Intelligent diagnosis system for full-life-cycle fault of display terminal

Abstract

The invention belongs to the technical field of display terminal fault diagnosis, and particularly relates to a display terminal full life cycle fault intelligent diagnosis system, which receives fault phenomenon description of a display terminal, retrieves a signal input end control object set associated with a fault phenomenon, and performs targeted detection on each control object in the set. Generating an original detection data set, carrying out instantaneous abnormal point elimination, obtaining an effective fault feature set carrying a potential fault identifier, fusing the effective fault feature set with a historical fault contribution degree weight, obtaining the fault reason confidence of each control object, judging the association strength of a fault mode of each control object and an environment factor through a preset environment association rule, and determining the fault mode of each control object according to the association strength. And generating an environment sensitivity evaluation label, binding the environment sensitivity evaluation label with the fault cause confidence coefficient, and revealing a coupling mechanism of fault root causes and environment inducements to output a fault positioning diagnosis report, thereby avoiding insufficient fault positioning granularity caused by complete machine generalization detection, and providing a more comprehensive basis for fault disposal.

Description

Technical Field

[0001] The present invention belongs to the technical field of display terminal fault diagnosis, and specifically relates to an intelligent fault diagnosis system for display terminals throughout their life cycle. Background Art

[0002] In the modern information society, display terminals are widely used in key scenarios such as production control, information display, and human-computer interaction. Their operational reliability is directly related to production efficiency, information transmission quality, and user experience.

[0003] However, display terminals have complex structures and numerous components. They are susceptible to environmental and aging factors throughout their lifecycles, making them prone to various faults such as screen distortion, flickering, black screens, and bright / dark lines. Traditional fault diagnosis relies on manual experience, resulting in low efficiency and high misjudgment rates. Repairs are often only performed reactively after a fault occurs, severely impacting production continuity and service availability. Therefore, developing an intelligent fault diagnosis system that covers the entire lifecycle of display terminals is crucial to rapidly diagnose and accurately locate faults.

[0004] In the prior art, there are also some solutions related to display terminal fault diagnosis. For example, the remote operation and maintenance method, system, terminal and storage medium of the LED display screen with Chinese patent publication number CN117993884A obtains the operating data of the LED display screen to be operated and maintained to determine whether there is a fault. If it is judged that there is a fault, the operating data is pre-processed to obtain a fault data model of the LED display screen to be operated and maintained, and the data is input into a fault model library containing historical fault data models and their corresponding fault solutions for comparison. The fault type and solution are determined and then issued for execution, thereby greatly improving the remote operation and maintenance efficiency and quality of the LED display screen.

[0005] Another Chinese patent, CN116703367A, discloses a maintenance method and system for an LED liquid crystal display screen. The system obtains first maintenance data and maintenance time, conducts an in-depth inspection of the first maintenance data to obtain second maintenance data, estimates second damage characteristics based on the first data, uses AI to display maintenance plans, and calculates the frequency of failure times. Failures that are below a preset life threshold or above a preset frequency are fed back to the factory, improving the efficiency of fault detection and maintenance and reducing maintenance time and costs.

[0006] Although the above solutions provide some solutions related to display terminal fault diagnosis, the existing technology still has the following limitations, specifically: 1. Existing technologies generally rely on static matching of historical fault cases or generalized detection of whole-machine performance parameters, ignoring the status monitoring and targeted diagnosis of the signal input terminal as a key hub of the display terminal. As a result, the diagnostic process needs to traverse the entire hardware topology, which is inefficient and unable to accurately locate the source of the fault at the signal link or interface level.

[0007] 2. Existing technologies lack a mechanism for verifying the temporal continuity of detection data, which can easily misjudge anomalies caused by transient interference as permanent faults, triggering redundant maintenance.

[0008] 3. Existing diagnostic logic does not incorporate the accelerating effect of environmental stress on hardware failure and ignores the quantitative correlation between temperature and humidity conditions and failure modes, which can easily lead to environmentally induced failures being mistakenly attributed to device defects. Summary of the Invention

[0009] In order to overcome the shortcomings of the background technology, an embodiment of the present invention provides a display terminal full life cycle fault intelligent diagnosis system, which can effectively solve the problems involved in the above background technology.

[0010] The purpose of the present invention can be achieved through the following technical solutions: a display terminal full life cycle fault intelligent diagnosis system, including: a fault information receiving module, a targeted detection execution module, a transient fault filtering module, a cause confidence output module, an environmental coupling analysis module and a diagnosis report generation module.

[0011] The fault information receiving module is connected to the targeted detection execution module, the targeted detection execution module is connected to the transient fault filtering module, the transient fault filtering module is connected to the cause confidence output module, the cause confidence output module is connected to the environmental coupling analysis module, and the environmental coupling analysis module is connected to the diagnostic report generation module.

[0012] The fault information receiving module receives a description of a fault phenomenon of the display terminal and retrieves a set of signal input terminal control objects associated with the fault phenomenon.

[0013] The targeted detection execution module performs targeted detection on each control object in the set to generate an original detection data set.

[0014] The instantaneous fault filtering module removes instantaneous abnormal points from the original detection data set based on density clustering and time series fluctuation analysis, and obtains a valid fault feature set that carries potential fault identification.

[0015] The cause confidence output module integrates the effective fault feature set and the historical fault contribution weights to output the fault cause confidence of each controlled object.

[0016] The environmental coupling analysis module extracts the ambient temperature and humidity data during the period when the fault phenomenon occurs, determines the correlation strength between the fault mode of each control object and the environmental factors through preset environmental association rules, and generates environmental sensitivity assessment labels.

[0017] The diagnosis report generation module integrates the environmental sensitivity assessment label with the fault cause confidence ranking result and outputs a fault location diagnosis report.

[0018] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention retrieves a set of signal input control objects associated with the display terminal fault phenomenon, conducts targeted detection on each control object in the set, and completes the initial screening of the status with the three domain parameters from power supply, clock signal to data bus, thereby avoiding the lack of granularity of fault location caused by generalized detection of the entire machine and focusing on the core hub of the display terminal video link.

[0019] (2) The present invention removes instantaneous anomalies from the original detection data set based on density clustering and time series fluctuation analysis, actively filters transient noise and interference from occasional events, ensures state reliability from the source of the data, significantly reduces the false alarm rate, and thus improves the accuracy of fault judgment and the credibility of the diagnosis report.

[0020] (3) The present invention integrates the effective fault feature set and the dynamically attenuated historical fault contribution weights to scientifically output the fault cause confidence of each control object, realize self-optimization of diagnostic logic, and thus improve the adaptability and accuracy of complex fault attribution.

[0021] (4) The present invention determines the correlation strength between the fault mode of each control object and the environmental factors through preset environmental association rules, generates environmental sensitivity assessment labels and binds them with the confidence level of the fault cause, reveals the coupling mechanism between the root cause of the fault and the environmental inducement, and outputs a fault location diagnosis report to provide a more comprehensive basis for fault handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0023] Figure 1 Schematic diagram of module connection of the present invention.

[0024] Figure 2 This is a logic flow chart for removing instantaneous abnormal points from an original detection data set in the instantaneous fault filtering module of the present invention.

[0025] Figure 3 A logical flow chart generated for the environmental sensitivity assessment tag in the environmental coupling analysis module of the present invention. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0027] Reference Figure 1 As shown, the present invention provides a display terminal full life cycle fault intelligent diagnosis system, including: a fault information receiving module, a targeted detection execution module, a transient fault filtering module, a cause confidence output module, an environmental coupling analysis module and a diagnosis report generation module.

[0028] The fault information receiving module is connected to the targeted detection execution module, the targeted detection execution module is connected to the transient fault filtering module, the transient fault filtering module is connected to the cause confidence output module, the cause confidence output module is connected to the environmental coupling analysis module, and the environmental coupling analysis module is connected to the diagnostic report generation module.

[0029] The fault information receiving module receives a description of a fault phenomenon of the display terminal and retrieves a set of signal input terminal control objects associated with the fault phenomenon.

[0030] It should be noted that the above-mentioned signal input terminal control object refers to a detectable hardware unit in the display terminal signal input link, including but not limited to an IC chip, a resistor and capacitor group, an interface module, etc.

[0031] In a preferred embodiment of the present invention, retrieving a set of signal input terminal control objects associated with the fault phenomenon includes: semantically parsing the fault phenomenon description to decompose it into fault behavior characteristic words and controlled unit identifiers.

[0032] It should be noted that the above semantic parsing process is: pre-loading the fault behavior vocabulary and the controlled unit vocabulary, performing part-of-speech tagging on the fault phenomenon description text, extracting verb phrases and object noun phrases, matching and verifying the marked verb phrases with the fault behavior vocabulary, outputting standardized fault behavior feature words, mapping and verifying the marked object noun phrases with the controlled unit vocabulary, and outputting standardized controlled unit identifiers.

[0033] Based on the grammatical dependency structure of the description content, the logical association between the fault behavior characteristic words and the controlled unit identifier is verified, and only the pairs that meet the following conditions are retained: the fault behavior characteristic words directly modify the controlled unit identifier in the grammatical structure, or the controlled unit identifier serves as the direct object of the fault behavior characteristic words.

[0034] The fault behavior vocabulary specifically includes display terminal-specific fault verbs, which can be generated by parsing historical maintenance manuals and work order statistics. The controlled unit vocabulary stores standardized names of hardware objects, which are derived from equipment BOM tables and circuit diagram annotations.

[0035] Based on a preset fault code library, a standard fault code that matches the combination of the fault behavior characteristic word and the controlled unit identifier is retrieved.

[0036] The dynamically updated mapping relationship table is accessed to extract candidate signal input terminal control objects mapped to the standard fault code corresponding to the fault phenomenon to generate a set.

[0037] The mapping relationship table is constructed based on a full life cycle fault case library of the display terminal, and records the corresponding relationship between the standard fault code and the signal input terminal control object associated with it.

[0038] It should be noted that the above-mentioned display terminal full life cycle failure case library is model-specific. It includes failure cases of a specific model of display terminal from commissioning to scrapping. When a new failure case is added, the trigger mapping relationship table is updated according to the following rules: Based on the signal flow tracing analysis of the newly added case, the signal input control object where the root cause of the fault is located is extracted. If the object does not exist in the original fault code mapping set, it is added to the associated object list of the corresponding fault code.

[0039] For existing control objects, adjust their fault contribution weights based on the verification results of new cases.

[0040] The targeted detection execution module performs targeted detection on each control object in the set to generate an original detection data set.

[0041] In a preferred embodiment of the present invention, the targeted detection execution process includes: performing multiple rounds of hierarchical detection on each control object in the set, and a single round of progressive detection from power supply, clock signal to data bus according to the hardware level.

[0042] The power supply response parameters, timing reference parameters and data integrity parameters of the control object are synchronously collected in each round.

[0043] Generate a raw detection dataset containing object identification, detection rounds, and layer parameters.

[0044] It should be noted that the basis for the above-mentioned single-round progressive detection from power supply, clock signal to data bus at the hardware level is: since all control objects at the signal input end rely on power supply to complete signal level analysis and rely on clock signal to achieve timing synchronization, these two constitute the common living environment of electronic circuits, and power supply or clock abnormalities will appear before data errors, so it is designed as a priority detection layer.

[0045] Data bus detection is based on the principle of physical layer signal integrity. Through protocol-independent indicators such as eye diagrams and bit error rates, it adapts to control objects of different interface standards and can form differentiated fault location capabilities.

[0046] The power supply response parameters include but are not limited to input pin voltage ripple coefficient, current transient response establishment time, etc.

[0047] The timing reference parameters include but are not limited to reference clock period jitter, clock-data skew, etc.

[0048] The data integrity parameters include but are not limited to bus eye opening, bit error rate, protocol packet CRC check failure count, etc.

[0049] The embodiment of the present invention retrieves a set of signal input control objects associated with the display terminal fault phenomenon, conducts targeted detection on each control object in the set, and completes initial status screening with three-domain parameters from power supply, clock signal to data bus, avoiding insufficient fault location granularity caused by generalized detection of the entire machine, and focusing on the core hub of the display terminal video link.

[0050] The instantaneous fault filtering module removes instantaneous abnormal points from the original detection data set based on density clustering and time series fluctuation analysis to obtain a valid fault feature set carrying potential fault identifiers.

[0051] Reference Figure 2 As shown, in a preferred embodiment of the present invention, the instantaneous outlier removal of the original detection data set includes: for the same type of parameters of the same control object, generating a parameter-specific time series according to the detection round.

[0052] The sliding analysis window of the time series is divided along the detection round direction, and the fluctuation dispersion index of the parameter detection value within the sliding analysis window is quantified. If the fluctuation dispersion index of a sliding analysis window exceeds the preset allowable threshold of the historical steady state, all parameter detection values ​​covered by the sliding analysis window are marked as candidate anomalies.

[0053] It should be noted that the fluctuation dispersion index of the parameter detection value in the above-mentioned sliding analysis window specifically refers to the standard deviation value. The historical steady-state preset permission threshold is mainly constructed based on the parameter fluctuation characteristics of the same model display terminal in the historical health state. The parameter standard deviation value in the historical health state can be calculated, and its sum with the preset permission error is used as the historical steady-state preset permission threshold.

[0054] The density clustering algorithm is used to identify clusters of candidate outliers that are continuously distributed on the time axis and to eliminate isolated candidate outliers.

[0055] It should be noted that the specific implementation process of the above-mentioned density clustering algorithm is as follows: each candidate outlier point is mapped to the time axis coordinate, a one-dimensional spatial point set is constructed, the coordinate spacing of each spatial point on the time axis is calculated, the median is taken as the clustering radius, the number of spatial points contained in the neighborhood constructed by the clustering radius of each spatial point is retrieved, and the spatial point whose neighborhood contains more spatial points than the preset number of points is taken as the core point. Starting from any core point, all points in its neighborhood are included in the current cluster. If the newly added point is a core point, its neighborhood points are merged into the cluster. When it is impossible to continue expanding, the generated continuous cluster is the outlier cluster, and the points not included in any cluster are determined to be isolated candidate outlier points.

[0056] Traverse the exclusive timing sequences of various types of parameters of the same control object, perform multi-domain collaborative verification on the retained abnormal clusters, and assign a potential fault mark to the control object when abnormal cluster conditions exist in at least two of the three dimensions of power supply response, timing benchmark, and data integrity.

[0057] Output the verified valid fault feature set, including the control object carrying the potential fault identification, the abnormal cluster data corresponding to each abnormal detection parameter that triggers the identification, and the start and end detection rounds.

[0058] The embodiment of the present invention removes instantaneous anomalies from the original detection data set based on density clustering and time series fluctuation analysis, actively filters transient noise and interference from occasional events, ensures status reliability from the source of the data, significantly reduces the false alarm rate, and thus improves the accuracy of fault judgment and the credibility of the diagnosis report.

[0059] The cause confidence output module integrates the effective fault feature set and the historical fault contribution weights to output the fault cause confidence of each controlled object.

[0060] In a preferred embodiment of the present invention, the process of obtaining the confidence level of the fault cause of each controlled object includes: setting the confidence level of the fault cause of the controlled object that does not carry a potential fault identifier as a reference lower limit value.

[0061] Record the control object with potential fault identification as the target object and perform the following steps: (a) Based on the abnormal cluster data corresponding to each anomaly detection parameter, the overall intensity characteristics of the target object are quantified, including the temporal persistence intensity, spatial correlation intensity, and fluctuation deviation intensity. The abnormal fault index is generated by fusing these three intensity characteristics.

[0062] It should be noted that the above-mentioned time persistence strength is obtained by taking the average value of the ratio of the start and end detection rounds of the abnormal cluster data corresponding to each anomaly detection parameter to the total detection rounds.

[0063] The spatial correlation strength is quantified based on the number of detection parameters that trigger anomalies and the dimensional categories involved. The ratio of the number of detection parameters that trigger anomalies to the total number of detection parameters, as well as the ratio of the number of dimensional categories involved to the total number of dimensional categories can be obtained as an example. The product of the two ratios is used as the spatial correlation strength.

[0064] The intensity of fluctuation deviation is obtained by screening the maximum value based on the fluctuation dispersion index of the abnormal cluster data corresponding to each anomaly detection parameter.

[0065] It should also be noted that the generation of the abnormal fault indicator by fusing the three strength features can be exemplarily obtained by accumulating the three strength features.

[0066] (b) Based on the hardware topology of the signal input end, if the target object's direct upstream control object carries a potential fault flag, the target object is assigned a cascading fault factor; otherwise, it is assigned a local fault factor. The confidence improvement of the local fault factor representation is higher than that of the cascading fault factor.

[0067] (c) Obtaining a basic fault source confidence level of the target object by applying the fault factor to the abnormal fault indicator.

[0068] It should be noted that the basic fault source confidence of the above-mentioned target object can be obtained by multiplying the corresponding fault factor and the abnormal fault index, where the cascade fault factor is less than 1, the local fault factor is greater than 1, and the excess value of both is within 0.3.

[0069] (d) Determine the historical fault contribution weight of the target object and add it to the basic fault source confidence to obtain the fault cause confidence of the target object.

[0070] In a preferred embodiment of the present invention, the process of determining the historical fault contribution weight of the target object includes: tracing back the historical fault records caused by the target object in the full life cycle fault case library of the display terminal, which contain the fault behavior characteristic words and controlled unit identifiers, and calculating the case frequency ratio of the target object as the root fault source.

[0071] Based on the duration of the most recent historical failure, a time decay factor is introduced to correct the case frequency ratio to generate the historical failure contribution weight of the target object.

[0072] The embodiment of the present invention integrates the effective fault feature set and the dynamically attenuated historical fault contribution weights to scientifically output the fault cause confidence of each controlled object, realize self-optimization of diagnostic logic, and thus improve the adaptability and accuracy of complex fault attribution.

[0073] The environmental coupling analysis module extracts the ambient temperature and humidity data during the period when the fault phenomenon occurs, determines the correlation strength between the fault mode of each control object and the environmental factors through preset environmental association rules, and generates an environmental sensitivity assessment label.

[0074] In a preferred embodiment of the present invention, the preset environment association rules include the following content: displaying the vulnerable temperature and humidity ranges of each control object at the terminal signal input end and the directional abnormal parameter set caused by the ranges.

[0075] Reference Figure 3 As shown, in a preferred embodiment of the present invention, the determination of the correlation strength between the failure mode of each control object and the environmental factors and the generation of corresponding environmental sensitivity assessment labels include: marking the control objects that do not carry potential fault identification with a no-environmental-relevance label.

[0076] For control objects with potential fault identification, perform environmental matching analysis: i. Compare the ambient temperature and humidity data during the period when the fault occurs to see if they fall within the target object's vulnerable temperature and humidity range.

[0077] ii. If not, the target object failure mode and environmental factors are marked as having no environmental relevance.

[0078] iii. If it falls into the category, then compare the target object's anomaly detection parameters to see if they belong to its directional anomaly parameter set, and quantify the target object's directional anomaly compliance and the number of continuous compliance rounds; It should be noted that the above-mentioned target object directional anomaly compliance specifically refers to the proportion of the target object anomaly detection parameters in the directional anomaly parameter set, and the continuous compliance round specifically refers to the minimum round of continuous compliance with the directional anomaly changes in the anomaly cluster data corresponding to the anomaly detection parameters.

[0079] The directional anomaly compliance greater than the preset compliance threshold and the continuous compliance rounds greater than the preset round threshold are both used as conditional items. If both conditional items are met, the target object failure mode and environmental factors are marked as high environmental sensitivity labels. If only one of the conditional items is met, the medium environmental sensitivity label is marked. If both are not met, the low environmental sensitivity label is marked.

[0080] It should be noted that the setting of the above-mentioned preset compliance threshold and preset round threshold is based on the statistical distribution characteristics of the display terminal full life cycle failure case library and is determined through the following technical paths: Extract all cases confirmed as environmentally induced faults from the case library, count the distribution of directional anomaly compliance of such cases, and select the inflection point as the preset compliance threshold.

[0081] The ratio of the minimum fault duration of all cases confirmed as environment-induced faults in the case library to the preset fixed duration of a single round is used as the preset round threshold.

[0082] In a preferred embodiment of the present invention, the comparing of each abnormality detection parameter of the target object to determine whether it belongs to its directional abnormality parameter set includes: performing parameter attribution verification on the abnormality detection parameter to confirm whether it belongs to the physical quantity type defined by the directional abnormality parameter set.

[0083] Based on the anomaly cluster data of the anomaly detection parameter, the actual change direction of the anomaly detection parameter relative to the historical steady-state arithmetic mean is calculated.

[0084] It should be noted that the above-mentioned historical steady-state arithmetic mean is based on the corresponding parameter monitoring collection of the same model display terminal in the historical health state, and is obtained by taking the arithmetic mean of all parameter monitoring values ​​in the collection. The actual change direction of the abnormal detection parameter relative to the historical steady-state arithmetic mean depends on the size comparison relationship between the abnormal detection parameter and the historical steady-state arithmetic mean. If the abnormal detection parameter is greater than the historical steady-state arithmetic mean, it indicates that the actual change direction is a positive change; if it is less than, it indicates a negative change; if it is equal, it indicates no change.

[0085] The expected change direction of each directional anomaly parameter of the target object within the vulnerable temperature and humidity range is extracted. When the actual change direction is consistent with the expected change direction, it is determined that the anomaly detection parameter passes the direction consistency verification.

[0086] Only when an anomaly detection parameter satisfies both the attribution verification and the direction matching verification, is it determined that the parameter belongs to the directional anomaly parameter set.

[0087] The diagnosis report generation module integrates the environmental sensitivity assessment label with the fault cause confidence ranking result and outputs a fault location diagnosis report.

[0088] In a preferred embodiment of the present invention, the integration process of the environmental sensitivity assessment label and the fault cause confidence ranking result includes: converting the environmental sensitivity assessment label into a grade value, and establishing a Cartesian coordinate system with the fault cause confidence as the horizontal axis and the environmental sensitivity assessment label as the vertical axis.

[0089] It should be noted that the above-mentioned conversion of environmental sensitivity assessment labels into numerical levels specifically refers to: converting labels with no environmental relevance, labels with low environmental sensitivity, labels with medium environmental sensitivity, and labels with high environmental sensitivity into 0, 1, 2, and 3 respectively.

[0090] A coordinate system partitioning rule is defined to standardize a high confidence-high sensitivity area and a medium confidence-medium sensitivity area in the coordinate system.

[0091] It should also be noted that the above coordinate system partitioning rules include the following: areas where the fault cause confidence is greater than a first preset confidence and the environmental sensitivity assessment label level value is 3 are regarded as high confidence-high sensitivity areas.

[0092] The area where the fault cause confidence is in the closed interval between the second preset confidence and the first preset confidence and the environmental sensitivity assessment label level value is 2 is regarded as the medium confidence-medium sensitivity area.

[0093] The control objects located in the high confidence-high sensitivity area are marked as the first-level fault sources, and corresponding core fault location entries are generated in the report.

[0094] For control objects located in the medium confidence-medium sensitivity zone, secondary fault sources are marked and corresponding auxiliary fault analysis entries are generated in the report.

[0095] The remaining control objects are marked with the fault sources to be verified, and a list of detection parameters is generated to recommend incremental detection rounds.

[0096] The embodiment of the present invention determines the correlation strength between the fault mode of each control object and the environmental factors through preset environmental association rules, generates an environmental sensitivity assessment label and binds it to the confidence level of the fault cause, reveals the coupling mechanism between the root cause of the fault and the environmental inducement, and outputs a fault location diagnosis report to provide a more comprehensive basis for fault handling.

[0097] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. A display terminal full life cycle fault intelligent diagnosis system, characterized by: include: A fault information receiving module receives a description of a fault phenomenon from a display terminal and retrieves a set of signal input terminal control objects associated with the fault phenomenon; a targeted detection execution module, performing targeted detection on each control object in the set to generate an original detection data set; The instantaneous fault filtering module removes instantaneous abnormal points from the original detection data set based on density clustering and time series fluctuation analysis to obtain a valid fault feature set that carries potential fault identification; The cause confidence output module integrates the effective fault feature set and the historical fault contribution weight to output the fault cause confidence of each controlled object; The environmental coupling analysis module extracts the ambient temperature and humidity data during the period when the fault occurs, determines the correlation strength between the fault mode of each control object and the environmental factors through preset environmental association rules, and generates an environmental sensitivity assessment label; The diagnosis report generation module integrates the environmental sensitivity assessment label with the fault cause confidence ranking result and outputs a fault location diagnosis report.

2. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 1, characterized in that: Retrieving a signal input terminal control object set associated with the fault phenomenon, including: Perform semantic analysis on the fault phenomenon description and break it down into fault behavior characteristic words and controlled unit identifiers; Retrieving a standard fault code that matches the combination of the fault behavior characteristic word and the controlled unit identifier based on a preset fault code library; Accessing the dynamically updated mapping relationship table, extracting candidate signal input terminal control objects mapped to the standard fault code corresponding to the fault phenomenon, to generate a set; The mapping relationship table is constructed based on a full life cycle fault case library of the display terminal, and records the corresponding relationship between the standard fault code and the signal input terminal control object associated with it.

3. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 1 is characterized by: The targeted detection execution process includes: Perform multiple rounds of hierarchical testing on each control object in the set, with each round progressively testing from power supply, clock signal to data bus according to the hardware level; Each round synchronously collects the power supply response parameters, timing reference parameters and data integrity parameters of the control object; Generate a raw detection dataset containing object identification, detection rounds, and layer parameters.

4. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 3, characterized in that: The instantaneous outlier removal from the original detection data set includes: For the same type of parameters of the same control object, a parameter-specific time series is generated according to the detection round; Divide the sliding analysis window of the time series along the detection round direction, quantify the fluctuation dispersion index of the parameter detection value within the sliding analysis window, and if the fluctuation dispersion index of a sliding analysis window exceeds the preset allowable threshold of the historical steady state, mark all parameter detection values ​​covered by the sliding analysis window as candidate anomaly points; A density clustering algorithm is used to identify clusters of candidate anomalies that are continuously distributed on the time axis and to eliminate isolated candidate anomalies. Traverse the exclusive timing sequences of various types of parameters of the same control object, perform multi-domain collaborative verification on the retained abnormal clusters, and assign a potential fault flag to the control object when abnormal cluster conditions exist in at least two of the three dimensions of power supply response, timing benchmark, and data integrity; Output the verified valid fault feature set, including the control object carrying the potential fault identification, the abnormal cluster data corresponding to each abnormal detection parameter that triggers the identification, and the start and end detection rounds.

5. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 4, characterized in that: The process of obtaining the confidence level of the fault cause of each control object includes: The confidence level of the fault cause of the control object that does not carry a potential fault identification is set as the reference lower limit value; Record the control object with potential fault identification as the target object and perform the following steps: (a) Based on the abnormal cluster data corresponding to each anomaly detection parameter, the overall intensity characteristics of the target object are quantified, including the temporal persistence intensity, spatial correlation intensity, and fluctuation deviation intensity. The abnormal fault index is generated by fusing these three intensity characteristics; (b) Based on the hardware topology of the signal input end, if the target object's direct upstream control object carries a potential fault flag, the target object is assigned a cascading fault factor; otherwise, it is assigned a local fault factor. The confidence level of the local fault factor is higher than that of the cascading fault factor. (c) obtaining a basic fault source confidence of the target object by applying the fault factor to the abnormal fault indicator; (d) Determine the historical fault contribution weight of the target object and add it to the basic fault source confidence to obtain the fault cause confidence of the target object.

6. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 5, characterized in that: The process of determining the historical fault contribution weight of the target object includes: Tracing back the historical fault records of the target object in the display terminal life cycle fault case library, which contain the fault behavior characteristic words and the controlled unit identifier, and calculating the frequency ratio of the case in which the target object is the root fault source; Based on the duration of the most recent historical failure, a time decay factor is introduced to correct the case frequency ratio to generate the historical failure contribution weight of the target object.

7. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 4, characterized in that: The preset environment association rules include the following: Displays the vulnerable temperature and humidity ranges of each control object at the terminal signal input end and the directional abnormal parameter set caused by the range.

8. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 7, characterized in that: Determining the correlation strength between the failure mode of each control object and the environmental factors and generating the corresponding environmental sensitivity assessment label includes: Marking control objects that do not carry potential fault identification with tags that have no environmental relevance; For control objects with potential fault indicators, perform environmental matching analysis: i. Compare the ambient temperature and humidity data during the fault period to see if they fall within the target object's vulnerable temperature and humidity range; ii. If not, the target object failure mode and environmental factors are marked as having no environmental relevance label; iii. If it falls into the category, then compare the target object's anomaly detection parameters to see if they belong to its directional anomaly parameter set, and quantify the target object's directional anomaly compliance and the number of continuous compliance rounds; The directional anomaly compliance greater than the preset compliance threshold and the continuous compliance rounds greater than the preset round threshold are both used as conditional items. If both conditional items are met, the target object failure mode and environmental factors are marked as high environmental sensitivity labels. If only one of the conditional items is met, the medium environmental sensitivity label is marked. If both are not met, the low environmental sensitivity label is marked.

9. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 8, characterized in that: The comparing of each anomaly detection parameter of the target object to determine whether it belongs to its directional anomaly parameter set includes: Perform parameter attribution verification on the anomaly detection parameter to confirm whether it belongs to the physical quantity type defined by the directional anomaly parameter set; Based on the anomaly cluster data of the anomaly detection parameter, calculate the actual change direction of the anomaly detection parameter relative to the historical steady-state arithmetic mean; Extracting the expected change direction of each directional anomaly parameter of the target object within the vulnerable temperature and humidity range. When the actual change direction is consistent with the expected change direction, it is determined that the anomaly detection parameter has passed the directional consistency verification; Only when an anomaly detection parameter satisfies both the attribution verification and the direction matching verification, is it determined that the parameter belongs to the directional anomaly parameter set.

10. The intelligent fault diagnosis system for display terminals throughout their life cycle according to claim 8, characterized in that: The integration process of the environmental sensitivity assessment label and the fault cause confidence ranking result includes: The environmental sensitivity assessment label is converted into a grade value, and a Cartesian coordinate system is established with the fault cause confidence as the horizontal axis and the environmental sensitivity assessment label as the vertical axis; Defining coordinate system partitioning rules to standardize high confidence-high sensitivity areas and medium confidence-medium sensitivity areas in the coordinate system; Mark the first-level fault source for control objects located in the high-confidence-high-sensitivity area, and generate corresponding core fault location items in the report; Mark the secondary fault source for the control object located in the medium confidence-medium sensitivity zone, and generate corresponding auxiliary fault analysis items in the report; The remaining control objects are marked with the fault sources to be verified, and a list of detection parameters is generated to recommend incremental detection rounds.

Images