Search

CN-121978592-A - Intelligent production quality detection system and method applied to capacitor

CN121978592ACN 121978592 ACN121978592 ACN 121978592ACN-121978592-A

Abstract

The invention discloses an intelligent detection system and method for the production quality of capacitors, which belong to the technical field of capacitor production quality detection, and are characterized by collecting the previous process data of each capacitor to be aged, calculating to obtain a state stability index, dividing the capacitors into a plurality of initial state categories, configuring an individualized initial aging parameter scheme comprising aging temperature, applied voltage and aging time for each capacitor by combining various statistical characteristics and the state stability index, carrying out aging detection on the capacitors according to the individualized scheme at an independent station, collecting leakage current and capacitance data in real time, extracting aging response characteristics, calculating aging matching degree by matching with category standard characteristics, carrying out category dynamic migration and parameter self-adaptive adjustment on the capacitors according to the matching degree, recording category migration track and parameter adjustment data in the aging process, forming an individualized aging file, and finally comprehensively judging by combining the aging file, the matching degree and electrical parameters, and outputting the quality grade of each capacitor.

Inventors

  • CHEN QI

Assignees

  • 南通南铭电子有限公司

Dates

Publication Date
20260505
Application Date
20260408

Claims (8)

  1. 1. The intelligent detection method for the production quality of the capacitor is characterized by comprising the following steps of: step 100, collecting previous process data of each capacitor to be aged, and uniquely identifying and documenting each capacitor; calculating a state stability index according to the previous process data, and dynamically dividing the capacitor into a plurality of initial state categories based on the state stability index; step 200, configuring a corresponding personalized initial aging parameter scheme for each capacitor by combining the statistical characteristics of each initial state category and the individual state stability index of each capacitor in the category, wherein the personalized initial aging parameter scheme comprises an aging temperature interval, an applied voltage range and a reference aging duration which are differentially set; step 300, based on a personalized initial aging parameter scheme, carrying out aging detection on a single capacitor at an independent station, continuously collecting leakage current and capacitance data in an aging process to obtain real-time aging response characteristics, matching the real-time aging response characteristics with standard response characteristics preset in the category, and calculating to obtain aging matching degree; Step 400, recording class migration tracks, parameter adjustment processes and execution time lengths of all stages of each capacitor to form an individual aging file, detecting final electrical parameters after aging, comprehensively judging by combining the individual aging file, the aging matching degree and the final electrical parameters, and outputting the quality grade of each capacitor.
  2. 2. The method for intelligently detecting the production quality of the capacitor according to claim 1, wherein the step S100 comprises the following steps: Step S101, acquiring technological parameters of each capacitor to be aged in a previous process stage, including a winding tension value, an impregnation saturation value and an oxide film thickness value, on a capacitor production line through a production management system, uniquely identifying and profiling each capacitor, endowing each capacitor with a unique serial number, and associating and storing acquired original data with the serial number; The method comprises the steps of providing n capacitors to be aged, wherein each capacitor corresponds to three technological parameter characteristics, using x ij to represent the jth original characteristic value of the ith capacitor, wherein j=1, 2 and 3 correspond to winding tension, impregnation saturation and oxide film thickness respectively, firstly calculating the mean mu j and standard deviation sigma j of the jth characteristic in all the capacitors, and then carrying out Z-score standardization processing on each characteristic, wherein x' ij =(x ij -μ j )/σ j ;x' ij represents the value of the ith capacitor standardized on the jth characteristic.
  3. 3. The method for intelligently detecting the production quality of a capacitor according to claim 2, wherein said step S100 further comprises: Step S102, mapping a normalized value x ij to a [0,1] interval through extremely poor transformation to obtain a probability value p ij =[x' ij -min(x' ij )]/[max(x' ij )-min(x' ij ], wherein p ij is the probability duty ratio of an ith capacitor on a jth feature, and min (x ' ij ) and max (x' ij ) respectively represent the minimum value and the maximum value of all capacitors normalized on the jth feature; calculating information entropy of the j-th feature: ; Wherein E j represents the information entropy of the jth feature; Calculating the weight of the j-th feature according to the information entropy: ; Wherein w j represents the weight of the jth feature; and carrying out weighted summation on the standardized characteristic value of each capacitor and the weight corresponding to the standardized characteristic value to obtain a state stability index: ; wherein Z i represents the state stability index of the i-th capacitor; Step 103, taking state stability indexes { Z 1 ,Z 2 ,…,Z n } corresponding to all capacitors as one-dimensional clustering sample points, and executing initial state classification by adopting a K-means clustering algorithm, wherein K state stability indexes are randomly selected as initial clustering centers according to preset clustering quantity K of production working conditions and product specifications; iteratively calculating Euclidean distance from each sample point to the corresponding cluster center, and dividing the samples into categories with the nearest distance; re-calculating the clustering centers of all the categories after each iteration until the clustering centers reach the maximum iteration times, and completing clustering after iteration convergence; an initial class label C i (0) epsilon {1, 2..once, K } is assigned to each capacitor, and the clustering center and sample standard deviation of each class are recorded to complete the classification of the initial state.
  4. 4. The method for intelligently detecting the production quality of a capacitor according to claim 3, wherein said step S200 comprises the steps of: Step S201, obtaining a state stability index corresponding to each capacitor, a clustering center of an initial state belonging to the capacitor and a class serial number K, and calculating a difference value between the individual state stability index and the clustering center of the class belonging to the capacitor to obtain an individual deviation degree D i =Z i -Z c,k , wherein D i is the stability state deviation degree of an ith capacitor in the class belonging to the capacitor, Z c,k represents the clustering center of the kth initial state class, k= {1,2,..; Step S202, configuring an individualized initial aging parameter scheme for each capacitor according to an initial class sequence number k, a class clustering center Z c,k and an individual deviation degree D i , presetting a reference aging temperature T base,k , a reference applied voltage U base,k and a reference aging time period T base,k for a kth class, correcting the reference parameters according to the individual deviation degree D i to obtain individualized parameters of the capacitor, namely U i =U base,k -λ U ·D i ;t i =t base,k +λ t ·D i , wherein U i is the individualized applied voltage of an ith capacitor, T i is the individualized aging time period of the ith capacitor, lambda U and lambda t are respectively preset voltage correction coefficients and time period correction coefficients, and finally forming the initial aging parameter scheme comprising an aging temperature interval, an applied voltage range and the individualized aging time period by adopting the class unified value T i =T base,k .
  5. 5. The method for intelligently detecting the production quality of the capacitor according to claim 1, wherein the step S300 comprises the following steps: Step 301, based on a personalized initial aging parameter scheme, each capacitor is placed in an independent aging station, aging detection is carried out according to an aging temperature interval, an applied voltage range and personalized aging time length set by the capacitor, and in the aging process, the leakage current value and the capacitance value of the capacitor are continuously collected at a fixed sampling frequency and respectively recorded as I (t) and C (t), so that a real-time leakage current curve and a capacitance curve are formed; Step S302, carrying out data preprocessing on a leakage current curve and a capacitance curve which are acquired in real time, wherein the preprocessing comprises the steps of removing abnormal points and moving average filtering, and then extracting real-time aging response characteristics of the internal state evolution of a capacitor from the preprocessed curve, wherein the real-time aging response characteristics comprise a leakage current stable value I s , a leakage current peak value I peak , a leakage current falling time constant tau, a capacitance recovery rate C h and a capacitance platform value C p ; The leakage current peak value I peak is obtained by taking the maximum value of a leakage current curve as a leakage current peak value in a preset time window at the initial stage of voltage application, the leakage current stable value is obtained by taking the average value of all sampling points in a time interval of which the leakage current curve is towards the last 10% as a leakage current stable value in the later stage of aging, the leakage current falling time constant tau is obtained by exponential fitting, namely, I (t) =i s +(I peak -I s )e -t/τ , t represents a time variable, namely, the time from the time when voltage application begins to time to the current sampling moment in the aging process, the capacitance plateau value C p is obtained by taking the average value of all sampling points in the same time interval as the leakage current stable value in the later stage of aging, the capacitance recovery rate C h is obtained by taking the data between the initial value of the capacitance curve and the plateau value for linear fitting, and the plateau value is the average value of the capacitance curve in the preset time interval after the capacitance curve enters the stable stage; Therefore, for the ith capacitor, the real-time aging response characteristic vector F i =[I s,i ,τ i ,C h,i ,C p,i ];F i 、I s,i 、τ i 、C h,i 、C p,i is constructed to represent the real-time aging response characteristic vector, the leakage current stable value, the leakage current falling time constant, the capacitance recovery rate, and the capacitance plateau value of the ith capacitor, respectively.
  6. 6. The intelligent detection method for production quality of capacitor according to claim 5, wherein said step S300 further comprises: Step S303, respectively solving the mean value of each feature according to the aging response data of the class history qualified sample to obtain a class standard response feature vector of the class aiming at the kth initial state class, respectively solving the standard deviation of each feature to obtain a corresponding feature standard deviation, and calculating the weighted Euclidean distance d i between the real-time aging response feature of the ith capacitor and the class standard response feature after normalizing each feature, wherein the calculation formula is as follows: ; Wherein w m represents the preset weight of the mth aging response characteristic, F i,m represents the mth real-time aging response characteristic value of the ith capacitor, F c,m represents the mth standard response characteristic value of the current category, sigma c,m represents the standard deviation of the mth characteristic of the current category, and then the aging matching degree M i =exp(-d i is calculated; Comparing the matching degree M i with a preset migration threshold M th , if M i ≥M th , judging that the aging process of the capacitor meets the current category expectations, and continuously maintaining the current category and the aging parameter scheme, if M i <M th , judging that the capacitor deviates from the current category track, and executing category dynamic migration; Step S304, calculating the matching degree between the real-time response feature vector F i and the standard response feature vectors of all the other classes of the capacitors to be migrated, selecting the class with the highest matching degree as a target class, automatically migrating the capacitor to the target class, switching to an aging parameter scheme corresponding to the target class, recalculating the aging parameters after switching according to the individual deviation degree of the capacitor in the target class, recording the time of migration, the original class, the target class and parameter adjustment data, continuously repeating the step S303 at fixed time intervals in the aging process until the personalized aging time of the capacitor is reached, and finally recording the complete class migration track of each capacitor in the whole aging process, including migration times, migration time, classes before and after migration and parameter adjustment histories of each stage.
  7. 7. The method for intelligently detecting the production quality of a capacitor according to claim 1, wherein the step S400 comprises the following steps: step S401, after the capacitor is subjected to personalized aging treatment, final electrical parameters of the capacitor are detected, wherein the final electrical parameters at least comprise capacitance, loss tangent and leakage current, the parameters are compared with the specification standard of the capacitor of the model, and if any parameter exceeds a qualified range, the capacitor is directly marked as a non-qualified product; Step S402, for all capacitors with qualified final electrical parameters, acquiring complete aging process data recorded in the step S300, wherein the complete aging process data comprises migration times N move , a lowest matching degree M min and an accumulated time length T move with the matching degree lower than a migration threshold M th ; Step S403, counting the comprehensive stability coefficients P i of all qualified capacitors, calculating the average value P avg and the standard deviation sigma P of the capacitors, and carrying out quality grade division according to the relation between the P i value and the overall distribution of each capacitor, wherein if the capacitor is P i ≤P avg -σ P , the capacitor is judged to be a superior product and is used for representing stable aging detection process, if the capacitor is P avg -σ P <P i ≤P avg +σ P , the capacitor is judged to be a qualified product and is used for representing the aging detection process in a normal fluctuation range, and if the capacitor is P i >P avg +σ P , the capacitor is judged to be a suspicious product and is used for representing that obvious abnormal fluctuation exists in the aging detection process, and retesting or short-term life test further confirmation is needed; And step S404, the quality grade judging result, the personalized aging file and the final electrical parameters of each capacitor are stored in a production quality database in a correlated manner. And when a plurality of unqualified products or suspicious products continuously appear in the subsequent production process and the pre-process characteristics are matched with the abnormal characteristics in the defect characteristic library, automatically triggering an early warning prompt to remind a craftsman to check whether the state or the process parameters of the pre-process equipment drift or not so as to adjust in time.
  8. 8. The intelligent detection system for the production quality of the capacitor is characterized by comprising an initial classification modeling module, a personalized parameter configuration module, an adaptive aging control module and a quality assessment module; the initial classification modeling module collects previous process data of each capacitor to be aged and carries out unique identification profiling on each capacitor; calculating a state stability index according to the previous process data, and dynamically dividing the capacitor into a plurality of initial state categories based on the state stability index; The personalized parameter configuration module configures a corresponding personalized initial aging parameter scheme for each capacitor by combining the statistical characteristics of each initial state category and the individual state stability index of each capacitor in the category, wherein the personalized initial aging parameter scheme comprises an aging temperature interval, an applied voltage range and a reference aging duration which are set differently; The self-adaptive aging control module is used for performing aging treatment on a single capacitor at an independent station based on a personalized initial aging parameter scheme, continuously collecting leakage current and capacitance data in an aging process to obtain real-time aging response characteristics, matching the real-time aging response characteristics with standard response characteristics preset in the category, and calculating to obtain aging matching degree; The quality evaluation module records the category migration track, the parameter adjustment process and the execution time of each stage of each capacitor to form an individual aging file, detects final electrical parameters after aging is finished, comprehensively judges by combining the individual aging file, the aging matching degree and the final electrical parameters, and outputs the quality grade of each capacitor.

Description

Intelligent production quality detection system and method applied to capacitor Technical Field The invention relates to the technical field of capacitor production quality detection, in particular to an intelligent production quality detection system and method applied to a capacitor. Background In the large-scale production process of the capacitor, the aging treatment is one of key procedures, and the oxidation film in the capacitor is repaired and perfected by applying rated voltage or slightly higher than rated voltage at a certain temperature, and meanwhile, the product with the congenital defect is caused to fail in advance so as to be removed. After aging treatment, the capacitor enters a final test link, and whether the capacitor is qualified or not is judged by measuring electric parameters such as capacitance, loss tangent, leakage current and the like. The existing aging detection method mainly adopts a batch processing mode, the capacitors with the same type are aged under the same temperature, voltage and time conditions, and then static electrical parameters after the aging is finished are used as qualification judgment basis. However, this approach ignores the dynamic effects of the state differences formed in the previous process for different capacitor individuals on the aging process. Although the inside of the partial capacitor has the potential defects of microcracks, winding looseness, insufficient impregnation and the like, the partial capacitor can still temporarily show that the electrical parameters are qualified under the uniform aging condition, and the actual internal damage has evolved in the aging process. Because the existing detection only pays attention to the final result, the dynamic characteristics of the leakage current change curve, capacitance recovery trend and the like reflecting the evolution of the individual state in the aging process can not be captured, so that the product is difficult to effectively identify, and the product becomes a main source of early failure in the use process of a user. Disclosure of Invention The invention aims to provide an intelligent detection system and method for production quality of a capacitor, which are used for solving the problems in the background technology. In order to solve the technical problems, the invention provides the following technical scheme that the intelligent detection method for the production quality of the capacitor comprises the following steps: step 100, collecting previous process data of each capacitor to be aged, and uniquely identifying and documenting each capacitor; calculating a state stability index according to the previous process data, and dynamically dividing the capacitor into a plurality of initial state categories based on the state stability index; step 200, configuring a corresponding personalized initial aging parameter scheme for each capacitor by combining the statistical characteristics of each initial state category and the individual state stability index of each capacitor in the category, wherein the personalized initial aging parameter scheme comprises an aging temperature interval, an applied voltage range and a reference aging duration which are differentially set; Step 300, based on a personalized initial aging parameter scheme, performing aging treatment on a single capacitor at an independent station, continuously collecting leakage current and capacitance data in an aging process to obtain real-time aging response characteristics, matching the real-time aging response characteristics with standard response characteristics preset in the category, and calculating to obtain aging matching degree; Step 400, recording class migration tracks, parameter adjustment processes and execution time lengths of all stages of each capacitor to form an individual aging file, detecting final electrical parameters after aging, comprehensively judging by combining the individual aging file, the aging matching degree and the final electrical parameters, and outputting the quality grade of each capacitor. Further, step S100 includes: Step S101, acquiring technological parameters of each capacitor to be aged in a previous process stage, including a winding tension value, an impregnation saturation value and an oxide film thickness value, on a capacitor production line through a production management system, uniquely identifying and profiling each capacitor, endowing each capacitor with a unique serial number, and associating and storing acquired original data with the serial number; The method comprises the steps of providing n capacitors to be aged, wherein each capacitor corresponds to three technological parameter characteristics, using x ij to represent the jth original characteristic value of the ith capacitor, wherein j=1, 2 and 3 correspond to winding tension, impregnation saturation and oxide film thickness respectively, firstly calculating the mean mu j and standard deviation sigma j of t