An electronic component fault monitoring module and method

Abstract

The present invention relates to the technical field of electrical variable measurement, and specifically relates to an electronic component fault monitoring module and method. The method determines the initial difference degree of each electronic component in each full life cycle stage, and obtains the periodic anomaly correlation based on the volatility of the initial difference degree in different full life cycle stages. Further, the monitoring anomaly degree of each electronic component in each full life cycle stage is obtained. Based on the distribution results in the clustering clusters, the stage sensitivity coefficient of each category in each full life cycle stage can be determined. According to the difference in the overall sensitivity coefficient between adjacent full life cycle stages and the anomaly of the periodic anomaly correlation of the electronic components, the monitoring consistency of each full life cycle stage is obtained and it is judged whether adjustment is required in this stage, and then the result of the faulty electronic component is obtained. The present invention ensures the accuracy of the fault detection of electronic components through the early warning of the full life cycle stage during the monitoring process.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical variable measurement, and particularly relates to an electronic component fault monitoring module and method. Background Art

[0002] Electronic components are the core components of various electronic devices, and their operating status directly affects the performance and lifespan of the devices. When developing the full-cycle intelligent testing and maintenance technology for electronic components, it is necessary to focus on the status monitoring, function testing, performance optimization, and fault repair of the entire life cycle of the components from design, production, operation to maintenance.

[0003] The full-life monitoring cycle of electronic components includes four stages: Design verification stage: At the component design stage, verify whether its electrical characteristics, thermal performance, reliability, etc. meet the design requirements. Production testing stage: Conduct function testing, screening testing, and consistency testing on the components during the production process to ensure the quality of the products leaving the factory. Operation monitoring stage: During the actual operation of the components, monitor their working status in real time, detect abnormalities and make diagnoses. Maintenance and repair stage: Conduct maintenance and optimization through historical data and prediction models, and perform fault repair or replacement when necessary.

[0004] During the process of monitoring the full-life cycle stage of electronic components, there is a certain forward and backward correlation in the fault manifestations of electronic components. The fault conditions are affected by monitoring methods or environmental factors in different full-life cycle stages. If the monitoring results of various types of electronic components in the full-life cycle stage are significantly unified, it indicates that the monitoring process in this full-life cycle stage is unreasonable, and it is impossible to distinguish the monitoring results of different types of electronic components, resulting in the inability to timely determine the faulty components in each full-life cycle stage. Summary of the Invention

[0005] In order to solve the technical problem in the prior art that due to unreasonable monitoring settings in a certain full-life cycle stage, the distinguishability of the monitoring results of different electronic components is not high, and thus the faulty components cannot be determined in a timely manner, the purpose of the present invention is to provide an electronic component fault monitoring module and method, and the specific technical solutions adopted are as follows:

[0006] The present invention proposes an electronic component fault monitoring method, and the method includes:

[0007] Obtain the monitoring data of electronic components in each full-life cycle stage in the monitoring scenario; the electronic components include multiple categories;

[0008] For each full life cycle stage, the monitoring data deviation between each electronic component and similar electronic components in the full life cycle stage is used as the initial difference of each electronic component; according to the volatility of the initial difference of the electronic components at different full life cycle stages, the periodic anomaly correlation is obtained; according to the periodic anomaly correlation, the initial difference is adjusted to obtain the monitoring anomaly of each electronic component at each full life cycle stage;

[0009] The time series of the whole life cycle stage is used as the horizontal axis, and the monitoring abnormality is used as the vertical axis to obtain the sample space, and all electronic components are mapped into the sample space to obtain the clustering results of the data points. According to the distribution concentration of each category of electronic components in the cluster cluster at each whole life cycle stage, the stage sensitivity coefficient of each category at each whole life cycle stage is obtained;

[0010] Based on the stage sensitivity coefficients, the overall sensitivity coefficients of the electronic components in each full life cycle stage are obtained; for each full life cycle stage, based on the differences in the overall sensitivity coefficients in adjacent full life cycle stages and the abnormalities of the periodic abnormality correlations of the electronic components, the monitoring consistency of each full life cycle stage is obtained; based on the monitoring consistency, whether the full life cycle stage needs to be adjusted is judged; based on the adjusted full life cycle stage, whether there are faulty electronic components is judged.

[0011] Furthermore, the method for obtaining the initial difference degree includes:

[0012] Obtain the average monitoring data of each electronic component in a full life cycle stage;

[0013] For any electronic component, obtain the mean difference of monitoring data between the electronic component and each other electronic component of the same type; and use the average value of the mean difference of the monitoring data as the initial difference degree of the electronic component in a full life cycle stage.

[0014] Furthermore, the method for obtaining the period anomaly correlation includes:

[0015] For each electronic component, the initial difference sequences in all life cycle stages are arranged into an initial difference sequence; a differential sequence of the initial difference sequence is obtained; and the ratio of the element mean in the initial difference sequence to the element mean in the differential sequence is normalized to obtain the periodic anomaly correlation.

[0016] Furthermore, the method for acquiring the monitoring abnormality degree includes:

[0017] Multiply the cycle anomaly correlation by the initial difference degree to obtain the monitoring anomaly degree of each electronic component in each full life cycle stage.

[0018] Further, in the sample space, use the density clustering algorithm to obtain each clustering cluster in the clustering result.

[0019] Further, the method for obtaining the stage sensitivity coefficient includes:

[0020] Optionally select an electronic component of a certain category as the target category and a full life cycle stage as the target stage;

[0021] In each clustering cluster, obtain the sample point ratio of the sample points of the electronic components of the target category in the target stage; count the standard deviation of the sample point ratios of the target category in all clustering clusters in the target stage;

[0022] In each clustering cluster, obtain the maximum difference distance between the sample points of the electronic components of the target category; obtain the average maximum difference distance of the target category in all clustering clusters;

[0023] Normalize the ratio of the average maximum difference distance to the standard deviation to obtain the stage sensitivity coefficient of the target category in the target stage.

[0024] Further, the method for obtaining the overall sensitivity coefficient includes:

[0025] Take the average value of the stage sensitivity coefficients of all categories in each full life cycle stage as the overall sensitivity coefficient.

[0026] Further, the method for obtaining the monitoring consistency includes:

[0027] For any full life cycle stage, obtain the difference in the overall sensitivity coefficient between the full life cycle stage and the previous full life cycle stage; obtain the average cycle anomaly correlation of all electronic components in each category, and obtain the data deviation of all average cycle anomaly correlations relative to the average value; normalize the ratio of the overall sensitivity coefficient difference to the data deviation to obtain the monitoring consistency.

[0028] Further, the judgment of whether the full life cycle stage needs to be adjusted according to the monitoring consistency includes:

[0029] If the monitoring consistency is greater than the preset consistency threshold, it is determined that the full life cycle stage needs to be adjusted.

[0030] The present invention also proposes an electronic component fault monitoring module, and the module includes:

[0031] A monitoring data acquisition module, configured to obtain the monitoring data of electronic components in each full life cycle stage of a monitoring scenario; the electronic components include multiple categories;

[0032] A monitoring abnormality degree acquisition module, for a full life cycle stage, taking the monitoring data deviation between each electronic component and its same - type electronic components within the full life cycle stage as the initial difference degree of each electronic component; obtaining the periodic abnormality correlation according to the volatility of the initial difference degrees of the electronic components in different full life cycle stages; adjusting the initial difference degree according to the periodic abnormality correlation to obtain the monitoring abnormality degree of each electronic component in each full life cycle stage;

[0033] A stage sensitivity coefficient analysis module, configured to use the time sequence of the full life cycle stage as the abscissa and the monitoring abnormality degree as the ordinate to obtain a sample space, map all electronic components into the sample space and obtain the clustering result of data points, and obtain the stage sensitivity coefficient of each category in each full life cycle stage according to the distribution concentration of the electronic components of each category in the clustering clusters in each full life cycle stage;

[0034] A full life cycle stage judgment module, configured to obtain the overall sensitivity coefficient of the electronic components in each full life cycle stage according to the stage sensitivity coefficient; for each full life cycle stage, obtain the monitoring consistency of each full life cycle stage according to the difference in the overall sensitivity coefficients in adjacent full life cycle stages and the abnormality of the monitoring abnormality degree in the full life cycle stage; judge whether the full life cycle stage needs to be adjusted according to the monitoring consistency.

[0035] The present invention has the following beneficial effects:

[0036] The present invention first determines the initial difference degree of each electronic component in each full life cycle stage. By using the volatility of the initial difference degree in different full life cycle stages, it can characterize the degree of abnormal state representation of the electronic component in different full life cycle stages. The greater the obtained cycle anomaly correlation, the more abnormal the subsequent stages are after an anomaly occurs, so a larger anomaly quantification result should be obtained. Therefore, the monitoring anomaly degree of each electronic component in each full life cycle stage can be further obtained. Based on the monitoring anomaly degree, all data can be classified, and based on the distribution results in the clustering clusters, the stage sensitivity coefficient of each category in each full life cycle stage can be determined. The stage sensitivity coefficient can reflect the monitoring and discrimination of the current full life cycle stage for the electronic components of each category. If there is a unified cycle anomaly correlation among the electronic components of different categories, and there is a large difference in the overall sensitivity coefficient between adjacent full life cycle stages, it indicates that there are significant problems in the current full life cycle stage and targeted adjustments are needed. Therefore, an adjustment signal can be fed back, and the effective monitoring of the electronic components can be achieved by adjusting the parameters of this full life cycle stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0038] Figure 1 It is a flowchart of a method for monitoring faults of electronic components provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, describe in detail the specific implementation manners, structures, features, and effects of a module and method for monitoring faults of electronic components proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0041] The following will specifically describe the specific solutions of a module and method for monitoring faults of electronic components provided by the present invention with reference to the accompanying drawings.

[0042] Please refer to Figure 1 , which shows a flowchart of a method for monitoring electronic component failures provided by an embodiment of the present invention. The method includes:

[0043] Step S1: Obtain the monitoring data of electronic components in each full life cycle stage of the monitoring scenario; the electronic components include multiple categories.

[0044] Common electronic components on a circuit board include surface mount capacitors, overload protectors, resistors, diodes, transistors, inductors, and integrated circuits, etc. They jointly undertake key functions such as filtering, overload protection, signal regulation, rectification, amplification, energy storage, and logic control. These components may experience performance degradation or failure due to environmental changes, long-term use, or electrical stress during actual operation. As the core component for filtering and energy storage, common problems of surface mount capacitors include capacitance attenuation, short circuit or open circuit, and increased leakage current caused by high temperature; the overload protector protects the circuit by limiting the current and may have problems such as premature triggering, failure to trigger in time, or drift of thermal characteristics; resistors may experience resistance drift, overheating and burning, or poor soldering due to aging, temperature change, or overload during long-term use; diodes are mainly used for unidirectional conduction and rectification, and common problems include increased leakage current, abnormal forward voltage drop, and thermal breakdown; as the core component for signal amplification and switch control, transistors may experience switch failure, gain reduction, or overheating failure; inductors are used for energy storage and filtering and may experience performance degradation due to coil breakage, inductance value drift, or core damage; integrated circuits (ICs), as the core of complex logic and control functions, common faults include internal circuit failure, overheating damage, and poor pin soldering, etc.

[0045] In order to ensure the safety of various electronic components on the circuit board, it is necessary to monitor the data of different stages of the full life cycle of various electronic components, evaluate the status of electronic components by monitoring key parameters, and conduct quality management. The main monitoring contents may include: the capacitance and leakage current of surface mount capacitors, the current characteristics and temperature response of overload protectors, the resistance change and power load of resistors, the forward voltage drop and reverse leakage current of diodes, the current gain and switch performance of transistors, the inductance value stability and core status of inductors, and the power consumption, logic function, and thermal management of integrated circuits. By combining intelligent diagnostic algorithms and fault prediction technologies, the operating status of electronic components can be deeply analyzed, their operating parameters can be optimized, potential problems can be warned in time, the risk of equipment failure can be reduced, and the overall reliability and safety of the system can be improved.

[0046] As described in the background art, the full life cycle stage has multiple phases, and there is a certain correlation between the fault manifestations of electronic components at different full life cycle stages. If the fault correlation is strong, there will be no significant difference in monitoring sensitivity between adjacent full life cycle stages, and consistent monitoring characteristics will exist in both stages. However, if the monitoring data shows a difference in monitoring sensitivity between adjacent full life cycle stages at this time, it indicates that the test settings for a certain full life cycle stage are unreasonable and need to be adjusted to determine the correct test settings to avoid missed detection of electronic components.

[0047] Therefore, in the embodiments of the present invention, first, the monitoring data of electronic components at each full life cycle stage in the monitoring scenario is obtained, and further analysis is performed on different types of electronic components and different full life cycle stages in the subsequent steps.

[0048] It should be noted that in the embodiments of the present invention, the parameter types monitored by different electronic components are different, so the value ranges and dimensions of the obtained monitoring data are different. To facilitate subsequent data processing, after obtaining the monitoring data, it needs to be further normalized to eliminate the influence of dimensions and value ranges. The monitoring system of the embodiments of the present invention includes:

[0049] (1) Sensor module: including a temperature sensor for real-time monitoring of the working temperature of components; an electrical sensor for monitoring electrical parameters such as voltage, current, and power; a vibration and stress sensor for detecting the impact of mechanical stress or environmental vibration on components; and an environmental sensor for detecting the potential impact of external environmental factors such as humidity and air pressure on components.

[0050] (2) Data acquisition and processing module: integrating a high-precision analog-to-digital converter to collect and quantify sensor data. The data preprocessing unit filters, denoises, and normalizes the collected data. An embedded AI chip or microprocessor is used to execute the electronic component fault monitoring method in the embodiments of the present invention.

[0051] (3) Communication module: supporting multiple communication protocols (such as CAN, I2C, SPI, wireless communication, etc.) to facilitate interaction with the upper computer or cloud platform. It has a real-time alarm function and sends alarm information through a buzzer, indicator light, or remote push.

[0052] (4) Power management module: adopting a low-power design, supporting external power supply or self-powered mode to ensure long-term stable operation.

[0053] (5) Storage and interface module: built-in non-volatile memory for recording the historical operation data of components, supporting offline analysis; providing standard interfaces (such as USB, UART) for easy debugging and expansion.

[0054] Step S2: For each full life cycle stage, take the monitoring data deviation between each electronic component and its peer components within the full life cycle stage as the initial difference degree of each electronic component; obtain the cycle anomaly correlation based on the volatility of the initial difference degrees of the electronic components in different full life cycle stages; adjust the initial difference degree according to the cycle anomaly correlation to obtain the monitoring anomaly degree of each electronic component in each full life cycle stage.

[0055] For the same full life cycle stage, different types of electronic components will all be subject to certain tests and monitoring. The content to be monitored in different life cycle stages is different, and the monitoring equipment and methods used are also different. For each electronic component, the smaller the deviation from its peer components in a full life cycle stage, the smaller the probability of failure of the electronic component. Therefore, first, within the full life cycle stage, obtain the detection data deviation between each electronic component and its peer components to obtain the initial difference degree of the electronic component.

[0056] Preferably, in an embodiment of the present invention, the method for obtaining the initial difference degree includes:

[0057] Obtain the mean value of the monitoring data of each electronic component in a full life cycle stage;

[0058] For any electronic component, obtain the difference in the mean value of the monitoring data between the electronic component and each other peer component of the same type; take the average of the differences in the mean value of the monitoring data as the initial difference degree of the electronic component in a full life cycle stage.

[0059] When a quality problem occurs in an electronic component, there will be an abnormal monitoring correlation in the full life cycle stage, that is, certain abnormalities will also occur in other full life cycle stages. Therefore, the cycle anomaly correlation can be obtained based on the volatility of the initial difference degrees of the electronic components in different full life cycle stages. That is, the smaller the volatility, the greater the cycle anomaly correlation.

[0060] Preferably, in an embodiment of the present invention, the method for obtaining the cycle anomaly correlation includes:

[0061] For each electronic component, arrange the sequence of the initial difference degrees in all full life cycle stages into an initial difference degree sequence; obtain the difference sequence of the initial difference degree sequence; normalize the ratio of the mean value of the elements in the initial difference degree sequence to the mean value of the elements in the difference sequence to obtain the cycle anomaly correlation.

[0062] The mean value of the elements in the initial difference degree sequence represents the fault performance of an electronic component in all life cycle stages. That is, the larger the mean value of the elements, the more unified and larger the abnormal performance of the electronic component in multiple life cycle stages. The difference sequence characterizes the change rate of the initial difference degree between consecutive life cycle stages. The smaller the difference sequence, the smaller the change fluctuation of the initial difference degree of the electronic component in different life cycle stages, and the more consistent the initial difference degree of the electronic component in different life cycle stages, indicating a greater cycle anomaly correlation.

[0063] It should be noted that the normalization method in the embodiments of the present invention uses range normalization. In other embodiments of the present invention, other normalization methods can also be selected, which will not be elaborated here.

[0064] The cycle anomaly correlation characterizes the degree of abnormal correlation of an electronic component in all life cycle stages, which reflects the characteristics of the electronic component in the entire monitoring process. Therefore, it can be used to adjust the initial difference degree to obtain the monitoring anomaly degree of each electronic component in each life cycle stage.

[0065] In the embodiments of the present invention, the cycle anomaly correlation is multiplied by the initial difference degree to obtain the monitoring anomaly degree of each electronic component in each life cycle stage.

[0066] Step S3: Use the time sequence of the life cycle stage as the abscissa and the monitoring anomaly degree as the ordinate to obtain a sample space. Map all electronic components into the sample space and obtain the clustering result of the data points. According to the distribution concentration of the electronic components of each category in the clustering cluster in each life cycle stage, obtain the stage sensitivity coefficient of each category in each life cycle stage.

[0067] The difference performance of different electronic components in different cycle stages is more affected by the setting of the life cycle stage. In an ideal state, due to their own uses and the technical difficulties involved, different types of electronic components will have different abnormal conditions extracted in different stages. For example, ordinary simple components such as resistors, capacitors, diodes, transistors, etc. have simple structures, and it is not easy to have faults in the design and testing stages. Their fault monitoring mainly focuses on the actual use feedback stage. Therefore, for the design of the monitoring stage, it should be possible to monitor the faults of different electronic components in this stage as much as possible and have distinctiveness.

[0068] Therefore, for any full life cycle stage, there should be an obvious distinction in the monitoring abnormality degree of electronic components between all categories at this stage. The greater this distinction is, the stronger the stage sensitivity coefficient is at this stage. Therefore, in the embodiments of the present invention, the data of all electronic components at all full life cycle stages are analyzed uniformly. Taking the time sequence of the full life cycle stage as the abscissa and the monitoring abnormality degree as the ordinate, a sample space is obtained. All electronic components are mapped into the sample space and the clustering result of data points is obtained. That is, there are sample points formed by the data of each electronic component in the sample space. For the clustering result, the more unified the distribution of the sample points of the electronic component samples of a certain category at a certain life cycle stage in the clustering cluster is, and the more dispersed the distribution of the sample points of the category is, the stronger the stage sensitivity is at this life cycle stage. A large number of electronic components of the same type in the same clustering cluster can be effectively distinguished. Therefore, in the embodiments of the present invention, the stage sensitivity coefficient of each category at each full life cycle stage can be obtained according to the concentration of the distribution of the electronic components of each category in the clustering cluster at each full life cycle stage.

[0069] In the embodiments of the present invention, a density clustering algorithm is used to obtain each clustering cluster in the clustering result.

[0070] Preferably, in an embodiment of the present invention, the method for obtaining the stage sensitivity coefficient includes:

[0071] Optionally select the electronic components of a certain category as the target category, and a full life cycle stage as the target stage;

[0072] In each clustering cluster, obtain the sample point ratio of the sample points of the electronic components of the target category at the target stage; count the standard deviation of the sample point ratio of the target category at the target stage in all clustering clusters. The greater this standard deviation is, the greater the volatility of the sample point ratio distribution is, and the lower the distinguishability of the target category at the target stage is.

[0073] In each clustering cluster, obtain the maximum difference distance between the sample points of the electronic components of the target category; obtain the average maximum difference distance of the target category in all clustering clusters. Because the greater the maximum difference distance is compared with the average maximum difference distance, it indicates that the electronic components of this category show obvious distinguishing features within multiple clustering clusters.

[0074] Therefore, the greater the average maximum difference distance is and the smaller the standard deviation is, the greater the distinguishability of the target category at the target stage is. Therefore, the ratio of the average maximum difference distance to the standard deviation is normalized to obtain the stage sensitivity coefficient of the target category at the target stage. In the embodiments of the present invention, the normalization method for this ratio is the hyperbolic tangent function mapping method.

[0075] Step S4: Obtain the overall sensitivity coefficient of electronic components in each full life cycle stage according to the stage sensitivity coefficient; for each full life cycle stage, obtain the monitoring consistency of each full life cycle stage according to the difference in the overall sensitivity coefficient between adjacent full life cycle stages and the abnormality of the cycle anomaly correlation of the electronic components; judge whether the full life cycle stage needs to be adjusted according to the monitoring consistency; judge whether there are faulty electronic components according to the adjusted full life cycle stage.

[0076] For a full life cycle stage, if the stage sensitivity coefficients of each category in this full life cycle stage are all relatively large, it indicates that the overall sensitivity coefficient in this full life cycle stage is relatively large.

[0077] When the fault characteristics of multiple categories in the full life cycle stage are relatively similar, it indicates that from the perspective of the monitoring stage setting, this stage has no significance in improving the monitoring accuracy, and the detection consistency is relatively high. For a certain category of electronic components, because the fault manifestations have a stage correlation tendency, that is, obvious abnormal state correlations will be reflected between adjacent stages. If there is an obvious difference in the overall sensitivity coefficient between adjacent stages at this time, it indicates that the monitoring setting in this stage is unreasonable, has a strong monitoring characteristic consistency, and needs to be adjusted for this stage. Therefore, in the embodiments of the present invention, for each full life cycle stage, the monitoring consistency of each full life cycle stage is obtained according to the difference in the overall sensitivity coefficient between adjacent full life cycle stages and the abnormality of the cycle anomaly correlation of the electronic components.

[0078] Judge whether the full life cycle stage needs to be adjusted according to the monitoring consistency. Judge whether there are faulty electronic components according to the adjusted full life cycle stage. By adjusting the monitoring settings within the stage, the accuracy of fault monitoring can be significantly improved, the reliability of the components can be enhanced, and intelligent monitoring and maintenance can be realized. On the basis of eliminating the influence of quality problems, by distinguishing between conventional abnormal manifestations and potential faults, combining big data and artificial intelligence technologies, and dynamically adjusting the monitoring frequency, not only can potential risks be accurately captured, but also the fault trend can be effectively predicted and early warning can be realized, thus avoiding sudden failures. In the design verification stage, potential design defects are discovered by optimizing the test plan, and the weak links of the components are identified by using the connection of the full life cycle data, providing data support for reliability improvement. At the same time, the Internet of Things and digital twin technologies are introduced to monitor the operating status of the components in real time, quickly respond to anomalies in combination with the predictive maintenance model, and repair or replace the components that may fail in advance, thereby reducing the probability of sudden failures and comprehensively improving the performance and operating stability of the components. It should be noted that the specific method for determining faulty electronic components is a well-known technical means for those skilled in the art and will not be elaborated here.

[0079] Preferably, in the embodiment of the present invention, the method for obtaining the monitoring consistency includes:

[0080] For any one full life cycle stage, obtain the overall sensitivity coefficient difference between the current full life cycle stage and the previous full life cycle stage. Obtain the average cycle anomaly correlation of all electronic components in each category, and obtain the data deviation of all average cycle anomaly correlations relative to the average value. The smaller the data deviation, the more unified the cycle anomaly correlations among all categories, and the greater the monitoring consistency. If the cycle anomaly correlations are relatively unified and there is a large overall sensitivity coefficient difference, it indicates that the monitoring consistency of this life cycle stage is greater. Therefore, normalize the ratio of the overall sensitivity coefficient difference to the data deviation to obtain the monitoring consistency.

[0081] In an embodiment of the present invention, the method for obtaining the data deviation is: obtain the average value of all average cycle anomaly correlations, take the absolute value of the difference between the average cycle anomaly correlation of each category and the average value as the initial data deviation, and take the average value of the initial data deviations of all categories as the data deviation.

[0082] In the embodiment of the present invention, if the monitoring consistency is greater than the preset consistency threshold, it is determined that the full life cycle stage needs to be adjusted. Since the monitoring consistency is a normalized value, the consistency threshold can be set to 0.8. When the monitoring consistency of a stage is greater than this consistency threshold, it indicates that different types of electronic components cannot be effectively monitored and distinguished at this stage, and a feedback warning signal needs to be sent to remind the staff to adjust the settings at this stage.

[0083] In summary, the embodiment of the present invention determines the initial difference degree of each electronic component in each full life cycle stage, and obtains the cycle anomaly correlation based on the volatility of the initial difference degrees in different full life cycle stages. Further obtain the monitoring anomaly degree of each electronic component in each full life cycle stage. Based on the distribution results in the clustering clusters, the stage sensitivity coefficient of each category in each full life cycle stage can be determined. According to the difference in the overall sensitivity coefficients between adjacent full life cycle stages and the abnormality of the cycle anomaly correlations of the electronic components, obtain the monitoring consistency of each full life cycle stage and determine whether this stage needs to be adjusted, thereby obtaining the results of faulty electronic components.

[0084] Based on the same inventive concept, the present invention also proposes an electronic component fault monitoring module, and the module includes:

[0085] A monitoring data acquisition module, configured to obtain the monitoring data of the electronic components in each full life cycle stage in the monitoring scenario; the electronic components include multiple categories;

[0086] The monitoring anomaly acquisition module, for a full life cycle stage, takes the monitoring data deviation between each electronic component and similar electronic components in the full life cycle stage as the initial difference of each electronic component; obtains the periodic anomaly correlation according to the volatility of the initial difference of the electronic component at different full life cycle stages; adjusts the initial difference according to the periodic anomaly correlation to obtain the monitoring anomaly of each electronic component at each full life cycle stage;

[0087] The stage sensitivity coefficient analysis module is used to obtain the sample space by taking the time sequence of the whole life cycle stage as the horizontal coordinate and the monitoring abnormality as the vertical coordinate, mapping all electronic components into the sample space and obtaining the clustering result of the data points, and obtaining the stage sensitivity coefficient of each category in each whole life cycle stage according to the distribution concentration of each category of electronic components in the clustering cluster at each whole life cycle stage;

[0088] The whole life cycle stage judgment module is used to obtain the overall sensitivity coefficient of the electronic components in each whole life cycle stage according to the stage sensitivity coefficient; for each whole life cycle stage, according to the difference in the overall sensitivity coefficients in adjacent whole life cycle stages and the abnormality of the monitoring abnormality in the whole life cycle stage, obtain the monitoring consistency of each whole life cycle stage; according to the monitoring consistency, judge whether the whole life cycle stage needs to be adjusted.

[0089] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0090] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for monitoring faults of electronic components, characterized in that, The method comprises: Obtain monitoring data of electronic components at each full life cycle stage in a monitoring scenario; the electronic components include multiple categories; For each full life cycle stage, the monitoring data deviation between each electronic component and similar electronic components in the full life cycle stage is used as the initial difference of each electronic component; according to the volatility of the initial difference of the electronic components at different full life cycle stages, the periodic anomaly correlation is obtained; according to the periodic anomaly correlation, the initial difference is adjusted to obtain the monitoring anomaly of each electronic component at each full life cycle stage; The time series of the whole life cycle stage is used as the horizontal axis, and the monitoring abnormality is used as the vertical axis to obtain the sample space, and all electronic components are mapped into the sample space to obtain the clustering results of the data points. According to the distribution concentration of each category of electronic components in the cluster cluster at each whole life cycle stage, the stage sensitivity coefficient of each category at each whole life cycle stage is obtained; Based on the stage sensitivity coefficients, the overall sensitivity coefficients of the electronic components in each full life cycle stage are obtained; for each full life cycle stage, based on the differences in the overall sensitivity coefficients in adjacent full life cycle stages and the abnormalities of the periodic abnormality correlations of the electronic components, the monitoring consistency of each full life cycle stage is obtained; based on the monitoring consistency, whether the full life cycle stage needs to be adjusted is judged; based on the adjusted full life cycle stage, whether there are faulty electronic components is judged.

2. The method for monitoring faults of electronic components according to claim 1, wherein The method for obtaining the initial difference degree includes: Obtain the average monitoring data of each electronic component in a full life cycle stage; For any electronic component, obtain the mean difference of monitoring data between the electronic component and each other electronic component of the same type; and use the average value of the mean difference of the monitoring data as the initial difference degree of the electronic component in a full life cycle stage.

3. A method for monitoring faults of electronic components according to claim 1, characterized in that, The method for obtaining the period anomaly correlation includes: For each electronic component, the initial difference sequences in all life cycle stages are arranged into an initial difference sequence; a differential sequence of the initial difference sequence is obtained; and the ratio of the element mean in the initial difference sequence to the element mean in the differential sequence is normalized to obtain the periodic anomaly correlation.

4. A method for monitoring faults of electronic components according to claim 1, characterized in that, The method for obtaining the monitoring abnormality degree includes: The periodic anomaly correlation is multiplied by the initial difference to obtain the monitoring anomaly degree of each electronic component at each full life cycle stage.

5. A method for monitoring faults of electronic components according to claim 1, characterized in that, A density clustering algorithm is used in the sample space to obtain each clustering cluster in the clustering result.

6. The method for monitoring the failure of an electronic component according to claim 1, wherein The method for obtaining the stage sensitivity coefficient includes: Choose any category of electronic components as the target category and one life cycle stage as the target stage; In each cluster, the proportion of sample points of electronic components of the target category in the target stage is obtained; and the standard deviation of the proportion of sample points of the target category in all clusters in the target stage is calculated; In each cluster, the maximum difference distance between sample points of electronic components of the target category is obtained; the average maximum difference distance of the target category in all clusters is obtained; The ratio of the average maximum difference distance to the standard deviation is normalized to obtain the stage sensitivity coefficient of the target category at the target stage.

7. A method for monitoring faults of electronic components according to claim 1, characterized in that, The method for obtaining the overall sensitivity coefficient includes: The average value of the stage sensitivity coefficients of all categories in each life cycle stage is taken as the overall sensitivity coefficient.

8. A method for monitoring faults of electronic components according to claim 1, characterized in that, The method for acquiring the monitoring consistency includes: For any full life cycle stage, obtain the overall sensitivity coefficient difference between the full life cycle stage and the previous full life cycle stage; obtain the average periodic anomaly correlation of all electronic components in each category, and obtain the data deviation of all average periodic anomaly correlations relative to the average value; normalize the ratio of the overall sensitivity coefficient difference to the data deviation to obtain the monitoring consistency.

9. A method for monitoring faults of electronic components according to claim 1, characterized in that, The determining whether the full life cycle stage needs to be adjusted according to the monitoring consistency includes: If the monitoring consistency is greater than a preset consistency threshold, it is determined that the full life cycle stage needs to be adjusted.

10. An electronic component fault monitoring module, characterized in that, The module includes: A monitoring data acquisition module, used to obtain monitoring data of electronic components in each full life cycle stage in the monitoring scenario; the electronic components include multiple categories; The monitoring anomaly acquisition module, for a full life cycle stage, takes the monitoring data deviation between each electronic component and similar electronic components in the full life cycle stage as the initial difference of each electronic component; obtains the periodic anomaly correlation according to the volatility of the initial difference of the electronic component at different full life cycle stages; adjusts the initial difference according to the periodic anomaly correlation to obtain the monitoring anomaly of each electronic component at each full life cycle stage; The stage sensitivity coefficient analysis module is used to obtain the sample space by taking the time sequence of the whole life cycle stage as the horizontal coordinate and the monitoring abnormality as the vertical coordinate, mapping all electronic components into the sample space and obtaining the clustering result of the data points, and obtaining the stage sensitivity coefficient of each category in each whole life cycle stage according to the distribution concentration of each category of electronic components in the clustering cluster at each whole life cycle stage; The whole life cycle stage judgment module is used to obtain the overall sensitivity coefficient of the electronic components in each whole life cycle stage according to the stage sensitivity coefficient; for each whole life cycle stage, according to the difference in the overall sensitivity coefficients in adjacent whole life cycle stages and the abnormality of the monitoring abnormality in the whole life cycle stage, obtain the monitoring consistency of each whole life cycle stage; according to the monitoring consistency, judge whether the whole life cycle stage needs to be adjusted.

Images