaluated, and mark the supplier identification combination with the highest expected reliability of the whole lamp as the best cooperative supplier combination scheme of the current benchmark supplier. Iterate through each candidate supplier identifier of the target component identifier, and use them as the current benchmark supplier in turn. Repeat the above process to determine all the generated best cooperative supplier combinations as several supply combination schemes.

4. The product quality lifecycle management method for an automotive projection lamp according to claim 3, characterized in that, The specific process for generating the expected reliability of the entire lamp is as follows: A set of candidate usage duration points arranged in chronological order is pre-defined. For each candidate usage duration point, the corresponding cumulative failure ratio is read from the reliability coefficient curve corresponding to each component identifier in the combination to be evaluated. The component survival ratio is obtained by subtracting the cumulative failure ratio from one. Obtain the failure correlation coefficient of the combination to be evaluated at the candidate usage time point; subtract the failure correlation coefficient from one to obtain the coupling adjustment coefficient; multiply the component survival rate by the coupling adjustment coefficient to obtain the corrected survival rate; Multiply the corrected survival rates of all components in the assembly to be evaluated at the candidate usage time points to obtain the overall lamp system survival rate; iterate through all candidate usage time points to obtain a sequence of overall lamp system survival rates; Calculate the duration difference between each pair of adjacent candidate usage duration points in the sequence, and calculate the average survival rate of the two whole lamp systems corresponding to the pair of adjacent candidate usage duration points; multiply the duration difference by the average value to obtain the area of ​​the sub-interval corresponding to each pair of adjacent points; sum all the sub-interval areas, and label the summation result as the expected reliability of the whole lamp of the supplier identifier combination to be evaluated.

5. The product quality lifecycle management method for an automotive projection lamp according to claim 4, characterized in that, It also includes a pre-set set of candidate usage duration points arranged in chronological order, the largest candidate usage duration point value being less than the shortest time range end point covered by the reliability coefficient curves corresponding to all component identifiers in the combination to be evaluated; when any candidate usage duration point does not have a corresponding failure correlation coefficient for the combination to be evaluated, the failure correlation coefficient used to calculate the coupling adjustment coefficient is set to zero.

6. The product quality lifecycle management method for an automotive projection lamp according to claim 1, characterized in that, In step S5, the specific process for generating the target solution is as follows: The cost parameters and the expected reliability of the entire lamp are standardized numerically to obtain the corresponding standard cost score and standard reliability score. According to the preset weight allocation, the standard reliability score and the standard cost score are weighted and summed to obtain the priority score of each supplier combination scheme. The priority scores of all supplier combination schemes are compared, and the scheme with the highest priority score is marked as the target scheme.

7. The product quality lifecycle management method for an automotive projection lamp according to claim 1, characterized in that, S6 also includes reading the product traceability file based on the whole machine identification when the vehicle projection lamp is scrapped and recycled, adding a new recycling record to the recycling database based on the supplier combination information in the product traceability, and triggering the update of the reliability coefficient curve and failure correlation coefficient.

Images