An intelligent monitoring system and method applied to the PCBA board processing flow

Abstract

The present invention discloses an intelligent monitoring system and method applied to the PCBA board processing flow, which relates to the technical field of intelligent monitoring. The system of the present invention includes: a historical data management module, a data analysis and deviation calculation module, a performance evaluation and mapping relationship construction module, a real-time data acquisition and correlation analysis module, and an anomaly prompt and decision support module; the historical data management module sorts out the historical data of the sensor to obtain target historical data; the data analysis and deviation calculation module analyzes the target historical data to obtain a deviation coefficient and a deviation coefficient trend curve; the performance evaluation and mapping relationship construction module evaluates the performance of the sensor and constructs a mapping relationship between the deviation coefficient trend curve and the performance evaluation index; the real-time data acquisition and correlation analysis module acquires real-time data and calculates the correlation degree between the real-time data and the target historical data segment; the anomaly prompt and decision support module performs corresponding processing according to the correlation result of the real-time data.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent monitoring, and specifically to an intelligent monitoring system and method applied to the PCBA board processing flow. Background Art

[0002] With the rapid development of electronic products, as the core component of the electronic manufacturing industry, PCBA occupies a crucial position in the production process of modern electronic products. The manufacturing process of PCBA usually includes links such as printing, chip mounting, soldering, and testing, and the accuracy and process requirements of each link are very high. Therefore, ensuring the quality of each link, avoiding defects, and optimizing the production process have become important goals in the industry.

[0003] With the continuous development of industrial automation and intelligent technologies, traditional manual monitoring methods have gradually been replaced by intelligent monitoring systems. These intelligent monitoring systems can collect key parameters (such as temperature, pressure, humidity, solder joint position, etc.) in the production process in real time through sensors, and through data processing and analysis, feedback the production status in real time. However, in practical applications, the existing intelligent monitoring systems for PCBA processing flow have the following problems: In the PCBA processing process, sensors are widely used to collect key physical quantity data (such as soldering temperature, pressure, solder joint position, etc.). However, as the use time of the sensors increases, the accuracy of the sensors may gradually decline, and over time, the aging problem of the sensors may become more obvious. This may lead to deviations and inaccuracies in the monitoring data. Existing systems usually adopt regular calibration or replacement of sensors to deal with the sensor aging problem, but this method has certain limitations, such as frequent maintenance costs and operation complexity, and it is difficult to effectively restore the performance and correct errors of sensors in long-term use, further increasing the probability of false alarms and missed alarms. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent monitoring system and method applied to the PCBA board processing flow to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the present invention provides the following technical solutions:

[0006] An intelligent monitoring method applied to the PCBA board processing flow includes the following steps:

[0007] Step S100. Obtain the historical data of the entire life cycle of the sensors in each processing link of the PCBA production line, as well as the corresponding event records; analyze the historical data of each sensor according to the event records to screen out the target historical data; for each processing link, divide the corresponding target historical data into several data segments;

[0008] Step S200. According to the partitioning result of the target historical data, compare and analyze each data segment with the corresponding standard process data segment to obtain a deviation coefficient; for each processing link, summarize the deviation coefficients corresponding to all data segments, and obtain the deviation coefficient trend curve of the corresponding target historical data;

[0009] Step S300. Based on the deviation coefficient trend curve of the target historical data, evaluate the performance of the sensors in the corresponding processing link to obtain a performance evaluation index; combine the deviation coefficient trend curve of the target historical data in the corresponding processing link and the performance evaluation index of the sensors to construct a mapping relationship between the two;

[0010] Step S400. Obtain the real-time data of the sensors in each processing link of the PCBA production line, analyze the correlation degree between the real-time data of each processing link and the corresponding data segment of the target historical data, so as to obtain the corresponding correlation degree coefficient; match the data segment of the target historical data according to the correlation degree coefficient, and output the corresponding prompt information based on the matching result.

[0011] Further, step S100 includes:

[0012] S101. For the sensors in each processing link, obtain the historical data of their entire life cycle and the corresponding event records; the historical data refers to the measurement data of the sensors, and the event records refer to the calibration or fault repair records of the sensors; for each sensor, summarize the corresponding historical data set D and event record set E, and the historical data set D = {d1, d2,..., dn}, where d1 represents the first data point, d2 represents the second data point, and so on, dn represents the nth data point, and n represents the number of data point numbers; the event record set E = {e1, e2,..., em}, where e1 represents the first event record, e2 represents the second event record, and similarly, em represents the mth event record, and m represents the number of event record numbers;

[0013] S102. Obtain the time period T in which the corresponding event record is located, and the time period T = [t1, t2], where t1 represents the starting point of the corresponding event record and t2 represents the ending point of the corresponding event record; according to the time period T, find the historical data in the same time period and eliminate it to obtain the target historical data; for the target historical data, divide the target historical data at consecutive time points into a data segment to obtain several data segments of consecutive time points; according to the time period T in which the event record is located, obtain the time stamp t0 corresponding to the last data point of each data segment, and calculate the difference δt by subtracting the time stamp t0 from the starting point of the time period where all event records are located, and δt = t1 - t0; if the difference δt is less than or equal to the threshold t, then associate the data segment where the time stamp t0 is located with the corresponding event record; according to the corresponding event category of the event record, classify all data segments into two categories: calibration and fault repair, which are respectively represented as the target historical data set DA and the target historical data set DB; and the target historical data set DA = {Da1, Da2,..., Dax}, where Da1 represents the target historical data segment corresponding to the first calibration event, Da2 represents the target historical data segment corresponding to the second calibration event, and so on, Dax represents the target historical data segment corresponding to the xth calibration event; similarly, the target historical data set DB is analyzed in the same way as the target historical data set DA, and the data segments corresponding to the target historical data set DA and the target historical data set DB are arranged in chronological order.

[0014] Further, step S200 includes:

[0015] S201. For the sensors in each processing step, define a standard process data set P, and P = {p1, p2,..., pk}, where p1 represents the first data point in the standard process flow, p2 represents the second data point in the standard process flow, and so on, pk represents the kth data point in the standard process flow, and k represents the data point number in the standard process flow; where the data points corresponding to the standard process data set are the values under ideal conditions; for the target historical data of the sensors in each processing step, according to the corresponding classification results, perform the following analysis on each data segment in the target historical data set DA and the target historical data set DB: for each data point di in the data segment Di and the corresponding standard process data point pj, where Di ∈ DA ∪ DB; calculate the corresponding deviation coefficient α, and the specific calculation formula is: α = [|di - (di + 1)| / |pj - (pj + 1)|] × |di - pj|;

[0016] S202. Aggregate the deviation coefficient α corresponding to each data point of each data segment in the target historical dataset DA and the target historical dataset DB, and for each data segment, draw a deviation coefficient trend curve in a plane rectangular coordinate system in chronological order. The horizontal axis of the plane rectangular coordinate system represents the data point number i, and the vertical axis represents the deviation coefficient α; thus, the deviation coefficient trend curve LDa and the deviation coefficient trend curve LDb corresponding to each data segment in the target historical dataset DA and the target historical dataset DB are obtained.

[0017] Further, step S300 includes:

[0018] S301. For the sensors of each processing link, perform the same analysis according to the deviation coefficient trend curve LDa and the deviation coefficient trend curve LDb corresponding to each data segment in the target historical dataset DA and the target historical dataset DB. Among them, for the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, the specific analysis process is as follows:

[0019] Extract the maximum value of the deviation coefficient α corresponding to each data segment in the target historical dataset DA, calculate the average value μa and the standard deviation σa, and obtain the deviation coefficient threshold Ta corresponding to the calibration event, and Ta = μa - g×σa, where g represents a constant; combined with the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, calculate the performance evaluation index Ra of the corresponding sensor, and the specific calculation formula is:

[0020] Ra=(1 / N)×∑u∈[1,N],Ku×[1-(αu / Ta)],

[0021] where Ku represents the slope of the u-th data point of the deviation coefficient trend curve LDa corresponding to the data segment, αu represents the deviation coefficient of the u-th data point of the deviation coefficient trend curve LDa corresponding to the data segment, and N represents the total number of data points of the deviation coefficient trend curve LDa corresponding to the data segment;

[0022] Similarly, referring to the above analysis process for the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, obtain the deviation coefficient threshold Tb corresponding to the fault repair event and the performance evaluation index Rb of the corresponding sensor;

[0023] S302. For each data segment in the target historical dataset DA and the target historical dataset DB, respectively construct the mapping relationships between the deviation coefficient trend curve LDa and the deviation coefficient trend curve LDb and the performance evaluation index Ra and the performance evaluation index Rb, and are respectively expressed as: Ga(LDa,Ra) and Gb(LDb,Rb).

[0024] Further, step S400 includes:

[0025] S401. Obtain the real-time data of the sensors at each processing link of the PCBA production line. For the real-time data Dr of each sensor, perform an association degree analysis with each data segment corresponding to the corresponding target historical data set DA and target historical data set DB. The specific analysis process is as follows:

[0026] Calculate the distance metric dist(Dr, Di) between the real-time data Dr and each data segment Di in the corresponding target historical data set DA and target historical data set DB, and dist(Dr, Di)=[∑v∈[1, V],(rv - dv)2]^(1 / 2), where V represents the total number of data points of the real-time data, rv represents the v-th data point of the real-time data Dr, and dv represents the v-th data point of the data segment Di; According to the distance metric dist(Dr, Di), calculate the association degree coefficient P, and P = 1 / [1 + dist(Dr, Di)];

[0027] S402. Aggregate the association degree coefficients P between the real-time data Dr and each data segment Di in the corresponding target historical data set DA and target historical data set DB, and select the data segment Di with the largest association degree coefficient P as the matching result of the current real-time data Dr; and obtain the value of the corresponding association degree coefficient P. If the association degree coefficient P is less than the threshold P0, no notification information is output;

[0028] If the association degree coefficient P is greater than or equal to the threshold P0, according to the data segment Di corresponding to the matching result, obtain the mapping relationship between the corresponding deviation coefficient trend curve and the performance evaluation index, and combine the analysis method of historical data to draw the deviation coefficient trend curve Lr corresponding to the real-time data. Take any point on the deviation coefficient trend curve of the data segment Di as the origin, and coincide the starting point of the deviation coefficient trend curve Lr with the origin, so that the deviation coefficient trend curve Lr slides sequentially on the data points of the deviation coefficient trend curve of the data segment Di. Stop when the number of coincident data points between the two is the largest. Calculate the time span T0 between the last coincident data point and the end point of the data segment Di, and output the time span T0, the event number corresponding to the data segment Di, and the performance evaluation index to the relevant personnel for further processing by the relevant personnel. Among them, the time span T0 is used as the prediction time of the calibration event or the fault repair event. According to the output time span T0, the relevant personnel perform corresponding processing based on the prediction time of the calibration event or the fault repair event; The performance evaluation index is used as a reference basis for subsequent feedback to help the relevant personnel make timely decisions and adjustments.

[0029] An intelligent monitoring system applied to the PCBA board processing flow, the system includes: a historical data management module, a data analysis and deviation calculation module, a performance evaluation and mapping relationship construction module, a real-time data acquisition and correlation analysis module, and an anomaly prompt and decision support module;

[0030] The historical data management module obtains the full life cycle data of sensors at each processing link of the PCBA production line, including measurement data and event records; sorts and classifies the historical data of sensors, constructs a historical data set, and manages event records; performs time period analysis on historical data, eliminates the target historical data affected by events, and performs segmented processing on the target historical data;

[0031] The data analysis and deviation calculation module compares the target historical data set with the standard process data, calculates the deviation coefficient of each data segment, and summarizes and generates a deviation coefficient trend curve of the target historical data;

[0032] The performance evaluation and mapping relationship construction module evaluates the sensor performance based on the deviation coefficient trend curve of the target historical data, calculates the performance evaluation index, and constructs a mapping relationship between the deviation coefficient trend curve and the performance evaluation index;

[0033] The real-time data acquisition and correlation analysis module obtains the real-time data of sensors at each processing link of the PCBA production line and performs preprocessing; calculates the correlation degree between the real-time data and the target historical data segment, calculates the correlation degree coefficient between the real-time data and the target historical data segment based on distance measurement, summarizes the correlation degree between the real-time data and the historical data segment, and selects the data segment with the largest correlation degree for real-time matching;

[0034] The anomaly prompt and decision support module performs corresponding processing according to the correlation result of the real-time data.

[0035] Furthermore, the historical data management module includes a data acquisition unit, a data preprocessing unit, and an event correlation unit;

[0036] The data acquisition unit obtains the historical data and event records of the full life cycle of sensors at each processing link; the data preprocessing unit cleans and preliminarily processes the collected historical data; the event correlation unit correlates the historical data with the corresponding event records according to the time period of the event records, and classifies them into a calibration event data set and a fault repair event data set.

[0037] Furthermore, the data analysis and deviation calculation module includes a data segmentation unit, a deviation calculation unit, and a trend curve drawing unit;

[0038] The data segmentation unit divides the target historical data into several consecutive time periods according to the event recording time period, and marks and classifies each data segment; the deviation calculation unit calculates the deviation coefficient of each data point from the standard process data point for each data segment; the trend curve plotting unit plots the deviation coefficient trend curve of each data segment according to the calculated deviation coefficient, so as to obtain the deviation coefficient trend curves of the target historical data sets DA and DB.

[0039] Further, the performance evaluation and mapping relationship construction module includes a threshold calculation unit, a performance evaluation unit, and a mapping relationship construction unit;

[0040] The threshold calculation unit calculates the deviation coefficient threshold for calibration and fault repair events; the performance evaluation unit calculates the performance evaluation index of each sensor based on the deviation coefficient trend curve; the mapping relationship construction unit establishes a mapping relationship according to the deviation coefficient trend curve and the performance evaluation index of each data segment.

[0041] Further, the real-time data acquisition and correlation analysis module includes a real-time data acquisition unit, a data matching unit, and a correlation analysis unit;

[0042] The real-time data acquisition unit obtains the real-time data of sensors at each processing link of the PCBA production line; the data matching unit calculates the distance metric between the real-time data and each data segment in the target historical data sets DA and DB, and calculates the corresponding correlation degree coefficient according to the distance metric of the data segment; the correlation analysis unit selects the historical data segment with the highest matching degree with the real-time data based on the correlation degree coefficient, and compares and analyzes the trend curves of the real-time data and the matching historical data segment.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows: By deeply analyzing the historical data and event records of the entire life cycle of the sensor, the present invention can better understand and evaluate the performance degradation trend of the sensor, avoiding the simple processing methods that only rely on regular calibration or replacement of sensors in traditional methods; By establishing the association between the historical data and event records of the sensor, the system can accurately track the performance changes of the sensor and make targeted corrections or adjustments, thereby reducing the measurement errors caused by sensor aging. In each processing link, by combining the comparison and analysis of the target historical data and the standard process data segment, a deviation coefficient trend curve of the sensor can be generated; Such a trend curve not only helps to identify the performance changes of the sensor, but also can predict potential problems in the processing process at the data segment level, avoiding the limitations of simply relying on instantaneous data or static thresholds to judge whether the equipment is normal in traditional methods. By analyzing the correlation degree between the real-time data and the target historical data segment, the present invention can evaluate the working state of the sensor in real time based on the matching results of the real-time monitoring data and the historical data; This method can effectively cope with the dynamic changes in complex production environments, help to discover potential abnormal or failure risks, and issue prompts or warnings in a timely manner, greatly reducing the occurrence probability of missed reports or false alarms. By combining the deviation coefficient trend curve of the historical data with the performance evaluation index of the sensor, the present invention can more comprehensively reflect the comprehensive performance of the sensor; And by establishing the mapping relationship between the deviation coefficient trend curve and the performance evaluation index, the system can more accurately evaluate the performance of the real-time data, further improving the reliability and stability in the production process. Compared with the traditional method of regularly calibrating or replacing sensors, the present invention provides a data-driven dynamic monitoring and warning mechanism, which can intervene before the sensor performance drops to the critical value; Such a monitoring method based on data analysis can not only improve the accuracy and reliability of sensor use, but also effectively reduce the high maintenance costs caused by frequent calibration or replacement of sensors, and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0045] Figure 1 It is a schematic diagram of the intelligent monitoring system module applied to the PCBA board processing flow of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] Please refer to Figure 1 , the present invention provides a technical solution:

[0048] An intelligent monitoring system applied to the PCBA board processing flow, the system includes: a historical data management module, a data analysis and deviation calculation module, a performance evaluation and mapping relationship construction module, a real-time data acquisition and correlation analysis module, and an exception prompt and decision support module;

[0049] The historical data management module obtains the full life cycle data of sensors at each processing link of the PCBA production line, including measurement data and event records; sorts and classifies the historical data of the sensors, constructs a historical data set, and manages the event records; performs time period analysis on the historical data, eliminates the target historical data affected by events, and performs segmentation processing on the target historical data;

[0050] The data analysis and deviation calculation module compares the target historical data set with the standard process data, calculates the deviation coefficient of each data segment, and summarizes and generates a deviation coefficient trend curve of the target historical data;

[0051] The performance evaluation and mapping relationship construction module evaluates the sensor performance based on the deviation coefficient trend curve of the target historical data, calculates the performance evaluation index, and constructs a mapping relationship between the deviation coefficient trend curve and the performance evaluation index;

[0052] The real-time data acquisition and correlation analysis module obtains the real-time data of sensors at each processing link of the PCBA production line and performs preprocessing; by calculating the correlation degree between the real-time data and the target historical data segment, calculates the correlation degree coefficient between the real-time data and the target historical data segment based on distance measurement, summarizes the correlation degree between the real-time data and the historical data segment, and selects the data segment with the largest correlation degree for real-time matching;

[0053] The exception prompt and decision support module performs corresponding processing according to the correlation result of the real-time data.

[0054] The historical data management module includes a data acquisition unit, a data preprocessing unit, and an event correlation unit;

[0055] The data acquisition unit obtains the historical data and event records of the entire life cycle of the sensors in each processing link; the data preprocessing unit cleans and preliminarily processes the collected historical data; the event correlation unit correlates the historical data with the corresponding event records according to the time period of the event records, and classifies them into a calibration event data set and a fault repair event data set.

[0056] The data analysis and deviation calculation module includes a data segmentation unit, a deviation calculation unit, and a trend curve drawing unit;

[0057] The data segmentation unit divides the target historical data into several continuous time periods according to the event record time period, and marks and classifies each data segment; the deviation calculation unit calculates the deviation coefficient of each data point from the standard process data point for each data segment; the trend curve drawing unit draws the deviation coefficient trend curve of each data segment according to the calculated deviation coefficient, so as to obtain the deviation coefficient trend curves of the target historical data sets DA and DB.

[0058] The performance evaluation and mapping relationship construction module includes a threshold calculation unit, a performance evaluation unit, and a mapping relationship construction unit;

[0059] The threshold calculation unit calculates the deviation coefficient thresholds for calibration and fault repair events; the performance evaluation unit calculates the performance evaluation index of each sensor based on the deviation coefficient trend curve; the mapping relationship construction unit establishes a mapping relationship according to the deviation coefficient trend curve and the performance evaluation index of each data segment.

[0060] The real-time data acquisition and correlation analysis module includes a real-time data acquisition unit, a data matching unit, and a correlation analysis unit;

[0061] The real-time data acquisition unit obtains the real-time data of the sensors in each processing link of the PCBA production line; the data matching unit calculates the distance metric between the real-time data and each data segment in the target historical data sets DA and DB, and calculates the corresponding correlation degree coefficient according to the distance metric of the data segment; the correlation analysis unit selects the historical data segment with the highest matching degree with the real-time data based on the correlation degree coefficient, and compares and analyzes the trend curves of the real-time data and the matching historical data segment.

[0062] An intelligent monitoring method applied to the PCBA board processing process includes the following steps:

[0063] Step S100. Obtain the historical data of the entire life cycle of the sensors in each processing link of the PCBA production line, and the corresponding event records; analyze the historical data of each sensor according to the event records, so as to screen out the target historical data; for each processing link, divide the corresponding target historical data into several data segments;

[0064] Step S200. According to the division result of the target historical data, compare and analyze each data segment with the corresponding standard process data segment to obtain the deviation coefficient; for each processing link, summarize the deviation coefficients corresponding to all data segments and obtain the deviation coefficient trend curve of the corresponding target historical data;

[0065] Step S300. Based on the deviation coefficient trend curve of the target historical data, evaluate the performance of the sensors in the corresponding processing link to obtain the performance evaluation index; combine the deviation coefficient trend curve of the target historical data in the corresponding processing link and the performance evaluation index of the sensors to construct the mapping relationship between the two;

[0066] Step S400. Obtain the real-time data of the sensors in each processing link of the PCBA production line, analyze the correlation degree between the real-time data of each processing link and the corresponding data segments of the target historical data to obtain the corresponding correlation degree coefficient; match the data segments of the target historical data according to the correlation degree coefficient and output the corresponding prompt information based on the matching result.

[0067] Step S100 includes:

[0068] S101. For the sensors in each processing link, obtain the historical data of their entire life cycle and the corresponding event records; the historical data refers to the measurement data of the sensors, and the event records refer to the calibration or fault repair records of the sensors; for each sensor, summarize the corresponding historical data set D and event record set E, and the historical data set D = {d1, d2,..., dn}, where d1 represents the first data point, d2 represents the second data point, and so on, dn represents the nth data point, and n represents the number of data point numbers; the event record set E = {e1, e2,..., em}, where e1 represents the first event record, e2 represents the second event record, and similarly, em represents the mth event record, and m represents the number of event record numbers;

[0069] S102. Obtain the time period T in which the corresponding event record is located, and the time period T = [t1, t2], where t1 represents the starting point of the corresponding event record and t2 represents the ending point of the corresponding event record; according to the time period T, find the historical data in the same time period and eliminate it to obtain the target historical data; for the target historical data, divide the target historical data at consecutive time points into a data segment to obtain several data segments of consecutive time points; according to the time period T in which the event record is located, obtain the time stamp t0 corresponding to the last data point of each data segment, and calculate the difference between the time stamp t0 and the starting point of the time period where all event records are located respectively to obtain the difference δt, and δt = t1 - t0; if the difference δt is less than or equal to the threshold t, associate the data segment where the time stamp t0 is located with the corresponding event record; according to the corresponding event category of the event record, classify all data segments into two categories: calibration and fault repair, which are respectively represented as the target historical data set DA and the target historical data set DB; and the target historical data set DA = {Da1, Da2,..., Dax}, where Da1 represents the target historical data segment corresponding to the first calibration event, Da2 represents the target historical data segment corresponding to the second calibration event, and so on, Dax represents the target historical data segment corresponding to the xth calibration event; similarly, the target historical data set DB is analyzed in the same way as the target historical data set DA, and the data segments corresponding to the target historical data set DA and the target historical data set DB are all arranged in chronological order.

[0070] Step S200 includes:

[0071] S201. For the sensors in each processing step, define a standard process data set P, and P = {p1, p2,..., pk}, where p1 represents the first data point in the standard process flow, p2 represents the second data point in the standard process flow, and so on, pk represents the kth data point in the standard process flow, and k represents the data point number in the standard process flow; where the data points corresponding to the standard process data set are the values under ideal conditions; for the target historical data of the sensors in each processing step, according to the corresponding classification results, for each data segment in the target historical data set DA and the target historical data set DB, perform the following analysis: for each data point di in the data segment Di and the corresponding standard process data point pj, where Di ∈ DA ∪ DB; calculate the corresponding deviation coefficient α, and the specific calculation formula is: α = [|di - (di + 1)| / |pj - (pj + 1)|] × |di - pj|;

[0072] S202. Aggregate the deviation coefficients α of the corresponding data points of each data segment in the target historical dataset DA and the target historical dataset DB, and for each data segment, draw a deviation coefficient trend curve in a plane rectangular coordinate system in chronological order. The horizontal axis of the plane rectangular coordinate system represents the data point number i, and the vertical axis represents the deviation coefficient α; thus, the deviation coefficient trend curves LDa and LDb corresponding to each data segment in the target historical dataset DA and the target historical dataset DB are obtained.

[0073] Step S300 includes:

[0074] S301. For the sensors of each processing link, the same analysis is performed according to the deviation coefficient trend curves LDa and LDb corresponding to each data segment in the target historical dataset DA and the target historical dataset DB. Among them, for the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, the specific analysis process is as follows:

[0075] Extract the maximum value of the deviation coefficient α corresponding to each data segment in the target historical dataset DA, calculate the average value μa and the standard deviation σa, and obtain the deviation coefficient threshold Ta corresponding to the calibration event, and Ta = μa - g×σa, where g represents a constant; combined with the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, calculate the performance evaluation index Ra of the corresponding sensor, and the specific calculation formula is:

[0076] Ra=(1 / N)×∑u∈[1,N],Ku×[1-(αu / Ta)],

[0077] where Ku represents the slope of the u-th data point of the deviation coefficient trend curve LDa corresponding to the data segment, αu represents the deviation coefficient of the u-th data point of the deviation coefficient trend curve LDa corresponding to the data segment, and N represents the total number of data points of the deviation coefficient trend curve LDa corresponding to the data segment;

[0078] Similarly, referring to the above analysis process for the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, obtain the deviation coefficient threshold Tb corresponding to the fault repair event and the performance evaluation index Rb of the corresponding sensor;

[0079] S302. For each data segment in the target historical dataset DA and the target historical dataset DB, construct the mapping relationships between the deviation coefficient trend curves LDa and LDb and the performance evaluation indices Ra and Rb, and are respectively expressed as: Ga(LDa,Ra) and Gb(LDb,Rb).

[0080] Step S400 includes:

[0081] S401. Obtain the real-time data of the sensors at each processing link of the PCBA production line. For the real-time data Dr of each sensor, perform an association degree analysis with each data segment corresponding to the corresponding target historical data set DA and target historical data set DB. The specific analysis process is as follows:

[0082] Calculate the distance metric dist(Dr, Di) between the real-time data Dr and each data segment Di in the corresponding target historical data set DA and target historical data set DB, and dist(Dr, Di) = [∑v∈[1,V],(rv - dv) 2 ^(1 / 2), where V represents the total number of data points of the real-time data, rv represents the v-th data point of the real-time data Dr, and dv represents the v-th data point of the data segment Di; according to the distance metric dist(Dr, Di), calculate the association degree coefficient P, and P = 1 / [1 + dist(Dr, Di)];

[0083] S402. Aggregate the association degree coefficients P between the real-time data Dr and each data segment Di in the corresponding target historical data set DA and target historical data set DB, and select the data segment Di with the largest association degree coefficient P as the matching result of the current real-time data Dr; and obtain the value of the corresponding association degree coefficient P. If the association degree coefficient P is less than the threshold P0, no notification information is output;

[0084] If the association degree coefficient P is greater than or equal to the threshold P0, according to the data segment Di corresponding to the matching result, obtain the mapping relationship between the corresponding deviation coefficient trend curve and the performance evaluation index, and combine the analysis method of historical data to draw the deviation coefficient trend curve Lr corresponding to the real-time data. Take any point on the deviation coefficient trend curve of the data segment Di as the origin, and coincide the starting point of the deviation coefficient trend curve Lr with the origin, so that the deviation coefficient trend curve Lr slides sequentially on the data points of the deviation coefficient trend curve of the data segment Di. Stop when the number of coincident data points between the two is the largest. Calculate the time span T0 between the last coincident data point and the end point of the data segment Di, and output the time span T0, the event number corresponding to the data segment Di, and the performance evaluation index to the relevant personnel for further processing. Among them, the time span T0 is used as the prediction time of the calibration event or the fault repair event. According to the output time span T0, the relevant personnel perform corresponding processing based on the prediction time of the calibration event or the fault repair event; for the performance evaluation index, it is used as a reference basis for subsequent feedback to help the relevant personnel make timely decisions and adjustments.

[0085] In this embodiment, for the best matching historical data segment, specifically:

[0086] For each data segment Di in the real-time data Dr and the historical data sets DA and DB, calculate the correlation degree coefficient P of all data segments; select the historical data segment Di with the largest correlation degree coefficient P as the best data segment matching the real-time data Dr; if the largest correlation degree coefficient P is less than the preset threshold P0, it is considered that the correlation between the real-time data and the historical data is low, and no notification information is output. If the largest correlation degree coefficient P is greater than or equal to the threshold P0, perform the subsequent analysis steps; obtain the deviation coefficient trend curve corresponding to the selected historical data segment Di, slide the deviation coefficient trend curve Lr on the deviation coefficient trend curve Li of the historical data segment Di for matching, select a starting point and make it coincide with the deviation coefficient trend curve of the historical data segment, and through sliding, make the number of data points where Lr and Li coincide as large as possible. When the number of coincident data points between the two is the largest, stop sliding; calculate the time span T0 between the last coincident data point and the end point of the data segment Di.

[0087] Suppose on a PCBA production line, the real-time data monitored by a certain sensor corresponds to temperature changes. According to the matching between the deviation coefficient trend curve Lr of the historical data segment Di and the real-time data, the following results are obtained:

[0088] When the number of data points where the deviation coefficient trend curve Lr coincides with the deviation coefficient trend curve Li of the historical data segment Di is the largest, the historical data segment Di is a calibration event. Calculate the time span T0 between the last coincident data point and the end point of the data segment Di. Suppose the time span T0 is 4 hours, indicating that it is expected that the calibration event will occur after 4 hours. Through analysis, obtain the performance evaluation index Ra, and then output the time span T0 of 4 hours and the performance evaluation index Ra to the relevant personnel for corresponding handling by the relevant personnel.

[0089] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0090] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent monitoring method applied to the PCBA board processing flow, characterized in that: The method includes the following steps: Step S100. Obtain the historical data of each sensor in the entire life cycle in the PCBA production line, as well as the corresponding event records; analyze the historical data of each sensor according to the event records to filter out the target historical data; for each processing link, divide the corresponding target historical data into several data segments; Step S200. According to the division result of the target historical data, compare and analyze each data segment with the corresponding standard process data segment to obtain the deviation coefficient; for each processing link, summarize the deviation coefficients corresponding to all data segments and obtain the deviation coefficient trend curve of the corresponding target historical data; Step S300. Based on the deviation coefficient trend curve of the target historical data, evaluate the performance of the sensor in the corresponding processing link to obtain the performance evaluation index; combine the deviation coefficient trend curve of the target historical data in the corresponding processing link and the performance evaluation index of the sensor to construct the mapping relationship between the two; Step S400. Obtain the real-time data of the sensors in each processing link of the PCBA production line, analyze the correlation degree between the real-time data of each processing link and the corresponding data segments of the target historical data to obtain the corresponding correlation degree coefficient; match the data segments of the target historical data according to the correlation degree coefficient and output the corresponding prompt information based on the matching result.

2. The intelligent monitoring method applied to the PCBA board processing flow according to claim 1, wherein: The said step S100 includes: S101. For the sensors in each processing link, obtain their historical data in the entire life cycle and the corresponding event records; the historical data refers to the measurement data of the sensors, and the event records refer to the calibration or fault repair records of the sensors; for each sensor, summarize the corresponding historical data set D and event record set E, and the historical data set D = {d1, d2,..., dn}, where d1 represents the first data point, d2 represents the second data point, and so on, dn represents the nth data point, and n represents the number of data point numbers; the event record set E = {e1, e2,..., em}, where e1 represents the first event record, e2 represents the second event record, and similarly, em represents the mth event record, and m represents the number of event record numbers; S102. Obtain the time period T in which the corresponding event record is located, and the time period T = [t1, t2], where t1 represents the starting point of the corresponding event record and t2 represents the ending point of the corresponding event record; according to the time period T, find the historical data in the same time period and eliminate it to obtain the target historical data; for the target historical data, divide the target historical data at consecutive time points into a data segment to obtain several data segments of consecutive time points; according to the time period T in which the event record is located, obtain the time stamp t0 corresponding to the last data point of each data segment, and calculate the difference by subtracting the time stamp t0 from the starting point of the time period where all event records are located, resulting in the difference δt, and δt = t1 - t0; if the difference δt is less than or equal to the threshold t, then associate the data segment where the time stamp t0 is located with the corresponding event record; according to the corresponding event category of the event record, classify all data segments into two categories: calibration and fault repair, respectively represented as the target historical data set DA and the target historical data set DB; and the target historical data set DA = {Da1, Da2,..., Dax}, where Da1 represents the target historical data segment corresponding to the first calibration event, Da2 represents the target historical data segment corresponding to the second calibration event, and so on, Dax represents the target historical data segment corresponding to the xth calibration event; similarly, the target historical data set DB is analyzed in the same way as the target historical data set DA, and the data segments corresponding to the target historical data set DA and the target historical data set DB are arranged in chronological order.

3. An intelligent monitoring method applied to the PCBA board processing flow according to claim 2, characterized in that: The step S200 includes: S201. For the sensors in each processing step, define a standard process data set P, and P = {p1, p2,..., pk}, where p1 represents the first data point in the standard process flow, p2 represents the second data point in the standard process flow, and so on, pk represents the kth data point in the standard process flow, and k represents the data point number in the standard process flow; for the target historical data of the sensors in each processing step, according to the corresponding classification results, perform the following analysis on each data segment in the target historical data set DA and the target historical data set DB: for each data point di in the data segment Di and the corresponding standard process data point pj, where Di ∈ DA ∪ DB; calculate the corresponding deviation coefficient α, and the specific calculation formula is: α = [|di - (di + 1)| / |pj - (pj + 1)|] × |di - pj|; S202. Summarize the deviation coefficients α corresponding to the data points of each data segment in the target historical data set DA and the target historical data set DB, and draw a deviation coefficient trend curve in the plane rectangular coordinate system for each data segment in chronological order. The horizontal axis of the plane rectangular coordinate system represents the data point number i, and the vertical axis represents the deviation coefficient α; thus, the deviation coefficient trend curve LDa and the deviation coefficient trend curve LDb corresponding to each data segment in the target historical data set DA and the target historical data set DB are obtained.

4. An intelligent monitoring method applied to the PCBA board processing flow according to claim 3, characterized in that: The step S300 includes: S301. For the sensors in each processing step, the same analysis is performed according to the deviation coefficient trend curves LDa and LDb corresponding to each data segment in the target historical dataset DA and the target historical dataset DB. Among them, for the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, the specific analysis process is as follows: Extract the maximum value of the deviation coefficient α corresponding to each data segment in the target historical dataset DA, calculate the average value μa and the standard deviation σa, and obtain the deviation coefficient threshold Ta corresponding to the calibration event, and Ta = μa - g×σa, where g represents a constant; combine the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, and calculate the performance evaluation index Ra of the corresponding sensor. The specific calculation formula is: Ra=(1 / N)×∑u∈[1,N],Ku×[1-(αu / Ta)], where Ku represents the slope of the u-th data point of the deviation coefficient trend curve LDa corresponding to the data segment, αu represents the deviation coefficient of the u-th data point of the deviation coefficient trend curve LDa corresponding to the data segment, and N represents the total number of data points of the deviation coefficient trend curve LDa corresponding to the data segment; Similarly, referring to the above analysis process for the deviation coefficient trend curve LDa corresponding to each data segment in the target historical dataset DA, obtain the deviation coefficient threshold Tb corresponding to the fault repair event and the performance evaluation index Rb of the corresponding sensor; S302. For each data segment in the target historical dataset DA and the target historical dataset DB, establish the mapping relationships between the deviation coefficient trend curves LDa and LDb and the performance evaluation indexes Ra and Rb, and are respectively expressed as: Ga(LDa,Ra) and Gb(LDb,Rb).

5. An intelligent monitoring method applied to the PCBA board processing flow according to claim 4, characterized in that: The step S400 includes: S401. Obtain the real-time data of the sensors in each processing step of the PCBA production line. For the real-time data Dr of each sensor, perform the correlation degree analysis with each data segment corresponding to the corresponding target historical dataset DA and the target historical dataset DB. The specific analysis process is as follows: Calculate the distance metric dist(Dr, Di) between the real-time data Dr and each data segment Di in the corresponding target historical data sets DA and DB, and dist(Dr, Di) = [∑v∈[1, V], (rv - dv) 2 ^(1 / 2), where V represents the total number of data points in the real-time data, rv represents the v-th data point of the real-time data Dr, and dv represents the v-th data point of the data segment Di; according to the distance metric dist(Dr, Di), calculate the correlation degree coefficient P, and P = 1 / [1 + dist(Dr, Di)]; S402. Summarize the correlation degree coefficient P between the real-time data Dr and each data segment Di in the corresponding target historical dataset DA and the target historical dataset DB, and select the data segment Di with the largest correlation degree coefficient P as the matching result of the current real-time data Dr; and obtain the value of the corresponding correlation degree coefficient P. If the correlation degree coefficient P is less than the threshold P0, no notification information is output; If the correlation degree coefficient P is greater than or equal to the threshold P0, then according to the data segment Di corresponding to the matching result, obtain the mapping relationship between the corresponding deviation coefficient trend curve and the performance evaluation index, and combine the analysis method of historical data to draw the deviation coefficient trend curve Lr corresponding to the real-time data. Take any point on the deviation coefficient trend curve of the data segment Di as the origin, and make the starting point of the deviation coefficient trend curve Lr coincide with the origin, so that the deviation coefficient trend curve Lr slides sequentially on the data points of the deviation coefficient trend curve of the data segment Di. Stop when the number of coincident data points between the two is the largest. Calculate the time span T0 between the last coincident data point and the end point of the data segment Di, and output the time span T0, the event number corresponding to the data segment Di, and the performance evaluation index to the relevant personnel for further processing by the relevant personnel.

6. An intelligent monitoring system applied to the PCBA board processing flow, which is applied to an intelligent monitoring method for the PCBA board processing flow described in any one of claims 1-5, characterized in that: The system includes: a historical data management module, a data analysis and deviation calculation module, a performance evaluation and mapping relationship construction module, a real-time data acquisition and correlation analysis module, and an anomaly prompt and decision support module; The historical data management module obtains the full life cycle data of the sensors in each processing link of the PCBA production line, including measurement data and event records; sorts and classifies the historical data of the sensors, constructs a historical data set, and manages the event records; performs time period analysis on the historical data, eliminates the target historical data after the event impact, and performs segmentation processing on the target historical data; The data analysis and deviation calculation module compares the target historical data set with the standard process data, calculates the deviation coefficient of each data segment, and summarizes and generates the deviation coefficient trend curve of the target historical data; The performance evaluation and mapping relationship construction module evaluates the sensor performance based on the deviation coefficient trend curve of the target historical data, calculates the performance evaluation index, and constructs the mapping relationship between the deviation coefficient trend curve and the performance evaluation index; The real-time data acquisition and correlation analysis module obtains the real-time data of the sensors in each processing link of the PCBA production line and performs preprocessing; calculates the correlation degree between the real-time data and the target historical data segment, calculates the correlation degree coefficient between the real-time data and the target historical data segment based on distance measurement, summarizes the correlation degree between the real-time data and the historical data segment, and selects the data segment with the largest correlation degree for real-time matching; The anomaly prompt and decision support module performs corresponding processing according to the correlation result of the real-time data.

7. An intelligent monitoring system applied to the PCBA board processing flow according to claim 6, characterized in that: The historical data management module includes a data acquisition unit, a data preprocessing unit, and an event correlation unit; The data acquisition unit obtains the historical data and event records of the full life cycle of the sensors in each processing link; the data preprocessing unit cleans and preliminarily processes the collected historical data; The event correlation unit correlates the historical data with the corresponding event records according to the time period of the event records, and classifies them into a calibration event data set and a fault repair event data set.

8. An intelligent monitoring system applied to the PCBA board processing flow according to claim 6, characterized in that: The data analysis and deviation calculation module includes a data segmentation unit, a deviation calculation unit, and a trend curve drawing unit; The data segmentation unit divides the target historical data into several consecutive time periods according to the event recording time period, and marks and classifies each data segment; The deviation calculation unit calculates the deviation coefficient of each data point from the standard process data point for each data segment; the trend curve drawing unit draws the deviation coefficient trend curve of each data segment according to the calculated deviation coefficient, so as to obtain the deviation coefficient trend curves of the target historical data sets DA and DB.

9. An intelligent monitoring system applied to the PCBA board processing flow according to claim 6, characterized in that: The performance evaluation and mapping relationship construction module includes a threshold calculation unit, a performance evaluation unit, and a mapping relationship construction unit; The threshold calculation unit calculates the deviation coefficient threshold for calibration and fault repair events; the performance evaluation unit calculates the performance evaluation index of each sensor based on the deviation coefficient trend curve; The mapping relationship construction unit establishes a mapping relationship according to the deviation coefficient trend curve and the performance evaluation index of each data segment.

10. An intelligent monitoring system applied to the PCBA board processing flow according to claim 6, characterized in that: The real-time data acquisition and correlation analysis module includes a real-time data acquisition unit, a data matching unit, and a correlation analysis unit; The real-time data acquisition unit acquires the real-time data of sensors at each processing link of the PCBA production line; the data matching unit calculates the distance metric between the real-time data and each data segment in the target historical data sets DA and DB, and calculates the corresponding correlation degree coefficient according to the distance metric of the data segment; The correlation analysis unit selects the historical data segment with the highest matching degree with the real-time data based on the correlation degree coefficient, and compares and analyzes the trend curves of the real-time data and the matching historical data segment.

Images