Search

CN-121980438-A - Electronic component reliability evaluation system and method based on big data

CN121980438ACN 121980438 ACN121980438 ACN 121980438ACN-121980438-A

Abstract

The invention discloses a reliability evaluation system and method of electronic instruments based on big data, and relates to the technical field of reliability evaluation, wherein the evaluation method comprises the following steps that a monitoring system is arranged to collect performance data of any electronic instrument and to analyze the use performance of the electronic instrument; capturing root causes generating abnormality, dividing the types of the root causes, establishing a causal relation chain between the root causes, connecting the root causes contained in the electronic component, analyzing contribution degrees of the root causes directly generating abnormality, analyzing influence conditions of the abnormality indirectly influencing the other root causes in the relation chain, evaluating the reliability of the electronic component based on the influence and contribution conditions of the root causes, identifying the abnormality of the electronic component, breaking through the limitation of the traditional method on data scale and instantaneity, realizing early identification and dynamic tracking of the abnormality, and effectively reducing the risk of system faults.

Inventors

  • YUAN YONGBIN

Assignees

  • 知码芯(杭州)电子科技有限公司

Dates

Publication Date
20260505
Application Date
20251212

Claims (10)

  1. 1. The electronic appliance reliability evaluation method based on big data is characterized by comprising the following steps of: Step S1, a monitoring system is arranged to collect performance data of any electronic component, and the performance of the electronic component is unfolded and analyzed; S2, analyzing and comparing all captured root causes, and dividing root cause types; carrying out causal reasoning on different root causes based on root cause types, and establishing a causal relation chain between the root causes; step S3, carrying out relation chain connection on root causes contained in the electronic component, carrying out contribution analysis on the direct root causes which directly generate the abnormality, and analyzing the influence condition of the abnormality indirectly influenced by other root causes in the relation chain; And S4, carrying out reliability evaluation on the electronic component based on the influence and contribution condition of each root cause in the electronic component, and carrying out anomaly identification on the electronic component according to the reliability evaluation.
  2. 2. The method for evaluating reliability of electronic component based on big data according to claim 1, wherein said step S1 comprises the steps of: Step S11, presetting a plurality of performance indexes in a monitoring system, acquiring performance data of the electronic component at every unit time point, and dividing the acquired performance data into a plurality of performance data sets according to the performance indexes, wherein one performance index is matched with the corresponding performance data set, and sequencing the performance data sets according to the acquisition sequence; Step S12, a performance database is established in advance, a plurality of performance indexes are stored in the performance database, and a performance evaluation rule and a corresponding abnormal evaluation threshold value are preset for any one performance index; randomly selecting a performance data set of one performance index from the plurality of performance data sets, setting the performance index which is the same as the selected performance index in a performance database as a target index, calling a performance evaluation rule of the target index to evaluate the performance data set to obtain a performance evaluation value g of the selected performance index, setting an abnormal evaluation threshold value of the selected performance index as g th , setting the selected performance index as an abnormal index if g < g th , and carrying out abnormal marking on a unit time point corresponding to the last performance data of the performance data set; Step S13, extracting all unit time points with abnormal marks, setting the unit time points as abnormal time points, sequencing all the abnormal time points according to the sequence of the unit time points, arbitrarily selecting two adjacent abnormal time points to obtain a time interval delta t ex between the two adjacent abnormal time points, setting the time interval delta t between the two adjacent unit time points, and merging the two abnormal time points into an abnormal time interval if delta t=delta t ex ; Step S14, combining all adjacent abnormal time points meeting delta t=delta t ex to generate a plurality of abnormal time intervals to obtain the time length of each abnormal time interval, extracting abnormal indexes contained in each abnormal time point in a certain abnormal time interval if the time length of the abnormal time interval exceeds a preset time length threshold, arbitrarily selecting one abnormal index, counting the number of the abnormal time points corresponding to the selected abnormal index as m, presetting an abnormal number threshold m th , setting the selected abnormal index as a root cause if m is larger than or equal to m th , and counting all the abnormal indexes to generate a plurality of root causes.
  3. 3. The method for evaluating reliability of electronic component based on big data according to claim 2, wherein said step S2 comprises the steps of: Step S21, acquiring an abnormal time interval in which a root is extracted and setting the abnormal time interval as a target interval, arbitrarily selecting one root from the selected target interval, extracting abnormal time points of the selected root in the selected target interval, and generating an abnormal subinterval of the selected root in the selected target interval if a plurality of abnormal time points are adjacent and continuous; Step S22, counting the number a of abnormal sub-intervals of a root cause in a selected target interval, extracting the last abnormal time point T end1 of the selected target interval if a=1, acquiring the last abnormal time point T end of the selected target interval, calculating to obtain a first time interval of delta T end =t end -t end1 between the two abnormal time points, setting the interval length of the selected target interval to be T, calculating to obtain a first interval occupation ratio eta 1=delta T end /T, presetting a first occupation ratio threshold eta 1 th , setting the selected root cause as a direct root cause if eta 1 is less than or equal to eta 1 th , setting the selected root cause as an indirect root cause if eta 1 is less than or equal to eta 1 th , extracting the first abnormal time point T s1 of the first abnormal sub-interval if a >1, obtaining the first abnormal time point T s of the selected target interval, calculating to obtain a second time interval delta T s =t s1 -t s , obtaining a second interval occupation ratio eta 2=delta T s /T, setting a second occupation ratio eta 2 and setting eta 2 as a second occupation ratio 362, setting eta 2 as a direct root cause if eta 1 is less than or equal to eta 2 3738, and setting eta 2 is less than or equal to eta 2; step S23, dividing all root factors of the selected target interval into a direct root factor set and an indirect root factor set, arbitrarily selecting one direct root factor from the direct root factor set, and establishing causal relation between each indirect root factor in the indirect root factor set and the selected direct root factor; two indirect root factors are selected randomly from the indirect root factor set, the two indirect root factors are extracted from the abnormal subintervals in the selected target interval respectively, if the two indirect root factors are both one abnormal subinterval and the last abnormal time point of one abnormal subinterval is before the first abnormal time point of the other abnormal subinterval, the two indirect root factors are established in a causal relationship, if one indirect root factor has a plurality of abnormal subintervals and the other indirect root factor has only one abnormal subinterval, the abnormal subinterval of the other indirect root factor is compared with the first abnormal subinterval of the one indirect root factor, and if the abnormal subinterval of the other indirect root factor is positioned before the first abnormal subinterval, the two indirect root factors are established in a causal relationship; Step S24, setting two root causes of the established causal relationship as causal relationship groups, arbitrarily selecting the two causal relationship groups, connecting the two causal relationship groups through the same root causes to generate a causal relationship chain if the two causal relationship groups have the same root causes, and connecting all the causal relationship groups to generate a plurality of causal relationship chains of a selected target interval.
  4. 4. The method for evaluating reliability of electronic component based on big data according to claim 3, wherein said step S3 comprises the steps of: step S31, acquiring direct root causes and indirect root causes contained in all causal relation chains to respectively form a direct root cause set and an indirect root cause set, randomly selecting one direct root cause from the direct root cause set, extracting a target interval where the selected direct root cause is positioned, counting the number M1 of abnormal time points of the selected direct root cause in the target interval, setting the total number of the abnormal time points in the target interval as M total , calculating to obtain contribution degree C=m1/M total of the selected direct root cause, acquiring the contribution degree of the selected direct root cause in each abnormal interval, and calculating an average value to obtain comprehensive contribution degree Z of the selected direct root cause; Step S32, randomly selecting one indirect root from the indirect root set, extracting causal relation chains containing the selected indirect root and setting the extracted indirect root as influence relation chains, randomly selecting one influence relation chain, counting the number of root connections between the selected indirect root and the direct root in the selected influence relation chain as n, and calculating to obtain the path length H=n-1 of the selected influence relation chain; Step S33, counting the number M1 ' of abnormal time points of the selected indirect root cause in the target interval, setting the total number of the abnormal time points in the target interval as M total , calculating to obtain the contribution degree C=m1 ' /M total of the selected direct root cause, acquiring the contribution degree of the selected indirect root cause in each abnormal interval, carrying out average calculation to obtain the comprehensive contribution degree Z ' of the selected indirect root cause, setting the number of influence relation chains as L, and according to the formula: ; Wherein d is a positive integer and d E [1, L ], H d is the path length of the d-th influence relation chain, and the comprehensive influence value Y of the selected indirect root factor is obtained by calculation.
  5. 5. The method for evaluating reliability of electronic component based on big data according to claim 4, wherein said step S4 comprises the steps of: Step S41, acquiring each monitoring index of the electronic appliance in real time by utilizing a monitoring system, generating a plurality of real-time root causes, dividing the plurality of real-time root causes into direct root causes and indirect root causes, acquiring the comprehensive contribution degree of each direct root cause and the comprehensive influence value of each indirect root cause, setting the number of the direct root causes as J and the number of the indirect root causes as K, and according to the formula: ; Wherein, p and q are positive integers, p is E [1, J ], q is E [1, K ], Z p is the comprehensive contribution degree of the p-th direct root, Y q is the comprehensive influence value of the q-th direct root; Step S42, obtaining the total number of root causes N now , calculating to obtain reliability evaluation values U=1-S/N now of the electronic component, presetting an abnormality degree threshold S th and a reliability threshold U th , sending abnormality early warning to the electronic component if S is more than or equal to S th and U is more than or equal to U th , and sending maintenance prompt to the electronic component if S is more than or equal to S th and U is less than U th .
  6. 6. An electronic meta-appliance reliability evaluation system for executing the big data-based electronic meta-appliance reliability evaluation method according to any one of claims 1 to 5, characterized in that the evaluation system comprises an appliance root cause identification module, a root cause relation establishment module, a root cause influence contribution module and a reliability evaluation judgment module; The device root cause identification module is used for carrying out performance data acquisition on any electronic component device by installing a monitoring system, and carrying out unfolding analysis on the service performance of the electronic component device; The root relation establishing module is used for analyzing and comparing all captured root causes, dividing root cause types, carrying out causal reasoning on different root causes based on the root cause types, and establishing a causal relation chain among the root causes; The root factor influence contribution module is used for carrying out relation chain connection on the root factors contained in the electronic appliance, carrying out contribution degree analysis on the direct root factors which directly generate the abnormality, and analyzing the influence condition of the indirect influence abnormality of the rest root factors in the relation chain; The reliability evaluation judging module is used for evaluating the reliability of the electronic component based on the influence and contribution of each root cause in the electronic component, and identifying the abnormality of the electronic component according to the reliability evaluation.
  7. 7. The electronic component device reliability evaluation system according to claim 6, wherein the device root cause identification module comprises a device information acquisition unit and an abnormal root cause capturing unit; the device information acquisition unit is used for acquiring performance data of any electronic component device by installing a monitoring system and performing unfolding analysis on the service performance of the electronic component device, and the abnormal root cause capture unit is used for identifying abnormal conditions based on analysis results and capturing the root cause generating the abnormality.
  8. 8. The electronic component reliability evaluation system according to claim 6, wherein the root relation establishment module comprises a root relation type dividing unit and a root relation matching unit; the root cause type dividing unit is used for analyzing and comparing all captured root causes and dividing the root cause types, and the root cause relation matching unit is used for carrying out causal reasoning on different root causes based on the root cause types and establishing a causal relation chain among the root causes.
  9. 9. The electronic component reliability evaluation system according to claim 6, wherein the root cause contribution module comprises a direct root cause contribution unit and an indirect root cause contribution unit; The direct root cause influencing unit is used for carrying out relation chain connection on the root causes contained in the electronic appliance and carrying out contribution analysis on the direct root causes which directly generate abnormality; the indirect root contribution unit is used for analyzing the influence condition of the abnormality of the other root indirect influence in the relation chain.
  10. 10. The system of claim 6, wherein the reliability evaluation module comprises an appliance reliability evaluation unit and an abnormality recognition judgment unit; the device reliability evaluation unit is used for evaluating the reliability of the electronic component device based on the influence and contribution of each root cause in the electronic component device, and the abnormality recognition judgment unit is used for recognizing the abnormality of the electronic component device according to the reliability evaluation.

Description

Electronic component reliability evaluation system and method based on big data Technical Field The invention relates to the technical field of reliability evaluation, in particular to a reliability evaluation system and method for electronic appliances based on big data. Background Along with the development of information technology, the complexity of electronic equipment is increased, the failure mode and mechanism of the electronic equipment are more and more complex, and the traditional reliability assessment method faces serious challenges; When facing massive running big data, the traditional methods have obvious defects in root cause tracing, namely, most of the traditional methods only can identify surface phenomena or single factors directly related to faults and lack excavating capability of indirect and chain causal relationships to the faults, and secondly, the traditional methods cannot quantify specific contribution degrees of direct and indirect root causes to final reliability indexes, so that abnormal causes with the greatest influence on system reliability cannot be accurately identified, and preventive maintenance and design optimization lack clear data support. Disclosure of Invention The invention aims to provide a reliability evaluation system and method for electronic appliances based on big data, which are used for solving the problems in the prior art. In order to achieve the above purpose, the invention provides the following technical scheme that the electronic component reliability evaluation method based on big data comprises the following steps: Step S1, a monitoring system is arranged to collect performance data of any electronic component, and the performance of the electronic component is unfolded and analyzed; S2, analyzing and comparing all captured root causes, and dividing root cause types; carrying out causal reasoning on different root causes based on root cause types, and establishing a causal relation chain between the root causes; step S3, carrying out relation chain connection on root causes contained in the electronic component, carrying out contribution analysis on the direct root causes which directly generate the abnormality, and analyzing the influence condition of the abnormality indirectly influenced by other root causes in the relation chain; And S4, carrying out reliability evaluation on the electronic component based on the influence and contribution condition of each root cause in the electronic component, and carrying out anomaly identification on the electronic component according to the reliability evaluation. Further, step S1 includes the steps of: Step S11, presetting a plurality of performance indexes in a monitoring system, acquiring performance data of the electronic component at every other unit time point, and dividing the acquired performance data into a plurality of performance data sets according to the performance indexes, wherein one performance index is matched with the corresponding performance data set, and sorting the performance data sets according to the acquisition sequence; Step S12, a performance database is established in advance, a plurality of performance indexes are stored in the performance database, a performance evaluation rule and a corresponding abnormal evaluation threshold value are preset for any one performance index, a performance data set of one performance index is selected randomly from the performance data sets, the performance index which is the same as the selected performance index in the performance database is set as a target index, the performance evaluation rule of the target index is called to evaluate the performance data set, a performance evaluation value g of the selected performance index is obtained, the abnormal evaluation threshold value of the selected performance index is set as g th, if g is smaller than g th, the selected performance index is set as an abnormal index, an abnormal mark is carried out on a unit time point corresponding to the last performance data of the performance data set, the performance evaluation rule can be set as an average value of output voltage for output voltage, the performance evaluation rule can be set as the highest temperature for output current, the performance evaluation rule can be set as a standard deviation evaluation rule of current data, and the like; Step S13, extracting all unit time points with abnormal marks, setting the unit time points as abnormal time points, sequencing all the abnormal time points according to the sequence of the unit time points, arbitrarily selecting two adjacent abnormal time points to obtain a time interval delta t ex between the two adjacent abnormal time points, setting the time interval delta t between the two adjacent unit time points, and merging the two abnormal time points into an abnormal time interval if delta t=delta t ex; Step S14, combining all adjacent abnormal time points meeting delta t=delta t ex to generate a