Water purification control methods and systems in PCB manufacturing process

Abstract

This invention belongs to the field of PCB manufacturing technology, specifically relating to a water purification control method and system in the PCB manufacturing process. This invention achieves proactive control by combining production planning with wastewater reuse. Based on historical data and production plans, it predicts the water quality of each batch of wastewater, classifies the wastewater according to the prediction results, and allocates differentiated reuse ratios. Through a feedforward-feedback composite control algorithm, it accurately executes the predetermined reuse ratio, transforming the original passive and lagging mode into feedforward predictive control. This reduces system impact from the source, thereby improving the water resource reuse rate while ensuring the safety of the ultrapure water system. Furthermore, by assessing the impact of future production scheduling on wastewater reuse, it can reverse-optimize the production plan, achieving closed-loop collaborative optimization from wastewater reuse to production management. This avoids the isolation of water recycling and production planning information, facilitating optimization based on future production fluctuations.

Description

Technical Field

[0001] This invention belongs to the field of PCB manufacturing technology, specifically relating to a water purification control method and system in the PCB manufacturing process. Background Technology

[0002] PCBs, also known as printed circuit boards, require large amounts of ultrapure water for cleaning in multiple stages of their production, generating wastewater in the process. Water purification involves a series of purification processes to prepare ultrapure water. Since some processes generate wastewater of high quality, it is often returned to the ultrapure water preparation system for further purification, thus conserving resources.

[0003] Current water purification control models are passive and outdated. They adjust the mixing ratio of recycled wastewater through real-time monitoring after wastewater is generated, making it difficult to predict water quality fluctuations and posing a risk to the ultrapure water system. Furthermore, the recycling strategy is crude, and the use of a fixed recycling ratio makes it difficult to balance water conservation and equipment safety, resulting in low recycling rates, high system operation risks, and the isolation of wastewater recycling and production planning information, making it difficult to optimize based on future production fluctuations. Summary of the Invention

[0004] The purpose of this invention is to provide a water purification control method and system for PCB manufacturing processes. This system achieves proactive control by combining production planning with wastewater reuse. Based on historical data and production plans, it predicts the water quality of each batch of wastewater, classifies the wastewater according to the prediction results, and allocates differentiated reuse ratios. Through a feedforward-feedback composite control algorithm, it accurately executes the predetermined reuse ratio, transforming the original passive and lagging mode into feedforward predictive control. This reduces system impact at the source, thereby improving water resource reuse rates while ensuring the safety of the ultrapure water system. Furthermore, by assessing the impact of future production scheduling on wastewater reuse, it can optimize the production plan in reverse, achieving closed-loop collaborative optimization from wastewater reuse to production management. This avoids the isolation of water recycling and production planning information, facilitating optimization based on future production fluctuations.

[0005] The specific technical solution adopted by this invention is as follows: A water purification control method for PCB manufacturing process, wherein the method is applied to a cleaning section where each batch of cleaning wastewater can be discharged independently, and the cleaning section has an independent wastewater reuse pipeline; the method includes: Obtain the production plan schedule for a future preset period and analyze the process route information for each production batch. The process route information includes at least the process parameters of the preceding processing steps. Based on the correlation analysis of historical production data and measured water quality data, a correspondence model between process parameters and water quality parameters is established. Using the correspondence model, the water quality parameters of the cleaning wastewater generated in each batch are predicted according to the process route information. The predicted water quality parameters are compared with the corresponding preset thresholds. Based on the comparison results, the reuse level of each batch of cleaning wastewater is determined, and a reuse ratio is assigned to each reuse level. In the actual production process, the cleaning wastewater is transported to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch.

[0006] In a preferred embodiment, the steps for obtaining the production plan schedule for a future preset time period and parsing the process route information for each production batch are as follows: Retrieve the production plan schedule for the future preset time period from the production management system. The production plan schedule includes the identification information of each batch to be produced and its planned production time. Based on the identification information of each batch, retrieve the corresponding process route information from the process database of the production management system; Analyze the process route information to identify the process sequence of the preceding processing steps; Extract the process parameters for each preceding processing step from the process route information. The process parameters include the process type, the name of the main chemical solution used, the concentration range of the chemical solution, and the standard processing time.

[0007] In a preferred embodiment, the process involves establishing a correlation model between process parameters and water quality parameters based on the correlation analysis of historical production data and measured water quality data. Using this model, and based on the process route information, the water quality parameters of the cleaning wastewater generated in each batch are predicted. The specific steps are as follows: Collect historical production data, which includes the process parameters of the preceding processing steps of each production batch that has been completed in history. Collect historical measured water quality data corresponding to historical production data. The historical measured water quality data is obtained by sampling and testing the cleaning wastewater generated in the subsequent cleaning process after each production batch has completed the previous processing steps. Feature engineering is performed on process parameters in historical production data to construct feature variables. The feature engineering process includes: converting the process sequence into a numerical vector representing the composition of the process, converting the concentration range of the drug solution into a single numerical value representing the concentration level, and extracting the statistical features of the processing time. The constructed feature variables and the corresponding historical measured water quality data are used as training samples, and one or more algorithms selected from linear regression, support vector regression and decision tree regression are used to train the corresponding relationship model. For the production batch to be predicted, the process parameters of its preceding processing steps are extracted based on its process route information to form an input feature vector representing the process characteristics of the batch. The input feature vector is fed into the correspondence model, and the prediction calculation is performed to output the predicted values ​​of the water quality parameters of the cleaning wastewater of the production batch to be predicted.

[0008] In a preferred embodiment, the steps of comparing the predicted water quality parameters with corresponding preset thresholds, determining the reuse level of each batch of cleaning wastewater based on the comparison results, and assigning a reuse ratio to each reuse level are as follows: For each key water quality parameter, a safety threshold and a danger threshold are set, where the danger threshold is greater than the safety threshold. The steps to determine the reuse level include: If the predicted values ​​of all water quality parameters of a production batch are lower than their respective safety thresholds, then the reuse level of the wastewater in that batch is determined to be Level 1. If at least one of the predicted water quality parameters of a production batch is higher than or equal to its safety threshold, but all predicted values ​​are lower than their respective danger thresholds, then the reuse level of the wastewater in that batch is determined to be Level II. If at least one of the predicted water quality parameters for a production batch is higher than or equal to its danger threshold, the reuse level of that batch of wastewater is determined to be Level III. The recycling ratio is allocated as follows: 80% to 100% for Level 1 recycling, 30% to 70% for Level 2 recycling, and 0% to 20% for Level 3 recycling.

[0009] In a preferred embodiment, during the actual production process, the cleaning wastewater is transported to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch. The specific steps are as follows: After the main wastewater outlet pipe of the cleaning process, a branch pipeline is set up to divide the wastewater into two paths, one of which is connected to the raw water tank of the ultrapure water preparation system, and the other is connected to the conventional wastewater treatment system. Install an electric regulating valve on the pipeline leading to the raw water tank of the ultrapure water preparation system; Based on the current production batch's reuse level, the corresponding reuse ratio is determined. During the wastewater transport process, the total wastewater flow rate is monitored in real time, and the target flow rate to the ultrapure water preparation system is calculated based on the reuse ratio and the total wastewater flow rate. Based on the target flow rate and the real-time detected branch flow rate, a control signal is generated and sent to the electric regulating valve to dynamically adjust its opening, so that the deviation between the wastewater flow rate entering the ultrapure water preparation system and the target flow rate is maintained within the preset error allowable range. The remaining wastewater after diversion will be transported to a conventional wastewater treatment system.

[0010] In a preferred embodiment, the method further includes updating the correspondence model, with the following specific steps: After the actual production batch is completed, the actual water quality test data of the cleaning wastewater of that batch is collected in order to monitor the performance predicted by the model. The actual water quality data is compared with the predicted water quality parameters obtained for this batch, and the prediction deviation is calculated. When the number of consecutive batches reaches the first preset threshold and the average prediction deviation of these batches exceeds the second preset fault tolerance threshold, it is determined that the original correspondence model has become inaccurate, and the model update process is triggered. The model update process includes adding newly collected actual production data and water quality data as new samples to the historical training dataset, and retraining the model using one or more algorithms selected from linear regression, support vector regression, and decision tree regression to generate an updated correspondence model. During planned system downtime, or when the total wastewater flow rate of the system is lower than the preset flow threshold, the online correspondence model will be automatically switched to the updated model.

[0011] In a preferred embodiment, the step of generating a control signal based on the target flow rate and the real-time detected branch flow rate and sending it to the electric regulating valve to dynamically adjust its opening is as follows: Based on the calculated target flow rate, the feedforward control quantity is calculated using the pre-generated flow rate-valve opening correspondence. The monitoring value of the branch flow is acquired in real time, compared with the target flow value, and the feedback correction amount is calculated by the proportional-integral controller based on the deviation between the two. The feedforward control quantity and the feedback correction quantity are superimposed to calculate the final opening control signal and send it to the electric regulating valve to dynamically adjust its opening.

[0012] In one preferred embodiment, after determining the reuse level of each planned batch and before actual production, the production schedule is optimized for load balancing. The specific steps are as follows: Iterate through and parse the production planning and scheduling data within the future preset time period to identify all production batches that are judged to be of the third-level reuse level; Calculate the maximum consecutive sequence length formed by these three-level reuse batches in the time series; If the maximum consecutive sequence length exceeds the first preset threshold, it is determined that there is an excessive concentration of high-pollution load batches in the current schedule, and the optimization process is initiated. The optimization process involves finding an insertion point in a continuous sequence of three-level reuse batches that meets the requirements of production cycle time and equipment availability, and swapping the planned subsequent first- or second-level reuse batches to that position. Generate a production plan optimization scheme that includes optimized processes, and send this scheme to the production management system.

[0013] This invention also provides a water purification control system for PCB manufacturing processes, using the aforementioned water purification control method for PCB manufacturing processes, comprising: The scheduling parsing module is used to obtain the production plan schedule for a future preset time period and parse the process route information of each production batch. The process route information includes at least the process parameters of the preceding processing steps. The water quality prediction module is used to establish a correspondence model between process parameters and water quality parameters based on the correlation analysis between historical production data and measured water quality data. Using the correspondence model, the water quality parameters of the cleaning wastewater generated in each batch are predicted according to the process route information. The reuse decision module is used to compare the predicted water quality parameters with the corresponding preset thresholds, determine the reuse level of each batch of cleaning wastewater based on the comparison results, and assign a reuse ratio to each reuse level. The control and execution module is used to transport the cleaning wastewater to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch during the actual production process.

[0014] And, an electronic device, the electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the water purification control method for PCB production process according to any one of claims 1 to 8.

[0015] The technical effects achieved by this invention are as follows: This invention achieves forward-looking control by combining production planning with wastewater reuse. Based on historical data and production plans, it predicts the water quality of each batch of wastewater, classifies the wastewater according to the prediction results, and allocates differentiated reuse ratios. Through a feedforward-feedback composite control algorithm, it accurately executes the predetermined reuse ratio, transforming the original passive and lagging mode into feedforward predictive control. This reduces system impact from the source, thereby improving the water resource reuse rate while ensuring the safety of the ultrapure water system. Furthermore, by assessing the impact of future production scheduling on wastewater reuse, it can optimize the production plan in reverse, realizing closed-loop collaborative optimization from wastewater reuse to production management. This avoids the isolation of water recycling and production planning information, and facilitates optimization based on future production fluctuations. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system modules of the present invention; Figure 3 This is a schematic diagram of the electronic device structure of the present invention. Detailed Implementation

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0020] Please see Figure 1 As shown, this invention provides a water purification control method in PCB manufacturing. This method is applied to a cleaning section where each batch of cleaning wastewater can be discharged independently, and this cleaning section has an independent wastewater reuse pipeline. It should be noted that this method is preferably applied to cleaning sections in PCB production lines with independent water reuse pipelines, such as the final DI water cleaning process. In this type of cleaning section, each batch of cleaning wastewater is discharged independently, and there is no mixing or interference with wastewater from other processes. The method includes: S1. Obtain the production plan schedule for the future preset time period and parse the process route information for each production batch. The process route information includes at least the process parameters of the preceding processing steps.

[0021] In step S1, by obtaining the production plan schedule for a future preset time period and analyzing the process route information of each production batch, the process is transformed from passive to proactive, achieving foresight. This connects the two traditionally independent systems of wastewater reuse and production management, allowing for understanding the driving forces of water quality changes from the specific production process source. The specific steps are as follows: Retrieve the production plan schedule for the future preset time period from the production management system. The production plan schedule includes the identification information of each batch to be produced and its planned production time. Based on the identification information of each batch, retrieve the corresponding process route information from the process database of the production management system; Analyze the process route information to identify the process sequence of the preceding processing steps; Extract the process parameters for each preceding processing step from the process route information. The process parameters include the process type, the name of the main chemical solution used, the concentration range of the chemical solution, and the standard processing time.

[0022] Specifically, communication is established with the enterprise's Manufacturing Execution System (MES) through a pre-defined application programming interface (API) or database connection middleware. The system periodically reads the production schedule table for a pre-defined future time period (e.g., the next 24 hours). This table contains structured data, including key fields such as unique identifiers for each batch to be produced, product models, and planned production times. Then, using the batch identifier as an index, a query is initiated to the MES's process database to retrieve the corresponding standard process route file stored in XML or JSON format. Finally, a parsing module (e.g., an XML parser) loads and parses the file. This document first identifies the complete sequence of preceding processing steps (e.g., drilling, copper plating, full-board electroplating, pattern transfer). Finally, for each step in the sequence, the parsing module extracts the key process parameters necessary for water quality prediction from its corresponding node. These parameters include at least the process type (e.g., chemical copper plating), the name of the main chemical solution used (e.g., copper plating solution), the standard concentration range of the chemical solution (e.g., sulfuric acid concentration: 180 to 220 g / L), and the standard processing time. This forms a structured data package containing the complete preceding process characteristics for each production batch, which can then be used by the subsequent prediction module.

[0023] S2. Based on the correlation analysis of historical production data and measured water quality data, establish a correspondence model between process parameters and water quality parameters. Using the correspondence model, predict the water quality parameters of the cleaning wastewater generated in each batch according to the process route information.

[0024] In step S2, a correlation model between process parameters and water quality parameters is established based on the correlation analysis of historical production data and measured water quality data. Using this model, the water quality parameters of the cleaning wastewater generated in each batch are predicted according to the process route information. This upgrades water quality prediction from subjective guessing based on human experience to objective calculation based on data correlation analysis. This allows for the prediction of the wastewater quality to be generated before actual production begins, based solely on known "process route information." This represents a fundamental shift from passive detection after wastewater generation to proactive prediction before production begins. The specific steps are as follows: Collect historical production data, which includes the process parameters of the preceding processing steps of each production batch that has been completed in history. Collect historical measured water quality data corresponding to historical production data. The historical measured water quality data is obtained by sampling and testing the cleaning wastewater generated in the subsequent cleaning process after each production batch has completed the previous processing steps. Feature engineering is performed on process parameters in historical production data to construct feature variables. The feature engineering process includes: converting the process sequence into a numerical vector representing the composition of the process, converting the concentration range of the drug solution into a single numerical value representing the concentration level, and extracting the statistical features of the processing time. The constructed feature variables and the corresponding historical measured water quality data are used as training samples, and one or more algorithms selected from linear regression, support vector regression and decision tree regression are used to train the corresponding relationship model. For the production batch to be predicted, the process parameters of its preceding processing steps are extracted based on its process route information to form an input feature vector representing the process characteristics of the batch. The input feature vector is fed into the correspondence model, and the prediction calculation is performed to output the predicted values ​​of the water quality parameters of the cleaning wastewater of the production batch to be predicted.

[0025] Specifically, complete data corresponding to each completed production batch in history is collected from historical production records and wastewater testing databases. This includes detailed process parameters of all preceding processing steps for each batch as input features, and actual water quality parameters obtained from immediate sampling and testing of wastewater from the cleaning process after the batch is completed as target outputs. Subsequently, in the data preprocessing stage, feature engineering is performed on the collected historical process parameters: the process sequence is converted into a binary numerical vector representing the process composition using one-hot encoding; the concentration range of the chemical solution is determined by taking its median or calculating its deviation from the standard process point to obtain a single numerical value representing the concentration level; and statistical features such as the sum, average, maximum, and variance of the processing time of multiple preceding processes are extracted to represent the overall processing intensity. For the main chemical solution names, one-hot encoding or mapping to a preset identifier is also performed. For the chemical solution concentration range, its median or quantification into a specific numerical feature based on process knowledge is used. The collected historical process parameters are then processed... The preprocessed and measured water quality data are cleaned, aligned, and normalized to ensure data quality and consistency. Then, in the model training phase, the numerical feature vectors generated after preprocessing and feature engineering are used as feature variables, and the corresponding historical measured water quality data are used as target variables. These are input into a machine learning training framework, using one or more algorithms such as linear regression, support vector regression, or decision tree regression. The model parameters are trained by minimizing the error function between the predicted and true values, thus constructing a stable correspondence model from process parameters to water quality parameters. For a new batch to be predicted, based on the process route information in its production plan, the process parameters of each preceding process are extracted and constructed as input feature vectors in the same order and normalization method as in the training phase. Finally, in the prediction phase, this feature vector is input into the trained correspondence model for forward computation. The model output is the predicted value of the water quality parameters for the new batch of cleaning wastewater, completing the accurate inference from known process conditions to unknown water quality conditions.

[0026] Secondly, the correspondence model is updated, and the specific steps are as follows: After the actual production batch is completed, the actual water quality test data of the cleaning wastewater of that batch is collected in order to monitor the performance predicted by the model. The actual water quality data is compared with the predicted water quality parameters obtained for this batch, and the prediction deviation is calculated. When the number of consecutive batches reaches the first preset threshold and the average prediction deviation of these batches exceeds the second preset fault tolerance threshold, it is determined that the original correspondence model has become inaccurate, and the model update process is triggered. The model update process includes adding newly collected actual production data and water quality data as new samples to the historical training dataset, and retraining the model using one or more algorithms selected from linear regression, support vector regression, and decision tree regression to generate an updated correspondence model. During planned system downtime, or when the total wastewater flow rate of the system is lower than the preset flow threshold, the online correspondence model will be automatically switched to the updated model.

[0027] Specifically, after each batch of actual production is completed, a water quality data acquisition process is automatically triggered. This involves sampling the cleaning wastewater generated for that batch and using an online water quality analyzer to obtain actual water quality data. This data, along with the model's predicted value for that batch, is sent to the performance monitoring unit. The monitoring unit compares the actual value with the predicted value, calculates the absolute or relative deviation as the prediction deviation for the current batch, and continuously records and maintains a fixed-length prediction deviation queue. When the number of batches in the queue that consecutively meet the conditions reaches a first preset threshold (e.g., 5 consecutive batches) and the average prediction deviation of these batches exceeds a second preset fault tolerance threshold (e.g., the average relative deviation is consistently higher than 15%), the original model is automatically determined to be no longer suitable for the current production conditions, triggering a model update flag. The update process then begins, and the newly collected data is updated. The collected actual production data and its corresponding measured water quality data are used as new samples and merged with the historical training dataset to form an expanded dataset. Then, a preset training algorithm (such as the linear regression, support vector regression, or decision tree regression algorithm mentioned above) is called to retrain this expanded dataset to generate an updated correspondence model that better reflects the current production situation. After the new model is trained, it is not put into use immediately. Instead, the system scheduler automatically and smoothly switches the old model in use online to the updated model when it detects that the system is in a planned shutdown period or when the total wastewater flow of the system is lower than the preset safety switching threshold (e.g., lower than 30% of the normal flow) during a low-load operation period, so as to ensure that the model update process does not interfere with the stable operation of the production system.

[0028] S3. Compare the predicted water quality parameters with the corresponding preset thresholds, determine the reuse level of each batch of cleaning wastewater based on the comparison results, and assign a reuse ratio to each reuse level.

[0029] In step S3, by comparing the predicted water quality parameters with the corresponding preset thresholds, the reuse level of each batch of cleaning wastewater is determined based on the comparison results, and a reuse ratio is assigned to each reuse level. This avoids the limitations of a fixed ratio and implements a differentiated and precise control strategy based on subtle differences in the predicted water quality values. This transforms traditional extensive control into refined management, allocating different reuse ratios to wastewater of different reuse levels, thus balancing water conservation and risk reduction. The specific steps are as follows: For each key water quality parameter, a safety threshold and a danger threshold are set, where the danger threshold is greater than the safety threshold. The steps to determine the reuse level include: If the predicted values ​​of all water quality parameters of a production batch are lower than their respective safety thresholds, then the reuse level of the wastewater in that batch is determined to be Level 1. If at least one of the predicted water quality parameters of a production batch is higher than or equal to its safety threshold, but all predicted values ​​are lower than their respective danger thresholds, then the reuse level of the wastewater in that batch is determined to be Level II. If at least one of the predicted water quality parameters for a production batch is higher than or equal to its danger threshold, the reuse level of that batch of wastewater is determined to be Level III. The recycling ratio is allocated as follows: 80% to 100% for Level 1 recycling, 30% to 70% for Level 2 recycling, and 0% to 20% for Level 3 recycling.

[0030] Specifically, two thresholds with a clear magnitude relationship are set for each key water quality parameter (such as conductivity, total organic carbon (TOC), and copper ion concentration): a relatively lenient safety threshold and a more stringent danger threshold. The danger threshold must be greater than the safety threshold. These threshold values ​​are pre-set based on a comprehensive analysis of the feed water quality requirements provided by the core membrane element manufacturer of the ultrapure water preparation system, historical data from long-term stable system operation, and relevant industry water quality standards. For example, for copper ion concentration, the danger threshold can be set to not exceed 80% of the maximum permissible feed water concentration specified by the ultrapure water system reverse osmosis membrane element manufacturer. Each parameter is divided into three numerical ranges: a safe zone, a transition zone, and a danger zone. Then, based on the predicted values ​​of all key water quality parameters for a production batch, the system automatically determines its reuse level according to the following defined logical rules: If the predicted value of each water quality parameter in the batch falls within its respective safest range (i.e., all are below their corresponding safety threshold), the system determines the reuse level of the batch of wastewater as Level 1, allowing for the highest possible reuse rate. If the predicted value of at least one water quality parameter in the batch exceeds the safe range (i.e., is greater than or equal to its safety threshold), but the predicted values ​​of all parameters do not reach the danger level (i.e., all are below their danger threshold), the system classifies it as Level 1. If a batch of water is classified as Level II, requiring restricted reuse, and any predicted value of any water quality parameter reaches a dangerous level (i.e., greater than or equal to its danger threshold), the system directly classifies it as Level III, which should not be reused in principle. Finally, a preset, operable specific reuse ratio range is assigned to each reuse level. Level I, with excellent water quality, is allocated a high reuse ratio of 80% to 100% to maximize water conservation; Level II, with risky but controllable water quality, is allocated a medium reuse ratio of 30% to 70% to achieve limited reuse while ensuring safety; and Level III, with water quality exceeding standards, is allocated an extremely low reuse ratio of 0% to 20%. The setting of this ratio range is an optimization range derived from statistical analysis of a large amount of historical operating data. It aims to maximize the reuse rate while ensuring system safety. For example, the reuse ratio of Class I water quality is set above 80%, which is based on historical data verification that the system can still stably produce compliant ultrapure water when reusing wastewater of this quality at a high ratio for a long period of time. At this time, a small amount is reused intermittently or all of it is directly discharged into the conventional wastewater treatment system, thereby providing absolute safety guarantee for the ultrapure water system. By mapping continuous predicted water quality parameters to discrete reuse levels through a dual threshold mechanism, and matching differentiated reuse action instructions for each level, the reuse strategy is refined, standardized and automated.

[0031] Secondly, it also includes load balancing optimization of the production schedule after determining the reuse level of each planned batch and before actual production. The specific steps are as follows: Iterate through and parse the production planning and scheduling data within the future preset time period to identify all production batches that are judged to be of the third-level reuse level; Calculate the maximum consecutive sequence length formed by these three-level reuse batches in the time series; If the maximum consecutive sequence length exceeds the first preset threshold, it is determined that there is an excessive concentration of high-pollution load batches in the current schedule, and the optimization process is initiated. The optimization process involves finding an insertion point in a continuous sequence of three-level reuse batches that meets the requirements of production cycle time and equipment availability, and swapping the planned subsequent first- or second-level reuse batches to that position. Generate a production plan optimization scheme that includes optimized processes, and send this scheme to the production management system.

[0032] Specifically, the system iterates through all production schedule data within a preset future time period (e.g., the next 24 hours), analyzes the process information of each planned batch, and calls the water quality prediction and reuse level determination module to filter out all high-pollution load batches pre-determined as Level 3 reuse. Then, in order of planned production time, it analyzes the distribution of these Level 3 batches on the time axis, calculates the maximum consecutive segment length (e.g., 5 consecutive Level 3 batches), and compares this segment length with a preset allowable threshold (e.g., the first preset threshold is 3 consecutive batches). If the actual length exceeds the threshold, it determines that the current schedule has a risk of excessive concentration of high-pollution load, and automatically initiates an optimization process. The core of this optimization process is that, within the identified consecutive Level 3 batch sequences, it automatically optimizes the process based on the preset production cycle time, equipment maintenance schedule, and compatibility rules between processes in the Manufacturing Execution System (MES) (e.g., pattern plating should not be arranged immediately after copper plating, but drilling can be arranged). The process involves evaluating and identifying feasible insertion positions. This process involves scanning each gap in a continuous three-level batch sequence in chronological order, verifying whether each gap meets the constraints of production cycle time, equipment availability, and process compatibility. The first gap that satisfies all constraints is identified as a feasible insertion position. After identifying a feasible insertion position, one or more batches with low pollution loads, already classified as level 1 or 2 reuse, are retrieved from subsequent production plans. Their planned production times are adjusted, and these batches are swapped to the identified feasible insertion positions. This disperses the originally continuous high-pollution-load batch sequence over time. Finally, a production plan optimization scheme containing specific batch adjustment instructions (such as swapping the production times of batch A and batch B) is generated. This structured scheme is then sent to the scheduling interface of the Manufacturing Execution System (MES). The MES system can then generate a new, more balanced production schedule, thereby preventing the ultrapure water preparation system from being subjected to prolonged high-concentration wastewater input from the source.

[0033] S4. In the actual production process, the cleaning wastewater is transported to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch.

[0034] In step S4, by transporting the cleaning wastewater to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch during the actual production process, the forward-looking intelligent decision-making is accurately translated into stable and reliable physical execution, ensuring the safe and stable operation of the wastewater reuse and ultrapure water preparation systems. The specific steps are as follows: After the main wastewater outlet pipe of the cleaning process, a branch pipeline is set up to divide the wastewater into two paths, one of which is connected to the raw water tank of the ultrapure water preparation system, and the other is connected to the conventional wastewater treatment system. Install an electric regulating valve on the pipeline leading to the raw water tank of the ultrapure water preparation system; Based on the current production batch's reuse level, the corresponding reuse ratio is determined. During the wastewater transport process, the total wastewater flow rate is monitored in real time, and the target flow rate to the ultrapure water preparation system is calculated based on the reuse ratio and the total wastewater flow rate. Based on the target flow rate and the real-time detected branch flow rate, a control signal is generated and sent to the electric regulating valve to dynamically adjust its opening, so that the deviation between the wastewater flow rate entering the ultrapure water preparation system and the target flow rate is maintained within the preset error allowable range. The remaining wastewater after diversion will be transported to a conventional wastewater treatment system.

[0035] Specifically, a T-junction is installed downstream of the main wastewater collection pipe of the cleaning process to form a branch pipeline, dividing the wastewater into two paths. One path connects to the inlet of the raw water tank of the ultrapure water preparation system, while the other path connects to the conventional wastewater treatment system. A corrosion-resistant electric regulating valve is installed on the branch pipe leading to the ultrapure water preparation system. This valve receives a 4-20mA standard current signal from the control system to control its opening. During operation, it first calls the preset target reuse ratio (e.g., 90% for Level 1) based on the reuse level (Level 1, Level 2, or Level 3) of the currently processed production batch. Then, the total wastewater flow rate is monitored in real time by an electromagnetic flow meter installed on the main pipe, and this total flow rate value is multiplied by the target reuse ratio to calculate the target flow rate value to be sent to the ultrapure water preparation system. Simultaneously, the actual flow rate entering the ultrapure water system is monitored in real time by a flow meter installed on the branch pipe. For branch flow, the control system compares the real-time monitored value of the branch flow with the calculated target flow value, calculates the instantaneous flow deviation, and inputs this deviation into a proportional-integral-derivative (PID) controller. The PID controller calculates the deviation value in real time and outputs a control signal (such as a 4 to 20mA current signal) to the electric regulating valve, dynamically and precisely adjusting its opening. When the actual flow is lower than the target value, the opening is increased; when the actual flow is higher than the target value, the opening is decreased. Through this closed-loop feedback control, the actual wastewater flow entering the ultrapure water preparation system is stably maintained within a preset allowable error range (such as ±5%) near the target flow value, thereby accurately achieving the predetermined reuse ratio. At the same time, the remaining wastewater after diversion automatically flows into the conventional wastewater treatment system for treatment through another passage of the three-way fitting, ultimately achieving precise diversion and proportional control of cleaning wastewater.

[0036] Secondly, based on the target flow rate and the real-time detected branch flow rate, a control signal is generated and sent to the electric regulating valve to dynamically adjust its opening. The specific steps are as follows: Based on the calculated target flow rate, the feedforward control quantity is calculated using the pre-generated flow rate-valve opening correspondence. The monitoring value of the branch flow is acquired in real time, compared with the target flow value, and the feedback correction amount is calculated by the proportional-integral controller based on the deviation between the two. The feedforward control quantity and the feedback correction quantity are superimposed to calculate the final opening control signal and send it to the electric regulating valve to dynamically adjust its opening.

[0037] Specifically, based on the calculated target flow rate, a flow rate-valve opening correspondence data table or curve, pre-generated through experimental or system characteristic modeling, is retrieved. This correspondence reflects the valve opening reference value required to achieve a specific flow rate under stable operating conditions. A feedforward control variable is then directly calculated based on this value, providing a rapid response. Simultaneously, by acquiring high-frequency signals from flowmeters installed on the branch pipes, the actual monitored flow rate of the branch is obtained in real time and continuously compared with the target flow rate value to calculate the instantaneous flow rate deviation. This deviation signal is then sent to a proportional-integral (PI) controller. The controller calculates the current deviation and the accumulation of historical deviations based on the preset proportional gain and integral time constant, and outputs a feedback correction value in real time to eliminate steady-state error. Finally, in the signal synthesis stage, the opening value calculated by the feedforward channel and the corrected opening value calculated by the feedback channel are combined and superimposed to calculate the final total opening control signal used to drive the actuator. This standard control signal (such as a 4 to 20mA current signal) is sent to the servo driver of the electric regulating valve. By dynamically and precisely adjusting the valve core position, the controller achieves fast, high-precision and stable control of the flow rate without steady-state error.

[0038] Please see Figure 2 A water purification control system for PCB manufacturing process, using the aforementioned water purification control method for PCB manufacturing process, includes: The scheduling parsing module is used to obtain the production plan schedule for a future preset time period and parse the process route information of each production batch. The process route information includes at least the process parameters of the preceding processing steps. The water quality prediction module is used to establish a correspondence model between process parameters and water quality parameters based on the correlation analysis between historical production data and measured water quality data. Using the correspondence model, the water quality parameters of the cleaning wastewater generated in each batch are predicted according to the process route information. The reuse decision module is used to compare the predicted water quality parameters with the corresponding preset thresholds, determine the reuse level of each batch of cleaning wastewater based on the comparison results, and assign a reuse ratio to each reuse level. The control and execution module is used to transport the cleaning wastewater to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch during the actual production process.

[0039] In the above, the scheduling analysis module is responsible for obtaining the production plan schedule for the future preset period and analyzing the process route information of each production batch. By actively acquiring and analyzing the future production plan and process route information, it realizes the prediction of the driving force of wastewater quality changes from the source of production, and transforms wastewater reuse control from passive response to proactive anticipation. The water quality prediction module is responsible for establishing a correspondence model between process parameters and water quality parameters based on the correlation analysis of historical production data and measured water quality data. Using the correspondence model, it predicts the water quality parameters of the cleaning wastewater generated in each batch according to the process route information, realizing the quantitative prediction of wastewater quality and providing a reliable data foundation for decision-making. The reuse decision module is responsible for comparing the predicted water quality parameters with the corresponding preset thresholds, determining the reuse level of each batch of cleaning wastewater according to the comparison results, and allocating a reuse ratio for each reuse level to realize the refinement and standardization of the reuse strategy. Under the premise of ensuring system safety, it optimizes the wastewater reuse rate. The control execution module is responsible for transporting the cleaning wastewater to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch during the actual production process, thereby improving the efficiency of wastewater reuse.

[0040] Please see Figure 3 An electronic device, characterized in that: the electronic device comprises: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the water purification control method for PCB production process according to any one of claims 1 to 8.

[0041] The processor of the aforementioned monitoring device can be a high-performance central processing unit (CPU) or graphics processing unit (GPU), and the memory can include storage devices such as random access memory (RAM), read-only memory (ROM), solid-state drive (SSD), or hard disk drive. In addition, the monitoring device may also include an arithmetic unit, input devices, output devices, and a network interface. The arithmetic unit can be a logic unit used to perform various arithmetic and logical operations to assist the processor in completing complex data processing tasks. Input devices can include keyboards, mice, touch screens, etc., used to receive user input instructions and data. Output devices can include displays, printers, etc., used to display processing results and output reports. The network interface is used to enable network communication between the monitoring device and other systems or devices for data exchange and remote monitoring.

[0042] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0043] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. A water purification control method in PCB manufacturing process, characterized in that: The method is applied to a cleaning section where each batch of cleaning wastewater can be discharged independently, and this cleaning section has an independent wastewater reuse pipeline; the method includes: Obtain the production plan schedule for a future preset period and analyze the process route information for each production batch. The process route information includes at least the process parameters of the preceding processing steps. Based on the correlation analysis of historical production data and measured water quality data, a correspondence model between process parameters and water quality parameters is established. Using the correspondence model, the water quality parameters of the cleaning wastewater generated in each batch are predicted according to the process route information. The predicted water quality parameters are compared with the corresponding preset thresholds. Based on the comparison results, the reuse level of each batch of cleaning wastewater is determined, and a reuse ratio is assigned to each reuse level. In the actual production process, the cleaning wastewater is transported to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch.

2. The water purification control method in PCB manufacturing process according to claim 1, characterized in that: The specific steps for obtaining the production plan schedule for a future preset time period and parsing the process route information for each production batch are as follows: Retrieve the production plan schedule for the future preset time period from the production management system. The production plan schedule includes the identification information of each batch to be produced and its planned production time. Based on the identification information of each batch, retrieve the corresponding process route information from the process database of the production management system; Analyze the process route information to identify the process sequence of the preceding processing steps; Extract the process parameters for each preceding processing step from the process route information. The process parameters include the process type, the name of the main chemical solution used, the concentration range of the chemical solution, and the standard processing time.

3. The water purification control method in the PCB manufacturing process according to claim 1, characterized in that: The process involves establishing a correlation model between process parameters and water quality parameters based on the correlation analysis of historical production data and measured water quality data. Using this model, and based on the process route information, the water quality parameters of the cleaning wastewater generated in each batch are predicted. The specific steps are as follows: Collect historical production data, which includes the process parameters of the preceding processing steps of each production batch that has been completed in history. Collect historical measured water quality data corresponding to historical production data. The historical measured water quality data is obtained by sampling and testing the cleaning wastewater generated in the subsequent cleaning process after each production batch has completed the previous processing steps. Feature engineering is performed on process parameters in historical production data to construct feature variables. The feature engineering process includes: converting the process sequence into a numerical vector representing the composition of the process, converting the concentration range of the drug solution into a single numerical value representing the concentration level, and extracting the statistical features of the processing time. The constructed feature variables and the corresponding historical measured water quality data are used as training samples, and one or more algorithms selected from linear regression, support vector regression and decision tree regression are used to train the corresponding relationship model. For the production batch to be predicted, the process parameters of its preceding processing steps are extracted based on its process route information to form an input feature vector representing the process characteristics of the batch. The input feature vector is fed into the correspondence model, and the prediction calculation is performed to output the predicted values ​​of the water quality parameters of the cleaning wastewater of the production batch to be predicted.

4. The water purification control method in the PCB manufacturing process according to claim 1, characterized in that: The process involves comparing the predicted water quality parameters with corresponding preset thresholds, determining the reuse level of each batch of cleaning wastewater based on the comparison results, and assigning a reuse ratio to each reuse level. The specific steps are as follows: For each key water quality parameter, a safety threshold and a danger threshold are set, where the danger threshold is greater than the safety threshold. The steps to determine the reuse level include: If the predicted values ​​of all water quality parameters of a production batch are lower than their respective safety thresholds, then the reuse level of the wastewater in that batch is determined to be Level 1. If at least one of the predicted water quality parameters of a production batch is higher than or equal to its safety threshold, but all predicted values ​​are lower than their respective danger thresholds, then the reuse level of the wastewater in that batch is determined to be Level II. If at least one of the predicted water quality parameters for a production batch is higher than or equal to its danger threshold, the reuse level of that batch of wastewater is determined to be Level III. The recycling ratio is allocated as follows: 80% to 100% for Level 1 recycling, 30% to 70% for Level 2 recycling, and 0% to 20% for Level 3 recycling.

5. The water purification control method in the PCB manufacturing process according to claim 1, characterized in that: In the actual production process, the cleaning wastewater is transported to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch. The specific steps are as follows: After the main wastewater outlet pipe of the cleaning process, a branch pipeline is set up to divide the wastewater into two paths, one of which is connected to the raw water tank of the ultrapure water preparation system, and the other is connected to the conventional wastewater treatment system. Install an electric regulating valve on the pipeline leading to the raw water tank of the ultrapure water preparation system; Based on the current production batch's reuse level, the corresponding reuse ratio is determined. During the wastewater transport process, the total wastewater flow rate is monitored in real time, and the target flow rate to the ultrapure water preparation system is calculated based on the reuse ratio and the total wastewater flow rate. Based on the target flow rate and the real-time detected branch flow rate, a control signal is generated and sent to the electric regulating valve to dynamically adjust its opening, so that the deviation between the wastewater flow rate entering the ultrapure water preparation system and the target flow rate is maintained within the preset error allowable range. The remaining wastewater after diversion will be transported to a conventional wastewater treatment system.

6. The water purification control method in PCB manufacturing process according to claim 3, characterized in that: It also includes updating the correspondence model, with the following specific steps: After the actual production batch is completed, the actual water quality test data of the cleaning wastewater of that batch is collected in order to monitor the performance predicted by the model. The actual water quality data is compared with the predicted water quality parameters obtained for this batch, and the prediction deviation is calculated. When the number of consecutive batches reaches the first preset threshold and the average prediction deviation of these batches exceeds the second preset fault tolerance threshold, it is determined that the original correspondence model has become inaccurate, and the model update process is triggered. The model update process includes adding newly collected actual production data and water quality data as new samples to the historical training dataset, and retraining the model using one or more algorithms selected from linear regression, support vector regression, and decision tree regression to generate an updated correspondence model. During planned system downtime, or when the total wastewater flow rate of the system is lower than the preset flow threshold, the online correspondence model will be automatically switched to the updated model.

7. A water purification control method in PCB manufacturing process according to claim 5, characterized in that: The control signal, generated based on the target flow rate and the real-time detected branch flow rate, is sent to the electric regulating valve to dynamically adjust its opening. The specific steps are as follows: Based on the calculated target flow rate, the feedforward control quantity is calculated using the pre-generated flow rate-valve opening correspondence. The monitoring value of the branch flow is acquired in real time, compared with the target flow value, and the feedback correction amount is calculated by the proportional-integral controller based on the deviation between the two. The feedforward control quantity and the feedback correction quantity are superimposed to calculate the final opening control signal and send it to the electric regulating valve to dynamically adjust its opening.

8. A water purification control method in PCB manufacturing process according to claim 4, characterized in that: This also includes load balancing optimization of the production schedule after determining the reuse level of each planned batch and before actual production. The specific steps are as follows: Iterate through and parse the production planning and scheduling data within the future preset time period to identify all production batches that are judged to be of the third-level reuse level; Calculate the maximum consecutive sequence length formed by these three-level reuse batches in the time series; If the maximum consecutive sequence length exceeds the first preset threshold, it is determined that there is an excessive concentration of high-pollution load batches in the current schedule, and the optimization process is initiated. The optimization process involves finding an insertion point in a continuous sequence of three-level reuse batches that meets the requirements of production cycle time and equipment availability, and swapping the planned subsequent first- or second-level reuse batches to that position. Generate a production plan optimization scheme that includes optimized processes, and send this scheme to the production management system.

9. A water purification control system for PCB manufacturing process, characterized in that: The water purification control method for PCB manufacturing process according to any one of claims 1 to 8 includes: The scheduling parsing module is used to obtain the production plan schedule for a future preset time period and parse the process route information of each production batch. The process route information includes at least the process parameters of the preceding processing steps. The water quality prediction module is used to establish a correspondence model between process parameters and water quality parameters based on the correlation analysis between historical production data and measured water quality data. Using the correspondence model, the water quality parameters of the cleaning wastewater generated in each batch are predicted according to the process route information. The reuse decision module is used to compare the predicted water quality parameters with the corresponding preset thresholds, determine the reuse level of each batch of cleaning wastewater based on the comparison results, and assign a reuse ratio to each reuse level. The control and execution module is used to transport the cleaning wastewater to the ultrapure water preparation system according to the reuse ratio corresponding to the reuse level of each batch during the actual production process.

10. An electronic device, characterized in that: The electronic device includes: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the water purification control method for PCB production process according to any one of claims 1 to 8.

Images