A periodic test optimization method based on equipment failure mode management

Abstract

This invention discloses a method for optimizing periodic tests based on equipment failure mode management, relating to the field of pattern recognition technology. The method includes: calculating the remaining uncovered risk value of each failure mode in the real-time status table of failure modes, generating a remaining uncovered risk table; determining the risk acceptance of each periodic test object based on the remaining uncovered risk table and a failure mode test mapping table, generating a periodic test necessity table; screening low-necessity periodic test objects in the periodic test necessity table and performing a test omission counterfactual risk assessment, generating a periodic test optimization candidate table; and determining adjustment constraints and execution strategies for each periodic test object based on the periodic test necessity table and the periodic test optimization candidate table, generating a periodic test execution task table. This invention improves the accuracy of coverage evaluation, the safety of test adjustments, and the rationality of resource allocation through pattern recognition detection mapping, historical calibration, coverage resolution, and counterfactual verification.

Description

Technical Field

[0001] This invention relates to the field of pattern recognition technology, and in particular to a periodic test optimization method based on equipment failure mode management. Background Technology

[0002] With the development of equipment management platforms, condition monitoring systems, and maintenance closed-loop systems, the management of periodic tests in industrial equipment, energy equipment, rail transit equipment, and critical infrastructure is gradually shifting from manual ledger management to data-driven, object-oriented, and closed-loop management. Existing periodic tests are typically based on equipment procedures, maintenance cycles, test item lists, and historical fault records. Through fixed-cycle execution, abnormal result verification, and maintenance feedback closed loops, the availability of equipment functions and the ability to expose hidden faults are ensured. At the same time, failure mode and effect analysis, equipment health management, pattern recognition, and risk-oriented maintenance are gradually being introduced into the periodic test optimization process to identify high-consequence faults, hidden faults, and functional items requiring mandatory verification from fault manifestations, abnormal directions, test feedback, and closed-loop records.

[0003] However, existing periodic test optimization techniques still have shortcomings. Existing methods often struggle to accurately distinguish whether a periodic test directly detects, detects precursors, or indirectly detects a particular fault mode. They also struggle to calibrate effective hits, false triggers, and missed detections in the detection relationship using historical closed-loop records. This leads to test necessity being easily solidified by static procedures or human experience. Multiple test items may originate from the same test object, the same test conditions, or the same feedback source. Existing methods often simply accumulate coverage evidence, which can easily lead to duplicate coverage being mistakenly identified as independent coverage. Furthermore, when adjusting low-necessity tests, there is a lack of counterfactual verification of whether the fault mode coverage status deteriorates after temporarily excluding the test. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a periodic test optimization method based on equipment failure mode management to solve the problems of difficulty in accurately calibrating the failure mode detection coverage relationship and lack of counterfactual risk verification for low-necessity test adjustments.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a periodic test optimization method based on equipment failure mode management, comprising: using equipment code, component code, and failure mode code as indexes to record the failure mode management attributes of each failure mode, constructing an equipment failure mode library; retrieving the test closed-loop records of the equipment to be optimized and performing objectification processing to obtain periodic test objects; based on the equipment failure mode library, establishing an initial detection mapping relationship between failure modes and periodic test objects and performing historical calibration to generate a failure mode test mapping table; converting the test closed-loop records participating in historical calibration in the failure mode test mapping table into coverage evidence of the corresponding failure modes to form a failure mode coverage ledger; based on the failure mode coverage ledger, performing independence resolution on the coverage evidence to generate an effective coverage status table, and determining the effective coverage status based on the coverage status. The system receives real-time triggered events and associates them with corresponding failure modes, updating the current tendency, coverage evidence quality, and coverage evidence validity of the corresponding failure mode to generate a real-time failure mode status table. It calculates the remaining uncovered risk value for each failure mode in the real-time failure mode status table, generating a remaining uncovered risk table. Based on the remaining uncovered risk table and the failure mode experiment mapping table, it performs risk acceptance judgment on each periodic experiment object, generating a periodic experiment necessity table. It filters low-necessity periodic experiment objects from the periodic experiment necessity table and performs a counterfactual risk assessment of experiment omissions, generating a periodic experiment optimization candidate table. Based on the periodic experiment necessity table and the periodic experiment optimization candidate table, it performs adjustment constraint judgment and execution strategy tuning on each periodic experiment object, generating a periodic experiment execution task table.

[0007] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management described in this invention, the steps of recording the failure mode management attributes of each failure mode using equipment code, component code, and failure mode code as indexes, constructing an equipment failure mode library, retrieving the test closed-loop records of the equipment to be optimized, and performing object-oriented processing to obtain periodic test objects are as follows: The original records of fault modes of the equipment to be optimized are indexed and normalized. The equipment code, component code and fault mode code are combined into a fault mode index to form a fault mode object that corresponds one-to-one with each fault mode. Content related to fault identification, risk assessment, and test constraints is grouped into fault mode management attributes and associated with the corresponding fault mode objects to build an equipment fault mode library. Closed-loop endpoint records belonging to the same periodic test process are chained together to form a test closed-loop chain, and closed-loop features used to characterize test anomalies and closed-loop feedback relationships are extracted to generate periodic test objects.

[0008] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the failure mode test mapping table are as follows: Based on the fault mode management attributes in the equipment fault mode library, extract the fault identification features corresponding to each fault mode object and the test anomaly features corresponding to each periodic test object. The fault identification features are matched with the test anomaly features, and the fault mode objects and periodic test objects that meet the requirements of consistent anomaly objects, consistent anomaly directions, and compatible anomaly occurrence conditions are defined as candidate detection relationships. The candidate detection relationships are determined by the detection type and the detection location. The detection type of the failure mode object corresponding to the periodic test object is analyzed and the initial test detectability is assigned to form the initial detection mapping relationship. Based on the closed-loop characteristics of the periodic test objects, the initial detection mapping relationship is historically calibrated to obtain the effective hit relationship, false trigger relationship and missed detection relationship, and the initial test detectability of the corresponding initial detection mapping relationship is corrected to generate a failure mode test mapping table.

[0009] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for forming the failure mode coverage ledger are as follows: The closed-loop test records that participated in historical calibration in the failure mode test mapping table are associated with the corresponding failure mode objects, and coverage evidence processing is performed to generate coverage evidence for the corresponding failure mode objects, as well as corresponding coverage strength, time status and source dependency tags. The coverage evidence, coverage strength, timeliness status, and source dependency markers are ledgerized and collected according to the failure mode coding. Coverage evidence corresponding to the same failure mode object is written into the same failure mode ledger item to form a failure mode coverage ledger.

[0010] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the effective coverage status table are as follows: Coverage evidence in the failure mode coverage ledger that has at least one of the common test object, common test conditions, and common closed-loop feedback sources is classified as duplicate coverage evidence, and coverage evidence that does not have common source dependence is classified as independent coverage evidence. Repeated coverage evidence is processed to eliminate its coverage contribution, while independent coverage evidence is processed to retain its coverage contribution. Combining the coverage strength and timeliness of the coverage evidence, the corresponding failure modes are classified into sufficient coverage, insufficient coverage, and coverage failure states, and an effective coverage state table is generated.

[0011] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the real-time status table of failure modes are as follows: Based on the time-series index of the closed-loop test record and the current calculation cycle, the closed-loop record that has undergone state changes is processed into an event-based event, generating a real-time trigger event, and configuring an event description item for the real-time trigger event; The event description item is associated and matched with the fault mode object in the equipment fault mode library to determine the target fault mode corresponding to the real-time triggered event and the event association strength, and then linkage correction is performed to generate a real-time status table of fault modes.

[0012] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the remaining uncovered risk table are as follows: Based on the effective coverage status table, the fault mode management attributes, current tendency and effective coverage status of the fault mode corresponding to the real-time status table of the fault mode are processed by risk parameterization to obtain the fault consequence quantity, historical occurrence quantity, hidden dependency quantity, current tendency quantity and effective coverage quantity. The risk baseline level of the corresponding failure mode is divided by the failure consequence quantity and the hidden dependency quantity, and the risk baseline level is adjusted in real time according to the current tendency quantity to obtain the failure risk level. Based on the effective coverage, the fault risk level is reduced by coverage status to obtain the remaining uncovered risk value of the corresponding fault mode. The remaining uncovered risk values ​​of each fault mode are then collected according to the fault mode code to generate a remaining uncovered risk table.

[0013] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the periodic test necessity table are as follows: Based on the failure mode test mapping table, establish the risk transfer path between failure mode objects and periodic test objects, and perform risk transfer allocation processing on the risk transfer path corresponding to the same failure mode object, dividing it into primary transfer path and secondary transfer path. Based on the remaining uncovered risk value undertaken by each periodic trial object and the corresponding primary and secondary undertaking paths, the necessity level of each periodic trial object is classified, and a periodic trial necessity table is generated.

[0014] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the periodic test optimization candidate table are as follows: Based on the periodic test necessity table, periodic test objects that do not correspond to the main undertaking path and only correspond to the auxiliary undertaking path are sorted according to the remaining uncovered risk value undertaken, and low necessity periodic test objects are identified. In the failure mode coverage ledger, temporarily exclude coverage evidence corresponding to low-necessity periodic test objects, and re-perform independence resolution on the excluded coverage evidence to obtain counterfactual coverage status; Consistency is determined between the counterfactual coverage status and the current coverage status in the effective coverage status table, and a candidate table for periodic experiment optimization is generated.

[0015] As a preferred embodiment of the periodic test optimization method based on equipment failure mode management according to the present invention, the specific steps for generating the periodic test execution task table are as follows: Based on the necessity level and acceptance path type in the periodic test necessity table, and in conjunction with the periodic test optimization candidate table, the execution attribution of each periodic test object is tuned to obtain the enhanced execution set, the candidate adjustment set, and the maintenance execution set; By analyzing the mandatory test attributes, fault concealment attributes, fault consequence levels, and test constraint attributes of the corresponding fault mode objects in the candidate adjustment set, the allowable adjustment range of each periodic test object is divided, and the candidate adjustment set is adjusted according to the allowable adjustment range to obtain the adjustable test set. The execution strategies for the reinforcement execution set, the adjustable test set, and the maintenance execution set are respectively tuned to form a periodic test execution task table.

[0016] The beneficial effects of this invention are as follows: By performing pattern recognition and corresponding matching between fault identification features and test anomaly features, and combining detection type determination, detection location positioning, and historical closed-loop calibration, a fault mode test mapping table is formed. This transforms the static experience judgment of the fault mode detection capability of periodically tested objects into a calibrable mapping relationship, improving the accuracy and traceability of periodic test necessity evaluation. By eliminating the independence of coverage evidence, assuming the risk of remaining uncovered areas, and verifying the counterfactual coverage status of low-necessity tests, the invention eliminates evidence of duplicate coverage and predicts the risk degradation after test adjustments, reducing duplicate tests, improving the critical fault coverage assurance capability, and enhancing the rationality of periodic test resource allocation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a periodic testing optimization method based on equipment failure mode management.

[0019] Figure 2 This is a schematic diagram of the formation of the Failure Mode and Condition Mapping Table.

[0020] Figure 3 This diagram illustrates the connection between the coverage status and the remaining uncovered risks.

[0021] Figure 4 A schematic diagram for setting up the test to exclude counterfactual assessment and execution tasks. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a periodic test optimization method based on equipment failure mode management, comprising the following steps: S1. Using equipment code, component code, and fault mode code as indexes, record the fault mode management attributes of each fault mode, build an equipment fault mode library, retrieve the test closed-loop records of the equipment to be optimized, and perform object-oriented processing to obtain periodic test objects.

[0026] S1.1. Index and normalize the original records of the fault modes of the equipment to be optimized, and combine the equipment code, component code and fault mode code into a fault mode index to form a fault mode object that corresponds one-to-one with each fault mode.

[0027] It should be noted that the original records of the fault modes corresponding to the equipment to be optimized are collected. The original records of the fault modes come from the equipment ledger, component ledger, fault mode analysis documents, historical fault records, defect handling records and maintenance closed-loop records, and are retrieved through the data interface of the equipment management platform or entered according to the three object levels of equipment, component and fault mode.

[0028] The device identifier, component identifier, and fault mode identifier in the original fault mode record are indexed and normalized: the device code, component code, and fault mode code are combined in the order of "device code-component code-fault mode code" to form the fault mode index.

[0029] When multiple fault mode original records correspond to the same fault mode index, the multiple fault mode original records are merged into the same fault mode index to form a fault mode object; when the fault mode original record lacks equipment code, component code or fault mode code, no fault mode object is generated, but the corresponding fault mode original record and the source of the original record are retained.

[0030] S1.2. Collect the content related to fault identification, risk assessment and test constraints into fault mode management attributes, and associate them with the corresponding fault mode objects to build a device fault mode library.

[0031] It should be noted that, for each failure mode object, content related to failure identification, risk assessment, and test constraints is extracted from the original failure mode records merged into the same failure mode index. The extracted content is then aggregated into the failure mode management attributes of the corresponding failure mode object. Among them, the failure identification content includes the failure mode name, failure performance description, failure cause description, and observable symptoms; the risk assessment content includes the failure consequences, failure severity, historical occurrence frequency, and failure concealment attributes; and the test constraint content includes mandatory test attributes, permission period limits, prohibition of suspension markers, and manual review requirements.

[0032] Historical occurrence frequency is determined based on the number of occurrences of closed fault handling records under the same fault mode index and the statistical observation window. The statistical observation window is the time range between the attribute effective time and the historical occurrence frequency statistics cutoff time. The historical occurrence frequency statistics cutoff time is the system time when the historical occurrence frequency statistics action is executed. When the statistical observation window is less than the baseline execution cycle of the corresponding periodic test object, the baseline execution cycle is used as the statistical observation window. Historical occurrence frequency is expressed as the number of occurrences within a unit statistical observation window. When there are no closed fault handling records under the same fault mode index, the historical occurrence frequency is zero.

[0033] When there are multiple original records of the same failure mode object, the entry value of the failure mode management attribute is determined according to the source type and effective time of the original record. The priority of the source type is arranged in the order of approved failure mode analysis documents, effective periodic test procedures, closed failure handling records, and closed defect handling records. When there are multiple source values ​​for the same attribute, the source value with the highest priority is adopted. When there are multiple source values ​​for the same priority, the source value with the latest effective time is adopted.

[0034] The fault mode index, fault mode object, fault mode management attribute, attribute source type, and attribute effective time are written into the same record to form a device fault mode library.

[0035] S1.3 Connect the closed-loop point records belonging to the same periodic test process into a test closed-loop chain, and extract the closed-loop features used to characterize test anomalies and closed-loop feedback relationships to generate periodic test objects.

[0036] It should be noted that the test closed-loop record of the equipment to be optimized should be retrieved. The test closed-loop record consists of the closed-loop node record formed during the periodic test. The closed-loop node record includes the test item identifier, equipment code, component code, test step identifier, measurement point object, node type, node occurrence time, test judgment result, and feedback result. The node occurrence time adopts the original timestamp in the closed-loop node record.

[0037] Process attribution processing is performed on the closed-loop test records: Closed-loop milestone records with consistent equipment codes, component codes, and test item identifiers, and whose node occurrence times belong to the same periodic test process, are grouped into the same periodic test process; the starting point of a periodic test process is determined by the test task generation node, and if there is no test task generation node, it is determined by the earliest test execution node under the same test item identifier; the ending point of a periodic test process is determined by the review and feedback node, and if there is no review and feedback node, it is determined by the test execution completion node; Closed-loop milestone records located between the starting point and the ending point are connected in sequence according to the chronological order of node occurrence time to form a closed-loop test chain.

[0038] Extracting closed-loop features from the test closed-loop chain: Closed-loop features consist of anomaly object, anomaly direction, anomaly occurrence conditions, and closed-loop feedback results. Anomaly object is determined based on the measurement point object, test step identifier, and component code in the closed-loop node record. Anomaly direction is determined based on the test judgment results. Anomaly occurrence conditions are determined based on the test steps, triggering actions, and execution conditions in the corresponding closed-loop node record. Closed-loop feedback results are determined based on the feedback results in the anomaly confirmation node or review feedback node.

[0039] When the test result is an abnormal result, or the feedback result confirms the abnormality, the corresponding closed loop point record is identified as the abnormality source record, and a closed loop feature item is generated; if there is no abnormality source record in the same test closed loop chain, a qualified closed loop feature item is generated. The qualified closed loop feature item records the test item identifier, test step identifier, measurement point object, test completion time, and qualified feedback result.

[0040] The test closed-loop chain and closed-loop features are solidified into periodic test objects. Each periodic test object corresponds to one periodic test process and has a unique periodic test object identifier.

[0041] S2. Based on the equipment failure mode library, establish an initial detection mapping relationship between failure modes and periodically tested objects and perform historical calibration to generate a failure mode test mapping table. Then, convert the closed-loop test records that participated in the historical calibration in the failure mode test mapping table into coverage evidence of the corresponding failure modes to form a failure mode coverage ledger.

[0042] S2.1. Based on the fault mode management attributes in the equipment fault mode library, extract the fault identification features corresponding to each fault mode object and the test anomaly features corresponding to each periodic test object.

[0043] It should be noted that, for each fault mode object in the equipment fault mode library, the content describing the fault manifestation in the corresponding fault mode management attribute is read, and the content of the fault manifestation is characterized: the fault object item, fault direction item, and fault condition item are determined from the content of the fault manifestation; the fault object item includes the component, functional location, action object, measurement point object, or affected parameter to which the fault points; the fault direction item includes the abnormal change type; and the fault condition item includes the operating condition, trigger action, feedback status, or test condition; fault object items, fault direction items, and fault condition items that belong to the same fault manifestation content and have a corresponding relationship are combined into a fault identification feature item, and multiple fault manifestation contents form multiple fault identification feature items respectively.

[0044] For each periodic test object, identify closed-loop feature items generated from the anomaly source record from the closed-loop features. Use the anomaly object, anomaly direction, and anomaly occurrence condition in the closed-loop feature item as the test anomaly object item, test anomaly direction item, and test anomaly condition item, respectively, and combine them to generate test anomaly feature items. For periodic test objects formed by qualified closed-loop feature items, only the qualified closed-loop mark is retained, and no test anomaly feature items are generated.

[0045] S2.2 Perform feature correspondence matching between fault identification features and test anomaly features, and define fault mode objects and periodic test objects that meet the requirements of consistent anomaly objects, consistent anomaly directions, and compatible anomaly occurrence conditions as candidate detection relationships. It should be noted that the fault identification features and the test abnormal features are matched item by item, and the abnormal object consistency judgment, the abnormal direction consistency judgment, and the abnormal occurrence condition compatibility judgment are performed respectively.

[0046] The consistency of abnormal objects is determined by whether the fault object item and the test abnormal object item point to the same component, functional location, action object, measuring point object, or affected parameter; if the text names of the two are different but the object identifiers can be matched with the same object, they are still determined to be consistent abnormal objects; if only the equipment code is the same and the specific object cannot be determined to be consistent, the consistency of abnormal objects is not determined.

[0047] The consistency of the abnormal direction is determined by the fault direction item and the test abnormal direction item pointing to the same abnormal change type; if different textual expressions correspond to the same abnormal change type, it is determined that the abnormal direction is consistent; if the two point to opposite change types or different action results, it is not determined that the abnormal direction is consistent.

[0048] The condition for an anomaly to occur is considered compatible if there are no mutual exclusion requirements between the fault condition item and the test anomaly condition item in terms of operating conditions, test conditions, trigger actions, and feedback status. If there are conflicts between equipment input and output status, conflict between trigger action execution and prohibition, or conflict between the establishment and deactivation of the lockout status, the anomaly to occur is determined to be incompatible. If any condition item is not recorded, the unrecorded condition item will not be used as a mutual exclusion condition for judgment.

[0049] When the same fault identification feature and the same test anomaly feature simultaneously satisfy the conditions of consistent anomaly object, consistent anomaly direction, and compatible anomaly occurrence, a candidate detection relationship is established between the fault mode object to which the fault identification feature belongs and the periodic test object to which the test anomaly feature belongs.

[0050] If there are multiple matching results of feature terms that meet the conditions between the same fault mode object and the same periodic test object, the multiple matching results of feature terms will be merged into the same candidate detection relationship, and the fault identification feature terms, test abnormality feature terms and the corresponding three types of matching judgment results will be retained.

[0051] S2.3 Determine the detection type and locate the detection position for the candidate detection relationship, analyze the detection type of the fault mode object corresponding to the periodic test object, assign the initial test detectability, and form the initial detection mapping relationship.

[0052] It should be noted that for each candidate detection relationship, the fault identification feature items, test abnormality feature items, and corresponding matching judgment results involved in the matching are read. The detection type is determined based on the degree of correspondence between the abnormal object, abnormal direction, and abnormal occurrence conditions. Specifically, if the fault identification feature item and the test abnormality feature item point to the same measurement point object, action object, or affected parameter, it is determined to be a direct detection type. If the test abnormality feature item corresponds to the response delay, fluctuation abnormality, action lag, or leakage abnormality before the fault occurs, it is determined to be a precursor detection type. If the association is established only through functional location, feedback status, or trigger action, it is determined to be an indirect detection type.

[0053] Detection location of candidate detection relationships: Based on the closed-loop node record that generates abnormal test features, determine the detectable location of candidate detection relationships in periodic test objects according to the test step identifier, measurement point object and test item identifier in the closed-loop node record; if the same candidate detection relationship corresponds to multiple detectable locations, retain each detectable location separately.

[0054] After determining the detection type and locating the detection location, the initial test detectability is determined based on the completeness of the detection type, detection location, and abnormal occurrence conditions. The detection sequence of the direct detection type is higher than that of the precursor detection type, and the detection sequence of the precursor detection type is higher than that of the indirect detection type. The detection sequence corresponding to the indirect detection type is designated as the first detection sequence, the detection sequence corresponding to the precursor detection type is designated as the second detection sequence, and the detection sequence corresponding to the direct detection type is designated as the third detection sequence. The first, second, and third detection sequences are converted into the initial test detectability according to the ratio of the sequence number to the total number of detection sequences. The initial test detectability for the first detection sequence is one-third, for the second detection sequence it is two-thirds, and for the third detection sequence it is one.

[0055] If the detection location includes both the test step identifier and the measurement point object, and there are no unrestricted conditions for the anomaly, the detection sequence corresponding to the current detection type is maintained. If the detection location includes only one of the test step identifier and the measurement point object, or if there are unrestricted conditions for the anomaly, the detection sequence corresponding to the current detection type is downgraded by one position. If the current detection type already corresponds to the first detection sequence, it will not be downgraded further. If the detection location cannot locate both the test step identifier and the measurement point object, the detection sequence is determined as the detection sequence corresponding to the indirect detection type, and the corresponding initial test detectability is obtained according to the determined detection sequence. Among them, the higher the detection sequence, the greater the initial test detectability. The initial test detectability is not less than one-third of the value and not more than one value.

[0056] The fault mode object, the periodic test object, the fault identification feature items involved in the matching, the test abnormal feature items, the detection type, the detection location, and the initial test detectability are associated to form an initial detection mapping relationship.

[0057] S2.4 Based on the closed-loop characteristics of the periodic test objects, perform historical calibration on the initial detection mapping relationship, obtain the effective hit relationship, false trigger relationship and missed detection relationship, and correct the initial test detectability of the corresponding initial detection mapping relationship to generate a failure mode test mapping table.

[0058] It should be noted that historical calibration is performed on the initial detection mapping relationship: closed-loop feature items that are consistent with the detection location and whose abnormal object, abnormal direction, and abnormal occurrence conditions are consistent with the initial detection mapping relationship are used as calibration feature items. When the closed-loop feedback result of the calibration feature item confirms the abnormality and points to the corresponding fault mode object, it is recorded as a valid hit relationship; when the calibration feature item has a test abnormality, but the closed-loop feedback result does not confirm the abnormality or does not point to the corresponding fault mode object, it is recorded as a false trigger relationship; when the periodic test object forms a qualified closed-loop feature item, but the closed-loop feedback result of the same periodic test process points to the corresponding fault mode object, it is recorded as a missed detection relationship.

[0059] The sum of the number of valid hits, false triggers, and missed detections in the same initial detection mapping relationship is taken as the total number of calibrations. When the total number of calibrations is zero, the initial test detectability is taken as the corrected test detectability, and the calibration status is recorded as uncalibrated. When the total number of calibrations is not zero, the ratio of the number of valid hits to the total number of calibrations is taken as the historical calibration detectability. The historical calibration detectability is determined as the corrected test detectability, and the calibration status is recorded as calibrated. When the historical calibration detectability is lower than the corresponding initial test detectability, the historical calibration detectability is used to replace the initial test detectability, indicating that the historical closed-loop record reduces the actual detection capability of the periodic test object for the corresponding fault mode object. When the historical calibration detectability is higher than or equal to the corresponding initial test detectability, the historical calibration detectability is used to replace the initial test detectability, indicating that the historical closed-loop record confirms or improves the actual detection capability of the detection relationship.

[0060] The fault mode test mapping table is generated by associating the fault mode object, periodic test object, detection type, detection location, number of valid hits, number of false triggers, number of missed detections, corrected test detectability, calibration status, and closed-loop feature items involved in calibration with the same mapping record.

[0061] S2.5. Associate the closed-loop test records that participated in historical calibration in the failure mode test mapping table with the corresponding failure mode objects, and perform coverage evidence processing to generate coverage evidence for the corresponding failure mode objects, as well as corresponding coverage strength, time status and source dependency tags.

[0062] It should be noted that for each mapping record in the failure mode test mapping table, the test closed-loop record corresponding to the closed-loop feature item participating in the calibration is associated with the same failure mode object and serves as the evidence carrier for covering evidence.

[0063] Test closed-loop records corresponding to valid hit relationships are used as confirmation coverage evidence, test closed-loop records corresponding to false trigger relationships are used as coverage evidence to be verified, and test closed-loop records corresponding to missed detection relationships are used as failure coverage evidence. Each piece of coverage evidence retains the corresponding test closed-loop record identifier and records the corresponding failure mode object, periodic test object, detection location, and coverage evidence type.

[0064] The coverage strength of confirmed coverage evidence is determined using the corrected test detectability in the corresponding mapping record; the coverage strength of coverage evidence awaiting verification is recorded as awaiting verification; the coverage strength of failed coverage evidence is recorded as failed. Simultaneously, the latest node occurrence time in the closed-loop record corresponding to the coverage evidence is extracted as the evidence timestamp, and the evidence timestamp and the periodic test object identifier are written into the validity status; the node occurrence time uses the original timestamp in the closed-loop record; the validity status also includes valid validity status, invalid validity status, and pending verification status; the validity status is determined based on the evidence timestamp, the license period limit of the corresponding failure mode object, and the validity judgment time, where the validity judgment time is the status update time corresponding to the generation of the coverage evidence; if the time interval between the evidence timestamp and the validity judgment time is not greater than the license period limit, it is determined to be valid validity status; if the time interval between the evidence timestamp and the validity judgment time is greater than the license period limit, it is determined to be invalid validity status; if the evidence timestamp is missing, the evidence timestamp is later than the validity judgment time, or the license period limit is missing, it is determined to be pending verification status.

[0065] The periodic test object identifier, test item identifier, test procedure identifier, measurement point object, and closed-loop feedback source are included as components of the source dependency marker; when two pieces of coverage evidence have the same periodic test object identifier, test item identifier, test procedure identifier, measurement point object, or closed-loop feedback source, the two pieces of coverage evidence are determined to have common source dependency.

[0066] S2.6. According to the failure mode coding, the coverage evidence, coverage strength, time status and source dependency mark are ledgerized and collected. The coverage evidence corresponding to the same failure mode object is written into the same failure mode ledger item to form a failure mode coverage ledger.

[0067] It should be noted that the failure mode code corresponding to the coverage evidence is used as the ledger aggregation index. Coverage evidence with the same failure mode code is written into the same failure mode ledger item, and coverage evidence with different failure mode codes is written into different failure mode ledger items.

[0068] When writing to the failure mode ledger, the test closed-loop record identifier, periodic test object identifier, detection location, coverage evidence type, coverage strength, timeliness status, and source dependence flag corresponding to the coverage evidence are written as the same coverage evidence item. If two pieces of coverage evidence correspond to the same test closed-loop record identifier, the same periodic test object identifier, the same detection location, and the same coverage evidence type, the two pieces of coverage evidence are merged into one coverage evidence item, and the later timeliness status is retained; if any item is different, they are retained as different coverage evidence items, forming a failure mode coverage ledger.

[0069] S3. Based on the fault mode coverage ledger, perform independent resolution on the coverage evidence to generate a valid coverage status table. Receive real-time trigger events according to the valid coverage status table and associate them with the corresponding fault mode. Update the current tendency, coverage evidence quality, and coverage evidence validity of the corresponding fault mode to generate a real-time fault mode status table.

[0070] S3.1. Coverage evidence in the failure mode coverage ledger that has at least one of the common test object, common test conditions, and common closed-loop feedback sources is classified as duplicate coverage evidence, and coverage evidence that does not have common source dependence is classified as independent coverage evidence.

[0071] It should be noted that for any two coverage evidence items under the same failure mode ledger, a source dependency comparison shall be performed. The source dependency comparison shall be based on the source dependency marker: if any one of the periodic test object identifier, test item identifier, test step identifier, or measurement point object is consistent between the two coverage evidence items, it shall be determined that there is a common test object; if the test step identifier, triggering action, execution condition, and measurement point object are consistent between the two coverage evidence items, it shall be determined that there is a common test condition; if the abnormality confirmation node, review feedback node, or feedback record source is consistent between the two coverage evidence items, it shall be determined that there is a common closed-loop feedback source.

[0072] For coverage evidence items that have any one of the common source dependencies among common test objects, common test conditions, and common closed-loop feedback sources, the coverage evidence items are classified as duplicate coverage evidence, and duplicate coverage groups are formed according to the type of common source dependency; for coverage evidence items under the same failure mode ledger that do not have common test objects, common test conditions, and common closed-loop feedback sources with other coverage evidence items, the coverage evidence items are classified as independent coverage evidence.

[0073] If the source dependency markers of a coverage evidence item are missing some comparison content, then only the recorded comparison content will be used for source dependency comparison. If the source dependency markers cannot provide any of the following: periodic test object identifier, test item identifier, test step identifier, measurement point object, and closed-loop feedback source, then it will be treated as duplicate coverage evidence to avoid coverage evidence with unknown sources being counted as independent coverage evidence.

[0074] After the partitioning is completed, duplicate coverage evidence, duplicate coverage groups, and independent coverage evidence are obtained under the same failure mode ledger item.

[0075] S3.2 Perform coverage contribution reduction processing on duplicate coverage evidence and coverage contribution retention processing on independent coverage evidence. Combining the coverage strength and timeliness status of the coverage evidence, classify the corresponding failure modes into sufficient coverage state, insufficient coverage state, and coverage failure state, and generate an effective coverage state table.

[0076] It should be noted that, for coverage evidence under the same failure mode ledger item, the non-expired coverage evidence is first determined based on the validity status of each coverage evidence. Coverage evidence with a validity status of valid is read as non-expired coverage evidence. Coverage evidence with a validity status of expired or pending verification is retained in the failure mode coverage ledger and kept as a source record, but does not participate in the determination of the current valid coverage status. Among them, the validity status of expired means that the coverage evidence has exceeded the license period limit, and the validity status of pending verification means that the coverage evidence cannot complete the validity determination due to abnormal evidence timestamp, current calculation period end point, or license period limit.

[0077] For the portion of non-expired coverage evidence that is classified as duplicate coverage evidence, read the duplicate coverage group that has been divided according to common source dependency. Coverage evidence within the same duplicate coverage group is only allowed to form a coverage effect once. The coverage evidence with the highest coverage strength within the group forms the group coverage strength. Other coverage evidence in the same group only retains the source record and is not counted repeatedly in the effective coverage.

[0078] For the portion of non-expired coverage evidence that is classified as independent coverage evidence, the coverage strength of each independent coverage evidence is retained as independent coverage strength and participates in the generation of effective coverage together with the group coverage strength; coverage evidence with an expiration status of expiration or expiration pending verification is not included in the coverage contribution retention process.

[0079] The effective coverage of a fault mode is generated based on the group coverage strength and independent coverage strength under the same fault mode ledger item. Specifically, the portion of the group coverage strength and independent coverage strength that is a numerical coverage strength is used as the effective coverage input item. The pending state and the failure state are not included in the effective coverage generation as numerical coverage strength. The effective coverage input items are accumulated starting from zero. Accumulation stops when the accumulated result reaches one value, and the current accumulated result is determined as the effective coverage. If there is no effective coverage input item, the effective coverage is determined to be zero. The coverage status is classified according to the evidence type of the non-failed coverage evidence and the effective coverage: if there is no non-failed coverage evidence, it is classified as a coverage failure status; if there is non-failed coverage evidence but no effective coverage input item, it is classified as a coverage insufficiency status; if there is an effective coverage input item and the effective coverage reaches one value, it is classified as a coverage sufficient status; if there is an effective coverage input item but the effective coverage does not reach one value, it is classified as a coverage insufficiency status.

[0080] Fault mode codes, effective coverage, and coverage status are written into the effective coverage status table. Repeated coverage groups and independent coverage evidence are only retained as source records for the coverage status generation process and are not required output fields of the effective coverage status table.

[0081] S3.3. Based on the time-series index of the experimental closed-loop record and the current calculation cycle, perform event-based processing on the closed-loop record that has undergone state changes, generate real-time trigger events, and configure event description items for the real-time trigger events.

[0082] It should be noted that a time-series index is formed based on the occurrence time of each closed loop node record in the test closed loop record. The current calculation cycle is determined by the time from the last time the effective coverage state table was generated to the time of this state update. When the state update is performed for the first time, the time from the time the fault mode coverage ledger was formed to the time of this state update is used as the current calculation cycle.

[0083] Within the current calculation cycle, extract closed loop node records whose occurrence time falls within the current calculation cycle according to the time sequence index, as well as closed loop node records whose occurrence time is earlier than the current calculation cycle but whose feedback results, verification results, or test judgment results are updated within the current calculation cycle. The extracted closed loop node records are then identified as closed loop records that have undergone state changes.

[0084] When processing closed-loop records of state changes as events, the event type is determined based on the node type, test judgment result, feedback result, and node occurrence time in the closed-loop node record. The event types include test completion, test anomaly, anomaly confirmation, anomaly denial, review completion, and feedback update. An event description item is formed based on the equipment code, component code, test step identifier, measurement point object, test judgment result, feedback result, test step, triggering action, execution condition, and node occurrence time. The event description item includes the event object, event direction, event conditions, event time, and feedback result.

[0085] When generating real-time trigger events, each real-time trigger event is associated with at least one closed-loop record that has undergone a state change, and the original record identifier and original timestamp of the corresponding closed-loop record are retained; the changes of closed-loop records within the current calculation cycle are converted into real-time trigger events, and a set of real-time trigger events with event description items is output.

[0086] S3.4. Associate and match the event description item with the fault mode object in the equipment fault mode library to determine the target fault mode corresponding to the real-time triggered event and the event association strength, and perform linkage correction to generate a real-time status table of fault modes.

[0087] It should be noted that during the association matching process, if the event object in the event description item and the fault mode object point to the same component, the same functional location, the same action object, the same measurement point object, or the same affected parameter, they are determined to be consistent objects; if the event direction and the fault manifestation in the fault mode object belong to the same abnormal change type, they are determined to be of the same direction; if the event conditions and the fault occurrence conditions in the fault mode object do not have mutually exclusive requirements in terms of operating conditions, triggering actions, and feedback states, they are determined to be compatible conditions.

[0088] For events with identical objects, similar directions, and compatible conditions, the corresponding fault mode object is designated as the target fault mode, and the event association strength is set to the full association strength. For events with identical objects and similar directions but whose event conditions are not recorded, the corresponding fault mode object is designated as the target fault mode, and the event association strength is set to the conditional default association strength. If the object is not identical or the direction is not similar, the target fault mode is not determined. If the same real-time triggered event corresponds to multiple target fault modes, corresponding event association strengths are formed for each mode. The event association strength is an enumerated value field, and its value range only includes the full association strength and the conditional default association strength. The condition for the event association strength to be the full association strength is that the real-time triggered event and the target fault mode simultaneously satisfy the conditions of identical objects, similar directions, and compatible conditions. The condition for the event association strength to be the conditional default association strength is that the real-time triggered event and the target fault mode simultaneously satisfy the conditions of identical objects and similar directions, and the event conditions are not recorded. The completeness of the association with the full association strength is higher than that with the conditional default association strength. No numerical weight is set for the event association strength.

[0089] After determining the target fault mode, the coverage status of the corresponding target fault mode in the effective coverage status table is read, and the current tendency is updated according to the event type and feedback result of the real-time triggered event: if the event type or feedback result indicates that the anomaly is established, it is recorded as an anomaly pointing tendency; if the event type or feedback result indicates that the anomaly is not established, it is recorded as an anomaly resolving tendency; if it cannot indicate that the anomaly is established or not established, the original current tendency is maintained; when there are multiple real-time triggered events that can update the current tendency within the current calculation cycle for the same target fault mode, one real-time triggered event is selected as the source event of the current tendency in sequence according to the event time, feedback confirmation degree, and event correlation strength; Real-time triggered events with later times are given priority. If the event times are the same but the feedback confirmation levels are different, the real-time triggered event with the higher feedback confirmation level is given priority. If the event times and feedback confirmation levels are the same, the real-time triggered event with the full correlation strength is given priority over the real-time triggered event with the conditional default correlation strength. If a unique source event still cannot be selected, and multiple real-time triggered events respectively represent the establishment and non-establishment of an anomaly, the current tendency is maintained, and a tendency conflict flag is recorded. The event correlation strength is written as a record field in the real-time status item of the fault mode to identify the degree of correlation between the real-time triggered event and the target fault mode.

[0090] The quality of the coverage evidence is determined according to the degree of confirmation of the feedback results. Feedback results that have been reviewed and confirmed are recorded as reviewed and confirmed quality, feedback results that come from the completion of the experiment but have not yet been reviewed and confirmed are recorded as execution confirmation quality, and feedback results that have not yet formed a confirmation conclusion are recorded as pending confirmation quality. The validity period of the coverage evidence is determined by the original timestamp of the closed-loop point record corresponding to the real-time triggering event. If there are multiple real-time triggering events for the same target failure mode within the current calculation cycle, the validity period of the coverage evidence is determined by the real-time triggering event with the latest event time.

[0091] The fault mode code, target fault mode, real-time trigger event identifier, event association strength, current tendency, quality of covered evidence, validity period of covered evidence, and coverage status are written into the same fault mode real-time status item to generate a fault mode real-time status table. If there are multiple real-time trigger events for the same target fault mode within the current calculation period, the real-time trigger event identifier selected according to the event time, feedback confirmation degree, and event association strength is written into the tendency source event identifier. If a unique real-time trigger event still cannot be selected and multiple real-time trigger events respectively represent the establishment of an anomaly and the failure of an anomaly, the tendency conflict flag is written into the corresponding fault mode real-time status item.

[0092] S4. Calculate the remaining uncovered risk value of each fault mode in the real-time status table of fault modes, generate the remaining uncovered risk table, and determine the risk acceptance of each periodic test object based on the remaining uncovered risk table and the fault mode test mapping table, and generate the periodic test necessity table.

[0093] S4.1 Based on the effective coverage status table, the fault mode management attributes, current tendency and effective coverage status of the fault mode corresponding to the fault mode real-time status table are processed by risk parameterization to obtain the fault consequence quantity, historical occurrence quantity, hidden dependency quantity, current tendency quantity and effective coverage quantity.

[0094] It should be noted that for each real-time status item in the real-time status table of fault modes, the fault mode management attributes, current tendency, quality of covered evidence, validity period of covered evidence, and coverage status and effective coverage amount of the corresponding fault mode object are read according to the fault mode code. If the fault mode management attributes include historical occurrence frequency, the historical occurrence frequency is also read; if the fault mode management attributes do not record historical occurrence frequency, the historical occurrence frequency is set to zero.

[0095] Risk attribution analysis is performed on the fault mode management attributes to extract descriptive content that characterizes the scope of fault impact and form fault consequence attribution. Descriptive content that characterizes the fault exposure mode is extracted and formed fault exposure mode attribution. Historical occurrence attribution is formed based on the historical occurrence frequency of fault mode objects corresponding to the same equipment code and the same component code.

[0096] The attribution of fault consequences, the attribution of fault exposure methods, and the current trend are all categorized and numbered, and the historical occurrence frequency is normalized: the fault consequences are assigned to one of the following consequence categories: loss of safety function, loss of main equipment function, performance degradation, and warning / alarm, and a fault consequence quantity is generated. The fault consequence categories are arranged in order of risk impact from low to high: warning / alarm, performance degradation, loss of main equipment function, and loss of safety function, and correspond to the first consequence number, second consequence number, third consequence number, and fourth consequence number, respectively. The fault consequence quantity is the ratio of the corresponding consequence number to the total number of consequence categories. If a fault cannot be assigned to a consequence category, the corresponding fault mode object is marked as a consequence pending verification object, no fault consequence quantity is generated, and no risk baseline level classification is performed on the corresponding fault mode object. The ratio between the historical occurrence frequency and the maximum historical occurrence frequency of each fault mode object under the same equipment code and the same component code is used as the historical occurrence quantity. When the maximum historical occurrence frequency under the same equipment code and the same component code is zero, the historical occurrence quantity is set to zero.

[0097] The fault exposure method is categorized into one of the following: exposed only through periodic testing, detectable through operational data but requiring manual verification, automatically detectable through online alarms, or directly visible during operation. A hidden dependency quantity is generated. The fault exposure method categories are arranged in ascending order of their dependence on periodic testing: directly visible during operation, automatically detectable through online alarms, detectable through operational data but requiring manual verification, and exposed only through periodic testing. These categories correspond to the first hidden sequence number, the second hidden sequence number, the third hidden sequence number, and the fourth hidden sequence number, respectively. The hidden dependency quantity is the ratio of the corresponding hidden sequence number to the total number of hidden categories. If a fault cannot be categorized into an exposure category, the corresponding fault mode object is marked as an exposed object awaiting verification; no hidden dependency quantity is generated, and no risk baseline level classification is performed on the corresponding fault mode object.

[0098] The current tendency is categorized into one of the following tendency categories: anomaly pointing tendency, maintain tendency, or anomaly resolving tendency. A current tendency value is generated based on the tendency category of the current tendency, without reading the event association strength. The event association strength is only used to select the tendency source event when multiple real-time trigger events correspond to the same target failure mode, and is recorded as a field in the real-time status item of the failure mode. If the current tendency has a tendency conflict flag, the current tendency is categorized into the maintain tendency, and no anomaly pointing strength or anomaly resolving strength is generated for risk correction. The current tendency value includes the tendency category and the tendency trigger value. If the current tendency is an anomaly pointing tendency or anomaly resolving tendency, the tendency trigger value is set to one value. If the current tendency is a maintain tendency or has a tendency conflict flag, the tendency trigger value is set to zero.

[0099] Boundary verification is performed on the effective coverage: When the coverage status is in a coverage failure state, the effective coverage is classified as no effective coverage value; when the coverage status is a sufficient coverage state and the quality of the coverage evidence is either the review and confirmation quality or the execution confirmation quality, the effective coverage is classified as sufficient coverage value; when the coverage status is a sufficient coverage state but the quality of the coverage evidence is pending confirmation quality, the effective coverage is not classified as sufficient coverage value, and the effective coverage value in the effective coverage status table is retained according to the insufficient coverage state; when the coverage status is insufficient coverage state, the effective coverage value in the effective coverage status table is retained; when the event time corresponding to the coverage evidence statute is earlier than the timestamp of the evidence participating in the current effective coverage status determination, the real-time triggering event is not used to increase the effective coverage.

[0100] Write the failure consequence quantity, historical occurrence quantity, hidden dependency quantity, current tendency quantity, and effective coverage quantity into the same failure mode real-time status item.

[0101] S4.2. Divide the risk baseline level of the corresponding failure mode by the failure consequence quantity and the hidden dependency quantity, and make real-time situational correction of the risk baseline level according to the current tendency quantity to obtain the failure risk level.

[0102] It should be noted that, for each failure mode object, the geometric mean of the failure consequence quantity and the hidden dependency quantity is used as the inherent risk base value of the corresponding failure mode object. All inherent risk base values ​​are arranged in ascending order of value, and the difference between adjacent inherent risk base values ​​is calculated. When the inherent risk base values ​​involved in the sorting can form more than two adjacent differences, the two adjacent positions with the largest differences are selected as the risk baseline boundary positions. According to the order of the two risk baseline boundary positions in the sorting results, the first risk baseline boundary position and the second risk baseline boundary position are formed respectively. The first risk baseline boundary positions are then sorted. The leading inherent risk base value is determined as the first risk baseline boundary value, and the leading inherent risk base value in the second risk baseline boundary position is determined as the second risk baseline boundary value. When the inherent risk base values ​​participating in the ranking can only form one adjacent difference, the adjacent position is taken as the first risk baseline boundary position, and the leading inherent risk base value in the adjacent position is determined as the first risk baseline boundary value, and no second risk baseline boundary value is generated. When the inherent risk base values ​​participating in the ranking cannot form adjacent differences, no risk baseline boundary position is generated. When multiple adjacent differences are the same, the adjacent position with the later ranking position is selected first.

[0103] When a first risk baseline boundary value and a second risk baseline boundary value are generated, failure mode objects with an inherent risk baseline value not higher than the first risk baseline boundary value are classified as first risk baseline level; failure mode objects with an inherent risk baseline value higher than the first risk baseline boundary value but not higher than the second risk baseline boundary value are classified as second risk baseline level; and failure mode objects with an inherent risk baseline value higher than the second risk baseline boundary value are classified as third risk baseline level. When only a first risk baseline boundary value is generated, failure mode objects with an inherent risk baseline value not higher than the first risk baseline boundary value are classified as first risk baseline level; failure mode objects with an inherent risk baseline value higher than the first risk baseline boundary value are classified as second risk baseline level; and no third risk baseline level is generated. When no risk baseline boundary position is generated, the corresponding failure mode object is classified as second risk baseline level, and a risk baseline pending verification mark is recorded.

[0104] Historical occurrences are not included in the calculation of inherent risk baseline values; they are only used for boundary tuning of risk baseline level attribution. When a first risk baseline boundary value and a second risk baseline boundary value are generated, if the inherent risk baseline value of a failure mode object is equal to the first or second risk baseline boundary value, the historical occurrences of the corresponding failure mode object are read for attribution tuning. If only the first risk baseline boundary value is generated, if the inherent risk baseline value of a failure mode object is equal to the first risk baseline boundary value, the historical occurrences of the corresponding failure mode object are read for attribution tuning. If no risk baseline boundary position is generated, historical occurrence tuning is not performed.

[0105] When the historical occurrence volume is higher than the median of the historical occurrence volume of each failure mode object under the same equipment code and component code, the corresponding failure mode object will be assigned to the adjacent higher risk baseline level; when the historical occurrence volume is not higher than the corresponding median value, the risk baseline level determined according to the inherent risk base value will be maintained; the median of the historical occurrence volume is determined according to the sorting result of the historical occurrence volume of each failure mode object under the same equipment code and component code from low to high. When the number of historical occurrence volumes is odd, the historical occurrence volume at the middle position of the sorting will be taken as the median value; when the number of historical occurrence volumes is even, the average of the historical occurrence volumes at the middle two positions will be taken as the median value; the attribution adjustment is only allowed between adjacent risk baseline levels, and it is not allowed to directly adjust from the first risk baseline level to the third risk baseline level, nor is it allowed to reduce the risk baseline level determined according to the inherent risk base value.

[0106] The current tendency value is assigned a direction. When the current tendency value corresponds to an anomaly pointing tendency and the tendency trigger value is one, the tendency trigger value is taken as the positive tendency part and the reverse tendency part is set to zero. When the current tendency value corresponds to an anomaly cancellation tendency and the tendency trigger value is one, the tendency trigger value is taken as the reverse tendency part and the positive tendency part is set to zero. When the current tendency value corresponds to a maintained tendency or a tendency conflict flag, both the positive and reverse tendency parts are set to zero.

[0107] The expression for calculating the failure risk value is: ; in, Indicates the first Fault risk value for each fault mode object; Indicates the first The inherent risk baseline value of each failure mode object; Indicates the first The positive tendency portion of a fault mode object; Indicates the first The reverse tendency part of a fault mode object.

[0108] When a first and second risk baseline boundary value are generated, the fault risk level is determined using both values. If the fault risk value is not higher than the first risk baseline boundary value, it is classified as the first fault risk level. If the fault risk value is higher than the first risk baseline boundary value but not higher than the second risk baseline boundary value, it is classified as the second fault risk level. If the fault risk value is higher than the second risk baseline boundary value, it is classified as the third fault risk level. When only the first risk baseline boundary value is generated, the fault risk level is determined using the first risk baseline boundary value. If the fault risk value is not higher than the first risk baseline boundary value, it is classified as the first fault risk level. If the fault risk value is higher than the first risk baseline boundary value, it is classified as the second fault risk level, and no third fault risk level is generated. When no risk baseline boundary value is generated, the fault risk level remains the second fault risk level, and the risk baseline pending verification mark is written into the corresponding fault mode's real-time status item. The inherent risk base value, risk baseline level, fault risk value, and fault risk level are written into the corresponding fault mode's real-time status item.

[0109] It should also be noted that the expression adopts the linkage correction relationship between the inherent risk base value and the upper and lower boundary margins. The positive tendency part acts on the remaining upward margin between the inherent risk base value and the upper boundary, and the negative tendency part acts on the existing risk margin between the zero value and the inherent risk base value. The boundary constraint keeps the fault risk value between zero and one, and avoids the risk change of the same magnitude under different inherent risk base values ​​for real-time triggered events.

[0110] S4.3. Based on the effective coverage, reduce the coverage status of the fault risk level to obtain the remaining uncovered risk value of the corresponding fault mode, and collect the remaining uncovered risk value of each fault mode according to the fault mode code to generate a remaining uncovered risk table.

[0111] It should be noted that, for each fault mode object, the fault risk value is used as the benchmark value for coverage reduction, and the reduction is performed in combination with the effective coverage and coverage status: If the coverage status is coverage failure, the fault risk value is retained as the remaining uncovered risk value; if the coverage status is sufficient, the remaining uncovered risk value is set to zero; if the coverage status is insufficient, the fault risk value is reduced according to the coverage ratio represented by the effective coverage, and the proportion of the fault risk value corresponding to the effective coverage is determined as the covered risk deduction, and the remaining part after deducting the covered risk deduction is determined as the remaining uncovered risk value; when the effective coverage is zero, no covered risk deduction is formed, and the remaining uncovered risk value remains the fault risk value; when the effective coverage is greater than zero and less than one, a covered risk deduction is formed according to the proportion corresponding to the effective coverage, and the fault risk corresponding to the uncovered proportion is retained as the remaining uncovered risk value; when the effective coverage reaches one, the remaining uncovered risk value is set to zero.

[0112] When the effective coverage amount and coverage status are inconsistent, the coverage status determines whether to perform reduction and the boundary assignment after reduction. The effective coverage amount determines the deduction range under the insufficient coverage status. If the coverage status is insufficient but the effective coverage amount is missing, the effective coverage amount is less than zero, or the effective coverage amount is greater than one, the corresponding risk record will be marked as coverage amount pending verification status, and no coverage status reduction will be performed. The fault risk value will be retained as the remaining uncovered risk value. If the coverage status is sufficient but the quality of the coverage evidence is pending confirmation, the reduction will be performed according to the effective coverage amount corresponding to the insufficient coverage status.

[0113] The remaining uncovered risk values ​​of each fault mode object are collected according to the fault mode code. The fault mode code, fault risk level, coverage status, effective coverage, coverage pending verification status and remaining uncovered risk value are written into the same risk record to generate a remaining uncovered risk table. If no risk record for coverage pending verification status is generated, the coverage pending verification status is recorded as null.

[0114] S4.4. Based on the failure mode test mapping table, establish the risk transfer path between failure mode objects and periodic test objects, and perform risk transfer allocation processing on the risk transfer path corresponding to the same failure mode object, dividing it into primary transfer path and secondary transfer path.

[0115] It should be noted that for each mapping record in the failure mode test mapping table, the corresponding remaining uncovered risk value is searched in the remaining uncovered risk table according to the failure mode code. If the corresponding remaining uncovered risk value can be found, the failure mode object is used as the risk source end and the periodic test object is used as the risk receiving end to establish a risk receiving path. The failure mode code, periodic test object identifier, detection type, detection location, corrected test detectability, and remaining uncovered risk value are written into the same path record.

[0116] For each risk acceptance path, the product of the remaining uncovered risk value and the corrected test detectability is used as the path acceptance value. Then, the risk acceptance paths are grouped according to the failure mode code, and within the path group corresponding to the same failure mode object, they are sorted from high to low according to the path acceptance value. The risk acceptance path with the highest ranking is designated as the primary acceptance path, and the remaining risk acceptance paths are designated as auxiliary acceptance paths.

[0117] When multiple risk-bearing paths with the same acceptance value exist within a path group and are all ranked first, the primary acceptance path is determined sequentially according to the priority of detection type and the completeness of detection location. Among them, direct detection type takes precedence over precursor detection type, precursor detection type takes precedence over indirect detection type, and risk-bearing paths with both test step identifier and measurement point object at the detection location take precedence over risk-bearing paths with only one of them. If they still cannot be distinguished, the corresponding multiple risk-bearing paths are divided into parallel primary acceptance paths.

[0118] S4.5. Based on the remaining uncovered risk value undertaken by each periodic test object and the corresponding primary and secondary undertaking paths, classify the necessity level of each periodic test object and generate a periodic test necessity table.

[0119] It should be noted that risk transfer paths are aggregated according to the periodic trial object identifiers, and the primary and secondary transfer paths pointing to the same periodic trial object are written under the same periodic trial object name; when classifying the necessity level of each periodic trial object, if a primary transfer path exists, the corresponding periodic trial object is classified as primary transfer necessity level; if there is no primary transfer path but there is a secondary transfer path, the corresponding periodic trial object is classified as secondary transfer necessity level; periodic trial objects that are not aggregated into primary and secondary transfer paths are classified as maintenance necessity level.

[0120] Write the periodic test object identifier, necessity level, number of primary acceptance paths, number of secondary acceptance paths, fault mode codes corresponding to primary acceptance paths, fault mode codes corresponding to secondary acceptance paths, and acceptance values ​​of each path into the same record to generate a periodic test necessity table.

[0121] S5. Screen the low-necessity periodic test objects in the periodic test necessity table, perform a counterfactual risk assessment of test omissions, generate a periodic test optimization candidate table, and adjust the constraint judgment and execution strategy of each periodic test object according to the periodic test necessity table and the periodic test optimization candidate table to generate a periodic test execution task table.

[0122] S5.1. Based on the periodic test necessity table, periodic test objects that do not correspond to the main undertaking path and only correspond to the auxiliary undertaking path are sorted according to the remaining uncovered risk value and low necessity periodic test objects are identified.

[0123] It should be noted that each periodic test object and its corresponding undertaking path type are taken from the periodic test necessity table. Periodic test objects that do not have a main undertaking path but have an auxiliary undertaking path are selected as periodic test objects to be screened. Periodic test objects that have a main undertaking path are not included in the low necessity screening range.

[0124] The remaining uncovered risk values ​​in the auxiliary acceptance path corresponding to each subject to be screened for periodic testing are accumulated to obtain the subject acceptance risk value. Within the set of subjects to be screened for periodic testing corresponding to the same equipment code and the same component code, the subjects are sorted from low to high according to the subject acceptance risk value, and a low necessity decision boundary is generated based on the sorting result.

[0125] The low necessity determination boundary is determined by the median position of the risk values ​​of the sorted objects; when the number of sorts is odd, the risk value of the object corresponding to the median position is taken as the low necessity determination boundary; when the number of sorts is even, the average of the risk values ​​of the objects corresponding to the two middle positions is taken as the low necessity determination boundary.

[0126] Objects subject to periodic testing whose risk value is below the low necessity threshold are identified as low necessity periodic testing objects; if the number of objects subject to periodic testing under the same equipment code and the same component code is less than two, they are not considered low necessity periodic testing objects.

[0127] S5.2 In the fault mode coverage ledger, temporarily exclude the coverage evidence corresponding to the low necessity periodic test objects, and re-perform independence resolution on the excluded coverage evidence to obtain the counterfactual coverage status.

[0128] It should be noted that, according to the periodic test object identifier of the low-necessity periodic test object, the coverage evidence item that matches the periodic test object identifier is located in the failure mode coverage ledger and is used as the coverage evidence to be excluded; if the same low-necessity periodic test object corresponds to multiple failure mode ledger items, the coverage evidence to be excluded is located in each failure mode ledger item respectively.

[0129] During the current assessment process, the coverage evidence to be excluded will be temporarily removed from the current coverage determination scope of the corresponding failure mode ledger item, while the coverage evidence formed by other periodic test subjects will be retained; the temporary removal will not rewrite the original coverage evidence item in the failure mode coverage ledger.

[0130] The remaining coverage evidence after temporary removal is subjected to independent resolution processing consistent with the effective coverage status table generation process, and the coverage status of the corresponding failure mode is re-determined based on the coverage strength and timeliness status of the remaining coverage evidence.

[0131] The low-necessity periodic test object identification, corresponding failure mode code, coverage evidence items that were temporarily excluded, and the re-determined coverage sufficient state, coverage insufficient state, or coverage failure state are associated and recorded to form the corresponding counterfactual coverage state.

[0132] S5.3. Determine the consistency between the counterfactual coverage status and the current coverage status in the effective coverage status table, and generate a candidate table for periodic test optimization.

[0133] It should be noted that, for each low-necessity periodic test object, the consistency between the current coverage state and the counterfactual coverage state corresponding to the same fault mode code is determined. The coverage state is arranged in descending order of coverage state: sufficient coverage state, insufficient coverage state, and coverage failure state. If the counterfactual coverage state has not decreased compared to the current coverage state, it is determined to be consistent with the coverage state. If the counterfactual coverage state has decreased compared to the current coverage state, it is determined to be inconsistent with the coverage state.

[0134] When a low-necessity periodic test object corresponds to multiple fault mode codes, consistency determination is performed on each fault mode code. If all fault mode codes are determined to have consistent coverage, the corresponding low-necessity periodic test object is written into the periodic test optimization candidate table. If any fault mode code is determined to have inconsistent coverage, the corresponding low-necessity periodic test object is not written into the periodic test optimization candidate table, and the corresponding fault mode code and inconsistent status change are recorded.

[0135] Write the periodic test object identifier, associated fault mode code, current coverage status, counterfactual coverage status, consistency judgment result, and coverage evidence items that have been temporarily excluded into the same candidate record to generate a periodic test optimization candidate table.

[0136] S5.4. Based on the necessity level and the type of the receiving path in the periodic test necessity table, and in conjunction with the periodic test optimization candidate table, the execution attribution of each periodic test object is tuned to obtain the enhanced execution set, the candidate adjustment set, and the maintenance execution set.

[0137] It should be noted that the periodic test objects in the periodic test necessity table are used as the execution attribution tuning objects. The execution attribution tuning is carried out in the order of strengthening attribution, candidate adjustment attribution, and maintenance attribution. Periodic test objects whose necessity level is the primary acceptance necessity level, or whose acceptance path type includes the primary acceptance path, are assigned to the strengthening execution set; periodic test objects that are not assigned to the strengthening execution set but whose periodic test object identifier has been written into the periodic test optimization candidate table are assigned to the candidate adjustment set; periodic test objects that are not assigned to either the strengthening execution set or the candidate adjustment set are assigned to the maintenance execution set.

[0138] When the same periodic test subject meets multiple assignment criteria at the same time, a unique assignment is determined in the order of reinforcement assignment, candidate adjustment assignment, and maintenance assignment, and the subject is not repeatedly written into multiple sets.

[0139] S5.5 By analyzing the mandatory test attributes, fault concealment attributes, fault consequence levels, and test constraint attributes of the corresponding fault mode objects in the candidate adjustment set, the allowable adjustment range of each periodic test object is divided, and the candidate adjustment set is adjusted according to the allowable adjustment range to obtain the adjustable test set.

[0140] It should be noted that for each periodic test object in the candidate adjustment set, the fault mode management attribute is read according to the corresponding associated fault mode code, and the mandatory test attribute, fault concealment attribute, and fault consequence level are parsed from the fault mode management attribute; at the same time, the permission cycle limit, the prohibition of suspension mark, and the manual review requirement of the associated fault mode object are read, and the permission cycle limit, the prohibition of suspension mark, and the manual review requirement are used as test constraint attributes to participate in the division of the allowable adjustment range; if the permission cycle limit is not read, the extended cycle is not allowed; if the prohibition of suspension mark indicates that suspension is prohibited, the postponement of execution is not allowed; if the manual review requirement indicates that manual review is required, the allowable adjustment range retains the manual review constraint mark, and the fault consequence level is determined by the fault consequence and fault severity in the fault mode management attribute, and is assigned in the order of loss of safety function, loss of main equipment function, performance degradation, and warning alarm; when the same fault mode object corresponds to multiple fault consequence assignments at the same time, the fault consequence assignment ranked first in the order of level is used as the fault consequence level.

[0141] For each associated fault mode object, generate an allowable adjustment subset: If the mandatory test attribute characterization is enforced, the allowable adjustment subset is limited to merged execution; if the fault concealment attribute characterization can only be exposed through periodic testing and the fault consequence level corresponds to the loss of safety functions or the loss of main equipment functions, the allowable adjustment subset is limited to merged execution; if the fault concealment attribute characterization can only be exposed through periodic testing but the fault consequence level does not correspond to the loss of safety functions or the loss of main equipment functions, the allowable adjustment subset is limited to extended period and merged execution; if it does not fall into the above categories, the allowable adjustment subset is determined as extended period, merged execution, and deferred execution.

[0142] After obtaining the allowed adjustment subset, the constraints of the allowed adjustment subset are reduced according to the test constraint attributes: if the allowed adjustment subset contains extended periods and no permitted period limit is read, the extended periods are deleted from the allowed adjustment subset; if the allowed adjustment subset contains deferred execution and the prohibition of pause flag indicates that pause is prohibited, the deferred execution is deleted from the allowed adjustment subset; if the allowed adjustment subset contains merged execution and the manual review requirement indicates that manual review is required, merged execution is retained and a manual review constraint flag is added; if the allowed adjustment subset contains extended periods and the manual review requirement indicates that manual review is required, extended periods are retained and a manual review constraint flag is added.

[0143] When the same periodic test object corresponds to multiple associated fault mode objects, the intersection of multiple allowable adjustment subsets is taken to obtain the allowable adjustment range of the periodic test object. If the allowable adjustment range is empty, the periodic test object is not written into the adjustable test set. If the allowable adjustment range is not empty, the periodic test object is written into the adjustable test set, and the allowable adjustment range, permission cycle limit and manual review constraint mark are written into the adjustable test set together.

[0144] S5.6. Configure the execution strategies for the reinforcement execution set, the adjustable test set, and the maintenance execution set respectively to form a periodic test execution task table.

[0145] It should be noted that, for periodic test objects in the reinforced execution set, adjustable test set, and maintenance execution set, the test item identifier, test step identifier, measurement point object, most recent test completion time, and baseline execution cycle of the corresponding periodic test object are read. A historical execution interval sequence is generated based on the time interval between two adjacent test completions in the test closed-loop record for the same periodic test object. The historical execution interval sequence is arranged from most recent to furthest in time. If a time interval between two adjacent test completions cannot be formed, the corresponding periodic test object is marked as having insufficient historical intervals. If an effective test procedure cycle record exists, it is used as the baseline execution cycle. If no effective test procedure cycle record exists but a historical execution interval exists in the historical execution interval sequence, the most recent historical execution interval in the test closed-loop record for the same periodic test object is used as the baseline execution cycle. If no effective test procedure cycle record exists and the historical execution interval sequence is empty, the corresponding periodic test object is marked as awaiting core release on a weekly basis, and no shortening or extending cycle strategy is generated.

[0146] For the periodic test objects in the enhanced execution set, the enhanced execution strategy is tuned, and the execution strategy type is determined to be either a shortened cycle strategy or an increased priority strategy. When there is a historical execution interval in the historical execution interval sequence that is shorter than the baseline execution cycle, the most recent historical execution interval that is shorter than the baseline execution cycle is selected as the target execution cycle, and the execution strategy type is determined to be a shortened cycle strategy. When there is no historical execution interval in the historical execution interval sequence that is shorter than the baseline execution cycle, no shortened cycle strategy is generated, and the execution strategy type is determined to be an increased priority strategy. The increased priority strategy corresponds to a higher execution priority than the periodic test objects in the maintenance execution set, and the baseline execution cycle is maintained as the target execution cycle. After the execution strategy type and target execution cycle are determined, an enhanced execution task record is formed.

[0147] Optimization and adjustment strategies are tuned for periodic test objects in the adjustable test set, forming optimization and adjustment task records. Optimization and adjustment strategies must not exceed the allowed adjustment range. If the allowed adjustment range includes extending the cycle, merging execution, or postponing execution, extension cycle strategies, merging execution strategies, or postponing execution strategies are formed respectively. If the allowed adjustment range includes extending the cycle, the permitted cycle limit for the corresponding periodic test object in the adjustable test set is read. If the permitted cycle limit is greater than the baseline execution cycle, the smallest historical execution interval in the historical execution interval sequence that is not greater than the permitted cycle limit but greater than the baseline execution cycle is determined as the target execution cycle. If there is no historical execution interval that meets the conditions, the permitted cycle limit is determined as the target execution cycle. If the permitted cycle limit is not greater than the baseline execution cycle, no extension cycle strategy is generated. If the corresponding periodic test object has a weekly expected kernel state, no extension cycle strategy is generated, and the weekly expected kernel state is written into the optimization and adjustment task record.

[0148] For adjustments that allow for merged execution, after the execution time window is generated, multiple periodic test objects with identical equipment codes, component codes, and test item identifiers, and whose execution time windows overlap, will be generated into a merged execution group. Periodic test objects with inconsistent constraint markers after manual review will not be included in the same merged execution group. The test step identifier, measurement point object, and test judgment requirements of each periodic test object must not be changed within the same merged execution group. After the merged execution group is generated, the equipment code, component code, test item identifier, and the start and end times of the overlapping execution time window interval will be combined to form the merged execution group identifier. If the execution time windows do not overlap, no merged execution group will be generated, and the corresponding periodic test object will be retained as a separate optimization and adjustment task record, with the merged execution group identifier recorded as a null value.

[0149] If the scope of adjustment includes temporarily suspended execution, the recovery trigger conditions will be written into the optimization and adjustment task record. The recovery trigger conditions include any one of the following: the corresponding fault mode re-enters the main receiving path, the counterfactual coverage status changes from a fully covered state to an insufficiently covered state, and a new anomaly judgment result occurs.

[0150] For periodic test objects in the maintenance execution set, maintenance strategies are solidified, generating maintenance execution task records. The maintenance execution strategy maintains the baseline execution cycle and original execution time window for the corresponding periodic test objects, without changing the test step identifiers, measurement points, or test judgment requirements. Execution time windows are generated based on the most recent test completion time and the target execution cycle. If there are allowable early or late boundaries in the existing effective test procedure cycle records, the start and end times of the execution time window are determined according to these boundaries. If there are no allowable early or late boundaries, the execution time window is the time point obtained by adding the target execution cycle to the most recent test completion time. Execution priorities are determined based on the execution strategy type: reinforcement execution task records use reinforcement priority, optimization and adjustment task records use adjustment priority, and maintenance execution task records use maintenance priority.

[0151] The records of enhanced execution tasks, optimized adjustment tasks, and maintained execution tasks shall be written into the periodic test execution task table. The periodic test execution task table shall at least record the periodic test object identifier, execution strategy type, target execution cycle, execution time window, execution priority, merged execution group identifier, recovery trigger condition, license cycle limit, manual review constraint mark, weekly review status, and associated fault mode code.

[0152] In summary, this invention achieves the transformation of the ability of periodically tested objects to detect fault patterns from static experience-based judgment to a calibrable mapping relationship by: performing pattern recognition and correspondence matching between fault identification features and test anomaly features; and combining detection type determination, detection location positioning, and historical closed-loop calibration to form a fault mode test mapping table, thereby improving the accuracy and traceability of periodic test necessity evaluation; and by eliminating the independence of coverage evidence, assuming the risk of remaining uncovered areas, and verifying the counterfactual coverage status of low-necessity tests, it eliminates evidence of duplicate coverage and predicts the risk degradation after test adjustments, thereby reducing duplicate tests, improving the coverage assurance capability of critical faults, and improving the rationality of periodic test resource allocation.

[0153] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A periodic test optimization method based on equipment failure mode management, characterized in that, include: Using equipment codes, component codes, and fault mode codes as indexes, the fault mode management attributes of each fault mode are recorded to build an equipment fault mode library. The test closed-loop records of the equipment to be optimized are retrieved and objectified to obtain periodic test objects. Based on the equipment failure mode library, an initial detection mapping relationship between failure modes and periodically tested objects is established and historical calibration is performed to generate a failure mode test mapping table. The test closed-loop records that participated in historical calibration in the failure mode test mapping table are converted into coverage evidence of the corresponding failure modes to form a failure mode coverage ledger. Based on the failure mode coverage ledger, independent resolution is performed on the coverage evidence to generate a valid coverage status table. Real-time trigger events are received according to the valid coverage status table and associated with the corresponding failure mode. The current tendency, coverage evidence quality and coverage evidence validity of the corresponding failure mode are updated to generate a real-time failure mode status table. Calculate the remaining uncovered risk value of each fault mode in the real-time status table of fault modes, generate a remaining uncovered risk table, and determine the risk acceptance of each periodic test object based on the remaining uncovered risk table and the fault mode test mapping table, and generate a periodic test necessity table. Screen low-necessity periodic test objects from the periodic test necessity table, perform a counterfactual risk assessment of test omissions, generate a periodic test optimization candidate table, and adjust the constraint judgment and execution strategy for each periodic test object based on the periodic test necessity table and the periodic test optimization candidate table to generate a periodic test execution task table.

2. The periodic test optimization method based on equipment failure mode management as described in claim 1, characterized in that, The process involves using equipment codes, component codes, and fault mode codes as indexes to record the fault mode management attributes of each fault mode, constructing an equipment fault mode library, retrieving the test closed-loop records of the equipment to be optimized, and performing object-oriented processing to obtain periodic test objects. The specific steps are as follows: The original records of fault modes of the equipment to be optimized are indexed and normalized. The equipment code, component code and fault mode code are combined into a fault mode index to form a fault mode object that corresponds one-to-one with each fault mode. Content related to fault identification, risk assessment, and test constraints is grouped into fault mode management attributes and associated with the corresponding fault mode objects to build an equipment fault mode library. Closed-loop endpoint records belonging to the same periodic test process are chained together to form a test closed-loop chain, and closed-loop features used to characterize test anomalies and closed-loop feedback relationships are extracted to generate periodic test objects.

3. The periodic test optimization method based on equipment failure mode management as described in claim 1 or 2, characterized in that, The specific steps for generating the failure mode test mapping table are as follows: Based on the fault mode management attributes in the equipment fault mode library, extract the fault identification features corresponding to each fault mode object and the test anomaly features corresponding to each periodic test object. The fault identification features are matched with the test anomaly features, and the fault mode objects and periodic test objects that meet the requirements of consistent anomaly objects, consistent anomaly directions, and compatible anomaly occurrence conditions are defined as candidate detection relationships. The candidate detection relationships are determined by the detection type and the detection location. The detection type of the failure mode object corresponding to the periodic test object is analyzed and the initial test detectability is assigned to form the initial detection mapping relationship. Based on the closed-loop characteristics of the periodic test objects, the initial detection mapping relationship is historically calibrated to obtain the effective hit relationship, false trigger relationship and missed detection relationship, and the initial test detectability of the corresponding initial detection mapping relationship is corrected to generate a failure mode test mapping table.

4. The periodic test optimization method based on equipment failure mode management as described in claim 3, characterized in that, The specific steps for establishing a fault mode coverage ledger are as follows: The closed-loop test records that participated in historical calibration in the failure mode test mapping table are associated with the corresponding failure mode objects, and coverage evidence processing is performed to generate coverage evidence for the corresponding failure mode objects, as well as corresponding coverage strength, time status and source dependency tags. The coverage evidence, coverage strength, timeliness status, and source dependency markers are ledgerized and collected according to the failure mode coding. Coverage evidence corresponding to the same failure mode object is written into the same failure mode ledger item to form a failure mode coverage ledger.

5. The periodic test optimization method based on equipment failure mode management as described in claim 4, characterized in that, The specific steps for generating the valid coverage status table are as follows: Coverage evidence in the failure mode coverage ledger that has at least one of the common test object, common test conditions, and common closed-loop feedback sources is classified as duplicate coverage evidence, and coverage evidence that does not have common source dependence is classified as independent coverage evidence. Repeated coverage evidence is processed to eliminate its coverage contribution, while independent coverage evidence is processed to retain its coverage contribution. Combining the coverage strength and timeliness of the coverage evidence, the corresponding failure modes are classified into sufficient coverage, insufficient coverage, and coverage failure states, and an effective coverage state table is generated.

6. The periodic test optimization method based on equipment failure mode management as described in claim 5, characterized in that, The specific steps for generating the real-time status table of fault modes are as follows: Based on the time-series index of the closed-loop test record and the current calculation cycle, the closed-loop record that has undergone state changes is processed into an event-based event, generating a real-time trigger event, and configuring an event description item for the real-time trigger event; The event description item is associated and matched with the fault mode object in the equipment fault mode library to determine the target fault mode corresponding to the real-time triggered event and the event association strength, and then linkage correction is performed to generate a real-time status table of fault modes.

7. The periodic test optimization method based on equipment failure mode management as described in claim 6, characterized in that, The specific steps for generating the remaining uncovered risk table are as follows: Based on the effective coverage status table, the fault mode management attributes, current tendency and effective coverage status of the fault mode corresponding to the real-time status table of the fault mode are processed by risk parameterization to obtain the fault consequence quantity, historical occurrence quantity, hidden dependency quantity, current tendency quantity and effective coverage quantity. The risk baseline level of the corresponding failure mode is divided by the failure consequence quantity and the hidden dependency quantity, and the risk baseline level is adjusted in real time according to the current tendency quantity to obtain the failure risk level. Based on the effective coverage, the fault risk level is reduced by coverage status to obtain the remaining uncovered risk value of the corresponding fault mode. The remaining uncovered risk values ​​of each fault mode are then collected according to the fault mode code to generate a remaining uncovered risk table.

8. The periodic test optimization method based on equipment failure mode management as described in claim 1 or 7, characterized in that, The specific steps for generating the periodic test necessity table are as follows: Based on the failure mode test mapping table, establish the risk transfer path between failure mode objects and periodic test objects, and perform risk transfer allocation processing on the risk transfer path corresponding to the same failure mode object, dividing it into primary transfer path and secondary transfer path. Based on the remaining uncovered risk value undertaken by each periodic trial object and the corresponding primary and secondary undertaking paths, the necessity level of each periodic trial object is classified, and a periodic trial necessity table is generated.

9. The periodic test optimization method based on equipment failure mode management as described in claim 8, characterized in that, The specific steps for generating the candidate table for periodic experiment optimization are as follows: Based on the periodic test necessity table, periodic test objects that do not correspond to the main undertaking path and only correspond to the auxiliary undertaking path are sorted according to the remaining uncovered risk value undertaken, and low necessity periodic test objects are identified. In the failure mode coverage ledger, temporarily exclude coverage evidence corresponding to low-necessity periodic test objects, and re-perform independence resolution on the excluded coverage evidence to obtain counterfactual coverage status; Consistency is determined between the counterfactual coverage status and the current coverage status in the effective coverage status table, and a candidate table for periodic experiment optimization is generated.

10. The periodic test optimization method based on equipment failure mode management as described in claim 9, characterized in that, The specific steps for generating the periodic test execution task list are as follows: Based on the necessity level and acceptance path type in the periodic test necessity table, and in conjunction with the periodic test optimization candidate table, the execution attribution of each periodic test object is tuned to obtain the enhanced execution set, the candidate adjustment set, and the maintenance execution set; By analyzing the mandatory test attributes, fault concealment attributes, fault consequence levels, and test constraint attributes of the corresponding fault mode objects in the candidate adjustment set, the allowable adjustment range of each periodic test object is divided, and the candidate adjustment set is adjusted according to the allowable adjustment range to obtain the adjustable test set. The execution strategies for the reinforcement execution set, the adjustable test set, and the maintenance execution set are respectively tuned to form a periodic test execution task table.

Images