Intelligent monitoring system and method for circuit board electroplating production line

Abstract

The invention relates to the technical field of intelligent monitoring, and more specifically discloses an intelligent monitoring system and method for a circuit board electroplating production line, and the method comprises the steps: obtaining a current signal, temperature data and humidity data, extracting a corresponding feature vector and a feature matrix, constructing a circuit board electroplating key parameter feature vector, and then carrying out the calculation of a current signal, temperature data and humidity data; and analyzing the key parameter feature vectors of the circuit board electroplating by using a classifier to obtain a classification result which is used for representing whether the circuit board electroplating production process is abnormal or not. The intelligent monitoring system can monitor key parameters in the electroplating process in real time, carry out anomaly detection and early warning, help to prevent production line faults and quality problems, and improve the stability and reliability of a production line.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring technology, and more specifically, to an intelligent monitoring system and method for a circuit board electroplating production line. Background Art

[0002] A circuit board is a circuit board that can connect circuits and facilitate the performance of each component. The circuit board uses an insulating board as the base material, cut into a certain size, with at least one conductive pattern attached to it, and is provided with holes (such as component holes, fastening holes, metallized holes, etc.), which are used to replace the chassis of the previous electronic components and realize the interconnection between electronic components.

[0003] The production process of circuit boards involves cutting and grinding, drilling, outer layer graphics, gold plating, etching, inspection, silk screen solder mask, hot air leveling, silk screen characters, appearance processing and testing inspection. In the above processes, various technologies must be used to protect copper printed lines, vias and plated through holes. These technologies are mainly electroplating technologies. The essence of electroplating is to use a strong current to electrolyze a copper tank potion with a high copper content, so that the copper water in the potion undergoes a reduction reaction and is reduced to the copper surface and circuit of the circuit board to achieve the purpose of thickening the coating.

[0004] In the existing PCB electroplating, the current is mainly used to deposit metal ions from the electrolyte onto the surface of the PCB to form a metal coating. The magnitude of the current directly affects the thickness and uniformity of the metal coating, which may cause the performance of the PCB to deteriorate, reduce reliability, or malfunction.

[0005] Therefore, an intelligent monitoring system and method for a circuit board electroplating production line is desired, which can realize abnormality detection and early warning functions by analyzing and processing real-time data and using machine learning or data mining algorithms. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an intelligent monitoring system and method for a circuit board electroplating production line, and an intelligent monitoring system for a circuit board electroplating production line based on real-time data analysis and machine learning for abnormality detection and early warning.

[0007] Accordingly, according to one aspect of the present application, an intelligent monitoring system for a circuit board electroplating production line is provided, comprising:

[0008] A circuit board electroplating production information acquisition module is used to obtain the current signal of the circuit board electroplating process, and to obtain the temperature values ​​at multiple predetermined time points within a predetermined time period from the temperature sensor and to obtain the humidity data at multiple predetermined time points within a predetermined time period from the humidity sensor;

[0009] A circuit board electroplating production information processing module, used to extract a circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and to extract a circuit board electroplating temperature feature vector and a circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the plurality of predetermined time points, respectively;

[0010] A circuit board electroplating production information fusion module is used to construct the circuit board electroplating current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain a circuit board electroplating key parameter feature vector;

[0011] The circuit board electroplating production information analysis module is used to pass the circuit board electroplating key parameter feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process.

[0012] According to another aspect of the present application, there is provided an intelligent monitoring method for a circuit board electroplating production line, comprising:

[0013] Acquire the current signal of the circuit board electroplating process, and acquire the temperature values ​​at multiple predetermined time points within a predetermined time period from the temperature sensor and acquire the humidity data at multiple predetermined time points within the predetermined time period from the humidity sensor;

[0014] Extracting a circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and extracting a circuit board electroplating temperature feature vector and a circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the plurality of predetermined time points, respectively;

[0015] Constructing the circuit board electroplating current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain the circuit board electroplating key parameter feature vector;

[0016] The circuit board electroplating key parameter feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process.

[0017] Compared with the prior art, the present application provides an intelligent monitoring system and method for a circuit board electroplating production line, which obtains current signals, temperature data, and humidity data, and extracts corresponding feature vectors and feature matrices to construct circuit board electroplating key parameter feature vectors, and then uses a classifier to analyze the circuit board electroplating key parameter feature vectors to obtain classification results, which are used to indicate whether there are abnormalities in the circuit board electroplating production process. This intelligent monitoring system can monitor the key parameters of the electroplating process in real time, perform abnormality detection and early warning, help prevent production line failures and quality problems, and improve the stability and reliability of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 It is a block diagram of an intelligent monitoring system for a circuit board electroplating production line according to an embodiment of the present application.

[0020] Figure 2 It is a block diagram of a circuit board electroplating production information processing module in an intelligent monitoring system of a circuit board electroplating production line according to an embodiment of the present application.

[0021] Figure 3 It is a block diagram of a current signal encoding unit in an intelligent monitoring system of a circuit board electroplating production line according to an embodiment of the present application.

[0022] Figure 4 It is a block diagram of a humidity feature extraction unit in an intelligent monitoring system of a circuit board electroplating production line according to an embodiment of the present application.

[0023] Figure 5 It is a flow chart of an intelligent monitoring method for a circuit board electroplating production line according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0025] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0026] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0027] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0028] Figure 1 The figure shows a block diagram of an intelligent monitoring system for a circuit board electroplating production line according to an embodiment of the present application. Figure 1 As shown, according to the embodiment of the present application, the intelligent monitoring system 100 of the circuit board electroplating production line includes: a circuit board electroplating production information acquisition module 110, which is used to obtain the current signal of the circuit board electroplating process, and obtain the temperature value of multiple predetermined time points within a predetermined time period from the temperature sensor and obtain the humidity data of multiple predetermined time points within a predetermined time period from the humidity sensor; a circuit board electroplating production information processing module 120, which is used to extract the circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and respectively extract the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the multiple predetermined time points; a circuit board electroplating production information fusion module 130, which is used to construct the circuit board electroplating current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain the circuit board electroplating key parameter feature vector; a circuit board electroplating production information analysis module 140, which is used to pass the circuit board electroplating key parameter feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process.

[0029] In the embodiment of the present application, the circuit board electroplating production information acquisition module 110 is used to obtain the current signal of the circuit board electroplating process, and to obtain the temperature values ​​of multiple predetermined time points within a predetermined time period from the temperature sensor and the humidity data of multiple predetermined time points within a predetermined time period from the humidity sensor. It should be understood that fault detection and prevention are achieved by monitoring key parameters such as temperature, humidity and current. By monitoring these parameters, the system can promptly detect abnormal conditions and take corresponding maintenance or adjustment measures to avoid production line shutdown and quality problems. The system uses electroplating technology to electrolyze copper cylinder potion with a high copper content, so that the copper water in the potion undergoes a reduction reaction and is reduced to the copper surface and circuit of the circuit board, so as to achieve the purpose of thickening the coating. The magnitude of the current directly affects the thickness and uniformity of the metal coating. Therefore, the intelligent monitoring system can realize accurate control and monitoring of the circuit board electroplating process by monitoring parameters such as current, thereby improving production efficiency and product quality. Specifically, it is considered that the magnitude of the current directly affects the thickness and uniformity of the metal coating. A larger current will result in a thicker coating, while a smaller current will result in a thinner coating. By monitoring the current, the rate of the electroplating reaction can be controlled to obtain the desired coating thickness and uniformity. An increase in temperature will increase the rate of the electroplating reaction, while a decrease in temperature will slow down the reaction rate. By monitoring the temperature, the rate of the electroplating reaction can be controlled so that the coating thickness and uniformity can be controlled and adjusted. Changes in temperature and humidity can affect the chemical properties of the electroplating solution, thereby affecting the quality of the coating. Too high or too low temperature and humidity may cause bubbles, poor crystallization, uneven color, etc. in the coating. By monitoring the temperature and humidity, the conditions of the electroplating solution can be adjusted in time to ensure the stability and consistency of the coating quality. Abnormal changes in temperature and humidity may be a precursor to failures in the electroplating process, such as equipment failure, changes in electrolyte concentration, etc. By monitoring key parameters in real time, real-time feedback on the electroplating process can be obtained. If there is an abnormality in temperature, humidity or current, the operator can immediately take measures to make adjustments to keep the electroplating process within a good working range. This real-time feedback and control helps prevent the expansion of faults and keep the production line running normally.

[0030] Specifically, sensors and monitoring equipment are installed to monitor the key parameters of the circuit board electroplating production line in real time, such as temperature, humidity and current. These sensors can be distributed in different key locations to cover the monitoring needs of the entire production line. Current sensors are usually used to measure the magnitude of current. They can provide numerical information of the current, but usually cannot directly provide the waveform of the current signal. To obtain the waveform of the current signal, you need to use other equipment or instruments to sample and record the current. A common method is to use an oscilloscope, which can convert the current signal into a waveform and display it on the screen. The oscilloscope can capture the changes in the current signal and display the fluctuation of the current over time in a graphical form. Use an oscilloscope to monitor the waveform of the current signal and apply it to the real-time monitoring and analysis of the current data to evaluate the stability of the electroplating process and the quality of the metal coating. Observe the stability of the current signal waveform to detect whether there is anomaly or fluctuation. A stable current waveform indicates that the current in the electroplating process is stable, which helps to obtain a uniform coating. Analyze the frequency component of the current signal waveform to detect whether there is frequency anomaly or resonance. Abnormal frequency components may cause problems in the electroplating process, such as uneven coating or abnormal reaction of the electrolyte.

[0031] In the embodiment of the present application, the circuit board electroplating production information processing module 120 is used to extract the circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and to extract the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the multiple predetermined time points. It should be understood that by analyzing the morphology, amplitude, frequency and other characteristics of the current signal, the key features that can reflect the state of the electroplating process are extracted. These features may include current peak value, periodicity of waveform, symmetry of waveform, etc., which are used to represent the change of current and abnormal conditions in the electroplating process. The purpose of extracting the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at multiple predetermined time points is to analyze the change trend and fluctuation of temperature and humidity in the electroplating process. These features may include average temperature, amplitude of temperature change, stability of humidity, etc., which are used to evaluate the temperature and humidity control in the electroplating process, and detect whether there are abnormal conditions. By extracting these feature vectors, the environmental conditions in the electroplating process can be more fully understood, and the stability and quality of the electroplating process can be assisted in judging.

[0032] Specifically, in one embodiment of the present application, Figure 2 The figure shows a block diagram of a circuit board electroplating production information processing module in an intelligent monitoring system of a circuit board electroplating production line according to an embodiment of the present application. Figure 2As shown, in the intelligent monitoring system 100 of the above-mentioned circuit board electroplating production line, the circuit board electroplating production information processing module 120 includes: a current signal encoding unit 121, which is used to extract multiple sampling windows from the waveform diagram of the current signal along the time dimension with a predetermined sampling window, and then obtain the circuit board electroplating current signal feature vector through convolution encoding; a temperature feature extraction unit 122, which is used to arrange the temperature values ​​of the multiple predetermined time points into a temperature input vector according to the time dimension and then obtain the circuit board electroplating temperature feature vector by using a third convolutional neural network model with a one-dimensional convolution kernel; a humidity feature extraction unit 123, which is used to obtain the circuit board electroplating humidity feature vector by passing the humidity data of the multiple predetermined time points through a time series encoder including a one-dimensional convolution layer and a fully connected layer.

[0033] Accordingly, in a specific example of the present application, the current signal encoding unit 121 is used to extract multiple sampling windows from the waveform diagram of the current signal along the timing dimension with a predetermined sampling window, and then obtain the circuit board electroplating current signal feature vector through convolution coding. It should be understood that the current signal can be divided into multiple small fragments using a predetermined sampling window, and each sampling window represents the current waveform within a period of time. The advantage of doing so is that the changing characteristics of the current signal in different time periods can be captured, such as instantaneous peak value, frequency, etc. Through the extraction of multiple sampling windows, a series of current signal fragments can be obtained. Next, these current signal fragments are processed by convolution coding to extract higher-level features. Convolution coding is a method that can learn local features, which can capture spatial correlations and important patterns in current signals through convolution operations. Through convolution coding, each current signal fragment can be mapped to a lower-dimensional feature vector, representing the key features in the fragment. Finally, these feature vectors are combined into a circuit board electroplating current signal feature vector, which can more conveniently analyze and process the current signal. Such a feature matrix can be used for tasks such as anomaly detection, classification, and clustering to help understand the characteristics and changing patterns of current signals and further improve the monitoring and control capabilities of the electroplating process.

[0034] further, Figure 3 The figure shows a block diagram of a current signal encoding unit in an intelligent monitoring system of a circuit board electroplating production line according to an embodiment of the present application. Figure 3As shown, in the circuit board electroplating production information processing module 120 of the intelligent monitoring system 100 of the above-mentioned circuit board electroplating production line, the current signal encoding unit 121 includes: a sampling window subunit 1211, which is used to intercept multiple sampling windows from the waveform of the current signal along the timing dimension with a predetermined sampling window, wherein the time intervals between each two adjacent sampling windows in the multiple sampling windows are consistent; a first convolution encoding subunit 1212, which is used to pass the waveform of each of the sampling windows through a first convolution neural network model to obtain multiple circuit board electroplating current signal feature vectors, and the first convolution neural network model uses a two-dimensional convolution kernel to extract the local implicit correlation features of the waveform of each of the sampling windows; a current signal feature encoding subunit 1213, which is used to feature encode the multiple circuit board electroplating current signal feature vectors to obtain the circuit board electroplating current signal feature vector.

[0035] Specifically, the sampling window subunit 1211 is used to intercept multiple sampling windows from the waveform diagram of the current signal along the timing dimension with a predetermined sampling window, wherein the time interval between each adjacent two sampling windows in the multiple sampling windows is consistent. It should be understood that the current signal may have important features on different time scales. By keeping the time intervals between adjacent sampling windows consistent, it can be ensured that unnecessary deviations are not introduced due to differences in time scales when analyzing the current signal. The current signal is usually dynamically changing, and the interval of the sampling window determines the speed of change and frequency range that we can capture. If the time interval between the sampling windows is too large, some rapidly changing features may be missed. On the contrary, if the time interval is too small, redundant information may be introduced. By maintaining a consistent time interval, the dynamic changes of the current signal can be captured on an appropriate time scale. If the time intervals between adjacent sampling windows are inconsistent, the features extracted from different sampling windows may have different resolutions and expressive capabilities. This will lead to inconsistency of features, making it difficult to make effective comparisons and analyses. By maintaining a consistent time interval, it can be ensured that the features extracted from each sampling window have similar time resolution and expressive capabilities, making them more comparable and consistent.

[0036] Specifically, the first convolutional encoding subunit 1212 is used to pass the waveform of each sampling window through the first convolutional neural network model to obtain multiple circuit board electroplating current signal feature vectors, and the first convolutional neural network model uses a two-dimensional convolution kernel to extract the local implicit correlation features of the waveform of each sampling window. It should be understood that considering the extraction of local implicit correlation features in the sampling window through the convolution operation, the characteristic information of the current signal can be better represented and captured. The waveform of the sampling window can be effectively processed using a two-dimensional convolution kernel. The two-dimensional convolution kernel can slide in two dimensions of time and amplitude to capture the local features in the waveform. Furthermore, the waveform of the current signal has continuous correlation in the time dimension, and there may also be certain correlation in the amplitude dimension. The two-dimensional convolution kernel can simultaneously consider the correlation in these two dimensions, so as to better capture the spatial correlation between local features. The two-dimensional convolution kernel can extract local features at different positions by means of a sliding window. These local features can capture important patterns, instantaneous peaks or frequencies in the waveform. Through the convolution operation, these local features can be combined and integrated to obtain a higher level of feature representation. Convolutional neural networks have the characteristic of parameter sharing, that is, the convolution kernel shares weights on the entire input. This sharing can reduce the number of model parameters and improve the efficiency and generalization ability of the model. For the waveform of the current signal, since different sampling windows have similar features, parameter sharing can better utilize this similarity and reduce the risk of overfitting.

[0037] Specifically, the current signal feature encoding subunit 1213 is used to feature encode the multiple circuit board electroplating current signal feature vectors to obtain the circuit board electroplating current signal feature vectors. More specifically, after the multiple circuit board electroplating current signal feature vectors are arranged in two dimensions, the circuit board electroplating current signal feature vectors are obtained by using the second convolutional neural network model as a filter using the global conversion matrix. It should be understood that after the multiple circuit board electroplating current signal feature vectors are arranged into a feature matrix, the information of these feature vectors can be integrated together. Through the filter operation of the convolutional neural network, a sliding window operation can be performed on the feature matrix to extract local and global features in the feature matrix. This integration can help capture the interaction and dependency between features and improve the expressiveness of features. The second convolutional neural network with the global conversion matrix as a filter can focus on the global information of the entire feature matrix. In this way, the global pattern, long-term dependency and contextual information in the feature matrix can be captured. Through the operation of the global conversion matrix, the features in the feature matrix can be globally weighted and adjusted, so as to extract more representative and distinguishing features. The arrangement of the feature matrix and the second convolutional neural network of the global conversion matrix can expand and enhance the dimension of the feature matrix. Through the convolution operation, the dimension of the feature matrix can be mapped from low-level features to high-level features.

[0038] Accordingly, the current signal feature encoding subunit is used to: obtain the circuit board electroplating current signal feature vector by using the second convolutional neural network using the global conversion matrix as a filter after two-dimensionally arranging the plurality of circuit board electroplating current signal feature vectors into a feature matrix according to the following formula;

[0039] Wherein, the formula is:

[0040] f i =active(N i ×f i-1 +B i )

[0041] Among them, f i-1 is the input of the i-th layer neural network, f i is the output of the i-th layer neural network, N i is the transformation matrix of the i-th layer neural network, and B i is the bias matrix of the i-th layer neural network, and active represents the nonlinear activation function.

[0042] Accordingly, in a specific example of the present application, the temperature feature extraction unit 122 is used to arrange the temperature values ​​of the multiple predetermined time points as a temperature input vector according to the time dimension, and then obtain the circuit board electroplating temperature feature vector by using the third convolutional neural network model of the one-dimensional convolution kernel. It should be understood that the temperature value usually has time correlation, that is, there may be a certain correlation between the temperature values ​​of adjacent time points. By arranging the temperature values ​​of the predetermined time points as a temperature input vector according to the time dimension, the time information of the time series data can be retained, so that the model can model and learn the time correlation. The one-dimensional convolution kernel can perform a sliding window operation on the time dimension to extract local features in the time series data. These local features can capture information such as instantaneous changes, trends, and periodicity of temperature. Through the convolution operation, these local features can be combined and integrated to obtain a higher level of feature representation. Similar to the two-dimensional convolutional neural network, the one-dimensional convolutional neural network also has the characteristic of parameter sharing. This sharing can reduce the number of parameters of the model and improve the efficiency and generalization ability of the model. For the temperature input vector of time series data, since different time points have similar characteristics, parameter sharing can better utilize this similarity and reduce the risk of overfitting.

[0043] Furthermore, in the circuit board electroplating production information processing module 120 of the intelligent monitoring system 100 of the above-mentioned circuit board electroplating production line, the temperature feature extraction unit 122 includes: a temperature feature convolution subunit, which is used to perform convolution processing on the input data based on a one-dimensional convolution kernel to generate a convolution feature map; a temperature feature pooling subunit, which is used to perform global mean pooling processing on the convolution feature map based on a feature matrix to generate a pooling feature map; a temperature feature activation subunit, which is used to perform nonlinear activation on the feature values ​​of each position in the pooling feature map to generate an activation feature enhancement map; wherein the input of the first convolutional neural network model is the temperature input vector, the input from the second layer to the last layer of the third convolutional neural network model is the output of the previous layer, and the output of the last layer of the third convolutional neural network model is the circuit board electroplating temperature feature vector.

[0044] Accordingly, in a specific example of the present application, the humidity feature extraction unit 123 is used to obtain a circuit board electroplating humidity feature vector by passing the humidity data of the multiple predetermined time points through a time series encoder including a one-dimensional convolution layer and a fully connected layer. It should be understood that the one-dimensional convolution layer can perform a sliding window operation on the time dimension to extract local and global features from the humidity data. These features can capture information such as the change pattern, trend, and periodicity of humidity. Through the convolution operation, these features can be combined and integrated to obtain a higher level of feature representation. The time series encoder can model the time correlation of humidity data. Humidity data usually has time correlation, that is, there may be a certain correlation between the humidity values ​​at adjacent time points. Through the time series encoder including a one-dimensional convolution layer, the time information of the time series data can be retained, so that the model can model and learn the time correlation. The time series encoder can convert high-dimensional humidity data into a low-dimensional circuit board electroplating humidity feature vector. Through the operation of the one-dimensional convolution layer and the fully connected layer, the dimension of the humidity data can be compressed and converted to obtain a more compact and expressive circuit board electroplating humidity feature vector. This can reduce the number of model parameters and computational complexity, and improve the model's efficiency and generalization ability. The time series encoder can encode humidity data into feature vectors, allowing the model to better process and understand humidity data. Feature vectors can capture important features and patterns of humidity data, providing a basis for subsequent analysis, prediction, and decision-making.

[0045] further, Figure 4 The figure shows a block diagram of a temperature feature extraction unit in an intelligent monitoring system of a circuit board electroplating production line according to an embodiment of the present application. Figure 4 As shown, in the circuit board electroplating production information processing module 120 of the above-mentioned intelligent monitoring system 100 of the circuit board electroplating production line, the humidity feature extraction unit 123 includes: a humidity feature input vector subunit 1231, which is used to arrange the humidity data of the multiple predetermined time points into a humidity data input vector according to the time dimension; a humidity feature fully connected layer subunit 1232, which is used to use the fully connected layer of the time series encoder to fully connect the humidity data input vector to extract the high-dimensional implicit features of the eigenvalues ​​at each position in the humidity data input vector; a humidity feature one-dimensional convolutional layer subunit 1233, which is used to use the one-dimensional convolutional layer of the time series encoder to perform one-dimensional convolutional encoding on the humidity data input vector to extract the high-dimensional implicit correlation features of the correlation between the eigenvalues ​​at each position in the humidity data input vector.

[0046] In the embodiment of the present application, the circuit board electroplating production information fusion module 130 is used to construct the current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain the circuit board electroplating key parameter feature vector. It should be understood that parameters such as current signal, temperature and humidity usually represent the status and characteristics of different aspects of the monitored system. By integrating the features extracted from them separately, a comprehensive feature representation can be formed to more comprehensively reflect the status of the monitored system. Parameters such as current signal, temperature and humidity often have different scales and distributions. By constructing the circuit board electroplating key parameter feature vector, the features of these parameters can be fused to eliminate the scale differences between different parameters so that they are comparable in the same feature space. Constructing the circuit board electroplating key parameter feature vector can reduce and compress the dimensions of multiple parameters. This can reduce the dimension of the data, improve the calculation efficiency, remove some redundant information, and highlight the key features. The circuit board electroplating key parameter feature vector can promote the feature interaction between different parameters. By integrating the features of parameters such as current signal, temperature and humidity, the interaction and association between them can be found, and higher-level features can be further extracted.

[0047] In one embodiment of the present application, the circuit board electroplating production information fusion module includes: a vector fusion unit, which is used to fuse the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain a temperature-humidity correlation feature vector; a vector optimization unit, which is used to perform feature importance perception optimization based on multi-perspective feature quantization on the temperature-humidity correlation feature vector based on the circuit board electroplating current signal feature vector to obtain an optimized temperature-humidity correlation feature vector; a matrix fusion subunit, which is used to fuse the circuit board electroplating current signal feature vector and the optimized temperature-humidity correlation feature vector to obtain the circuit board electroplating key parameter feature vector. Specifically, the matrix fusion subunit is used to: fuse the circuit board electroplating current signal feature vector and the optimized temperature-humidity correlation feature vector using the following cascade formula to obtain the circuit board electroplating key parameter feature vector;

[0048] Wherein, the cascade formula is:

[0049] M=Concat[M1,M2]

[0050] Among them, M1 represents the characteristic vector of the circuit board electroplating current signal, M2 represents the optimized temperature-humidity correlation characteristic vector, M represents the circuit board electroplating key parameter characteristic vector, and Concat[·,·] represents the cascade function.

[0051] In the embodiment of the present application, the circuit board electroplating production information analysis module 140 is used to pass the circuit board electroplating key parameter feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process. It should be understood that various abnormal conditions may exist in the circuit board electroplating production process, such as too large or too small current, abnormal temperature, abnormal humidity, etc. By constructing the circuit board electroplating key parameter feature vector and using the classifier, the normal and abnormal production processes can be distinguished, thereby achieving the purpose of abnormality detection. The classifier can learn and identify the patterns and characteristics of normal and abnormal production processes. By training the classifier to use the circuit board electroplating key parameter feature vector for classification, a model of the normal production process can be established, and the new production process can be compared with the model to determine whether there is an abnormality. Inputting the circuit board electroplating key parameter feature vector into the classifier can achieve real-time monitoring and judgment. In the circuit board electroplating production process, by continuously collecting and updating the circuit board electroplating key parameter feature vector and inputting it into the classifier, abnormal conditions in the production process can be detected in time, and measures can be taken in time to deal with them. The classification results can be provided to operators or automation systems to indicate whether there are any abnormalities in the PCB electroplating production process. Based on the classification results, timely measures can be taken to adjust, correct or alarm to ensure the stability of the production process and product quality.

[0052] In the technical solution of the present application, it is considered that the current signal is a direct record of electrochemical activity in the electroplating process, reflecting the electrical characteristics of the electroplating process; while the temperature and humidity data are measures of environmental conditions, which may be related to the chemical reaction and physical state of the electroplating process, but are not direct indicators of electrochemical activity. This difference in data characteristics may lead to inconsistencies in the feature expression of the two. The circuit board electroplating current signal feature vector extracts local implicit correlation features through a two-dimensional convolution kernel, focusing on the waveform analysis of the current signal; while the temperature-humidity correlation feature vector is extracted through a one-dimensional convolution layer and a fully connected layer, focusing on the analysis of time series data. These two different feature extraction methods may lead to semantic differences in feature vectors. When fusing the circuit board electroplating current signal feature vector and the temperature-humidity correlation feature vector, if the fusion strategy does not fully consider the temporal characteristics and interactions of the two, it may lead to the lack of a posteriori-priori causal association. In order to solve this problem, in the technical solution of the present application, based on the circuit board electroplating current signal feature vector, the temperature-humidity correlation feature vector is optimized based on multi-view feature quantization to obtain an optimized temperature-humidity correlation feature vector.

[0053] Among them, based on the circuit board electroplating current signal feature vector, the temperature-humidity correlation feature vector is optimized based on the feature importance perception of multi-view feature quantization to obtain an optimized temperature-humidity correlation feature vector, including: calculating the position difference between the temperature-humidity correlation feature vector and the circuit board electroplating current signal feature vector to obtain a first viewpoint metric representation vector; calculating the position point multiplication between the temperature-humidity correlation feature vector and the circuit board electroplating current signal feature vector to obtain a second viewpoint metric representation vector; calculating the position sum between the temperature-humidity correlation feature vector and the circuit board electroplating current signal feature vector to obtain a third viewpoint metric representation vector. A quantity representation vector; calculating the gradient amplitude of each position in the first viewpoint measurement representation vector, the second viewpoint measurement representation vector and the third viewpoint measurement representation vector to obtain a first viewpoint measurement perception vector, a second viewpoint measurement perception vector and a third viewpoint measurement perception vector; inputting the first viewpoint measurement perception vector, the second viewpoint measurement perception vector and the third viewpoint measurement perception vector into a backward machine learning reverse multi-viewpoint perception optimization module to obtain a reverse multi-viewpoint perception optimization representation vector; calculating the position point multiplication between the reverse multi-viewpoint perception optimization representation vector and the temperature-humidity association feature vector to obtain the optimized temperature-humidity association feature vector.

[0054] The backward machine learning reverse multi-view perception optimization module is expressed as follows:

[0055]

[0056] in, and represents the eigenvalues ​​of the i-th position of the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector, respectively, and μ represents the three eigenvalues ​​of the corresponding positions in the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector , σ is the standard deviation of the three eigenvalues ​​at corresponding positions in the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector, represents the maximum value of the three eigenvalues ​​at the corresponding positions in the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector, ReLU represents a linear rectification function, softmax represents a normalized exponential function, and v ri Represents the eigenvalue of the i-th position of the inverse multi-view perceptual optimization representation vector.

[0057] Specifically, in the technical solution of the present application, based on the circuit board electroplating current signal feature vector, the temperature-humidity correlation feature vector is optimized based on the feature importance perception of multi-view feature quantization, and the process includes first calculating the position difference, point multiplication and sum between the temperature-humidity correlation feature vector and the circuit board electroplating current signal feature vector through multi-view analysis (Multi-View Analysis) to obtain three viewpoint metric representation vectors, and quantify the relationship between features from different angles. Then, the gradient amplitude of each position in these viewpoint metric representation vectors is calculated through backward error perception to identify key change points to provide a basis for feature optimization.

[0058] Then, the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector are input into the backward machine learning reverse multi-viewpoint perception optimization module, wherein the backward machine learning reverse multi-viewpoint perception optimization module captures the significant comprehensive ability factor of the feature importance perception (Feature Importance Perception) of the feature sub-dimension distribution of the corresponding position in the first to third viewpoint metric perception vectors by adopting a squeeze-incentive mechanism composed of -functions. Then, the position point multiplication between the reverse multi-viewpoint perception optimization representation vector and the temperature-humidity correlation feature vector is calculated to perform backward optimization perception (Backward Optimization Perception) on the temperature-humidity correlation feature vector to adjust the temperature-humidity correlation feature vector so as to better associate with the circuit board electroplating current signal feature vector, thereby enhancing the causal relationship between the temperature-humidity correlation feature vector and the circuit board electroplating current signal feature vector.

[0059] In one embodiment of the present application, the circuit board electroplating production information analysis module includes: a fully connected encoding unit, used to use the fully connected layer of the classifier to fully connect the circuit board electroplating key parameter feature vector to obtain a fully connected encoded feature vector; a probability acquisition unit, used to pass the fully connected encoded feature vector through the Softmax classification function of the classifier to obtain a first probability attributable to the existence of an abnormality in the circuit board electroplating production process and a second probability attributable to the absence of an abnormality in the circuit board electroplating production process; a classification result determination unit, used to determine the classification result based on a comparison between the first probability and the second probability.

[0060] In summary, the intelligent monitoring system and method of the circuit board electroplating production line according to the embodiment of the present application obtains current signals, temperature data and humidity data, and extracts corresponding feature vectors and feature matrices to construct circuit board electroplating key parameter feature vectors, and then uses a classifier to analyze the circuit board electroplating key parameter feature vectors to obtain classification results, which are used to indicate whether there are abnormalities in the circuit board electroplating production process. This intelligent monitoring system can monitor the key parameters in the electroplating process in real time, perform abnormality detection and early warning, help prevent production line failures and quality problems, and improve the stability and reliability of the production line.

[0061] As described above, the intelligent monitoring system 100 for the circuit board electroplating production line according to the embodiment of the present application can be implemented in various terminal devices, such as a server of the intelligent monitoring system for the circuit board electroplating production line. In one example, the intelligent monitoring system 100 for the circuit board electroplating production line can be integrated into the terminal device as a software module and / or a hardware module. For example, the intelligent monitoring system 100 for the circuit board electroplating production line can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the intelligent monitoring system 100 for the circuit board electroplating production line can also be one of the many hardware modules of the terminal device.

[0062] Alternatively, in another example, the intelligent monitoring system 100 of the circuit board electroplating production line and the terminal device may also be separate devices, and the intelligent monitoring system 100 of the circuit board electroplating production line may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0063] Figure 5 Flow chart of the intelligent monitoring method of the circuit board electroplating production line according to the embodiment of the present application. Figure 5As shown, the intelligent monitoring method of the circuit board electroplating production line according to the embodiment of the present application includes the steps of: S110, obtaining the current signal of the circuit board electroplating process, and obtaining the temperature values ​​at multiple predetermined time points within a predetermined time period from the temperature sensor and obtaining the humidity data at multiple predetermined time points within the predetermined time period from the humidity sensor; S120, extracting the circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and extracting the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the multiple predetermined time points respectively; S130, constructing the circuit board electroplating current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain the circuit board electroplating key parameter feature vector; S140, passing the circuit board electroplating key parameter feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process.

[0064] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned intelligent monitoring method for the circuit board electroplating production line have been described in the above reference. Figures 1 to 4 The description of the intelligent monitoring system for the circuit board electroplating production line has been introduced in detail, and therefore, its repeated description will be omitted.

[0065] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0066] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0067] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0068] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0069] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0070] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent monitoring system for a circuit board electroplating production line, characterized in that: include: A circuit board electroplating production information acquisition module is used to obtain the current signal of the circuit board electroplating process, and to obtain the temperature values ​​at multiple predetermined time points within a predetermined time period from the temperature sensor and to obtain the humidity data at multiple predetermined time points within a predetermined time period from the humidity sensor; A circuit board electroplating production information processing module, used to extract a circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and to extract a circuit board electroplating temperature feature vector and a circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the plurality of predetermined time points, respectively; A circuit board electroplating production information fusion module is used to construct the circuit board electroplating current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain a circuit board electroplating key parameter feature vector; The circuit board electroplating production information analysis module is used to pass the circuit board electroplating key parameter feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process.

2. The intelligent monitoring system for the circuit board electroplating production line according to claim 1 is characterized in that: The circuit board electroplating production information processing module includes: A current signal encoding unit, used to extract multiple sampling windows from the waveform diagram of the current signal along the time dimension with a predetermined sampling window and then perform convolution encoding to obtain a characteristic vector of the circuit board electroplating current signal; A temperature feature extraction unit, configured to arrange the temperature values ​​of the plurality of predetermined time points into a temperature input vector according to the time dimension and obtain a circuit board electroplating temperature feature vector by using a third convolutional neural network model with a one-dimensional convolution kernel; The humidity feature extraction unit is used to obtain a circuit board electroplating humidity feature vector by passing the humidity data at the plurality of predetermined time points through a temporal encoder including a one-dimensional convolutional layer and a fully connected layer.

3. The intelligent monitoring system for the circuit board electroplating production line according to claim 2 is characterized in that: The current signal encoding unit comprises: A sampling window subunit, used to intercept a plurality of sampling windows from the waveform diagram of the current signal along a timing dimension using a predetermined sampling window, wherein the time intervals between each two adjacent sampling windows in the plurality of sampling windows are consistent; A first convolutional encoding subunit, used for passing the waveform graphs of each of the sampling windows through a first convolutional neural network model to obtain a plurality of circuit board electroplating current signal feature vectors, wherein the first convolutional neural network model uses a two-dimensional convolution kernel to extract local implicit correlation features of the waveform graphs of each of the sampling windows; The current signal feature encoding subunit is used to perform feature encoding on the plurality of circuit board electroplating current signal feature vectors to obtain the circuit board electroplating current signal feature vector.

4. The intelligent monitoring system for the circuit board electroplating production line according to claim 3 is characterized in that: The current signal characteristic encoding subunit comprises: Arranging the plurality of circuit board electroplating current signal feature vectors in two dimensions to obtain a circuit board electroplating current signal correlation feature matrix; The circuit board electroplating current signal correlation feature matrix is ​​used to obtain the circuit board electroplating current signal feature vector through a second convolutional neural network model using a global conversion matrix as a filter.

5. The intelligent monitoring system for the circuit board electroplating production line according to claim 4 is characterized in that: The humidity feature extraction unit comprises: A humidity feature input vector subunit, used for arranging the humidity data of the plurality of predetermined time points into a humidity data input vector according to a time dimension; A humidity feature fully connected layer subunit, used for performing fully connected encoding on the humidity data input vector using the fully connected layer of the temporal encoder to extract high-dimensional implicit features that are feature values ​​at each position in the humidity data input vector; The humidity feature one-dimensional convolution layer subunit is used to use the one-dimensional convolution layer of the temporal encoder to perform one-dimensional convolution encoding on the humidity data input vector to extract the high-dimensional implicit correlation features of the correlation between the feature values ​​at each position in the humidity data input vector.

6. The intelligent monitoring system for the circuit board electroplating production line according to claim 5 is characterized in that: The circuit board electroplating production information fusion module includes: A vector fusion unit, used for fusing the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain a temperature-humidity correlation feature vector; A vector optimization unit, used for performing feature importance perception optimization based on multi-view feature quantization on the temperature-humidity correlation feature vector based on the circuit board electroplating current signal feature vector to obtain an optimized temperature-humidity correlation feature vector; The matrix fusion subunit is used to fuse the circuit board electroplating current signal feature vector and the optimized temperature-humidity correlation feature vector to obtain the circuit board electroplating key parameter feature vector.

7. The intelligent monitoring system for the circuit board electroplating production line according to claim 6 is characterized in that: The vector optimization unit comprises: Calculating the position difference between the temperature-humidity association feature vector and the circuit board electroplating current signal feature vector to obtain a first viewpoint metric representation vector; Calculate the position point multiplication between the temperature-humidity association feature vector and the circuit board electroplating current signal feature vector to obtain a second viewpoint metric representation vector; Calculating the positional sum of the temperature-humidity association feature vector and the circuit board electroplating current signal feature vector to obtain a third viewpoint metric representation vector; Calculating the gradient magnitude of each position in the first viewpoint metric representation vector, the second viewpoint metric representation vector, and the third viewpoint metric representation vector to obtain a first viewpoint metric perception vector, a second viewpoint metric perception vector, and a third viewpoint metric perception vector; Inputting the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector into a backward machine learning inverse multi-view perception optimization module to obtain a reverse multi-view perception optimization representation vector; The optimized temperature-humidity association feature vector is obtained by calculating the point-by-point multiplication between the inverse multi-view perception optimization representation vector and the temperature-humidity association feature vector.

8. The intelligent monitoring system for the circuit board electroplating production line according to claim 7 is characterized in that: The circuit board electroplating production information analysis module includes: A fully connected encoding unit, used for performing fully connected encoding on the circuit board electroplating key parameter feature vector using the fully connected layer of the classifier to obtain a fully connected encoding feature vector; A probability obtaining unit, used for passing the fully connected encoded feature vector through the Softmax classification function of the classifier to obtain a first probability attributable to the presence of an abnormality in the circuit board electroplating production process and a second probability attributable to the absence of an abnormality in the circuit board electroplating production process; A classification result determination unit is used to determine the classification result based on a comparison between the first probability and the second probability.

9. An intelligent monitoring method for a circuit board electroplating production line, characterized in that: include: Acquire the current signal of the circuit board electroplating process, and acquire the temperature values ​​at multiple predetermined time points within a predetermined time period from the temperature sensor and acquire the humidity data at multiple predetermined time points within the predetermined time period from the humidity sensor; Extracting a circuit board electroplating current signal feature vector from the waveform diagram of the current signal, and extracting a circuit board electroplating temperature feature vector and a circuit board electroplating humidity feature vector from the temperature values ​​and humidity data at the plurality of predetermined time points, respectively; Constructing the circuit board electroplating current signal feature vector, the circuit board electroplating temperature feature vector and the circuit board electroplating humidity feature vector to obtain the circuit board electroplating key parameter feature vector; The circuit board electroplating key parameter feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the circuit board electroplating production process.

Images