Circuit board production control method and system based on data monitoring

Abstract

The embodiment of the invention discloses a circuit board production control method and system based on data monitoring, and the method comprises the steps: collecting a real-time monitoring data set of a target circuit board production line, the real-time monitoring data set comprising a plurality of continuous time sequence process parameter monitoring data and equipment operation monitoring data; performing multi-dimensional time sequence feature extraction processing on the real-time monitoring data set, and generating a production time sequence state feature and a process fluctuation correlation feature of the target circuit board production line; calling a production decision model to perform state fusion processing on the production time sequence state characteristics and the process fluctuation correlation characteristics, and generating production control decision information of the target circuit board production line; and generating an equipment control instruction set according to the production control decision information, and transmitting the equipment control instruction set to an execution terminal of the target circuit board production line to realize production state adjustment operation. Therefore, accurate control over the circuit board production process can be achieved, and the stability and efficiency of production are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of data processing technology, and in particular to a circuit board production control method and system based on data monitoring. Background Art

[0002] In the field of circuit board production, the quality and production efficiency of circuit boards play a key role in the performance and cost of electronic products. With the diversification and high-performance demands of electronic products, the circuit board production process has become increasingly complex, and parameter control and equipment operating status monitoring during the production process have become increasingly important. However, existing circuit board production control technologies often only capture a single type of data during data collection, and are unable to comprehensively and continuously monitor the production line's process parameters and equipment operating status. This leads to an incomplete understanding of the production process and difficulty in identifying potential problems. During the production decision-making process, existing technologies lack effective multi-dimensional data analysis and fusion capabilities, and are unable to fully explore the inherent connections between data, resulting in inaccurate production decisions and difficulty adapting to complex and changing production needs. In addition, traditional technologies lack a dynamic adjustment mechanism based on real-time data analysis for production status adjustments, making it difficult to make timely and accurate adjustments based on the actual production process, affecting the stability and production efficiency of the production line. Summary of the Invention

[0003] The embodiments of the present invention provide a circuit board production control method and system based on data monitoring, which is used to achieve precise control of the circuit board production process and improve production stability and efficiency through comprehensive collection of real-time monitoring data, multi-dimensional feature extraction, state fusion decision-making, and precise equipment control instruction transmission.

[0004] In the first aspect, an embodiment of the present invention provides a circuit board production control method based on data monitoring, which is applied to a circuit board production control system, the method comprising: collecting a real-time monitoring data set of a target circuit board production line, the real-time monitoring data set comprising a plurality of continuous time series of process parameter monitoring data and equipment operation monitoring data; performing multi-dimensional time series feature extraction processing on the real-time monitoring data set to generate production time series state features and process fluctuation correlation features of the target circuit board production line; calling a production decision model to perform state fusion processing on the production time series state features and the process fluctuation correlation features to generate production control decision information of the target circuit board production line; generating an equipment control instruction set according to the production control decision information, and transmitting the equipment control instruction set to the execution terminal of the target circuit board production line to implement production state adjustment operations.

[0005] In a second aspect, an embodiment of the present invention provides a circuit board production control system, comprising: processor; a storage device having a computer program stored thereon, When the computer program is executed by the processor, the processor implements any one of the circuit board production control methods based on data monitoring.

[0006] An embodiment of the present invention provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, the steps of the circuit board production control method based on data monitoring are implemented.

[0007] It can be seen that the embodiments of the present invention have the following beneficial effects: The embodiments of the present invention perform multi-dimensional time series feature extraction processing on the collected real-time monitoring data set of the target circuit board production line, generate production time series state features and process fluctuation correlation features, can deeply analyze the operation characteristics of the production line from different dimensions, accurately extract key information, and provide a comprehensive and detailed basis for production decisions. Subsequently, the production decision model is called to perform state fusion processing on the above two features, fully mining the potential connections between the features, generating accurate production control decision information, which can comprehensively consider various factors and make the decision more scientific and reasonable. Finally, an equipment control instruction set is generated according to the production control decision information and transmitted to the execution terminal to realize the precise adjustment of the production state, effectively improving the stability and production efficiency of the production line, ensuring the smooth progress of the circuit board production process, and comprehensively optimizing the circuit board production control process.

[0008] In summary, the embodiments of the present invention aim to solve the problems in the prior art such as incomplete data collection, inaccurate decision-making, and untimely adjustment of the production state. Through comprehensive collection of real-time monitoring data, multi-dimensional feature extraction, state fusion decision-making, and accurate transmission of equipment control instructions, precise control of the circuit board production process is realized, and the stability and efficiency of production are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a flowchart of a circuit board production control method based on data monitoring provided by an embodiment of the present invention.

[0010] Figure 2 It is a schematic diagram of the basic structure of a circuit board production control system provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0011] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0012] See Figure 1 As shown, this figure is a flowchart of a circuit board production control method based on data monitoring provided by an embodiment of the present invention, and this method can be applied to a circuit board production control system. As Figure 1As shown, the method includes steps 110 - 140.

[0013] Step 110: Collect a real - time monitoring data set of the target circuit board production line, where the real - time monitoring data set includes process parameter monitoring data and equipment operation monitoring data of multiple consecutive time series.

[0014] In an embodiment of the present invention, taking the circuit board production line of an electronics factory as an example, this production line produces multiple types of circuit boards, and various parameters need to be monitored in real time during production. For the process parameter monitoring data, it covers parameters such as etching solution concentration, electroplating current intensity, drilling depth, etc. These parameters continuously change over time during production. For example, the etching solution concentration may be monitored every 10 seconds, and each monitoring obtains a value, forming a continuous time - series monitoring data sequence. For the equipment operation monitoring data, it includes the operation speed of the mounter, the temperature change of the reflow soldering furnace, etc. The operation speed of the mounter is measured by the number of components placed per minute and is also collected at regular time intervals to form the time - series data of the equipment operation state. By arranging sensors and data acquisition devices at key positions on the production line, these process parameters and equipment operation data are collected and summarized in real time, thereby obtaining a real - time monitoring data set that includes process parameter monitoring data and equipment operation monitoring data of multiple consecutive time series.

[0015] Step 120: Perform multi - dimensional time - series feature extraction processing on the real - time monitoring data set to generate the production time - series state features and process fluctuation correlation features of the target circuit board production line.

[0016] Next, the collected real - time monitoring data set is deeply processed to obtain the key feature information of the production line. This process involves analyzing the data from multiple dimensions to generate production time - series state features and process fluctuation correlation features that can reflect the operation state of the production line.

[0017] As an optional embodiment, step 120 includes: Step 121: Perform data cleaning processing on the real - time monitoring data set to obtain an optimized monitoring data set; the data cleaning processing includes: replacing abnormal data points in the process parameter monitoring data with adjacent time - series data, and aligning non - continuous acquisition data in the equipment operation monitoring data with a time window.

[0018] In an embodiment of the present invention, for process parameter monitoring data, such as in an etching solution concentration data sequence, if the concentration value monitored at a certain moment differs significantly from the values at the previous and subsequent moments and clearly does not fall within the normal production fluctuation range, the data point is determined to be an abnormal data point. At this time, it is replaced with the value of the previous or subsequent normal data point adjacent to the abnormal data point. For equipment operation monitoring data, such as reflow oven temperature acquisition data, there may be a situation where the acquisition time intervals are inconsistent due to certain reasons. At this time, the time window alignment method is adopted, a suitable time window is set, and the data collected at different times is aligned according to the time window to ensure the consistency and comparability of the data in the time dimension, thereby obtaining an optimized monitoring data set.

[0019] Step 122: Perform a time series fluctuation analysis process on the optimized monitoring data set to identify the process parameter fluctuation characteristics and equipment operation state fluctuation characteristics of the target circuit board production line in multiple consecutive production cycles.

[0020] In this step, a time series fluctuation analysis is further performed on the optimized monitoring data set to gain a deeper understanding of the operation changes of the production line in different production cycles.

[0021] In one implementation, step 122 includes: Step 1221: Divide the optimized monitoring data set into a process parameter time series and an equipment operation state time series.

[0022] In an embodiment of the present invention, based on the attributes of the data, the optimized monitoring data set is clearly divided into two series. The data related to process parameters, such as etching solution concentration, electroplating current intensity, etc., are combined to form a process parameter time series; the data related to the equipment operation state, such as the running speed of the mounter, the temperature of the reflow oven, etc., are combined to form an equipment operation state time series, which facilitates subsequent targeted analysis.

[0023] Step 1222: Perform a sliding window segmentation process on the process parameter time series to obtain multiple process parameter window subsequences, and perform trend fitting on each process parameter window subsequence to determine the fluctuation direction and fluctuation amplitude of the process parameter window subsequence.

[0024] For the process parameter time series, a sliding window with a fixed size is set. For example, the window size is 20 time points. The window starts from the beginning of the sequence and moves one time point each time, successively dividing the sequence to obtain multiple process parameter window subsequences. For each window subsequence, a curve fitting algorithm, such as least squares fitting, is used to determine the change trend of the subsequence, thereby judging whether the fluctuation direction is rising, falling or stable, and calculating the fluctuation amplitude. For example, for a window subsequence of a certain etching solution concentration, a straight line is obtained through fitting. If the slope of the straight line is positive, the fluctuation direction is rising, and the fluctuation amplitude can be calculated by combining the slope value with the data range.

[0025] Step 1223: Perform state segmentation processing on the device operation state time series to obtain multiple device operation state sub-intervals, and perform stability evaluation on each device operation state sub-interval to determine the operation offset and offset duration of the device operation state sub-interval.

[0026] Optionally, for the device operation state time series, state segmentation is performed according to the change situation of the device operation state. For example, for the placement machine operation speed sequence, when the speed remains relatively stable for a period of time, it is divided into a sub-interval; when the speed changes significantly, a new sub-interval is re-divided. Stability evaluation is performed on each sub-interval, and the stability is judged by calculating statistical quantities such as the variance of the data. If the variance is small, it indicates that the device operation in this sub-interval is relatively stable; if the variance is large, there is an operation offset. The operation offset can be calculated by comparing the difference between the mean value of the data in the sub-interval and the normal operation value, and the offset duration is determined according to the time span of the sub-interval.

[0027] Step 1224: Generate the process parameter fluctuation characterization and the device operation state fluctuation characterization based on the fluctuation direction, the fluctuation amplitude, the operation offset, and the offset duration.

[0028] Optionally, after obtaining the fluctuation direction, fluctuation amplitude of the process parameter window subsequence, and the operation offset and offset duration of the device operation state sub-interval, these information are integrated. For example, the fluctuation directions and fluctuation amplitudes of different process parameters are summarized to form a process parameter fluctuation characterization that can comprehensively reflect the process parameter fluctuation situation; the operation offsets and offset durations of each device operation state sub-interval are sorted out to generate a device operation state fluctuation characterization, so as to clearly present the fluctuation characteristics of the process parameters and the device operation state in multiple consecutive production cycles of the production line.

[0029] Step 1: Generate the process fluctuation correlation characteristics of the target printed circuit board production line based on the process parameter fluctuation characterization, and generate the production time series state characteristics based on the device operation state fluctuation characterization.

[0030] Optionally, based on the process parameter fluctuation characterization obtained above, further mine the associated information therein to generate process fluctuation association features. For example, analyze the correlation between different process parameter fluctuations, such as the relationship between the etching solution concentration fluctuation and the electroplating quality fluctuation, and quantitatively represent this associated information to form process fluctuation association features. For the equipment operation state fluctuation characterization, comprehensively consider the changes in the equipment operation state at different time points to generate production timing state features that can reflect the overall production timing state of the production line, providing a strong basis for subsequent production decisions.

[0031] Step 130: Invoke the production decision model to perform state fusion processing on the production timing state features and the process fluctuation association features, and generate production control decision information for the target circuit board production line.

[0032] This step aims to use the production decision model to deeply fuse the two features extracted previously to formulate production control decision information that meets the actual requirements of the production line.

[0033] In an exemplary embodiment, step 130 includes: Step 131: Map the production timing state features to a first state vector, map the process fluctuation association features to a second state vector, and perform vector concatenation processing on the first state vector and the second state vector to obtain a fused state vector.

[0034] In the embodiment of the present invention, a feature mapping algorithm is adopted to transform the production timing state features into a vector form, that is, the first state vector. Each dimension of this vector represents different aspect information of the production timing state. Similarly, the process fluctuation association features are mapped to the second state vector. Then, these two vectors are concatenated according to a certain rule. For example, the dimensions of the first state vector and the dimensions of the second state vector are arranged in sequence to form a new fused state vector for subsequent model processing.

[0035] Step 132: Invoke the multi-layer perceptron in the production decision model to perform non-linear transformation processing on the fused state vector, and generate a state prediction result for the target circuit board production line.

[0036] Among them, the multi-layer perceptron in the production decision model consists of multiple neuron layers. Input the fused state vector into the multi-layer perceptron, and the neurons in the perceptron process the input data through weighted summation and non-linear activation functions. For example, the neurons multiply the values of each dimension of the fused state vector by the corresponding weights and then sum them, and then perform transformation through non-linear activation functions such as ReLU. After passing through the processing of multiple neuron layers in sequence, finally, a state prediction result for the target circuit board production line is generated, and this result contains prediction information about the future operation state of the production line.

[0037] Step 133: Match the predefined production control strategy library according to the state prediction result, and determine the production control decision information.

[0038] Optionally, match the state prediction result with the predefined production control strategy library, find the most suitable production control strategy from it, and then determine the production control decision information.

[0039] In one implementation, step 133 includes: Step 1331: Analyze the process parameter fluctuation amplitude set, equipment operation offset set, and production line efficiency deviation index included in the state prediction result, and generate a current prediction parameter set.

[0040] Optionally, perform a detailed analysis of the state prediction result and extract key information from it. For example, organize the fluctuation amplitude values of different process parameters in the prediction result into a process parameter fluctuation amplitude set, form an equipment operation offset set with the offset values of equipment operation, and calculate the production line efficiency deviation index at the same time. Integrate this information to form a current prediction parameter set, providing a basis for subsequent strategy matching.

[0041] Step 1332: Obtain the historical process parameter fluctuation amplitude set, historical equipment operation offset set, and historical production line efficiency deviation index corresponding to each historical control strategy in the predefined production control strategy library, and perform dimension alignment processing on the historical process parameter fluctuation amplitude set and the historical equipment operation offset set with the current prediction parameter set to generate a normalized historical parameter set; wherein, the dimension alignment processing includes: performing a linear conversion of the temperature parameter in the historical process parameter fluctuation amplitude set from Celsius to the standard temperature scale, performing a decimal scaling of the pressure parameter from kilopascals to megapascals, and performing a unit unification of the time dimension parameter in the historical equipment operation offset set from milliseconds to seconds.

[0042] In this step, take out the relevant parameter sets corresponding to each historical control strategy from the predefined production control strategy library. To make these historical parameters comparable with the current prediction parameters, perform dimension alignment processing. For the temperature parameter in the historical process parameter fluctuation amplitude set, convert Celsius to the standard temperature scale according to a certain linear conversion formula; for the pressure parameter, convert kilopascals to megapascals through decimal scaling; for the time dimension parameter in the historical equipment operation offset set, unify milliseconds to seconds. After these processes, generate a normalized historical parameter set.

[0043] Step 1333: Calculate the multi-dimensional similarity values between the current prediction parameter set and each of the normalized historical parameter sets. The multi-dimensional similarity values include the process parameter fluctuation direction consistency coefficient, the equipment operation offset timing matching degree, and the efficiency deviation weight distance. Among them, the calculation of the multi-dimensional similarity values satisfies the following conditions: The process parameter fluctuation direction consistency coefficient is determined by the slope symbol matching degree of the corresponding process parameters in the current prediction parameter set and the normalized historical parameter set; the equipment operation offset timing matching degree is calculated by the time warping algorithm to obtain the minimum path distance between two equipment operation offset sequences; the efficiency deviation weight distance is calculated by the weighted Euclidean distance formula to fuse the normalized differences of the production line efficiency deviation indicators.

[0044] In this step, calculate the similarity values of multiple dimensions between the current prediction parameter set and each normalized historical parameter set respectively. For the process parameter fluctuation direction consistency coefficient, compare the slope symbols of the corresponding process parameters in the current prediction parameter set and the normalized historical parameter set. If the symbols are the same, the consistency coefficient is higher. For the equipment operation offset timing matching degree, use the time warping algorithm to find the minimum path distance between two equipment operation offset sequences. The smaller the distance, the higher the matching degree. For the efficiency deviation weight distance, through the weighted Euclidean distance formula, considering the weights of different factors, fuse the normalized differences of the production line efficiency deviation indicators, calculate the corresponding distance value, and comprehensively measure the similarity degree between the two by integrating these similarity values.

[0045] Step 1334: Sort the historical control strategies in the predefined production control strategy library according to the multi-dimensional similarity values, and filter out the target historical control strategy that meets the similarity threshold and has the highest priority.

[0046] Optionally, sort the historical control strategies in the production control strategy library according to the calculated multi-dimensional similarity values. Set a similarity threshold, and only the strategies whose similarity values reach or exceed this threshold are considered. Among the strategies that meet the threshold, according to the pre-set priority rules, filter out the target historical control strategy with the highest priority, which is the most in line with the prediction state of the current production line.

[0047] Step 1335: Extract the process adjustment parameter mapping table, the equipment control instruction sequence, and the execution timing constraint conditions in the target historical control strategy, and combine them with the real-time production line efficiency deviation indicator in the current prediction parameter set to perform parameter correction, and generate the production control decision information.

[0048] Optionally, key information such as the process adjustment parameter mapping table, the equipment control instruction sequence, and the execution timing constraint conditions is retrieved from the target historical control strategy. Combining with the real-time production line efficiency deviation index in the current prediction parameter set, these information are corrected for parameters. For example, according to the efficiency deviation situation, some parameters in the process adjustment parameter mapping table are slightly adjusted, and the instruction execution time in the equipment control instruction sequence is appropriately adjusted, so as to generate the final production control decision information to ensure that the decision can effectively guide the operation of the production line.

[0049] In an alternative embodiment, the training method of the production decision model includes: Step 210: Obtain the historical monitoring data set of the historical production line and the corresponding historical control decision set; perform multi-dimensional time series feature extraction processing on the historical monitoring data set to generate historical production time series state features and historical process fluctuation correlation features.

[0050] In this step, the monitoring data of the historical production line in the past period of time are collected, and these data cover the process parameter monitoring data and the equipment operation monitoring data to form the historical monitoring data set. At the same time, the historical control decision set corresponding to these historical data is obtained, that is, the control decision made for the production line state at that time. Then, according to the same method as the real-time monitoring data processing before, the historical monitoring data set is subjected to multi-dimensional time series feature extraction processing to generate historical production time series state features and historical process fluctuation correlation features, providing a data basis for model training.

[0051] Step 220: Input the historical production time series state features and the historical process fluctuation correlation features as training samples into the initial production decision model to generate predicted control decision information.

[0052] Optionally, the generated historical production time series state features and historical process fluctuation correlation features are used as input data and sent into the initial production decision model. The initial production decision model processes these input features using its own structure and algorithm and outputs predicted control decision information, which is the prediction of the control decision that should be made in the case of historical data.

[0053] Step 230: Construct a loss function according to the difference between the predicted control decision information and the historical control decision set, and perform iterative training on the initial production decision model based on the loss function until convergence to obtain the production decision model.

[0054] In this step, a loss function is constructed by comparing the differences between the predictive control decision information and the historical control decision set. The loss function is used to measure the gap between the model prediction results and the actual historical decisions. Based on this loss function, an optimization algorithm, such as the stochastic gradient descent algorithm, is adopted to iteratively adjust the parameters of the initial production decision model. In each iteration, the model parameters are adjusted according to the value of the loss function, so that the prediction results of the model gradually approach the historical control decisions until the loss function converges to a smaller value, and at this time, the trained production decision model is obtained.

[0055] Further, constructing the loss function according to the difference between the predictive control decision information and the historical control decision set includes: Step 231: Perform vectorized encoding processing on the predictive control decision information to obtain a predictive decision vector, and perform vectorized encoding processing on the historical control decision set to obtain a historical decision vector.

[0056] Optionally, a feature encoding algorithm is adopted to convert the predictive control decision information into a vector form, that is, a predictive decision vector, and each dimension of the vector represents different aspects of the decision information. Similarly, the historical control decision set is vectorized and encoded to obtain a historical decision vector for subsequent comparison and calculation.

[0057] Step 232: Calculate the cosine similarity between the predictive decision vector and the historical decision vector, determine the first loss component according to the cosine similarity, calculate the Euclidean distance between the predictive decision vector and the historical decision vector, and determine the second loss component according to the Euclidean distance.

[0058] Further, using the vector calculation method, calculate the cosine similarity between the predictive decision vector and the historical decision vector. The cosine similarity reflects the similarity degree of the two vectors in direction. According to this similarity value, determine the first loss component. The higher the similarity, the smaller the first loss component. At the same time, calculate the Euclidean distance between the two vectors. The Euclidean distance measures the distance of the vectors in space. According to the Euclidean distance, determine the second loss component. The smaller the distance, the smaller the second loss component.

[0059] Step 233: Perform weighted summation on the first loss component and the second loss component to obtain the loss function.

[0060] In this step, according to the actual requirements, different weights are assigned to the first loss component and the second loss component. For example, if more attention is paid to the consistency of the decision direction, a larger weight is assigned to the first loss component determined by the cosine similarity; if more attention is paid to the specific numerical differences of the decisions, a larger weight is assigned to the second loss component determined by the Euclidean distance. The two loss components are summed according to the weights to obtain the loss function for measuring the model prediction error.

[0061] Step 140: Generate a set of device control instructions according to the production control decision information, and transmit the set of device control instructions to the execution terminal of the target printed circuit board production line to implement the production status adjustment operation.

[0062] Optionally, according to the generated production control decision information, formulate a specific set of device control instructions and send them to the execution terminal of the production line, so as to adjust the operating state of the production line.

[0063] In an alternative embodiment, the generating a set of device control instructions according to the production control decision information includes: Step 141: Analyze the device adjustment type and device adjustment range in the production control decision information; match the corresponding execution device identifier in the target printed circuit board production line according to the device adjustment type, generate the device control quantization feature of the execution device identifier according to the device adjustment range; perform instruction encapsulation processing on the execution device identifier and the device control quantization feature to generate the set of device control instructions.

[0064] Deeply analyze the production control decision information, extract the device adjustment type from it, such as the opening, closing, speed adjustment, etc. of the device, and the specific value of the device adjustment range. According to the device adjustment type, find the corresponding execution device identifier in the device list of the target printed circuit board production line. For example, if the decision information requires adjusting the running speed of the mounter, determine the unique identifier of the mounter by looking up the device list. According to the device adjustment range, generate the device control quantization feature for the execution device identifier, such as the specific value of the speed adjustment. Finally, encapsulate the execution device identifier and the device control quantization feature according to the set instruction format to form a set of device control instructions.

[0065] In another alternative embodiment, the transmitting the set of device control instructions to the execution terminal of the target printed circuit board production line to implement the production status adjustment operation includes: Step 142: Perform instruction priority sorting on the set of device control instructions to determine the execution order of each device control instruction; transmit the set of device control instructions to the execution terminal of the target printed circuit board production line in sequence according to the execution order; monitor the response status of the execution terminal to the set of device control instructions, and when a response anomaly is detected, regenerate a backup set of device control instructions and transmit them.

[0066] Optionally, according to factors such as the importance and urgency of the equipment control instructions, prioritize the set of equipment control instructions. For example, the equipment control instructions related to the key links of the production line have a higher priority. After determining the execution order of each instruction, send the set of equipment control instructions to the execution terminal of the target circuit board production line in this order. During the transmission process, monitor the response status of the execution terminal to the set of equipment control instructions in real time. If the execution terminal fails to execute the instructions as expected, such as abnormal situations like the equipment not responding or execution errors, regenerate a backup set of equipment control instructions according to the production control decision information and transmit it again to ensure that the production line can adjust the production status according to the correct control instructions.

[0067] As an alternative embodiment, the method further includes: Step 310: Convert the production control decision information into visual chart data, where the visual chart data includes a process parameter adjustment trend chart, an equipment operation status comparison chart, and a production line efficiency prediction chart.

[0068] In the embodiments of the present invention, in order to more intuitively present the production control decision information and facilitate production management personnel to analyze and make decisions, it is necessary to convert the production control decision information. For the process parameter adjustment trend chart, extract the adjustment information of each process parameter from the production control decision information, such as the change in the adjustment value of the etching solution concentration over time. Use the time as the horizontal axis and the adjustment value of the etching solution concentration as the vertical axis, and use a drawing algorithm to draw a curve reflecting the concentration adjustment trend. Draw the adjustment trends of multiple process parameters on the same chart to form a process parameter adjustment trend chart. For the equipment operation status comparison chart, select the operation status parameters of key equipment, such as the operation speed of the mounter and the temperature of the reflow oven. Compare the current operation status parameters of the equipment in the decision information with the historical normal operation status parameters, and use different graphic elements (such as bar charts, line charts, etc.) to represent the current and historical states respectively to visually display the differences in the equipment operation status and generate an equipment operation status comparison chart. For the production line efficiency prediction chart, according to the production line efficiency deviation index in the production control decision information and related prediction models, predict the efficiency change of the production line in a future period of time. Use the time as the horizontal axis and the predicted production line efficiency value as the vertical axis to draw a prediction curve to form a production line efficiency prediction chart. Through these visual chart data, production management personnel can quickly understand the operation status and future development trend of the production line.

[0069] Step 320: Send the visual chart data to the production management terminal for display, and when detecting a decision correction instruction responded by the production management terminal, re - execute the multi - dimensional time - series feature extraction process and the state fusion process to update the production control decision information.

[0070] In an embodiment of the present invention, the generated visual chart data is sent to a production management terminal through network transmission or other means. The production management terminal can be a device such as a computer or a tablet. After viewing these visual charts, production management personnel may issue decision correction instructions based on actual experience and production requirements.

[0071] In a preferred embodiment, when detecting the decision correction instruction responded by the production management terminal in step 320, re - executing the multi - dimensional time - series feature extraction process and the state fusion process to update the production control decision information includes: Step 321: Analyze the process parameter adjustment amount set, equipment operation offset compensation coefficient set, and production cycle time - series constraint condition set in the decision correction instruction to generate a correction parameter set.

[0072] Optionally, perform a detailed analysis of the decision correction instruction responded by the production management terminal. Extract the process parameter adjustment amount set from the instruction, such as the specific values that different process parameters (etchant concentration, electroplating current intensity, etc.) need to be adjusted; extract the equipment operation offset compensation coefficient set, that is, the coefficients for compensating the equipment operation offset situation; and the production cycle time - series constraint condition set, such as information on the time limit of the production cycle. Integrate this information together to form a correction parameter set, providing a basis for subsequent data processing.

[0073] Step 322: Superimpose the process parameter adjustment amount set of the correction parameter set onto the process parameter monitoring data of the optimized monitoring data set according to the corresponding time window to generate a corrected process parameter sequence; perform sliding window filtering fusion on the equipment operation offset compensation coefficient set of the correction parameter set and the equipment operation monitoring data of the optimized monitoring data set to generate a corrected equipment operation sequence.

[0074] Optionally, for the process parameter adjustment amount set in the correction parameter set, according to the corresponding time window, superimpose the adjustment amount onto the process parameter monitoring data of the optimized monitoring data set. For example, if the decision correction instruction requires increasing the etchant concentration by a certain value within a certain time period, add this increase amount to the etchant concentration monitoring data of the corresponding time period, thereby generating a corrected process parameter sequence. For the equipment operation offset compensation coefficient set, use the method of sliding window filtering fusion to process it with the equipment operation monitoring data of the optimized monitoring data set. Set an appropriate sliding window size, slide the window on the equipment operation monitoring data sequence, and perform weighted fusion on the data according to the compensation coefficient within each window to generate a corrected equipment operation sequence, making the equipment operation data more in line with the corrected requirements.

[0075] Step 323: Intercept the time series intervals of the corrected process parameter sequence and the corrected equipment operation sequence according to the set of production cycle time series constraint conditions to generate a set of corrected monitoring data; perform the time series fluctuation analysis process on the set of corrected monitoring data to generate an updated process parameter fluctuation characterization and an equipment operation state fluctuation characterization; generate an updated process fluctuation correlation feature based on the updated process parameter fluctuation characterization, and generate an updated production time series state feature based on the updated equipment operation state fluctuation characterization.

[0076] In the embodiment of the present invention, according to the set of production cycle time series constraint conditions, the time series intervals that meet the time requirements are intercepted from the corrected process parameter sequence and the corrected equipment operation sequence to form a set of corrected monitoring data. Then, the set of corrected monitoring data is processed according to the time series fluctuation analysis method described above to determine the fluctuation direction and amplitude of the process parameters, as well as the operation offset and offset duration of the equipment operation state, and generate an updated process parameter fluctuation characterization and an equipment operation state fluctuation characterization. Based on these new fluctuation characterizations, an updated process fluctuation correlation feature and an updated production time series state feature are further generated to provide new feature data for subsequent state fusion.

[0077] Step 324: Perform time series dimension alignment processing on the updated process fluctuation correlation feature and the process fluctuation correlation feature, and use endpoint replication or mean filling to extend them to the same time step when the feature lengths are inconsistent; perform feature channel matching processing on the updated production time series state feature and the production time series state feature, and map them to the same feature space through a fully connected layer when the number of channels is inconsistent.

[0078] In the embodiment of the present invention, time series dimension alignment is performed on the updated process fluctuation correlation feature and the process fluctuation correlation feature. If the time lengths of the two features are different, for example, the time step of the updated process fluctuation correlation feature is shorter, the endpoint replication method is used to repeat the endpoint value at the shorter end, or the mean filling method is used to fill the missing part with the mean of the front and back data to make them have the same time step. For the updated production time series state feature and the production time series state feature, when the number of their feature channels is inconsistent, a fully connected layer is used for mapping. Each neuron in the fully connected layer is connected to all neurons of the input feature. By learning the weight matrix, the features with different numbers of channels are mapped to the same feature space for subsequent vector splicing.

[0079] Step 325: Perform vector splicing on the updated process fluctuation correlation feature after time series dimension alignment and the updated production time series state feature after feature channel matching to generate a fused updated feature vector; call the production decision model to perform the state fusion process on the fused updated feature vector to generate intermediate control decision information.

[0080] Optionally, the updated process fluctuation correlation feature vector aligned in the time sequence dimension and the updated production time sequence state feature vector after feature channel matching are concatenated. The respective dimensions of the two vectors are arranged together in a certain order to form a fused updated feature vector. Then, this fused updated feature vector is input into the production decision model, and the production decision model performs state fusion processing on it, and performs operations such as non-linear transformation through structures such as multi-layer perceptrons inside the model to generate intermediate control decision information, which is the preliminary decision result generated based on the new feature data.

[0081] Step 326: Calculate the set of absolute differences of the corresponding process adjustment parameters and the set of cosine similarities of the equipment control instructions in the intermediate control decision information and the production control decision information; screen out the conflict control items in the set of absolute differences that exceed the preset process threshold and the set of cosine similarities that are lower than the preset equipment threshold; perform weighted moving average correction on the process adjustment parameters in the conflict control items, and the weight is adjusted according to the time sequence fluctuation amplitude of the corresponding process parameters.

[0082] In this step, calculate the absolute differences of the corresponding process adjustment parameters in the intermediate control decision information and the production control decision information, and form a set of absolute differences with these differences. At the same time, calculate the cosine similarities of the equipment control instructions in the two decision-making informations to form a set of cosine similarities. Set the preset process threshold and the preset equipment threshold, screen out the process parameter differences greater than the preset process threshold from the set of absolute differences, and screen out the equipment control instruction similarity conditions lower than the preset equipment threshold from the set of cosine similarities to determine the conflict control items. For the process adjustment parameters in the conflict control items, use the method of weighted moving average correction. Determine the weight according to the time sequence fluctuation amplitude of the corresponding process parameter. The parameter with a larger fluctuation amplitude has a larger weight, and the process adjustment parameter is weighted and averaged and corrected within the sliding window to make the corrected process parameter more reasonable.

[0083] Step 327: Perform priority sorting on the equipment control instructions in the conflict control items, and determine the coverage order according to the time matching degree between the equipment response delay parameter and the production cycle time sequence constraint condition; merge the corrected process adjustment parameters and the equipment control instructions after priority coverage into the production control decision information to generate the updated production control decision information.

[0084] Optionally, for the device control instructions in the conflict control items, priority sorting is performed according to the time matching degree between the device response delay parameter and the production cycle timing constraint conditions. If a device has a long response delay, but the production cycle has urgent time requirements for the operation of this device, then the priority of this device control instruction is relatively high. After determining the priority, the device control instructions are overwritten in the order of priority. The corrected process adjustment parameters and the device control instructions after priority overwriting are merged with the original production control decision information to form updated production control decision information, ensuring that the decision information can better adapt to the actual situation of the production line.

[0085] Step 328: Correct the timing matching between the temperature adjustment step of the process adjustment parameters and the voltage change rate of the device control instructions in the updated production control decision information, so that the millisecond-level timestamp of the temperature adjustment window covers the voltage change interval.

[0086] Optionally, check the time matching between the temperature adjustment step of the process adjustment parameters and the voltage change rate of the device control instructions in the updated production control decision information. If the temperature adjustment window and the voltage change interval do not match in time, for example, the start time of temperature adjustment is later than the start time of voltage change, or the end time of temperature adjustment is earlier than the end time of voltage change, correct the timestamp according to the specific situation. By adjusting the execution time of the temperature adjustment step or the voltage change rate, ensure that the millisecond-level timestamp of the temperature adjustment window can completely cover the voltage change interval, making the process adjustment and device control coordinated in time and ensuring the stable operation of the production line.

[0087] In a non-limiting embodiment, after transmitting the set of device control instructions to the execution terminal of the target circuit board production line to implement the production status adjustment operation, it further includes: collecting the response feedback data set of the execution terminal, where the response feedback data set includes the actual adjustment value sequence of process parameters and the updated sequence of device operation status; extracting the execution effect characteristics from the response feedback data set to generate the actual process fluctuation characteristics and the actual device state timing characteristics; calculating the difference after the time series alignment of the actual process fluctuation characteristics and the process fluctuation correlation characteristics to generate a set of process adjustment deviation coefficients; performing window sliding correlation analysis on the actual device state timing characteristics and the production timing state characteristics to generate a set of device control lag time parameters; adjusting the control magnitude and execution timing interval in the set of device control instructions according to the set of process adjustment deviation coefficients and the set of device control lag time parameters, generating a compensation control instruction set and appending it to be transmitted to the execution terminal.

[0088] In this embodiment, after the set of device control instructions is transmitted to the execution terminal and executed, the data fed back by the execution terminal is collected. Among them, the actual adjustment value sequence of process parameters records the values of various process parameters after actual adjustment, and the device operation state update sequence reflects the change in the operation state of the device after executing the instructions. Feature extraction is performed on these response feedback data sets to analyze the actual fluctuations of process parameters and generate actual process fluctuation features; analyze the change law of the device operation state over time and generate actual device state time series features. Align the actual process fluctuation features with the previously generated process fluctuation correlation features in time series, calculate the difference between the two, and form a process adjustment deviation coefficient set, which reflects the deviation degree between the actual adjustment and the expected adjustment of process parameters. Perform window sliding correlation analysis on the actual device state time series features and the production time series state features. For example, set a sliding window, move the window on the two feature sequences, and calculate the correlation of the data within the window to determine the time parameter of device control lag and generate a set of device control lag time parameters. According to the situations reflected by these two sets, adjust the control magnitude in the set of device control instructions, such as increasing or decreasing the control force of certain devices; adjust the execution time interval, such as extending or shortening the time interval for instruction execution. After generating the compensation control instruction set, append and transmit it to the execution terminal to correct the deviation that occurred during the execution of the previous device control instructions and make the operation of the production line more in line with expectations.

[0089] In a non-limiting embodiment, after transmitting the set of device control instructions to the execution terminal of the target printed circuit board production line to implement the production state adjustment operation, it further includes: obtaining in real time the target monitoring data stream of the target printed circuit board production line, where the target monitoring data stream includes a derivative sequence of process parameters after execution and a device operation state change rate sequence; performing abnormal fluctuation detection on the target monitoring data stream to identify abnormal parameter identifiers that do not match the preset process threshold range in the production control decision information and the corresponding abnormal time windows; extracting the execution records of the device control instructions corresponding to the abnormal parameter identifiers, intercepting the historical operation offset sequence of the associated device according to the abnormal time windows, and generating an abnormal traceability feature vector; inputting the abnormal traceability feature vector into a pre-trained compensation strategy generation model to output a set of compensation control parameters and a priority adjustment coefficient for the abnormal parameter identifiers; covering the parameter values of the corresponding instructions in the original set of device control instructions according to the set of compensation control parameters, and reordering the instruction transmission queue based on the priority adjustment coefficient to generate an optimized control instruction set and perform a secondary transmission.

[0090] In this embodiment, the target monitoring data stream of the target circuit board production line is obtained in real time, where the post-execution process parameter derivative sequence records the derivative data of the process parameters after the execution of the equipment control instructions, such as the etching accuracy of the circuit board after the etching process; the equipment operation state change rate sequence reflects the change rate of the equipment operation state over time. Anomaly fluctuation detection is performed on the target monitoring data stream. Through the set detection algorithm, the monitoring data is compared with the preset process threshold range in the production control decision information. If it is found that the values of some process parameters exceed the preset range, the identifiers of these abnormal parameters and the time window of the anomaly occurrence are determined. The execution records of the equipment control instructions corresponding to these abnormal parameter identifiers are found, and according to the abnormal time window, the historical operation offset sequences of the associated equipment within this time period are intercepted from the equipment operation data, and these sequences are integrated to generate an anomaly tracing feature vector. The anomaly tracing feature vector is input into a pre-trained compensation strategy generation model, which outputs a set of compensation control parameters and a priority adjustment coefficient for these abnormal parameter identifiers according to the previously learned knowledge and patterns. According to the set of compensation control parameters, the parameter values of the corresponding instructions in the original equipment control instruction set are overwritten and updated; according to the priority adjustment coefficient, the instruction transmission queue is re-sorted to generate an optimized control instruction set and transmitted to the execution terminal again to solve the anomalies occurring in the production line operation process and ensure the stable and efficient operation of the production line.

[0091] Specifically, based on the above technical solution, in the actual application process, the dynamic time warping algorithm (DTW) in the prior art can be combined to calculate the matching degree of the equipment operation offset time series. The minimum path distance is found by constructing an accumulated distance matrix, and at the same time, the unit conversion process of the historical process parameters is carried out in combination with the equipment control parameter dimension specification in the industrial standard ISA-88. For example, when converting Celsius temperature to Kelvin scale, the linear conversion formula T(K)=T(℃)+273.15 is used, and a decimal scaling from kilopascal to megapascal is implemented for the pressure parameter (1MPa = 1000kPa).

[0092] In the actual application process of the embodiment of the present invention, for the sliding window segmentation process, the window size can be determined according to the Nyquist-Shannon sampling theorem to ensure that the window can cover at least two complete process fluctuation cycles. When designing the multi-layer perceptron, a three-layer network structure with the ReLU activation function is adopted, the number of neurons in the hidden layer is configured according to the 2 / 3 rule of the input feature dimension, and parameter update is based on the Adam optimizer. For the weight assignment of the weighted Euclidean distance, the entropy weight method can be combined to calculate the index weights of each process parameter to ensure that the weight coefficients objectively reflect the importance of the parameters. The time series alignment process can adopt the dynamic time warping algorithm for non-equal-length sequence matching, and anomaly detection is achieved through a dynamic threshold mechanism that sets the 3σ principle.

[0093] It can be understood that in the visualization chart generation stage, the pyplot interface of the Matplotlib library is used for multi-dimensional data visualization rendering, and the millisecond-level timestamp is accurately aligned through the time series database InfluxDB. The priority sorting of device control instructions can be implemented according to the device criticality classification system in the ANSI / ISA-95 standard to ensure that critical device instructions are executed first. Such a design can ensure the accurate extraction of process parameter fluctuations and the effective training of production decision-making models, and realize the accurate generation and execution of production line control decisions.

[0094] It can be understood that a corresponding process control logic chain is constructed between the input and output of the above neural network model (such as the production decision-making model) through a multi-level feature abstraction and state evolution mechanism. The input end takes the time series data of process parameters such as the etching solution concentration and electroplating current intensity collected in real time by the target production line, and the device status monitoring data such as the running speed of the mounter and the temperature of the reflow soldering furnace as the original input. After the abnormal point replacement and time window alignment processing in data cleaning, an optimized monitoring data set with time series continuity is formed; this data extracts the fluctuation direction / amplitude characteristics of process parameters through sliding window segmentation and trend fitting, and combines the stability evaluation of device status segmentation to generate a running offset characteristic, forming a primary feature layer representing the microscopic fluctuation characteristics of the production line.

[0095] The intermediate processing layer analyzes the time series correlation between the process parameter fluctuation direction and the device running offset, mines the mesoscopic correlation characteristics such as the coupling relationship between the change of etching process parameters and the vibration of the drilling equipment, and constructs a process fluctuation correlation matrix reflecting the collaborative working state of multiple devices; at the same time, it maps the device running state fluctuation representation into a production time series state vector including macroscopic indicators such as the production line beat synchronization and process connection efficiency.

[0096] In the decision-making generation stage, the multi-layer perceptron fuses the process fluctuation correlation matrix and the production time series state vector into a state prediction vector in the high-dimensional feature space through non-linear transformation. This vector calculates the multi-dimensional similarity with the historical control strategy through the policy library matching mechanism, and finally transforms the feature correlations such as the matching degree of the slope symbol of the abstract process parameter and the distance of the device offset time series path into specific executable control outputs such as the adjustment instruction of the etching solution supply amount and the acceleration parameter of the mounter, forming a cross-time scale mapping from millisecond-level sensor data to minute-level control instructions, which not only maintains the real-time responsiveness of process parameter adjustment, but also ensures the collaborative consistency of multi-device control instructions within the production line beat cycle, realizing the layer-by-layer semantic improvement and logical connection of data features from the original monitoring value to the device action signal.

[0097] In the embodiments of the present invention, multi-dimensional time-series feature extraction processing is performed on the collected real-time monitoring data set of the target circuit board production line to generate production time-series state features and process fluctuation correlation features, which can deeply analyze the operation features of the production line from different dimensions, accurately extract key information, and provide a comprehensive and detailed basis for production decision-making. Subsequently, the production decision-making model is called to perform state fusion processing on the above two features, fully explore the potential connections between the features, and generate accurate production control decision-making information, which can comprehensively consider various factors and make the decision more scientific and reasonable. Finally, an equipment control instruction set is generated according to the production control decision-making information and transmitted to the execution terminal to achieve accurate adjustment of the production state, effectively improve the stability and production efficiency of the production line, ensure the smooth progress of the circuit board production process, and comprehensively optimize the circuit board production control process.

[0098] In summary, the embodiments of the present invention aim to solve the problems in the prior art such as incomplete data collection, inaccurate decision-making, and untimely adjustment of the production state. By comprehensively collecting real-time monitoring data, multi-dimensional feature extraction, state fusion decision-making, and accurate transmission of equipment control instructions, accurate control of the circuit board production process is realized, and the stability and efficiency of production are improved.

[0099] See Figure 2 As shown in the figure, the figure is a schematic diagram of the basic structure of a circuit board production control system 200 provided by the embodiments of the present invention. The circuit board production control system 200 includes: A processor 201; A storage device 202, on which a computer program 2020 is stored; When the computer program 2020 is executed by the processor 201, the processor 201 implements any of the circuit board production control methods based on data monitoring.

[0100] On this basis, a readable storage medium is provided. Programs or instructions are stored on the readable storage medium, and when the programs or instructions are executed by a processor, the steps of the above method are implemented.

[0101] It should be noted that the embodiments in this specification are described in a progressive manner. The key points of each embodiment are the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the systems or devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and the relevant parts can be referred to the description of the method part.

Claims

1. A circuit board production control method based on data monitoring, characterized in that, Including: Collecting a real-time monitoring data set of a target circuit board production line, where the real-time monitoring data set includes process parameter monitoring data and equipment operation monitoring data in multiple consecutive time series; Performing multi-dimensional time series feature extraction processing on the real-time monitoring data set to generate production time series state features and process fluctuation correlation features of the target circuit board production line; Invoking a production decision model to perform state fusion processing on the production time series state features and the process fluctuation correlation features to generate production control decision information for the target circuit board production line; Generating a set of equipment control instructions according to the production control decision information and transmitting the set of equipment control instructions to an execution terminal of the target circuit board production line to implement production state adjustment operations.

2. The method according to claim 1, characterized in that, The performing multi-dimensional time series feature extraction processing on the real-time monitoring data set to generate production time series state features and process fluctuation correlation features of the target circuit board production line includes: Performing data cleaning processing on the real-time monitoring data set to obtain an optimized monitoring data set; the data cleaning processing includes: replacing abnormal data points in the process parameter monitoring data with adjacent time series data, and aligning non-consecutive acquisition data in the equipment operation monitoring data with a time window; Performing time series fluctuation analysis processing on the optimized monitoring data set to identify process parameter fluctuation characteristics and equipment operation state fluctuation characteristics of the target circuit board production line in multiple consecutive production cycles; Generating process fluctuation correlation features of the target circuit board production line based on the process parameter fluctuation characteristics, and generating production time series state features based on the equipment operation state fluctuation characteristics.

3. The method according to claim 2, wherein The performing time series fluctuation analysis processing on the optimized monitoring data set to identify process parameter fluctuation characteristics and equipment operation state fluctuation characteristics of the target circuit board production line in multiple consecutive production cycles includes: Dividing the optimized monitoring data set into a process parameter time series and an equipment operation state time series; Performing sliding window segmentation processing on the process parameter time series to obtain multiple process parameter window subsequences, and performing trend fitting on each process parameter window subsequence to determine the fluctuation direction and fluctuation amplitude of the process parameter window subsequence; Performing state segmentation processing on the equipment operation state time series to obtain multiple equipment operation state sub-intervals, and performing stability evaluation on each equipment operation state sub-interval to determine the operation offset amount and offset duration of the equipment operation state sub-interval; Generating the process parameter fluctuation characteristics and the equipment operation state fluctuation characteristics according to the fluctuation direction, the fluctuation amplitude, the operation offset amount, and the offset duration.

4. The method according to claim 3, wherein The invoking a production decision model to perform state fusion processing on the production time series state features and the process fluctuation correlation features to generate production control decision information for the target circuit board production line includes: Map the production timing state characteristics to a first state vector, map the process fluctuation correlation characteristics to a second state vector, and perform vector splicing on the first state vector and the second state vector to obtain a fused state vector; Call the multi-layer perceptron in the production decision model to perform non-linear transformation on the fused state vector to generate a state prediction result of the target circuit board production line; Match the predefined production control strategy library according to the state prediction result to determine the production control decision information.

5. The method according to claim 4, characterized in that, The matching of the predefined production control strategy library according to the state prediction result to determine the production control decision information includes: Analyze the process parameter fluctuation amplitude set, equipment operation offset set and production line efficiency deviation index included in the state prediction result to generate a current prediction parameter set; Obtain the historical process parameter fluctuation amplitude set, historical equipment operation offset set and historical production line efficiency deviation index corresponding to each historical control strategy in the predefined production control strategy library, and perform dimension alignment processing on the historical process parameter fluctuation amplitude set and the historical equipment operation offset set with the current prediction parameter set to generate a normalized historical parameter set; Calculate the multi-dimensional similarity values between the current prediction parameter set and each of the normalized historical parameter sets, where the multi-dimensional similarity values include a process parameter fluctuation direction consistency coefficient, an equipment operation offset timing matching degree, and an efficiency deviation weight distance; Sort the historical control strategies in the predefined production control strategy library according to the multi-dimensional similarity values, and filter out the target historical control strategy that meets the similarity threshold and has the highest priority; Extract the process adjustment parameter mapping table, equipment control instruction sequence and execution timing constraint conditions in the target historical control strategy, and perform parameter correction in combination with the real-time production line efficiency deviation index in the current prediction parameter set to generate the production control decision information; Among them, the dimension alignment processing includes: performing a linear conversion of the temperature parameter in the historical process parameter fluctuation amplitude set from Celsius to the standard temperature scale, performing a decimal scaling of the pressure parameter from kilopascals to megapascals, and performing a unit unification of the time dimension parameter in the historical equipment operation offset set from milliseconds to seconds; Among them, the calculation of the multi-dimensional similarity value satisfies the following conditions: the process parameter fluctuation direction consistency coefficient is determined by the slope symbol matching degree of the corresponding process parameters in the current prediction parameter set and the normalized historical parameter set; the equipment operation offset timing matching degree calculates the minimum path distance of two equipment operation offset sequences through a time warping algorithm; the efficiency deviation weight distance fuses the normalized difference of the production line efficiency deviation index through a weighted Euclidean distance formula.

6. The method according to claim 4, characterized in that, The training method of the production decision model includes: Obtain the historical monitoring data set of the historical production line and the corresponding historical control decision set; Perform multi-dimensional time series feature extraction processing on the historical monitoring data set to generate historical production timing state characteristics and historical process fluctuation correlation characteristics; Input the historical production time - series state features and the historical process fluctuation correlation features as training samples into the initial production decision - making model to generate predictive control decision information; Construct a loss function based on the difference between the predictive control decision information and the historical control decision set, and iteratively train the initial production decision - making model based on the loss function until convergence to obtain the production decision - making model; The constructing a loss function based on the difference between the predictive control decision information and the historical control decision set includes: Perform vectorized encoding processing on the predictive control decision information to obtain a predictive decision vector, and perform vectorized encoding processing on the historical control decision set to obtain a historical decision vector; Calculate the cosine similarity between the predictive decision vector and the historical decision vector, determine the first loss component according to the cosine similarity, calculate the Euclidean distance between the predictive decision vector and the historical decision vector, and determine the second loss component according to the Euclidean distance; Perform weighted summation on the first loss component and the second loss component to obtain the loss function.

7. The method according to claim 1, wherein The generating an equipment control instruction set according to the production control decision information includes: Analyze the equipment adjustment type and equipment adjustment amplitude in the production control decision information; Match the corresponding execution equipment identifier in the target printed circuit board production line according to the equipment adjustment type, and generate the equipment control quantization feature of the execution equipment identifier according to the equipment adjustment amplitude; Perform instruction encapsulation processing on the execution equipment identifier and the equipment control quantization feature to generate the equipment control instruction set; The transmitting the equipment control instruction set to the execution terminal of the target printed circuit board production line to implement the production state adjustment operation includes: Perform instruction priority sorting processing on the equipment control instruction set to determine the execution order of each equipment control instruction; Transmit the equipment control instruction set to the execution terminal of the target printed circuit board production line in sequence according to the execution order; Monitor the response status of the execution terminal to the equipment control instruction set, and when a response anomaly is detected, regenerate a backup equipment control instruction set and transmit it.

8. The method according to claim 1, characterized in that, The method further includes: Convert the production control decision information into visual chart data, where the visual chart data includes a process parameter adjustment trend chart, an equipment operation status comparison chart, and a production line efficiency prediction chart; Send the visual chart data to the production management terminal for display, and when a decision - correction instruction responded by the production management terminal is detected, re - execute the multi - dimensional time - series feature extraction process and the state fusion process to update the production control decision information.

9. The method according to claim 8, wherein The re - executing the multi - dimensional time - series feature extraction process and the state fusion process to update the production control decision information when a decision - correction instruction responded by the production management terminal is detected includes: Analyze the process parameter adjustment amount set, the equipment operation offset compensation coefficient set, and the production cycle time - series constraint condition set in the decision - correction instruction to generate a correction parameter set; Superimpose the process parameter adjustment amount set of the correction parameter set on the process parameter monitoring data of the optimized monitoring data set according to the corresponding time series window to generate a corrected process parameter sequence; perform sliding window filtering fusion on the equipment operation offset compensation coefficient set of the correction parameter set and the equipment operation monitoring data of the optimized monitoring data set to generate a corrected equipment operation sequence; Intercept the time series intervals of the corrected process parameter sequence and the corrected equipment operation sequence according to the production cycle time series constraint condition set to generate a corrected monitoring data set; perform the time series fluctuation analysis process on the corrected monitoring data set to generate an updated process parameter fluctuation characterization and an equipment operation state fluctuation characterization; generate an updated process fluctuation correlation feature based on the updated process parameter fluctuation characterization, and generate an updated production time series state feature based on the updated equipment operation state fluctuation characterization; Perform time series dimension alignment processing on the updated process fluctuation correlation feature and the process fluctuation correlation feature, and use endpoint replication or mean filling to extend to the same time step when the feature lengths are inconsistent; perform feature channel matching processing on the updated production time series state feature and the production time series state feature, and map to the same feature space through a fully connected layer when the number of channels is inconsistent; Perform vector splicing on the updated process fluctuation correlation feature after time series dimension alignment and the updated production time series state feature after feature channel matching to generate a fused updated feature vector; call the production decision model to perform the state fusion process on the fused updated feature vector to generate intermediate control decision information; Calculate the absolute difference set of the corresponding process adjustment parameters and the cosine similarity set of the equipment control instructions in the intermediate control decision information and the production control decision information; screen the conflict control items that exceed the preset process threshold in the absolute difference set and the cosine similarity set is lower than the preset equipment threshold; perform weighted sliding average correction on the process adjustment parameters in the conflict control items, and the weight is adjusted according to the time series fluctuation amplitude of the corresponding process parameters; Perform priority sorting on the equipment control instructions in the conflict control items, and determine the coverage order according to the time matching degree between the equipment response delay parameter and the production cycle time series constraint condition; merge the corrected process adjustment parameters and the equipment control instructions after priority coverage into the production control decision information to generate an updated production control decision information; Correct the time series matching of the temperature adjustment step of the process adjustment parameters and the voltage change rate of the equipment control instructions in the updated production control decision information so that the millisecond-level time stamp of the temperature adjustment window covers the voltage change interval.

10. A circuit board production control system, characterized in that, Including: A processor; A storage device on which a computer program is stored, and when the computer program is executed by the processor, the processor implements the printed circuit board production control method based on data monitoring as described in any one of claims 1-9.

Images