Evaluation method, device, equipment, medium and program product for car-grade chip detection report

Abstract

The application relates to an evaluation method, device, equipment, medium and program product of a car-grade chip detection report. The method comprises the following steps: obtaining a car-grade chip detection report to be evaluated, performing standardization processing on the chip detection report to obtain structured data; performing quality detection, project scoring and risk matching on the structured data by using a judgment model constructed based on a standard evaluation rule library to obtain a report judgment result; wherein the quality detection comprises compliance detection, integrity detection and accuracy detection, and the risk matching identifies the failure risk of an untested item based on a mapping relationship between a test item and a potential failure risk. The application realizes automatic analysis, verification and risk evaluation of test technical data in a multi-source heterogeneous car-grade chip detection report, improves the processing efficiency and evaluation accuracy of the technical data, and solves the problems that report evaluation in the industry relies on manual experience and standards are not unified.

Description

Technical Field

[0001] This application relates to the field of automotive electronics technology, and in particular to an evaluation method, apparatus, equipment, medium, and program product for automotive-grade chip test reports. Background Technology

[0002] With the rapid development of new energy technologies, the adoption rate of domestically produced automotive chips in vehicles is gradually increasing. However, there is still a significant gap between the chip quality and reliability and those of mature international supply chain companies. In practical application scenarios, OEMs and Tier 1 suppliers mainly rely on AEC-Q test reports to conduct chip quality assessments, material selection, and supply chain access.

[0003] Standards such as AEC-Q100 verify the quality and reliability of devices through a series of stress tests, prevent various conditions or potential failure states that may occur in products, guide component suppliers to use chips that comply with the specifications during the development process, and conduct rigorous quality and reliability verification for each chip case.

[0004] However, the test reports issued by various chip companies come in diverse formats (including PDF, Word, scanned copies, etc.), with varying content structures and inconsistent quality. The technical data recorded in these reports, such as test items, test parameters, test conditions, and test results, are multi-sourced, heterogeneous, and unstructured, making it difficult for computers to uniformly parse and automate the processing of this type of technical data. Currently, the evaluation of test reports mainly relies on manual page-by-page verification, which is not only inefficient and lacks standardized evaluation criteria, but also makes it difficult to perform full-dimensional automated verification of the test data in the reports, and cannot automatically identify potential failure risks corresponding to untested items.

[0005] Therefore, how to automate the analysis, verification, and risk assessment of test technical data in multi-source heterogeneous automotive-grade chip test reports has become an urgent technical problem to be solved. Summary of the Invention

[0006] Therefore, it is necessary to provide an evaluation method, device, equipment, medium, and program product for automotive-grade chip test reports to address the aforementioned technical issues, so as to achieve automated analysis, intelligent verification, quantitative scoring, and failure risk identification of test technical data in automotive-grade chip test reports.

[0007] Firstly, this application provides a method for evaluating automotive-grade chip testing reports, including:

[0008] Obtain the automotive-grade chip test report to be evaluated, standardize the chip test report, and obtain structured data;

[0009] The pre-built evaluation model is used to perform quality inspection, project scoring, and risk matching on structured data to obtain the report evaluation results. The evaluation model is built on a standard evaluation rule base, which includes quality inspection rules, project scoring rules, and risk matching rules.

[0010] Quality inspection includes compliance inspection, integrity inspection, and accuracy inspection. Based on the results of quality inspection, qualitative evaluation conclusions and a list of key issues are determined.

[0011] The quantitative scoring conclusions and detailed scores for each item are determined based on the project scoring results.

[0012] Risk matching includes: filtering and processing structured data to obtain data on untested items, performing risk matching on the untested item data, and obtaining risk item alert information;

[0013] The report evaluation results include multiple components such as qualitative evaluation conclusions, quantitative scoring conclusions, detailed breakdown of scores for each item, a list of key issues, and risk item reminders.

[0014] In one embodiment, the automotive-grade chip test report is an AEC-Q series standard test report or an AQG324 standard test report; wherein, the AEC-Q series standards include AEC-Q100, AEC-Q101, AEC-Q102, AEC-Q103, AEC-Q104 and AEC-Q200.

[0015] In one embodiment, the structured data undergoes quality inspection, and qualitative evaluation conclusions and a list of key issues are determined based on the results of the quality inspection, including:

[0016] The structured data is subjected to compliance checks to obtain the first check result; the first check result is used to characterize whether the structured data is compliant.

[0017] The structured data is subjected to integrity checks to obtain a second check result; the second check result is used to characterize whether the structured data is complete.

[0018] The structured data is subjected to an accuracy test to obtain a third test result; the third test result is used to characterize whether the structured data is accurate.

[0019] The qualitative evaluation conclusions and a list of key issues are determined based on the first, second, and third test results.

[0020] In one embodiment, the structured data is scored, and a quantitative scoring conclusion and detailed itemized scores are determined based on the scoring results, including:

[0021] Based on the mapping relationship between items and scores in the standard evaluation rule base, structured data is matched with items, and the scores corresponding to successfully matched items are determined as the sub-item scores of the structured data, generating detailed sub-item score details;

[0022] The quantitative scoring conclusion is obtained by weighted summation of all sub-item scores in the detailed score breakdown.

[0023] In one embodiment, the mapping relationship between items and scores includes a quantitative scoring rule consisting of a base score and bonus points;

[0024] The basic score is 100 points, including 30 points for report quality and 70 points for test item coverage. Report quality includes three sub-items: equipment effectiveness, completeness of test result information, and completeness of test condition information, each worth 10 points. Test item coverage is scored according to the AEC-Q100 standard for the coverage of four core test groups: accelerated environmental stress test group, accelerated life simulation test group, package integrity test group, and electrical performance verification test group. The original score of each core test group is 105 points, which is then normalized to 70 points at a ratio of 70 / 105.

[0025] Bonus points are awarded for process management (20 points), including 6 points for fault classification and characterization, 10 points for chip manufacturing reliability, and 4 points for defect screening.

[0026] In one embodiment, scoring items from structured data further includes:

[0027] Obtain structural, packaging, process, and material characteristics of automotive-grade chips. Based on these characteristics, remove inapplicable items from all test items to obtain the set of test items.

[0028] The project scoring is based on the coverage of the test item set by structured data; among which, the quantitative scoring conclusion includes the detection coverage rate, which is the ratio between the number of test items covered by structured data and the total number of test items in the test item set.

[0029] In one embodiment, determining the score corresponding to a successfully matched item as a sub-item score of structured data includes: for each item, determining a scoring coefficient based on the test execution status; wherein, the test execution status includes test conditions, test order, and batch and quantity of test samples; if the test execution status fully meets the standard requirements, the scoring coefficient is the first coefficient; if the test execution status partially meets the standard requirements, the scoring coefficient is the second coefficient, which is less than the first coefficient; if the item is not executed, the scoring coefficient is zero; and the product of the score corresponding to the item and the scoring coefficient is determined as the sub-item score.

[0030] In one embodiment, the structured data is filtered to obtain untested item data, including:

[0031] Extract the test item number or identifier from the structured data, match the test item number or identifier with the mandatory test items in the risk matching rules, and filter out the mandatory test items that have not been executed as untested item data;

[0032] Risk matching is performed on untested project data to obtain risk item alerts, including: matching the untested project data with the mapping relationship between test items and potential failure risks in the risk matching rules to obtain risk item alerts.

[0033] In one embodiment, the mapping relationship between test items and potential failure risks includes at least:

[0034] When temperature and humidity bias or high accelerated stress testing is not performed, the corresponding potential failure risks are metallization corrosion, package leakage and bond point damage.

[0035] When high-pressure accelerated aging or unbiased high-accelerated stress testing is not performed, the corresponding potential failure risks are ion contamination corrosion and the inability to expose packaging hermeticity defects.

[0036] When temperature cycling testing is not performed, the potential failure risks are bond wire breakage, chip cracking, and delamination between the molding compound and the chip.

[0037] When power temperature cycling tests are not performed, the potential failure risks are bond point fatigue, electromigration, and via fracture.

[0038] When high-temperature storage life testing is not performed, the potential failure risks are intermetallic compound embrittlement and silicone grease drying.

[0039] In one embodiment, the method for constructing a standard evaluation rule base includes:

[0040] Obtain standard chip test reports, extract core elements from these reports, and generate quality testing rules. The core elements include test items, test judgment conditions, and quality reliability requirements.

[0041] Obtain the preset report quality scoring rules and test item scoring rules, and generate project scoring rules based on the report quality scoring rules and test item scoring rules;

[0042] Establish a correspondence between test items and potential failure risks when they are not executed, and generate risk matching rules;

[0043] A standard evaluation rule library is constructed by combining quality inspection rules, project scoring rules, and risk matching rules.

[0044] In one embodiment, the chip testing report is standardized to obtain structured data, including:

[0045] The chip test report is converted to its original format to obtain initial text data;

[0046] Feature extraction is performed on the initial text data using regular expressions and / or keyword matching to obtain structured data.

[0047] In one embodiment, the method further includes:

[0048] The report evaluation results are converted into visual infographics; the visual infographics include test coverage statistics charts, quality evaluation radar charts, failure risk heat maps, and score bar charts for each test group.

[0049] Secondly, this application also provides an evaluation device for automotive-grade chip testing reports, comprising:

[0050] The acquisition module is used to acquire the automotive-grade chip test report to be evaluated, and to standardize the chip test report to obtain structured data;

[0051] The evaluation module utilizes a pre-built evaluation model to perform quality checks, project scoring, and risk matching on structured data, resulting in a report evaluation. The evaluation model is built upon a standard evaluation rule base, which includes quality check rules, project scoring rules, and risk matching rules. Quality checks include compliance checks, completeness checks, and accuracy checks, determining qualitative evaluation conclusions and a list of key issues based on the results. Quantitative scoring conclusions and detailed sub-item scores are determined based on the project scoring results. Risk matching involves filtering structured data to obtain untested project data, performing risk matching on this untested data, and generating risk item alerts. The report evaluation results include multiple options from qualitative evaluation conclusions, quantitative scoring conclusions, detailed sub-item scores, a list of key issues, and risk item alerts.

[0052] In one embodiment, the apparatus further includes a building module for:

[0053] Obtain standard chip test reports, extract core elements from these reports, and generate quality testing rules. The core elements include test items, test judgment conditions, and quality reliability requirements.

[0054] Obtain the preset report quality scoring rules and test item scoring rules, and generate project scoring rules based on the report quality scoring rules and test item scoring rules;

[0055] Establish a correspondence between test items and potential failure risks when they are not executed, and generate risk matching rules;

[0056] A standard evaluation rule library is constructed by combining quality inspection rules, project scoring rules, and risk matching rules.

[0057] In one embodiment, the acquisition module is specifically used to: convert the format of the chip test report to obtain initial text data; and extract features from the initial text data using regular expressions and / or keyword matching to obtain structured data.

[0058] In one embodiment, the apparatus further includes a conversion module for: converting the report evaluation results into a visual infographic; the visual infographic includes a test coverage statistics chart, a quality evaluation radar chart, a failure risk heat map, and a bar chart of scores for each test group.

[0059] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method of any of the above embodiments.

[0060] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the methods of any of the above embodiments.

[0061] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the methods of any of the above embodiments.

[0062] The aforementioned evaluation methods, devices, equipment, media, and program products for automotive-grade chip test reports convert multi-source, heterogeneous, and unstructured chip test reports into computer-processable structured technical data through standardized processing, solving the technical problem that test technical data is difficult to uniformly parse and automatically process. Utilizing an evaluation model built on a standard evaluation rule base, the structured data undergoes three-dimensional quality inspection, quantitative scoring, and risk matching, including compliance, completeness, and accuracy. The correspondence between the test items and potential failure risks used in risk matching is established based on the chip failure physical mechanism (such as metallization corrosion, electromigration, and bond point fatigue), enabling the algorithm processing process and chip reliability technical characteristics to functionally support each other. The above solution replaces the traditional manual page-by-page verification method, improving the processing efficiency and evaluation accuracy of test technical data in chip test reports. It achieves automated identification of failure risks for untested items and solves the problems of reliance on human experience and inconsistent standards in report evaluation within the industry. By eliminating inapplicable items based on the structure, packaging, process, and material characteristics of automotive-grade chips and calculating test coverage, and by determining scoring coefficients based on test execution for tiered scoring, the accuracy and precision of quantitative scoring are further improved. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 This is a block diagram of a computer device that uses an evaluation method for automotive-grade chip testing reports in one embodiment.

[0065] Figure 2 This is a flowchart illustrating the evaluation method for automotive-grade chip testing reports in one embodiment;

[0066] Figure 3 This is a flowchart illustrating the process of using an evaluation model to perform quality inspection, project scoring, and risk matching on structured data in one embodiment.

[0067] Figure 4 This is a flowchart illustrating the quality inspection process for structured data in one embodiment.

[0068] Figure 5 This is a flowchart illustrating the process of scoring items from structured data in one embodiment.

[0069] Figure 6This is a flowchart illustrating the risk matching process for structured data in one embodiment;

[0070] Figure 7 This is a flowchart illustrating the process of constructing a standard evaluation rule base in one embodiment;

[0071] Figure 8 This is a schematic diagram illustrating the standardization process for chip testing reports in one embodiment.

[0072] Figure 9 This is a flowchart illustrating the evaluation method for automotive-grade chip testing reports in yet another embodiment;

[0073] Figure 10 This is a structural block diagram of an evaluation device for automotive-grade chip testing reports in one embodiment. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0075] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0076] With the rapid development of new energy technologies, the adoption rate of domestically produced automotive chips in vehicles is gradually increasing. Automakers and Tier 1 suppliers primarily rely on automotive-grade testing reports such as AEC-Q for chip quality assessment, material selection, and supply chain access. However, the testing reports issued by various chip companies have diverse formats and varying content structures. The technical data in these reports, including test items, parameters, conditions, and results, are multi-sourced, heterogeneous, and unstructured. Computers struggle to uniformly parse and automate this data, while manual evaluation suffers from low efficiency, inconsistent standards, and the inability to automatically identify potential failure risks from untested items.

[0077] In view of the above-mentioned technical problems, this application provides an evaluation method for automotive-grade chip test reports. The following embodiments will specifically illustrate the evaluation method for automotive-grade chip test reports.

[0078] The evaluation method for automotive-grade chip testing reports provided in this application embodiment can be applied to, for example... Figure 1 The computer device shown may be a server, and its internal structure diagram may be as follows. Figure 1 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores chip testing reports, evaluation rule bases, structured data extracted from reports, and evaluation results. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an evaluation method for automotive-grade chip testing reports. In one alternative implementation, the software program in the computer program is developed using a B / S architecture (browser / server architecture), adapted to mainstream server operating systems such as Windows and Linux, and supported by mainstream browsers; the development language is, for example, Python + Java, the front-end visualization is, for example, ECharts or Highcharts, the OCR recognition in the standardization process is, for example, PaddleOCR, the evaluation model is, for example, built based on the Qwen deep learning model framework, and the database is, for example, MySQL.

[0079] Those skilled in the art will understand that Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0080] In one exemplary embodiment, such as Figure 2 As shown, an evaluation method for automotive-grade chip testing reports is provided. This embodiment illustrates the method by applying it to computer equipment. In this embodiment, the method includes:

[0081] S201: Obtain the automotive-grade chip test report to be evaluated, standardize the chip test report, and obtain structured data.

[0082] The automotive-grade chip test report can be an AEC-Q series standard test report or an AQG324 standard test report, such as the AEC-Q100 automotive-grade chip test report, the AEC-Q101 automotive-grade discrete device test report, the AEC-Q102 automotive-grade optoelectronic component test report, the AEC-Q103 automotive-grade sensor test report, the AEC-Q104 automotive-grade multi-chip component test report, the AEC-Q200 automotive-grade passive component test report, and the AQG324 automotive-grade power module test report. These standards are merely examples, and this application does not limit the specific standards upon which the automotive-grade chip test report is based. The chip test report may be in formats including, but not limited to, PDF, Word, and scanned copies.

[0083] In the embodiments of this application, the computer device can obtain the chip test report from the chip company's network interface in advance and store the chip test report in the database. When an evaluation request is obtained, the chip test report related to the evaluation request can be retrieved from the database. Alternatively, data can be crawled from web page links using web crawling technology to obtain the chip test report to be evaluated. Alternatively, the chip test report to be evaluated can be obtained by having the user upload the chip test report to the front end of the computer device.

[0084] In some embodiments, the chip testing report to be evaluated is standardized, that is, the report is converted to an editable text format; and features are extracted from the editable text format to obtain structured data. Feature extraction methods include, but are not limited to, natural language processing, machine learning, regular expressions, data modeling, document template methods, decision trees, and deep learning methods.

[0085] S202 utilizes a pre-built evaluation model to perform quality checks, project scoring, and risk matching on structured data, resulting in a report evaluation.

[0086] The evaluation model is built upon a standard evaluation rule base and can be either a rule-engine-based model or an AI-based model trained using deep learning. The standard evaluation rule base includes quality inspection rules, project scoring rules, and risk matching rules. The evaluation report results include various types of information, such as qualitative evaluation conclusions, quantitative scoring conclusions, detailed breakdowns of scores for each item, a list of key issues, and risk item alerts. The evaluation results corresponding to quality inspection rules can include qualitative evaluation conclusions and a list of key issues; the evaluation results corresponding to project scoring rules can include quantitative scoring conclusions and detailed breakdowns of scores for each item; and the evaluation results corresponding to risk matching rules can be risk item alerts.

[0087] In the embodiments of this application, after obtaining structured data, the structured data is input into the evaluation model for quality inspection, project scoring, and risk matching. Specifically, the quality inspection rules, project scoring rules, and risk matching rules from the standard evaluation rule library can be called simultaneously to match the structured data and obtain the report evaluation results. Alternatively, the quality inspection rules from the standard evaluation rule library can be called first to perform quality inspection on the structured data, obtaining qualitative evaluation conclusions and a list of key issues; then, the project scoring rules from the standard evaluation rule library can be called to score the structured data, obtaining quantitative scoring conclusions and detailed scores for each item; finally, the risk matching rules from the standard evaluation rule library can be called to perform risk matching on the structured data, obtaining risk item alert information. It should be noted that the order in which the various evaluation rules are called is not limited and can be adapted to the actual scenario.

[0088] The aforementioned evaluation method for automotive-grade chip test reports transforms multi-source, heterogeneous, and unstructured chip test reports into structured technical data through standardized processing. It then utilizes an evaluation model built on a standard evaluation rule base to perform three-dimensional quality inspection, quantitative scoring, and risk matching on the structured data. This replaces the traditional method of manual page-by-page verification, improving the processing efficiency and evaluation accuracy of test technical data. It also enables the automated identification of failure risks for untested items and solves the problem of industry report evaluation relying on human experience and lacking standardized criteria.

[0089] In one exemplary embodiment, such as Figure 3 As shown, a pre-built evaluation model is used to perform quality checks, project scoring, and risk matching on structured data to obtain report evaluation results, including:

[0090] S301. Input structured data into the evaluation model, perform quality checks on the structured data, and obtain qualitative evaluation conclusions and a list of key issues.

[0091] Quality testing includes compliance testing, completeness testing, and accuracy testing. Qualitative evaluation conclusions can be based on whether the test items in the chip testing report are compliant, complete, and accurate. The list of key issues can include lists of non-compliant, incomplete, and inaccurate items.

[0092] In the embodiments of this application, after obtaining structured data, the computer device inputs the structured data into an evaluation model to perform quality checks on the structured data. In some embodiments, the compliance condition in the quality check rules can be called first, and compliance-related data can be extracted from the structured data and substituted into the compliance condition for compliance checks; then, the integrity condition in the quality check rules can be called, and integrity-related data can be extracted from the structured data and substituted into the integrity condition for integrity checks; finally, the accuracy condition in the quality check rules can be called, and accuracy-related data can be extracted from the structured data and substituted into the accuracy condition for accuracy checks. It should be noted that the execution order of compliance checks, integrity checks, and accuracy checks is not limited, and the above description is only for illustrative purposes and can be adapted to the actual scenario.

[0093] S302, score the structured data to obtain quantitative scoring conclusions and detailed scores for each item.

[0094] The quantitative scoring conclusion describes the overall score of the chip testing report, while the detailed breakdown of individual scores can be a list obtained by combining the scores of each item in the chip testing report.

[0095] In the embodiments of this application, after obtaining structured data, the computer device can input the structured data into an evaluation model to score the items, or it can perform quality checks on the structured data before scoring the items. When scoring the items, scoring rules can be called from a standard evaluation rule library, and for each scoring rule, corresponding feature data is extracted from the structured data for scoring, resulting in a sub-item score. After all scoring rules have been scored, detailed sub-item scores are obtained. Finally, the detailed sub-item scores are integrated to obtain a quantitative scoring conclusion.

[0096] S303 performs risk matching on structured data to obtain risk item alert information.

[0097] Among them, the risk warning information can be the potential failure risk that exists when a certain test item is not performed in the chip testing report.

[0098] In the embodiments of this application, after obtaining structured data, the computer device can input the structured data into the evaluation model to perform risk matching on the structured data; alternatively, risk matching can be performed after the structured data has undergone quality inspection.

[0099] In some embodiments, when performing risk matching, test item numbers or identifiers can be extracted from structured data, and mandatory test items in the risk matching rules can be determined. The test item numbers or identifiers can be matched with mandatory test items to filter out untested items in the chip test report. The untested items can then be used as search units to match the project risk mapping relationship in the risk matching rules to obtain risk item reminder information.

[0100] In some embodiments, when performing risk matching, after performing integrity checks on structured data, the untested items in the chip test report can be determined from the integrity check results; and the untested items can be used as search units to match the item risk mapping relationship of the risk matching rules to obtain risk item reminder information.

[0101] By performing quality inspection, project scoring, and risk matching on structured data, it can cover the entire chain of chip testing and process management: three-dimensional quality inspection focuses on the evaluation quality dimensions of the test results in the report, quantitative scoring strengthens the objectivity and comparability of the evaluation results, and risk matching automatically identifies potential failure risks of untested items based on the physical mechanism of chip failure, providing enterprises with comprehensive report value and risk identification basis, and adapting to the high reliability and long cycle quality requirements of automotive chips.

[0102] In one exemplary embodiment, such as Figure 4 As shown, quality checks were performed on the structured data to obtain qualitative evaluation conclusions and a list of key issues, including:

[0103] S401 performs a compliance check on the structured data and obtains the first check result.

[0104] The first detection result is used to characterize whether the structured data is compliant, including whether the test project is compliant or non-compliant. The first detection result is the summary result of compliance detection for all test projects.

[0105] In the embodiments of this application, when inputting structured data into the evaluation model for compliance testing, test items and related operational parameters can be extracted from the structured data. Compliance condition expressions in the quality inspection rules are then called, and the extracted data is substituted into these expressions for compliance testing. The compliance condition expressions can be regular expressions. For each test item and related operational parameter, if all extracted parameters satisfy the compliance condition expressions, the test item is determined to be compliant; if at least one extracted parameter does not meet the compliance condition expressions, the test item is determined to be non-compliant. After all test items in the structured data have undergone compliance testing, the compliance test results for all test items are summarized, and the issues of all non-compliant test items are summarized to obtain the first detection result. In some embodiments, the compliance testing further includes: verifying whether the version of the standard used for testing is the currently valid version; verifying whether the execution order of each test item meets the standard requirements; verifying whether the sampling plan and sample allocation of the test samples meet the standard requirements; and verifying whether the metrological validity period of the instruments and equipment used in the test covers the test time and whether the accuracy of the instruments and equipment meets the standard requirements.

[0106] S402, perform integrity checks on the structured data to obtain the second check result.

[0107] The second test result is used to characterize whether the structured data is complete, including whether the test items are complete or incomplete.

[0108] In the embodiments of this application, when structured data is input into the evaluation model for integrity detection, test conditions and test results of test items can be extracted from the structured data, and the integrity condition expression in the quality inspection rules can be called. The data is then substituted into the integrity condition expression for integrity detection. The integrity condition expression includes mandatory test items for automotive-grade chips and can be a regular expression. For the test conditions and test results of each test item, the extracted test conditions and test results are matched with the mandatory test items of the integrity condition expression to obtain a second detection result. When the extracted test item completely matches the mandatory test items of the integrity condition expression, the test item is determined to be complete; when the extracted test item does not completely match the mandatory test items of the integrity condition expression, that is, at least one mandatory test item is missing from the chip inspection report, the test item is determined to be incomplete. After all test items in the structured data have undergone integrity detection, the integrity detection results of all test items are summarized, and the untested test items are also summarized to obtain a second detection result.

[0109] S403 performs an accuracy test on the structured data to obtain the third test result.

[0110] The third test result is used to characterize whether the structured data is accurate, including whether the test items are accurate or inaccurate. The third test result is the summary result of the accuracy test of all test items.

[0111] In the embodiments of this application, when inputting structured data into the evaluation model for accuracy testing, test results of test items can be extracted from the structured data. The accuracy condition expression in the quality inspection rules is then invoked, and the extracted test results are substituted into the accuracy condition expression for accuracy testing. The accuracy condition expression can be the value range of the test results for each test item. For each test item's test result, if the extracted test result falls within the value range of the test result in the accuracy condition expression, the test item is determined to be accurate; if at least one of the extracted test results is not within the value range of the test result, the test item is determined to be inaccurate. After all test items in the structured data have undergone accuracy testing, the accuracy test results of all test items are summarized, and the problems of all inaccurate test items are summarized to obtain a third detection result. In some embodiments, accuracy testing also includes verifying the consistency between the experimental data and the reported conclusions: if the experimental data does not support the reported conclusions, or if there are logical contradictions or abnormal fluctuations between the experimental data, the corresponding test item is determined to be inaccurate, and the data anomalies are summarized in the third detection result.

[0112] S404. Determine the qualitative evaluation conclusion and key issue list based on the first test result, the second test result, and the third test result.

[0113] In the embodiments of this application, for the calculated first, second, and third detection results, the fields indicating whether the test item is compliant (e.g., test item A, compliant), whether the test item is complete (e.g., test item A, complete), and whether the test item is accurate (e.g., test item B, inaccurate) are summarized to generate a qualitative evaluation conclusion. Issues related to non-compliant test items, untested test items, and inaccurate test items are summarized to generate a list of key issues. In some embodiments, the issues in the critical issues list are output in a tiered manner according to their impact on the report's validity, including important and general items: Important items are defects that directly affect the report's validity, including tests not performed in the standard order and affecting test results, incorrect test conclusions, sampling plans or sample allocation not meeting standard requirements, and the use of non-compliant instruments and equipment in the tests (such as equipment with insufficient accuracy or used beyond its metrological range), suggesting supplementary testing and verification; General items are defects that affect the report's standardization, including incomplete report preparation, review, and approval signatures, incomplete test record information, lack of original data with only test results, absence of the testing organization's official seal, missing or incorrect testing basis, incomplete sample information, incomplete testing plan content, discrepancies between instrument and equipment information or metrological validity period and related records, incomplete or missing report numbers and page numbers, suggesting optimization and improvement, etc. The tiered output of important and general items provides priority guidance for report rectification.

[0114] The above method uses a three-dimensional evaluation framework of compliance, completeness, and accuracy to qualitatively evaluate structured data, which can focus on the evaluation quality dimension of the report test results and effectively verify the test technical data of the chip test report.

[0115] In one exemplary embodiment, such as Figure 5 As shown, the structured data is scored to obtain quantitative scoring conclusions and detailed scores for each item, including:

[0116] The S500 acquires information on the structure, packaging, process, and material properties of automotive-grade chips, eliminates inapplicable items from all test items, and determines the set of items to be tested.

[0117] In the embodiments of this application, the computer device can also acquire structural characteristic information, packaging characteristic information, process characteristic information, and material characteristic information of automotive-grade chips. Based on the above characteristic information, inapplicable items are eliminated from all test items specified in the standard to obtain a set of testable items. For example, the integrity test of air-sealed devices is only applicable to hermetically sealed devices, and this test item is not included in the test item set for plastic-sealed devices; some electrical characteristic test sub-items are only applicable to chips with specific functions or structures, and chips without the corresponding functions or structures can be eliminated as inapplicable items. When scoring the structured data, the matching score is performed based on the coverage of the structured data to the test item set, and the ratio between the number of test items covered by the structured data and the total number of test items in the test item set is calculated to obtain the detection coverage rate. By eliminating inapplicable items, test items that are not involved in the chip itself are avoided from being misjudged as unexecuted items, thus improving the accuracy of project scoring and risk matching.

[0118] S501, based on the mapping relationship between items and scores in the standard evaluation rule base, matches structured data with items, determines the scores corresponding to successfully matched items as sub-item scores of structured data, and generates detailed sub-item score details.

[0119] The mapping relationship between items and scores can be represented by a quantitative scoring rule consisting of a base score (100 points) and bonus points (20 points). The base score is the testing score, with a total value of 100 points, including two main sections: report quality (30 points) and test item coverage (70 points, normalized). The report quality is further subdivided into three sub-items: equipment effectiveness, test result information completeness, and test condition information completeness (10 points each). The test item coverage is scored according to the AEC-Q100 standard, focusing on the coverage of the four core test groups: Group A accelerated environmental stress test, Group B accelerated life simulation test, Group C package integrity test, and Group E electrical performance verification test. Specifically, the original scores for each core test group are set according to the number and importance of the test items included in each test group in the AEC-Q100 standard. For example, Group A has 42 points, Group B has 19 points, Group C has 12 points, and Group E has 32 points, with a total original score of 105 points. When scoring the structured data, the original scores of the covered test items in each test group are calculated first, and then the sum of the original scores of each test group is normalized according to the ratio of 70 / 105 to obtain the score of test item coverage (with a full score of 70 points).

[0120] In the embodiments of this application, for each test item, the computer device determines a scoring coefficient based on the test execution status, and the product of the score corresponding to the test item and the scoring coefficient is used as the sub-item score for that test item. Test execution status includes test conditions, test sequence, and the batch and quantity of test samples: if the test execution status fully meets the standard requirements, the scoring coefficient is 1; if the test execution status partially meets the standard requirements, the scoring coefficient is an intermediate value between 0 and 1, such as when the test sample comes from only one production batch while the standard requires test samples to come from multiple non-continuous production batches, or when the sample quantity or test time is lower than the standard requirements; if the test item is not executed, the scoring coefficient is 0. In some embodiments, the computer device extracts batch information of the test samples from structured data, verifies whether the test samples come from multiple non-continuous production batches, and substitutes the verification result into the process of determining the scoring coefficient. By using tiered scoring, the quantitative scoring conclusion can distinguish different degrees of compliance with test execution, improving the precision of the quantitative scoring.

[0121] Bonus points are awarded for process management, worth 20 points. This section is designed for the entire chip manufacturing process and includes three sub-items: E6 / E7 (6 points), Group D chip manufacturing reliability (10 points), and Group F defect screening (4 points), aligning with the quality requirements for traceability in automotive chip processes. Specifically, the E6 / E7 sub-items correspond to the process management information related to fault classification (E6) and characteristic characterization (E7) in Group E of the AEC-Q100 standard. The detailed breakdown of sub-item scores includes scores for each item related to report quality and scores for each item related to test coverage.

[0122] In the embodiments of this application, when scoring structured data, quantitative scoring rules can be extracted from a standard evaluation rule base. The test items extracted from the structured data are matched with these quantitative scoring rules, and the scores corresponding to successfully matched items are used as the test item coverage scores. Qualitative evaluation conclusions are also matched with the quantitative scoring rules to obtain a report quality score. Specifically, the proportion of compliant test items among all test items is determined, and this proportion is multiplied by the equipment effectiveness score to obtain the actual equipment effectiveness score. Similarly, the proportion of complete test items among all test items is determined, and this proportion is multiplied by the test result information completeness and test condition information completeness scores to obtain the actual test result information completeness and test condition information completeness scores, respectively. Finally, the report quality score and the test item coverage score are summarized to generate a detailed breakdown of sub-items.

[0123] S502, weighted summation of all sub-item scores in the sub-item score details to obtain a quantitative score conclusion.

[0124] In the embodiments of this application, the computer device is pre-configured with weights for report quality, test item coverage, and process management, which are set according to the actual scenario. After obtaining the detailed scores for each item, the weight corresponding to each item score is determined, and the weighted sum of each item score and its weight is obtained to arrive at a quantitative scoring conclusion. It should be noted that the weights of the items in each type of quantitative scoring rule are the same. For example, the weights of equipment effectiveness and test result information completeness in report quality are the same.

[0125] In the embodiments of this application, the computer device further determines the evaluation level of the chip testing report based on the quantitative scoring conclusion: if the quantitative scoring conclusion is greater than or equal to the first threshold, the evaluation level is A, indicating that the report elements are complete, the test items are complete, and the test samples are sufficient; if the quantitative scoring conclusion is greater than or equal to the second threshold but less than the first threshold, the evaluation level is B, indicating that the report elements are incomplete, the test items are missing, and the test samples are insufficient; if the quantitative scoring conclusion is less than the second threshold, the evaluation level is C, indicating that the report elements are significantly incomplete and the test items are missing a lot. For example, the first threshold is 80 points, and the second threshold is 60 points. The evaluation level is output along with the quantitative scoring conclusion, making it convenient for users to quickly judge the quality and reliability level of the chip testing report.

[0126] The above-mentioned quantitative scoring rules are used to evaluate structured data, which solves the problem of industry report evaluation relying on human experience and inconsistent standards, and promotes the industry to shift from experience-based screening to data-driven approaches.

[0127] In one exemplary embodiment, such as Figure 6 As shown, risk matching is performed on structured data to obtain risk item alert information, including:

[0128] S601 filters and processes the structured data to obtain data on untested items.

[0129] Among them, the data for untested items can be the untested test items and the test conditions or test results related to the test items.

[0130] In the embodiments of this application, after obtaining structured data, the computer device can determine test items from the structured data and match the test items with the mandatory test items in the risk matching rules. If at least one mandatory test item fails to match, the mandatory test item is filtered to obtain untested item data.

[0131] In some embodiments, the computer device can also perform quality checks on structured data, and after the structured data has undergone compliance checks, integrity checks, and accuracy checks, it can identify untested data items from the integrity check results.

[0132] S602 performs risk matching on untested project data to obtain risk item alert information.

[0133] The untested test items are correlated with potential failure risks. These test items include accelerated environmental stress testing, accelerated life simulation testing, package assembly integrity testing, wafer manufacturing reliability testing, electrical characteristic verification testing, defect screening and monitoring testing, and unsealed device integrity testing. Each test item comprises several sub-items; the absence of different sub-items will result in different potential failure risks. This correlation is established based on AEC-Q100 standard clauses and chip failure physical mechanisms. For example, accelerated environmental stress testing includes temperature and humidity bias or high accelerated stress testing, high voltage accelerated aging or unbiased HAST, temperature cycling, power temperature cycling, and high-temperature storage life: The lack of temperature and humidity bias or high accelerated stress testing may lead to potential failure risks such as metallization corrosion, package leakage, and bond point damage; the lack of high voltage accelerated aging or unbiased HAST may lead to potential failure risks such as ion contamination corrosion and failure to expose insufficient package hermeticity; the lack of temperature cycling may lead to potential failure risks such as bond wire breakage, chip cracking, and delamination between molding compound and chip; the lack of power temperature cycling may lead to potential failure risks such as bond point fatigue, electromigration, and via breakage; and the lack of high-temperature storage life may lead to potential failure risks such as intermetallic compound embrittlement and silicone grease drying.

[0134] In the embodiments of this application, the screened untested project data is matched with the untested test items in the mapping relationship, and the potential failure risks of the successfully matched potential failure risks are taken as the potential failure risks corresponding to the actual untested project data. The potential failure risks of all actual untested project data are summarized to obtain risk item reminder information.

[0135] The above method provides enterprises with comprehensive reporting value and risk identification basis by automatically identifying potential failure risks of untested items, thus meeting the high reliability and long cycle quality requirements of automotive chips.

[0136] In one exemplary embodiment, such as Figure 7 As shown, the method for constructing the standard evaluation rule base includes:

[0137] S701: Obtain standard chip test reports, extract core elements from standard chip test reports, and generate quality test rules.

[0138] The standard chip test report can be a chip test report that conforms to the AEC-Q100 standard. Its core elements include test items, test judgment conditions, and quality and reliability requirements.

[0139] In the embodiments of this application, experts in the field of automotive chip quality assessment are organized to deconstruct the AEC-Q100 standard clause by clause, extracting core elements such as 43 test items, test judgment conditions, equipment requirements, and data accuracy requirements of the AG group. Computer equipment acquires AEC-Q100 chip test reports from different chip companies. After information extraction and standardization of the reports, the most frequently used elements are selected as core elements. Compliance condition formulas are generated based on the quality reliability requirements of the core elements; integrity condition formulas are generated based on the test judgment conditions and test items; and accuracy condition formulas are generated based on the test items, test judgment conditions, and quality reliability requirements. The compliance, integrity, and accuracy condition formulas are then summarized to generate quality inspection rules.

[0140] S702, obtain the preset report quality scoring rules and test item scoring rules, and generate project scoring rules based on the report quality scoring rules and test item scoring rules.

[0141] In the embodiments of this application, the report quality scoring rules and test item scoring rules can be preset by domain experts according to the AEC-Q100 standard. In the implementation using a deep learning model for the evaluation model, the preset report quality scoring rules and test item scoring rules can also be used as initial scoring rules. Chip testing reports from different chip companies (including valid reports, missing reports, non-compliant reports, etc.) are collected, and after information extraction and standardization, domain experts label the quality evaluation conclusions and scoring results of the reports to form a training dataset. The training dataset is then input into the deep learning model for training. During training, the model parameters corresponding to the initial scoring rules are iteratively adjusted based on the results of the loss function. When the model tends to converge, the scoring rule parameters corresponding to the model in the converged state are obtained to obtain the item scoring rules. Through this method, the scoring rules inherit the experience of domain experts and can continuously optimize the judgment accuracy and scoring precision through training with batch report samples.

[0142] S703 establishes the correspondence between test items and potential failure risks when they are not executed, and generates risk matching rules.

[0143] In the embodiments of this application, the computer device can identify missing items and existing problems in batch chip inspection reports. For each missing item, the most frequently occurring problem is identified from the existing problems as a potential failure risk, and a correspondence is established between the missing item and the potential failure risk. After establishing a correspondence between all missing items and potential failure risks, all correspondences are summarized to generate risk matching rules. The above correspondences can also be verified and supplemented by domain experts based on AEC-Q100 standard clauses and chip failure physical mechanisms.

[0144] S704 combines quality inspection rules, project scoring rules, and risk matching rules to construct a standard evaluation rule library.

[0145] In the embodiments of this application, after obtaining the quality inspection rules, project scoring rules, and risk matching rules, the various rules are stored in blocks to generate a standard evaluation rule library. The standard evaluation rule library supports online updates and can iterate the rules according to the version upgrade of the AEC-Q100 standard and changes in industry quality requirements.

[0146] By establishing a standard evaluation rule base and constructing an evaluation model, a three-dimensional evaluation framework and quantitative scoring rules are formed, enabling automated extraction, intelligent evaluation, and quantitative scoring of chip testing report information. This effectively verifies the test technical data in chip testing reports, replacing the traditional method of manual page-by-page verification, and improving the processing efficiency and evaluation accuracy of test technical data. At the same time, it can solve the problem of industry report evaluation relying on human experience and inconsistent standards.

[0147] In one exemplary embodiment, such as Figure 8 As shown, the chip test report is standardized to obtain structured data, including:

[0148] S801 converts the chip test report to obtain the initial text data.

[0149] In the embodiments of this application, the user logs into the software program through a browser on a computer device and uploads the chip testing report to be evaluated. The chip testing report can be in any format such as Word, PDF, scanned document, or image. The chip testing report is then converted into initial text data using OCR (Optical Character Recognition) technology, allowing the text to be edited.

[0150] S802 uses regular expressions and / or keyword matching to extract features from the initial text data to obtain structured data.

[0151] The structured data can include key information such as test items, test parameters, test results, equipment information, and test conditions. It can also include report number, testing organization information and signature information, test start and end dates, testing standards and versions, sample information (such as sample name, quantity, number, production batch number, and sample batch), testing environment information, and remarks (such as reference instructions for general data of the same spectrum of products). The above fields correspond to the report cover, report overview, testing plan, report text, and attachments of the test report, providing a data foundation for quality testing, project scoring, and risk matching.

[0152] In embodiments of this application, a computer device can use regular expressions to extract features from initial text data to obtain structured data. In some embodiments, keyword matching can also be used to extract features from initial text data to obtain structured data. In some embodiments, both regular expressions and keyword matching can be used to extract features from initial text data to obtain structured data.

[0153] The above method converts the extracted unstructured information into a standardized structured data table, while labeling the text features required for qualitative evaluation and the data dimensions required for visualization, thereby achieving unified processing of report information in different formats and providing a standardized data foundation for subsequent evaluation.

[0154] In one exemplary embodiment, the method further includes: converting the report evaluation results into a visual infographic; the visual infographic includes a test coverage statistics chart, a quality evaluation radar chart, a failure risk heat map, and a score bar chart for each test group.

[0155] In the embodiments of this application, the computer device uses a built-in visualization engine to convert report evaluation results into visual infographics, supporting the export and printing of visual information and text reports, thus achieving an intuitive presentation of evaluation results and facilitating quick understanding and decision-making by users. Transforming abstract evaluation results into intuitive graphical data allows OEMs, Tier 1 suppliers, and chip companies to quickly grasp report quality, enabling horizontal supplier comparisons and scientific decision-making. In some embodiments, the computer device also supports batch uploading and batch evaluation of multiple reports, storing evaluation data and reports in a database, supporting historical report queries, supplier report comparisons, and statistical analysis of evaluation data.

[0156] In some embodiments, the computer device also supports the fusion evaluation of multiple test reports for the same automotive-grade chip product: after performing the above evaluation method on each of the multiple test reports to obtain their respective report evaluation results, the scores of each report are merged and statistically analyzed to obtain a comprehensive quantitative scoring conclusion, overcoming the limitation that a single report only covers a portion of the batch tests. In some embodiments, the computer device also identifies common data of the same product line referenced in the report from the remarks information of the structured data, and verifies the validity of the common data, including verifying whether the common data contains failure records and whether there have been changes in the chip product's manufacturing process that affect the applicability of the common data; when the common data contains unrectified failure records or there are process changes that affect applicability, corresponding prompt information is output in the critical issue list.

[0157] In addition to the methods described in all the above embodiments, a method for evaluating automotive-grade chip testing reports is also provided, such as... Figure 9 As shown, the method includes:

[0158] S901: Obtain the automotive-grade chip test report to be evaluated, convert the chip test report to a new format, and obtain the initial text data.

[0159] S902, use regular expressions and / or keyword matching to extract features from the initial text data to obtain structured data;

[0160] S903: Input structured data into the evaluation model, perform compliance checks on the structured data, and obtain the first detection result; the first detection result is used to characterize whether the structured data is compliant;

[0161] S904, Perform integrity checks on the structured data to obtain a second check result; the second check result is used to characterize whether the structured data is complete;

[0162] S905, perform accuracy testing on the structured data to obtain a third test result; the third test result is used to characterize whether the structured data is accurate;

[0163] S906, Determine the qualitative evaluation conclusions and key issue list based on the first test results, the second test results, and the third test results;

[0164] S907, based on the mapping relationship between items and scores in the standard evaluation rule base, matches structured data with items, determines the scores corresponding to successfully matched items as the sub-item scores of structured data, and generates detailed sub-item score details;

[0165] S908, weighted summation of all sub-item scores in the sub-item score details to obtain a quantitative score conclusion;

[0166] S909, the structured data is filtered and processed to obtain the data of untested items;

[0167] S910 performs risk matching on untested project data and obtains risk item alert information.

[0168] Each of the above steps has been described in the foregoing embodiments. For details, please refer to the foregoing content. They will not be repeated here.

[0169] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0170] Based on the same inventive concept, this application also provides an evaluation device for automotive-grade chip testing reports, which implements the evaluation method for automotive-grade chip testing reports described above. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations of one or more embodiments of the evaluation device for automotive-grade chip testing reports provided below can be found in the limitations of the evaluation method for automotive-grade chip testing reports described above, and will not be repeated here.

[0171] In one exemplary embodiment, such as Figure 10 As shown, an evaluation device for automotive-grade chip testing reports is provided, comprising: an acquisition module 101 and an evaluation module 102, wherein:

[0172] The acquisition module 101 is used to acquire the automotive-grade chip test report to be evaluated, and to standardize the chip test report to obtain structured data.

[0173] The evaluation module 102 is used to perform quality inspection, project scoring, and risk matching on the structured data using a pre-built evaluation model to obtain the report evaluation results. The evaluation model is built based on a standard evaluation rule library, which includes quality inspection rules, project scoring rules, and risk matching rules. The quality inspection includes compliance inspection, completeness inspection, and accuracy inspection; based on the results of the quality inspection, a qualitative evaluation conclusion and a list of key issues are determined. Based on the results of the project scoring, a quantitative scoring conclusion and detailed sub-item scores are determined. The risk matching includes: filtering the structured data to obtain untested project data; performing risk matching on the untested project data to obtain risk item reminder information. The report evaluation results include multiple options from the qualitative evaluation conclusion, the quantitative scoring conclusion, the detailed sub-item scores, the list of key issues, and the risk item reminder information.

[0174] In an exemplary embodiment, the above-described apparatus further includes a construction module 103, configured to: acquire a standard chip testing report; extract core elements from the standard chip testing report to generate quality testing rules; the core elements include test items, test judgment conditions, and quality reliability requirements; acquire preset report quality scoring rules and test item scoring rules; generate project scoring rules based on the report quality scoring rules and the test item scoring rules; establish a correspondence between the test items when they are not executed and potential failure risks, and generate risk matching rules; and combine the quality testing rules, the project scoring rules, and the risk matching rules to construct the standard evaluation rule library.

[0175] In an exemplary embodiment, the acquisition module 101 is specifically used to: convert the format of the chip detection report to obtain initial text data; and extract features from the initial text data using regular expressions and / or keyword matching to obtain the structured data.

[0176] In an exemplary embodiment, the above-described apparatus further includes a conversion module 104, configured to: convert the report evaluation results into a visual infographic; the visual infographic includes a test coverage statistics chart, a quality evaluation radar chart, a failure risk heat map, and a bar chart of scores for each test group.

[0177] The acquisition module 101, evaluation module 102, construction module 103 and conversion module 104 in the above-mentioned device correspond to the report information extraction and format unification module, AI intelligent evaluation and quantitative scoring module, evaluation system construction module and result output and visualization display module in the technical disclosure scheme, respectively.

[0178] The modules in the aforementioned evaluation device for automotive-grade chip testing reports can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0179] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program that, when executed by the processor, implements the steps in the above-described method embodiments.

[0180] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0181] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0182] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0183] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0184] The above embodiments only illustrate several implementation methods of this application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of this application's patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. For example, the technical solutions of this application can be adapted to other AEC-Q series automotive-grade testing standards such as AEC-Q101, AEC-Q102, AEC-Q103, AEC-Q104, and AEC-Q200, as well as the AQG324 automotive-grade power module standard; knowledge graphs can also be introduced to associate failure cases and test data, or interfaced with chip simulation tools to incorporate evaluation results into virtual vehicle testing, forming a full-chain quality control system. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An evaluation method for automotive-grade chip testing reports, characterized in that, The method includes: Obtain the automotive-grade chip test report to be evaluated, and standardize the chip test report to obtain structured data; The structured data is subjected to quality inspection, project scoring, and risk matching using a pre-built evaluation model to obtain the report evaluation results; wherein, the evaluation model is built based on a standard evaluation rule base, which includes quality inspection rules, project scoring rules, and risk matching rules; The quality inspection includes compliance inspection, integrity inspection and accuracy inspection, and qualitative evaluation conclusions and a list of key issues are determined based on the results of the quality inspection; Based on the results of the project scoring, determine the quantitative scoring conclusions and detailed scores for each item; The risk matching includes: filtering the structured data to obtain untested item data, performing risk matching on the untested item data, and obtaining risk item reminder information; The report evaluation results include multiple aspects such as the qualitative evaluation conclusions, the quantitative scoring conclusions, the detailed breakdown of sub-items scores, the list of key issues, and the risk item reminder information.

2. The method according to claim 1, characterized in that, The automotive-grade chip test report is an AEC-Q series standard test report or an AQG324 standard test report; wherein, the AEC-Q series standards include AEC-Q100, AEC-Q101, AEC-Q102, AEC-Q103, AEC-Q104 and AEC-Q200.

3. The method according to claim 1, characterized in that, The process of performing quality inspection on the structured data, and determining qualitative evaluation conclusions and a list of key issues based on the results of the quality inspection, includes: The structured data is subjected to a compliance check to obtain a first check result; the first check result is used to characterize whether the structured data is compliant. The structured data is subjected to integrity checks to obtain a second check result; the second check result is used to characterize whether the structured data is complete. The structured data is subjected to an accuracy test to obtain a third test result; the third test result is used to characterize whether the structured data is accurate. The qualitative evaluation conclusion and the list of key issues are determined based on the first test result, the second test result, and the third test result.

4. The method according to claim 1, characterized in that, The structured data is scored on items, and the quantitative scoring conclusion and the detailed sub-item scores are determined based on the results of the item scoring, including: Based on the mapping relationship between items and scores in the standard evaluation rule base, the structured data is matched with the items, and the scores corresponding to the successfully matched items are determined as the sub-item scores of the structured data, generating detailed sub-item scores. The quantitative scoring conclusion is obtained by weighted summation of all the sub-item scores in the detailed sub-item score list.

5. The method according to claim 4, characterized in that, The mapping relationship between the items and the scores includes a quantitative scoring rule consisting of a base score and bonus items; The basic score is 100 points, including 30 points for report quality and 70 points for test item coverage. The report quality includes three sub-items: equipment effectiveness, completeness of test result information, and completeness of test condition information, each worth 10 points. The test item coverage is scored according to the AEC-Q100 standard for the coverage of four core test groups: accelerated environmental stress test group, accelerated life simulation test group, package integrity test group, and electrical performance verification test group. The original scores for each of the core test groups total 105 points, which are then normalized to 70 points at a ratio of 70 / 105. The bonus points are 20 points for process management, including 6 points for fault classification and characteristic characterization, 10 points for chip manufacturing reliability, and 4 points for defect screening.

6. The method according to claim 5, characterized in that, Scoring projects based on the structured data also includes: Obtain the structural characteristics, packaging characteristics, process characteristics, and material characteristics of the automotive-grade chip. Based on the structural characteristics, packaging characteristics, process characteristics, and material characteristics, remove inapplicable items from all test items to obtain the set of test items. The coverage of the test item set is scored based on the structured data; wherein the quantitative scoring conclusion includes the detection coverage rate, which is the ratio between the number of test items covered by the structured data and the total number of test items in the test item set.

7. The method according to claim 6, characterized in that, The score corresponding to the successfully matched item is the sub-item score of the structured data, including: for each item, determining a scoring coefficient based on the test execution status; wherein, the test execution status includes test conditions, test order, and batch and quantity of test samples; if the test execution status fully meets the standard requirements, the scoring coefficient is a first coefficient; if the test execution status partially meets the standard requirements, the scoring coefficient is a second coefficient less than the first coefficient; if the item is not executed, the scoring coefficient is zero; the product of the score corresponding to the item and the scoring coefficient is determined as the sub-item score.

8. The method according to claim 1, characterized in that, The process of filtering the structured data to obtain untested item data includes: Extract the test item number or identifier from the structured data, match the test item number or identifier with the mandatory test items in the risk matching rules, and filter out the mandatory test items that have not been executed as the untested item data; The step of performing risk matching on the untested project data to obtain the risk item alert information includes: matching the untested project data with the mapping relationship between test items and potential failure risks in the risk matching rules to obtain the risk item alert information.

9. The method according to claim 8, characterized in that, The mapping relationship between the test items and potential failure risks includes at least the following: When temperature and humidity bias or high accelerated stress testing is not performed, the corresponding potential failure risks are metallization corrosion, package leakage and bond point damage. When high-pressure accelerated aging or unbiased high-accelerated stress testing is not performed, the corresponding potential failure risks are ion contamination corrosion and the inability to expose packaging hermeticity defects. When temperature cycling testing is not performed, the potential failure risks are bond wire breakage, chip cracking, and delamination between the molding compound and the chip. When power temperature cycling tests are not performed, the potential failure risks are bond point fatigue, electromigration, and via fracture. When high-temperature storage life testing is not performed, the potential failure risks are intermetallic compound embrittlement and silicone grease drying.

10. The method according to any one of claims 1-9, characterized in that, The method for constructing the standard evaluation rule base includes: Obtain a standard chip test report, extract core elements from the standard chip test report, and generate quality testing rules; the core elements include test items, test judgment conditions, and quality reliability requirements. Obtain preset report quality scoring rules and test item scoring rules, and generate project scoring rules based on the report quality scoring rules and test item scoring rules; Establish a correspondence between the test items and potential failure risks when they are not executed, and generate risk matching rules; The quality inspection rules, the project scoring rules, and the risk matching rules are combined to construct the standard evaluation rule library.

11. The method according to claim 1, characterized in that, The standardization process of the chip testing report to obtain structured data includes: The chip test report is format-converted to obtain initial text data; The structured data is obtained by extracting features from the initial text data using regular expressions and / or keyword matching.

12. The method according to any one of claims 1-9, characterized in that, The method further includes: The report evaluation results are converted into visual infographics; the visual infographics include test coverage statistics charts, quality evaluation radar charts, failure risk heat maps, and score bar charts for each test group.

13. An evaluation device for automotive-grade chip testing reports, characterized in that, The device includes: The acquisition module is used to acquire the automotive-grade chip test report to be evaluated, and to standardize the chip test report to obtain structured data. The evaluation module is used to perform quality checks, project scoring, and risk matching on the structured data using a pre-built evaluation model to obtain the report evaluation results. The evaluation model is built based on a standard evaluation rule library, which includes quality check rules, project scoring rules, and risk matching rules. The quality check includes compliance checks, completeness checks, and accuracy checks; based on the results of the quality check, qualitative evaluation conclusions and a list of key issues are determined. Based on the project scoring results, quantitative scoring conclusions and detailed sub-item scores are determined. The risk matching includes: filtering the structured data to obtain untested project data; performing risk matching on the untested project data to obtain risk item alert information. The report evaluation results include multiple options from the qualitative evaluation conclusions, quantitative scoring conclusions, detailed sub-item scores, the list of key issues, and the risk item alert information.

14. The apparatus according to claim 13, characterized in that, The device further includes a construction module for: Obtain a standard chip test report, extract core elements from the standard chip test report, and generate quality testing rules; the core elements include test items, test judgment conditions, and quality reliability requirements. Obtain preset report quality scoring rules and test item scoring rules, and generate project scoring rules based on the report quality scoring rules and test item scoring rules; Establish a correspondence between the test items and potential failure risks when they are not executed, and generate risk matching rules; The quality inspection rules, the project scoring rules, and the risk matching rules are combined to construct the standard evaluation rule library.

15. The apparatus according to claim 13, characterized in that, The acquisition module is specifically used for: converting the format of the chip test report to obtain initial text data; and extracting features from the initial text data using regular expressions and / or keyword matching to obtain the structured data.

16. The apparatus according to any one of claims 13-15, characterized in that, The device also includes a conversion module for converting the report evaluation results into a visual information graph; the visual information graph includes a test coverage statistical graph, a quality evaluation radar graph, a failure risk heat map, and a bar chart of scores for each test group.

17. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 12.

18. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 12.

19. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 12.

Images