Electronic component and appliance reliability evaluation system and method based on big data

Abstract

The invention discloses an electronic component device reliability evaluation system and method based on big data, and relates to the technical field of reliability evaluation, and the evaluation method comprises the following steps: a monitoring system is installed to carry out the performance data collection of any electronic component device, and carries out the analysis of the use performance of the electronic component device; capturing the abnormal root causes, and dividing the types of the root causes; establishing a causal relationship chain among root causes; performing relation chain connection on the root causes contained in the electronic element appliance, and performing contribution degree analysis on the direct root causes which directly generate abnormities; analyzing influence conditions of indirect influence anomalies of other root causes in the relation chain; based on the influence and contribution condition of each root cause, carrying out reliability evaluation on the electronic component and appliance, and carrying out abnormity identification on the electronic component and appliance; the method breaks through the limitation of a traditional method in data scale and real-time performance, achieves the early recognition and dynamic tracking of abnormal conditions, and effectively reduces the fault risk of a system.

Description

Technical Field

[0001] This invention relates to the field of reliability assessment technology, specifically to a big data-based reliability assessment system and method for electronic components. Background Technology

[0002] In the field of electronic manufacturing and system maintenance, the reliability of electronic components is the core factor determining product quality and system stability. With the development of information technology, the complexity of electronic devices has increased dramatically, and their failure modes and mechanisms have become increasingly complex, posing a serious challenge to traditional reliability assessment methods. When faced with massive amounts of operational big data, traditional methods have significant shortcomings in root cause analysis: First, they can mostly only identify surface phenomena or single factors directly related to the failure, lacking the ability to uncover indirect and chain causal relationships; Second, existing technologies cannot quantify the specific contribution of each direct and indirect root cause to the final reliability index, resulting in the inability to accurately identify the abnormal triggers that have the greatest impact on system reliability, thus making preventive maintenance and design optimization lack clear data support. Summary of the Invention

[0003] The purpose of this invention is to provide a big data-based reliability assessment system and method for electronic components to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for reliability assessment of electronic components based on big data, the assessment method comprising the following steps: Step S1: The monitoring system is installed to collect performance data of any electronic component and analyze the performance of the electronic component; based on the analysis results, abnormal situations are identified and the root causes of the abnormalities are captured. Step S2: Analyze and compare all captured root causes and classify them into root cause types; perform causal reasoning based on root cause type for different root causes and establish causal relationship chains between root causes. Step S3: Connect the root causes contained in the electronic components and analyze the contribution of the direct root causes that directly cause the abnormality; and analyze the influence of the other root causes in the relationship chain on the abnormality. Step S4: Based on the impact and contribution of each root cause in the electronic device, conduct a reliability assessment of the electronic device; identify anomalies in the electronic device based on the reliability assessment.

[0005] Furthermore, step S1 includes the following steps: Step S11: Several performance indicators are pre-set in the monitoring system. The performance data of electronic components are collected at each unit time point, and the collected performance data is divided into several performance datasets according to the performance indicators. Each performance indicator matches a corresponding performance dataset, and the datasets are sorted according to the collection order. The performance indicators of electronic components may include output voltage, temperature, humidity, output current, and power consumption, etc. Step S12: A performance database is pre-established, storing several performance metrics. Each performance metric has pre-set performance evaluation rules and a corresponding anomaly evaluation threshold. A performance dataset for any one performance metric is randomly selected from the datasets. The performance metric in the performance database that matches the selected metric is set as the target metric. The performance evaluation rules for the target metric are retrieved to evaluate the performance dataset, obtaining the performance evaluation value g for the selected metric. The anomaly evaluation threshold for the selected performance metric is set as g. th If g < g th If the selected performance index is set as an anomaly index, an anomaly will be marked at the unit time point corresponding to the last performance data in the performance dataset; for output voltage, the performance evaluation rule can be set to the average value of the output voltage; for temperature, the performance evaluation rule can be set to the highest temperature; for output current, the performance evaluation rule can be set to evaluate the volatility of the current data, and so on. Step S13: Extract all unit time points with anomaly markers and set them as anomalous time points. Sort all anomalous time points according to their chronological order. Randomly select two adjacent anomalous time points and obtain the time interval between the two anomalous time points as Δt. ex Let Δt be the time interval between two adjacent unit time points. If Δt = Δt ex Then the two abnormal time points will be merged into one abnormal time interval; Step S14: For all conditions satisfying Δt=Δt ex Adjacent abnormal time points are merged to generate several abnormal time intervals, and the time length of each abnormal time interval is obtained. If the time length of any abnormal time interval exceeds a preset time length threshold, the abnormal indicators contained in each abnormal time point of the abnormal time interval are extracted. An abnormal indicator is arbitrarily selected, and the number of abnormal time points corresponding to the selected abnormal indicator is counted as m. A preset abnormal number threshold m is set. th If m≥m th If so, an abnormal indicator will be selected as a root cause; statistical analysis will be performed on all abnormal indicators to generate several root causes.

[0006] Furthermore, step S2 includes the following steps: Step S21: Obtain the abnormal time interval of the root cause and set it as the target interval; arbitrarily select a target interval, arbitrarily select a root cause from the selected target interval, extract the abnormal time point of the selected root cause in the selected target interval, if there are several abnormal time points that are adjacent to each other and continuous, then generate the abnormal sub-interval of the selected root cause in the selected target interval. Step S22: Count the number 'a' of abnormal sub-intervals in the selected target interval that identify the root cause; if a=1, extract the last abnormal time point 't' in the selected target interval. end1 And obtain the last abnormal time point t in the selected target interval. end The first time interval between the two abnormal time points was calculated to be Δt. end =t end -t end1 Let the length of the selected target interval be T, and calculate the proportion of the first interval η1 = Δt. end / T, with a preset first proportion threshold η1 th If η1≤η1 th If η1 < η1, then the selected root cause will be set as the direct root cause. th If a > 1, then the selected root cause will be set as an indirect root cause; if a > 1, then the first anomalous time point t of the first anomalous sub-interval will be extracted. s1 Obtain the first abnormal time point t in the selected target interval. s The second time interval Δt is calculated. s =t s1 -t s And obtain the proportion of the second interval η2=Δt s / T, with a preset second proportion threshold η2 th If η2≤η2 th If η2 > η2, then the selected root cause will be set as an indirect root cause. th If so, the root cause will be set as the direct root cause; When there is an abnormal sub-interval, it is necessary to determine the time period and position of the abnormal sub-interval. For example, if it occurs at the beginning of the abnormal interval and there is no more abnormality afterward, it means that the abnormality in the abnormal sub-interval induced and transferred the abnormality, so it is set as an indirect root cause. If it occurs at the end of the interval, it means that the abnormality is directly caused by the abnormal sub-interval, so it is set as a direct root cause. Step S23: Divide all root causes in the selected target interval into a set of direct root causes and a set of indirect root causes. Randomly select one direct root cause from the set of direct root causes and establish a causal relationship between each indirect root cause in the set of indirect root causes and the selected direct root cause. Randomly select two indirect root causes from the set of indirect root causes and extract the abnormal sub-intervals of the two indirect root causes in the selected target interval. If both indirect root causes are abnormal sub-intervals, and the last abnormal time point of one abnormal sub-interval is before the first abnormal time point of the other abnormal sub-interval, then establish a causal relationship between the two indirect root causes. If one indirect root cause has several abnormal sub-intervals and the other indirect root cause has only one abnormal sub-interval, then compare the abnormal sub-interval of the other indirect root cause with the first abnormal sub-interval of one indirect root cause. If the abnormal sub-interval of the other indirect root cause is before the first abnormal sub-interval, then establish a causal relationship between the two indirect root causes. Step S24: Set the two root causes that establish a causal relationship as a causal relationship group. Randomly select two causal relationship groups. If there is a common root cause in the two causal relationship groups, then connect the two causal relationship groups through the common root cause to generate a causal relationship chain. Connect all causal relationship groups to generate several causal relationship chains for the selected target interval.

[0007] Furthermore, step S3 includes the following steps: Step S31: Obtain the direct and indirect root causes contained in all causal chains, forming a direct root cause set and an indirect root cause set respectively; arbitrarily select a direct root cause from the direct root cause set, extract the target interval containing the selected direct root cause, and count the number m1 of abnormal time points in the target interval for the selected direct root cause, setting the total number of abnormal time points in the target interval as M. total The contribution of the selected direct root cause was calculated as C = m1 / M. total The contribution of the selected direct cause in each abnormal interval will be obtained, and the average value will be calculated to obtain the comprehensive contribution Z of the selected direct cause. Step S32: Randomly select an indirect root cause from the set of indirect root causes, extract the causal relationship chain containing the selected indirect root cause, and set them as influence relationship chains; randomly select an influence relationship chain, count the number of root cause connections between the selected indirect root cause and the direct root cause in the selected influence relationship chain as n, and calculate the path length H=n-1 of the selected influence relationship chain. Step S33: Count the number m1 of abnormal time points in the target interval where indirect root causes are selected. ’ Set the total number of abnormal time points in the target interval to M. total The contribution of the selected direct root cause, C=m1, was calculated. ’ / Mtotal The contribution of selected indirect causes in each abnormal interval is obtained, and the average value is calculated to obtain the comprehensive contribution Z of the selected indirect causes. ’ Let L be the number of influencing relationship chains, according to the formula: ; Where d is a positive integer and d∈[1,L], H d Let d be the path length of the dth influence relationship chain; calculate the comprehensive influence value Y of the selected indirect root cause; the influence value of the indirect root cause mainly reflects the influence of the indirect root cause on the direct root cause, as well as the abnormal occurrence range of the indirect root cause. Combining the two can yield a more accurate influence situation.

[0008] Furthermore, step S4 includes the following steps: Step S41: Utilize the monitoring system to collect real-time data on various monitoring indicators of electronic components and generate several real-time root causes. These root causes are then categorized into direct and indirect root causes. The comprehensive contribution of each direct root cause and the comprehensive impact value of each indirect root cause are obtained. The number of direct root causes is set as J, and the number of indirect root causes as K, according to the formula: ; Where p and q are positive integers and p∈[1,J], q∈[1,K], Z p Y represents the overall contribution of the p-th direct cause. q Let be the comprehensive impact value of the qth direct cause; calculate the degree of abnormality S of the electronic components; Step S42: Obtain the total number of root causes, N. now The reliability assessment value of the electronic components was calculated as U = 1 - S / N. now A preset anomaly threshold S is defined. th and reliability threshold U th If S≥S th And U≥U th Then an abnormal warning is sent to the electronic components. If S≥S th And U<U th Then, a maintenance reminder will be sent to the electronic components.

[0009] To better implement the above methods, an electronic component reliability assessment system is also proposed. The assessment system includes a component root cause identification module, a root cause relationship establishment module, a root cause impact contribution module, and a reliability assessment judgment module. The appliance root cause identification module is used to collect performance data of any electronic appliance with a monitoring system and analyze the performance of the electronic appliance; based on the analysis results, it identifies abnormal situations and captures the root causes of the abnormalities. The root cause relationship establishment module is used to analyze and compare all captured root causes and classify them into root cause types; it performs causal reasoning based on root cause type for different root causes and establishes causal relationship chains between root causes. The Root Cause Influence Contribution Module is used to connect the root causes contained in electronic components and appliances into a relationship chain, analyze the contribution of direct root causes that directly cause abnormalities, and analyze the influence of other root causes in the relationship chain on the abnormalities. The reliability assessment module is used to assess the reliability of electronic components based on the impact and contribution of each root cause in the electronic components; and to identify anomalies in the electronic components based on the reliability assessment.

[0010] Furthermore, the appliance root cause identification module includes an appliance information acquisition unit and an abnormal root cause capture unit; The appliance information acquisition unit is used to collect performance data of any electronic appliance equipped with a monitoring system and to analyze the performance of the electronic appliance. The anomaly root cause capture unit is used to identify abnormal situations based on the analysis results and capture the root causes of the abnormalities.

[0011] Furthermore, the root cause relationship establishment module includes a root cause type classification unit and a root cause relationship matching unit; The root cause type classification unit is used to analyze and compare all captured root causes and classify them into root cause types; the root cause relationship matching unit is used to perform causal reasoning based on root cause type for different root causes and establish causal relationship chains between root causes.

[0012] Furthermore, the root cause contribution module includes direct root cause contribution units and indirect root cause contribution units; The direct root cause influence unit is used to connect the root causes contained in electronic components and devices, and to analyze the contribution of the direct root causes that directly cause the abnormality; the indirect root cause contribution unit is used to analyze the influence of the remaining root causes in the relationship chain on the abnormality.

[0013] Furthermore, the reliability assessment module includes an equipment reliability assessment unit and an anomaly identification and judgment unit; The device reliability assessment unit is used to assess the reliability of electronic devices based on the influence and contribution of each root cause in the device; the anomaly identification and judgment unit is used to identify anomalies in electronic devices based on the reliability assessment.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention achieves in-depth mining of abnormal root causes by constructing a root cause type classification and causal relationship chain. It can not only identify direct causes, but also reveal the complex relationship between indirect factors, thereby comprehensively analyzing the fault mechanism and improving the hierarchy and logic of fault diagnosis. 2. This invention introduces a contribution and impact quantification mechanism to accurately measure the role of each root cause in system reliability; by calculating the contribution of direct root causes and the combined impact of indirect root causes, the evaluation process is more objective and comparable, helping users identify key risk points, achieve precise resource allocation and priority management, and improve maintenance efficiency. 3. By integrating big data analysis and real-time monitoring technologies, this invention significantly improves the comprehensiveness and accuracy of reliability assessment of electronic components; it breaks through the limitations of traditional methods in terms of data scale and real-time performance, enabling early identification and dynamic tracking of abnormal situations, and effectively reducing the risk of system failure. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the structure of a big data-based electronic component reliability assessment system; Figure 2 This is a schematic diagram illustrating the steps of a big data-based reliability assessment method for electronic components. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example: Figures 1 to 2 As shown, this invention provides a method for reliability assessment of electronic components based on big data. The assessment method includes the following steps: Step S1: The monitoring system is installed to collect performance data of any electronic component and analyze the performance of the electronic component; based on the analysis results, abnormal situations are identified and the root causes of the abnormalities are captured. Step S1 includes the following steps: Step S11: Several performance indicators are pre-set in the monitoring system. The performance data of electronic components are collected at every unit time point. The collected performance data is divided into several performance datasets according to the performance indicators. Each performance indicator matches a corresponding performance dataset. The performance datasets are sorted according to the collection order. Step S12: A performance database is pre-established, storing several performance metrics. Each performance metric has pre-set performance evaluation rules and a corresponding anomaly evaluation threshold. A performance dataset for any one performance metric is randomly selected from the datasets. The performance metric in the performance database that matches the selected metric is set as the target metric. The performance evaluation rules for the target metric are retrieved to evaluate the performance dataset, obtaining the performance evaluation value g for the selected metric. The anomaly evaluation threshold for the selected performance metric is set as g. th If g < g th If so, the selected performance metric will be set as the abnormal metric, and an anomaly will be marked at the unit time point corresponding to the last performance data in the performance dataset. Step S13: Extract all unit time points with anomaly markers and set them as anomalous time points. Sort all anomalous time points according to their chronological order. Randomly select two adjacent anomalous time points and obtain the time interval between the two anomalous time points as Δt. ex Let Δt be the time interval between two adjacent unit time points. If Δt = Δt ex Then the two abnormal time points will be merged into one abnormal time interval; Step S14: For all conditions satisfying Δt=Δt ex Adjacent abnormal time points are merged to generate several abnormal time intervals, and the time length of each abnormal time interval is obtained. If the time length of any abnormal time interval exceeds a preset time length threshold, the abnormal indicators contained in each abnormal time point of the abnormal time interval are extracted. An abnormal indicator is arbitrarily selected, and the number of abnormal time points corresponding to the selected abnormal indicator is counted as m. A preset abnormal number threshold m is set. th If m≥m th If so, an abnormal indicator will be selected as a root cause; statistical analysis will be performed on all abnormal indicators to generate several root causes.

[0018] Step S2: Analyze and compare all captured root causes and classify them into root cause types; perform causal reasoning based on root cause type for different root causes and establish causal relationship chains between root causes. Step S2 includes the following steps: Step S21: Obtain the abnormal time interval of the root cause and set it as the target interval; arbitrarily select a target interval, arbitrarily select a root cause from the selected target interval, extract the abnormal time point of the selected root cause in the selected target interval, if there are several abnormal time points that are adjacent to each other and continuous, then generate the abnormal sub-interval of the selected root cause in the selected target interval. Step S22: Count the number 'a' of abnormal sub-intervals in the selected target interval that identify the root cause; if a=1, extract the last abnormal time point 't' in the selected target interval. end1 And obtain the last abnormal time point t in the selected target interval. end The first time interval between the two abnormal time points was calculated to be Δt. end =t end -t end1 Let the length of the selected target interval be T, and calculate the proportion of the first interval η1 = Δt. end / T, with a preset first proportion threshold η1 th If η1≤η1 th If η1 < η1, then the selected root cause will be set as the direct root cause. th If a > 1, then the selected root cause will be set as an indirect root cause; if a > 1, then the first anomalous time point t of the first anomalous sub-interval will be extracted. s1 Obtain the first abnormal time point t in the selected target interval. s The second time interval Δt is calculated. s =t s1 -t s And obtain the proportion of the second interval η2=Δt s / T, with a preset second proportion threshold η2 th If η2≤η2 th If η2 > η2, then the selected root cause will be set as an indirect root cause. th If so, the root cause will be set as the direct root cause; Example 1: Select a root cause as the output voltage, set the number of abnormal sub-intervals of the output voltage a=1, and extract the last abnormal time point t of the abnormal sub-interval. end1 =107 seconds, the last abnormal time point t in the target interval end =107 seconds, calculate Δt end =107-107=0 seconds, target interval length T=3 seconds, first interval proportion η1=0 / 3=0. Preset first proportion threshold η1. th =0.1, since η1≤η1 th Therefore, the output voltage is set as the direct root cause; another root cause is selected as temperature, and the number of anomalous sub-intervals a=2. The first anomalous time point t of the first anomalous sub-interval is extracted. s1 =105 seconds, the first abnormal time point t in the target interval s =105 seconds, calculate Δt s =105-105=0 seconds, the second interval percentage η2=0 / 3=0. The preset second percentage threshold η2. th =0.2, since η2≤η2 th Therefore, temperature is set as an indirect root cause. Step S23: Divide all root causes in the selected target interval into a set of direct root causes and a set of indirect root causes. Randomly select one direct root cause from the set of direct root causes and establish a causal relationship between each indirect root cause in the set of indirect root causes and the selected direct root cause. Randomly select two indirect root causes from the set of indirect root causes and extract the abnormal sub-intervals of the two indirect root causes in the selected target interval. If both indirect root causes are abnormal sub-intervals, and the last abnormal time point of one abnormal sub-interval is before the first abnormal time point of the other abnormal sub-interval, then establish a causal relationship between the two indirect root causes. If one indirect root cause has several abnormal sub-intervals and the other indirect root cause has only one abnormal sub-interval, then compare the abnormal sub-interval of the other indirect root cause with the first abnormal sub-interval of one indirect root cause. If the abnormal sub-interval of the other indirect root cause is before the first abnormal sub-interval, then establish a causal relationship between the two indirect root causes. Example 2: Set the direct root cause set to include output voltage, and the indirect root cause set to include temperature and humidity; select the output voltage from the direct root cause set and establish causal relationships with temperature and humidity from the indirect root cause set respectively; for the two indirect root causes of temperature and humidity, set the temperature abnormal sub-interval to [103, 105], if the humidity abnormal sub-interval is [106, 107], since 105 < 106, then a causal relationship is established between the two; if the humidity abnormal sub-interval is [104, 106], then a causal relationship cannot be established. Step S24: Set the two root causes that establish a causal relationship as a causal relationship group. Randomly select two causal relationship groups. If there is a common root cause in the two causal relationship groups, then connect the two causal relationship groups through the common root cause to generate a causal relationship chain. Connect all causal relationship groups to generate several causal relationship chains for the selected target interval.

[0019] Step S3: Connect the root causes contained in the electronic components and analyze the contribution of the direct root causes that directly cause the abnormality; and analyze the influence of the other root causes in the relationship chain on the abnormality. Step S3 includes the following steps: Step S31: Obtain the direct and indirect root causes contained in all causal chains, forming a direct root cause set and an indirect root cause set respectively; arbitrarily select a direct root cause from the direct root cause set, extract the target interval containing the selected direct root cause, and count the number m1 of abnormal time points in the target interval for the selected direct root cause, setting the total number of abnormal time points in the target interval as M. total The contribution of the selected direct root cause was calculated as C = m1 / M. totalThe contribution of the selected direct cause in each abnormal interval will be obtained, and the average value will be calculated to obtain the comprehensive contribution Z of the selected direct cause. Step S32: Randomly select an indirect root cause from the set of indirect root causes, extract the causal relationship chain containing the selected indirect root cause, and set them as influence relationship chains; randomly select an influence relationship chain, count the number of root cause connections between the selected indirect root cause and the direct root cause in the selected influence relationship chain as n, and calculate the path length H=n-1 of the selected influence relationship chain. Step S33: Count the number m1 of abnormal time points in the target interval where indirect root causes are selected. ’ Set the total number of abnormal time points in the target interval to M. total The contribution of the selected direct root cause, C=m1, was calculated. ’ / M total The contribution of selected indirect causes in each abnormal interval is obtained, and the average value is calculated to obtain the comprehensive contribution Z of the selected indirect causes. ’ Let L be the number of influencing relationship chains, according to the formula: ; Where d is a positive integer and d∈[1,L], H d Let be the path length of the d-th influencing relationship chain; calculate the comprehensive influence value Y of the selected indirect root cause; Example 3: For indirect root cause temperature, the number of anomalous time points m1 in the target interval ’ =2, Total abnormal time points M total =3, Contribution C ’ =2 / 3=0.67; Assume that the overall contribution of temperature Z in several intervals is... ’ =0.65; Set the number of influencing relationship chains L=2, and the path length H d If the values ​​are 2 and 1 respectively, then the comprehensive impact value Y = 0.65 × 1 / 2 × (1 + 1 / 2) = 0.4875.

[0020] Step S4: Based on the impact and contribution of each root cause in the electronic components, conduct a reliability assessment of the electronic components; identify anomalies in the electronic components based on the reliability assessment. Step S4 includes the following steps: Step S41: Utilize the monitoring system to collect real-time data on various monitoring indicators of electronic components and generate several real-time root causes. These root causes are then categorized into direct and indirect root causes. The comprehensive contribution of each direct root cause and the comprehensive impact value of each indirect root cause are obtained. The number of direct root causes is set as J, and the number of indirect root causes as K, according to the formula: ; Where p and q are positive integers and p∈[1,J], q∈[1,K], Z p Y represents the overall contribution of the p-th direct cause. q Let be the comprehensive impact value of the qth direct cause; calculate the degree of abnormality S of the electronic components; Step S42: Obtain the total number of root causes, N. now The reliability assessment value of the electronic components was calculated as U = 1 - S / N. now A preset anomaly threshold S is defined. th and reliability threshold U th If S≥S th And U≥U th Then an abnormal warning is sent to the electronic components. If S≥S th And U<U th Then, a maintenance reminder will be sent to the electronic components.

[0021] An electronic component reliability assessment system, the assessment system includes a root cause identification module, a root cause relationship establishment module, a root cause impact contribution module, and a reliability assessment judgment module; The appliance root cause identification module is used to collect performance data of any electronic appliance with a monitoring system and analyze the performance of the electronic appliance; based on the analysis results, it identifies abnormal situations and captures the root causes of the abnormalities. The root cause relationship establishment module is used to analyze and compare all captured root causes and classify them into root cause types; it performs causal reasoning based on root cause type for different root causes and establishes causal relationship chains between root causes. The Root Cause Influence Contribution Module is used to connect the root causes contained in electronic components and appliances into a relationship chain, analyze the contribution of direct root causes that directly cause abnormalities, and analyze the influence of other root causes in the relationship chain on the abnormalities. The reliability assessment module is used to assess the reliability of electronic components based on the impact and contribution of each root cause in the electronic components; and to identify anomalies in the electronic components based on the reliability assessment.

[0022] The appliance root cause identification module includes an appliance information acquisition unit and an abnormal root cause capture unit. The appliance information acquisition unit is used to collect performance data of any electronic appliance equipped with a monitoring system and to analyze the performance of the electronic appliance. The anomaly root cause capture unit is used to identify abnormal situations based on the analysis results and capture the root causes of the abnormalities.

[0023] The root cause relationship establishment module includes a root cause type division unit and a root cause relationship matching unit; The root cause type classification unit is used to analyze and compare all captured root causes and classify them into root cause types; the root cause relationship matching unit is used to perform causal reasoning based on root cause type for different root causes and establish causal relationship chains between root causes.

[0024] The root cause contribution module includes direct root cause contribution units and indirect root cause contribution units. The direct root cause influence unit is used to connect the root causes contained in electronic components and devices, and to analyze the contribution of the direct root causes that directly cause the abnormality; the indirect root cause contribution unit is used to analyze the influence of the remaining root causes in the relationship chain on the abnormality.

[0025] The reliability assessment and judgment module includes an equipment reliability assessment unit and an anomaly identification and judgment unit. The device reliability assessment unit is used to assess the reliability of electronic devices based on the influence and contribution of each root cause in the device; the anomaly identification and judgment unit is used to identify anomalies in electronic devices based on the reliability assessment.

[0026] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for reliability assessment of electronic components based on big data, characterized in that: The evaluation method includes the following steps: Step S1: The monitoring system is installed to collect performance data of any electronic component and analyze the performance of the electronic component; based on the analysis results, abnormal situations are identified and the root causes of the abnormalities are captured. Step S2: Analyze and compare all captured root causes and classify them into root cause types; perform causal reasoning based on root cause type for different root causes and establish causal relationship chains between root causes. Step S3: Connect the root causes contained in the electronic components and analyze the contribution of the direct root causes that directly cause the abnormality; and analyze the influence of the other root causes in the relationship chain on the abnormality. Step S4: Based on the impact and contribution of each root cause in the electronic device, conduct a reliability assessment of the electronic device; identify anomalies in the electronic device based on the reliability assessment.

2. The method for reliability assessment of electronic components based on big data according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Several performance indicators are pre-set in the monitoring system. The performance data of electronic components are collected at every unit time point. The collected performance data is divided into several performance datasets according to the performance indicators. Each performance indicator matches a corresponding performance dataset. The performance datasets are sorted according to the collection order. Step S12: A performance database is pre-established, storing several performance metrics. Each performance metric has pre-set performance evaluation rules and a corresponding anomaly evaluation threshold. A performance dataset for any one performance metric is randomly selected from the datasets. The performance metric in the performance database that matches the selected metric is set as the target metric. The performance evaluation rules for the target metric are retrieved to evaluate the performance dataset, obtaining the performance evaluation value g for the selected metric. The anomaly evaluation threshold for the selected performance metric is set as g. th If g < g th If so, the selected performance metric will be set as the abnormal metric, and an anomaly will be marked at the unit time point corresponding to the last performance data in the performance dataset. Step S13: Extract all unit time points with anomaly markers and set them as anomalous time points. Sort all anomalous time points according to their chronological order. Randomly select two adjacent anomalous time points and obtain the time interval between the two anomalous time points as Δt. ex Let Δt be the time interval between two adjacent unit time points. If Δt = Δt ex Then the two abnormal time points will be merged into one abnormal time interval; Step S14: For all conditions satisfying Δt=Δt ex Adjacent abnormal time points are merged to generate several abnormal time intervals, and the time length of each abnormal time interval is obtained. If the time length of any abnormal time interval exceeds a preset time length threshold, the abnormal indicators contained in each abnormal time point of the abnormal time interval are extracted. An abnormal indicator is arbitrarily selected, and the number of abnormal time points corresponding to the selected abnormal indicator is counted as m. A preset abnormal number threshold m is set. th If m≥m th If so, an abnormal indicator will be selected as a root cause; statistical analysis will be performed on all abnormal indicators to generate several root causes.

3. The method for reliability assessment of electronic components based on big data according to claim 2, characterized in that: Step S2 includes the following steps: Step S21: Obtain the abnormal time interval of the root cause and set it as the target interval; arbitrarily select a target interval, arbitrarily select a root cause from the selected target interval, extract the abnormal time point of the selected root cause in the selected target interval, if there are several abnormal time points that are adjacent to each other and continuous, then generate the abnormal sub-interval of the selected root cause in the selected target interval. Step S22: Count the number 'a' of abnormal sub-intervals in the selected target interval that identify the root cause; if a=1, extract the last abnormal time point 't' in the selected target interval. end1 And obtain the last abnormal time point t in the selected target interval. end The first time interval between the two abnormal time points was calculated to be Δt. end =t end -t end1 Let the length of the selected target interval be T, and calculate the proportion of the first interval η1 = Δt. end / T, with a preset first proportion threshold η1 th If η1≤η1 th If η1 < η1, then the selected root cause will be set as the direct root cause. th If a > 1, then the selected root cause will be set as an indirect root cause; if a > 1, then the first anomalous time point t of the first anomalous sub-interval will be extracted. s1 Obtain the first abnormal time point t in the selected target interval. s The second time interval Δt is calculated. s =t s1 -t s And obtain the proportion of the second interval η2=Δt s / T, with a preset second proportion threshold η2 th If η2≤η2 th If η2 > η2, then the selected root cause will be set as an indirect root cause. th If so, the root cause will be set as the direct root cause; Step S23: Divide all root causes in the selected target interval into a set of direct root causes and a set of indirect root causes. Randomly select one direct root cause from the set of direct root causes and establish a causal relationship between each indirect root cause in the set of indirect root causes and the selected direct root cause. Randomly select two indirect root causes from the set of indirect root causes and extract the abnormal sub-intervals of the two indirect root causes in the selected target interval. If both indirect root causes are abnormal sub-intervals, and the last abnormal time point of one abnormal sub-interval is before the first abnormal time point of the other abnormal sub-interval, then establish a causal relationship between the two indirect root causes. If one indirect root cause has several abnormal sub-intervals and the other indirect root cause has only one abnormal sub-interval, then compare the abnormal sub-interval of the other indirect root cause with the first abnormal sub-interval of one indirect root cause. If the abnormal sub-interval of the other indirect root cause is before the first abnormal sub-interval, then establish a causal relationship between the two indirect root causes. Step S24: Set the two root causes that establish a causal relationship as a causal relationship group. Randomly select two causal relationship groups. If there is a common root cause in the two causal relationship groups, then connect the two causal relationship groups through the common root cause to generate a causal relationship chain. Connect all causal relationship groups to generate several causal relationship chains for the selected target interval.

4. The method for reliability assessment of electronic components based on big data according to claim 3, characterized in that: Step S3 includes the following steps: Step S31: Obtain the direct and indirect root causes contained in all causal chains, forming a direct root cause set and an indirect root cause set respectively; arbitrarily select a direct root cause from the direct root cause set, extract the target interval containing the selected direct root cause, and count the number m1 of abnormal time points in the target interval for the selected direct root cause, setting the total number of abnormal time points in the target interval as M. total The contribution of the selected direct root cause was calculated as C = m1 / M. total The contribution of the selected direct cause in each abnormal interval will be obtained, and the average value will be calculated to obtain the comprehensive contribution Z of the selected direct cause. Step S32: Randomly select an indirect root cause from the set of indirect root causes, extract the causal relationship chain containing the selected indirect root cause, and set them as influence relationship chains; randomly select an influence relationship chain, count the number of root cause connections between the selected indirect root cause and the direct root cause in the selected influence relationship chain as n, and calculate the path length H=n-1 of the selected influence relationship chain. Step S33: Count the number m1 of abnormal time points in the target interval where indirect root causes are selected. ’ Set the total number of abnormal time points in the target interval to M. total The contribution of the selected direct root cause, C=m1, was calculated. ’ / M total The contribution of selected indirect causes in each abnormal interval is obtained, and the average value is calculated to obtain the comprehensive contribution Z of the selected indirect causes. ’ Let L be the number of influencing relationship chains, according to the formula: ; Where d is a positive integer and d∈[1,L], H d Let d be the path length of the d-th influencing relationship chain; calculate the comprehensive influence value Y of the selected indirect root cause.

5. The method for reliability assessment of electronic components based on big data according to claim 4, characterized in that: Step S4 includes the following steps: Step S41: Utilize the monitoring system to collect real-time data on various monitoring indicators of electronic components and generate several real-time root causes. These root causes are then categorized into direct and indirect root causes. The comprehensive contribution of each direct root cause and the comprehensive impact value of each indirect root cause are obtained. The number of direct root causes is set as J, and the number of indirect root causes as K, according to the formula: ; Where p and q are positive integers and p∈[1,J], q∈[1,K], Z p Y represents the overall contribution of the p-th direct cause. q Let be the comprehensive impact value of the qth direct cause; calculate the degree of abnormality S of the electronic components; Step S42: Obtain the total number of root causes, N. now The reliability assessment value of the electronic components was calculated as U = 1 - S / N. now A preset anomaly threshold S is defined. th and reliability threshold U th If S≥S th And U≥U th Then an abnormal warning is sent to the electronic components. If S≥S th And U<U th Then, a maintenance reminder will be sent to the electronic components.

6. An electronic component reliability assessment system, used to execute the big data-based electronic component reliability assessment method according to any one of claims 1-5, characterized in that: The evaluation system includes an appliance root cause identification module, a root cause relationship establishment module, a root cause impact contribution module, and a reliable evaluation and judgment module. The appliance root cause identification module is used to collect performance data of any electronic appliance by installing a monitoring system, and to analyze the performance of the electronic appliance; based on the analysis results, it identifies abnormal situations and captures the root causes of the abnormalities. The root cause relationship establishment module is used to analyze and compare all captured root causes and classify them into root cause types; and to perform causal reasoning based on root cause type for different root causes to establish causal relationship chains between root causes. The root cause influence contribution module is used to connect the root causes contained in electronic components and devices into a relationship chain, analyze the contribution of direct root causes that directly cause abnormalities, and analyze the influence of other root causes in the relationship chain on the abnormalities. The reliability assessment module is used to assess the reliability of electronic components based on the influence and contribution of each root cause in the electronic components; and to identify anomalies in the electronic components based on the reliability assessment.

7. The electronic component reliability assessment system according to claim 6, characterized in that: The device root cause identification module includes a device information acquisition unit and an abnormal root cause capture unit; The device information acquisition unit is used to acquire performance data of any electronic device by installing a monitoring system, and to analyze the performance of the electronic device; the abnormal root cause capture unit is used to identify abnormal situations based on the analysis results, and to capture the root causes of the abnormalities.

8. The electronic component reliability assessment system according to claim 6, characterized in that: The root cause relationship establishment module includes a root cause type classification unit and a root cause relationship matching unit; The root cause type classification unit is used to analyze and compare all captured root causes and classify them into root cause types; the root cause relationship matching unit is used to perform causal reasoning on different root causes based on root cause types and establish causal relationship chains between root causes.

9. The electronic component reliability assessment system according to claim 6, characterized in that: The root cause influence contribution module includes a direct root cause influence unit and an indirect root cause contribution unit. The direct root cause influence unit is used to connect the root causes contained in electronic components and analyze the contribution of the direct root causes that directly cause the abnormality; the indirect root cause contribution unit is used to analyze the influence of the remaining root causes in the relationship chain on the abnormality.

10. The electronic component reliability assessment system according to claim 6, characterized in that: The reliability assessment and judgment module includes an equipment reliability assessment unit and an anomaly identification and judgment unit; The device reliability assessment unit is used to assess the reliability of electronic components based on the influence and contribution of each root cause in the electronic components; the anomaly identification and judgment unit is used to identify anomalies in the electronic components based on the reliability assessment.

Images