Cleaning control method and system based on electronic cleaning agent

Abstract

The invention discloses a cleaning control method and system based on an electronic cleaning agent, and the method comprises the steps: obtaining the distribution characteristics of pollutants and a material sensitivity judgment result through obtaining the micron-sized particle distribution data, chemical residue concentration data and surface roughness parameters of the pollutants on the surface of an electronic workpiece; according to the pollutant distribution characteristics and the material sensitivity judgment result, matching a candidate formula from a preset cleaning agent formula database, calculating pollutant removal rate predicted values of different concentration and time combinations through a linear regression algorithm, and determining optimized cleaning agent formula parameters and cleaning strength control parameters; and in the cleaning process, the residual quantity of pollutants and the surface damage degree are collected in real time, if the residual quantity of the pollutants exceeds a preset threshold value T1 or the surface damage degree exceeds a preset threshold value T2, cleaning agent injection pressure and cleaning time parameters are dynamically adjusted, and updated cleaning execution parameters are generated. Accurate and efficient cleaning of the surface of the electronic workpiece is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic cleaning, and in particular discloses a cleaning control method and system based on an electronic cleaning agent. Background Art

[0002] Electronic cleaning technology plays a crucial role in modern manufacturing. Especially in the production of precision electronic components, its cleanliness directly affects product performance and reliability.

[0003] With the development of miniaturization and high integration of electronic products, cleaning technology needs to meet higher precision and environmental protection requirements, becoming a key link to ensure production yield and sustainable development. However, existing cleaning methods have obvious limitations when dealing with complex workpieces and diverse production requirements.

[0004] Many traditional cleaning processes rely on a single cleaning agent formula and are difficult to adapt to different material and pollutant types, resulting in poor cleaning effects or workpiece damage.

[0005] At the same time, some systems lack intelligent control, and the adjustment of cleaning parameters depends on manual experience, with low efficiency and poor consistency.

[0006] In addition, the problems of low wastewater recovery rate and high energy consumption are widespread, and it is difficult to meet the strict environmental protection standards and cost control requirements.

[0007] These limitations reflect the core challenges in the field of cleaning technology, mainly concentrated in three aspects: precise control of the cleaning process, efficient utilization of resources, and intelligent operation.

[0008] Firstly, the precise control of the cleaning process requires the efficient removal of pollutants in micron-level cleaning while avoiding damage to the surface of precision workpieces, which requires a balance between cleaning intensity and workpiece protection.

[0009] Secondly, the efficient utilization of resources involves the recycling of cleaning agents and wastewater. How to maximize the recovery rate and reduce energy consumption while maintaining the cleaning effect is a technical difficulty.

[0010] Finally, intelligent operation needs to achieve unmanned production through real-time monitoring and adaptive adjustment, but existing systems are still insufficient in sensor integration and control logic optimization.

[0011] Therefore, how to design a cleaning control method and system that can precisely control cleaning parameters, achieve efficient resource recycling, and support intelligent operation has become a key issue in promoting the development of electronic cleaning technology. Summary of the Invention

[0012] The present invention provides a cleaning control method and system based on an electronic cleaning agent, aiming to solve at least one of the defects existing in the above-mentioned existing electronic cleaning technology.

[0013] One aspect of the present invention relates to a cleaning control method based on an electronic cleaning agent, comprising the following steps: Obtain the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters of the contaminants on the surface of the electronic workpiece, and use a convolutional neural network algorithm to extract and classify the data to obtain the contaminant distribution characteristics and the material sensitivity judgment result; According to the contaminant distribution characteristics and the material sensitivity judgment result, match candidate formulas from a preset cleaning agent formula database, calculate the predicted values of the contaminant removal rate for different concentration and time combinations through a linear regression algorithm, and determine the optimized cleaning agent formula parameters and cleaning intensity control parameters; During the cleaning process, real-time collect the contaminant residue amount and the surface damage degree. If the contaminant residue amount exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, then dynamically adjust the cleaning agent injection pressure and the cleaning time parameters to generate updated cleaning execution parameters; Conduct a chemical composition analysis of the wastewater generated by cleaning, use a support vector machine algorithm to identify the recyclable cleaning agent components, and determine the wastewater recovery parameters and the concentration value of the recyclable cleaning agent; According to the wastewater recovery parameters and the concentration value of the recyclable cleaning agent, establish an energy consumption optimization model, calculate the optimal power distribution scheme for the cleaning system and the wastewater treatment equipment, and generate the system operation parameters with the lowest total energy consumption.

[0014] Further, the step of obtaining the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters of the contaminants on the surface of the electronic workpiece, and using a convolutional neural network algorithm to extract and classify the data to obtain the contaminant distribution characteristics and the material sensitivity judgment result includes: Obtain the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters of the surface of the electronic workpiece through a scanning electron microscope, and use a data fusion method to integrate the data to obtain a fusion data set including particle distribution, chemical residue, and roughness characteristics; Use a convolutional neural network algorithm to extract features from the fusion data set, generate a feature vector describing the contaminant distribution law and surface texture characteristics, and obtain a set of feature vectors available for classification; If the spatial distribution value of the feature vector is greater than a preset threshold, then classify the feature vector through a clustering algorithm to obtain the high-density and low-density distribution categories of the contaminants; According to the high-density and low-density distribution categories of the contaminants and the chemical residue concentration value, use a decision tree algorithm to evaluate the data and judge the sensitivity level of the workpiece material.

[0015] Further, according to the pollutant distribution characteristics and the material sensitivity judgment results, the steps of matching candidate formulas from the preset cleaning agent formula database, calculating the predicted values of pollutant removal rates for different concentration and time combinations through the linear regression algorithm, and determining the optimized cleaning agent formula parameters and cleaning intensity control parameters include: Receive the pollutant distribution characteristic data and the material sensitivity data, classify the pollutant distribution characteristic data by using the clustering algorithm, and obtain the pollutant distribution categories and the material sensitivity levels; According to the pollutant distribution categories and the material sensitivity levels, query and match the candidate formulas from the preset cleaning agent formula database. If the number of candidate formulas is greater than the preset threshold, sort the formulas according to the material sensitivity levels to determine the priority formula list.

[0016] For each formula in the priority formula list, calculate the predicted values of pollutant removal rates for different concentration and cleaning time combinations through the linear regression algorithm to obtain the pollutant removal rate prediction data set; Extract the concentration and time combination corresponding to the highest removal rate from the pollutant removal rate prediction data set, and use logical judgment to process and optimize the cleaning intensity control parameters to determine the final cleaning agent formula parameters and cleaning intensity control parameters.

[0017] Further, during the cleaning process, the pollutant residue amount and the surface damage degree are collected in real time. If the pollutant residue amount exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, the steps of dynamically adjusting the cleaning agent injection pressure and the cleaning time parameters to generate the updated cleaning execution parameters include: Collect the pollutant residue amount and the surface damage degree data in real time during the cleaning process, and perform denoising and formatting processing on the original data through the data processing module to obtain the cleaning status data; If the pollutant residue amount in the cleaning status data exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, perform iterative calculation on the cleaning agent injection pressure and the cleaning time parameters through the parameter optimization algorithm to obtain the adjusted parameter values; According to the adjusted parameter values, generate the updated cleaning execution parameters through the data processing module to determine the operation instruction of the cleaning equipment; Transmit the cleaning execution parameters to the cleaning equipment through the cleaning process control module, obtain the operation status data of the equipment, and judge whether the cleaning process meets the preset conditions.

[0018] Further, the steps of analyzing the chemical components of the wastewater generated by cleaning, identifying the recyclable cleaning agent components by using the support vector machine algorithm, and determining the wastewater recovery parameters and the recyclable cleaning agent concentration values include: The chemical components of the wastewater generated by cleaning are detected by a liquid chromatograph, and wastewater component data is obtained from the detection results. The principal component analysis method is used to perform dimensionality reduction processing on the wastewater component data to obtain a wastewater component feature set after feature extraction; According to the wastewater component feature set, the support vector machine classification method is used to identify the components of the recyclable cleaning agent, determine whether there are components of the recyclable cleaning agent in the wastewater, and obtain the classification result of the recyclable cleaning agent; If the classification result indicates the existence of components of the recyclable cleaning agent, the wastewater recovery parameters matching the classification result are obtained through a preset wastewater recovery parameter database, and the linear regression method is used to optimize the wastewater recovery parameters to determine a wastewater recovery parameter set; According to the wastewater recovery parameter set, the mass balance calculation method is used to calculate the concentration value of the recyclable cleaning agent, and by comparing with a preset concentration threshold, the concentration value of the recyclable cleaning agent is obtained.

[0019] Furthermore, according to the wastewater recovery parameters and the concentration value of the recyclable cleaning agent, an energy consumption optimization model is established, and the steps of calculating the optimal power distribution scheme for the cleaning system and the wastewater treatment equipment and generating the system operation parameters with the lowest total energy consumption include: Obtain the wastewater flow rate, pollutant concentration, and cleaning agent concentration data, and calculate the initial distribution ratio of the cleaning equipment power and the wastewater treatment equipment power through a preset linear regression algorithm to obtain a preliminary power distribution scheme; According to the preliminary power distribution scheme, if the circulating water quality is lower than the preset threshold, adjust the cleaning agent replenishment rate, obtain a new cleaning agent concentration, and determine the updated circulating water quality parameters; Use the updated circulating water quality parameters and cleaning agent concentration to calculate the operation time of the cleaning equipment and the wastewater treatment equipment through the operation time calculation formula to obtain an adjusted power distribution scheme, where the operation time calculation formula is T = Q / (P1 + P2), T represents the operation time, Q represents the wastewater flow rate, P1 represents the cleaning equipment power, and P2 represents the wastewater treatment equipment power; Obtain the adjusted power distribution scheme, calculate the total energy consumption in combination with the wastewater flow rate and pollutant concentration. If the total energy consumption is higher than the preset threshold, determine the final operation parameters through iterative optimization of the power distribution ratio.

[0020] Another aspect of the present invention relates to a cleaning control system based on an electronic cleaning agent for implementing the above-mentioned cleaning control method based on an electronic cleaning agent. The cleaning control system based on an electronic cleaning agent includes: An acquisition module for acquiring the micron-level particle distribution data, chemical residue concentration data, and surface roughness parameters of the pollutants on the surface of the electronic workpiece, and using a convolutional neural network algorithm to perform feature extraction and classification on the data to obtain the pollutant distribution characteristics and the material sensitivity judgment result; The first determination module is used to match candidate formulas from a preset cleaning agent formula database according to the pollutant distribution characteristics and the material sensitivity judgment result, calculate the predicted values of the pollutant removal rate for different concentration and time combinations through a linear regression algorithm, and determine the optimized cleaning agent formula parameters and cleaning intensity control parameters; The first generation module is used to collect the pollutant residue amount and the surface damage degree in real time during the cleaning process. If the pollutant residue amount exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, the cleaning agent injection pressure and the cleaning time parameters are dynamically adjusted to generate updated cleaning execution parameters; The second determination module is used to analyze the chemical components of the wastewater generated by cleaning, identify the recyclable cleaning agent components by using a support vector machine algorithm, and determine the wastewater recycling parameters and the concentration value of the recyclable cleaning agent; The second generation module is used to establish an energy consumption optimization model according to the wastewater recycling parameters and the concentration value of the recyclable cleaning agent, calculate the optimal power distribution scheme for the cleaning system and the wastewater treatment equipment, and generate the system operation parameters with the lowest total energy consumption.

[0021] Furthermore, the acquisition module includes: The first acquisition unit is used to obtain the micron-scale particle distribution data, chemical residue concentration data and surface roughness parameters on the surface of the electronic workpiece through a scanning electron microscope, and integrate the data by using a data fusion method to obtain a fusion data set including particle distribution, chemical residue and roughness characteristics; The second acquisition unit is used to extract features from the fusion data set by using a convolutional neural network algorithm, generate feature vectors describing the pollutant distribution law and surface texture characteristics, and obtain a set of feature vectors available for classification; The third acquisition unit is used to classify the feature vectors by using a clustering algorithm if the spatial distribution value of the feature vectors is greater than a preset threshold, and obtain the high-density and low-density distribution categories of pollutants; The first judgment unit is used to evaluate the data by using a decision tree algorithm according to the high-density and low-density distribution categories of pollutants and the chemical residue concentration value, and judge the sensitivity level of the workpiece material.

[0022] Furthermore, the first determination module includes: The fourth acquisition unit is used to receive the pollutant distribution characteristic data and the material sensitivity data, classify the pollutant distribution characteristic data by using a clustering algorithm, and obtain the pollutant distribution category and the material sensitivity level; The first determination unit is used to query and match candidate formulas from a preset cleaning agent formula database according to the pollutant distribution category and the material sensitivity level. If the number of candidate formulas is greater than a preset threshold, the formulas are sorted according to the material sensitivity level to determine the priority formula list.

[0023] A fifth acquisition unit, configured to calculate predicted values of pollutant removal rates for different combinations of concentrations and cleaning times for each formulation in the priority formulation list through a linear regression algorithm, and obtain a pollutant removal rate prediction data set; A second determination unit, configured to extract the concentration and time combination corresponding to the highest removal rate from the pollutant removal rate prediction data set, optimize the cleaning intensity control parameter by using a logical judgment process, and determine the final cleaning agent formulation parameter and the cleaning intensity control parameter.

[0024] Further, the first generation module includes: A sixth acquisition unit, configured to collect data on pollutant residue and surface damage degree in real time during the cleaning process, and perform denoising and formatting processing on the original data through a data processing module to obtain cleaning state data; A seventh acquisition unit, configured to, if the pollutant residue in the cleaning state data exceeds a preset threshold T1 or the surface damage degree exceeds a preset threshold T2, perform iterative calculation on the cleaning agent injection pressure and the cleaning time parameter through a parameter optimization algorithm to obtain adjusted parameter values; A third determination unit, configured to generate updated cleaning execution parameters through a data processing module according to the adjusted parameter values, and determine the operation instruction of the cleaning device; A second judgment unit, configured to transmit the cleaning execution parameters to the cleaning device through a cleaning process control module, obtain the operation state data of the device, and judge whether the cleaning process meets the preset conditions.

[0025] The beneficial effects obtained by the present invention are as follows: The present invention provides a cleaning control method and system based on an electronic cleaning agent. First, micro data of pollutants on the surface of a workpiece is collected, and a convolutional neural network is used for feature extraction and classification to obtain pollutant distribution characteristics and material sensitivity. According to these results, a cleaning agent formulation is matched from a preset database, and the cleaning parameters are optimized through a linear regression algorithm; during the cleaning process, the pollutant residue and the surface damage degree are monitored in real time, and the cleaning parameters are dynamically adjusted according to preset thresholds; the components of the wastewater are analyzed, and a support vector machine algorithm is used to identify recyclable components and determine the wastewater recycling parameters; finally, an energy consumption optimization model is established, and an optimal power distribution scheme is calculated to achieve the operation parameters with the lowest total energy consumption. The cleaning control method and system based on an electronic cleaning agent provided by the present invention realize precise and efficient cleaning of the surface of an electronic workpiece through intelligent algorithms and real-time monitoring, maximize resource recycling and utilization, reduce energy consumption, and have significant economic and environmental benefits. Description of the Drawings

[0026] Figure 1 It is a schematic flowchart of an embodiment of a cleaning control method based on an electronic cleaning agent of the present invention. Detailed implementation manners

[0027] To better understand the above technical solution, the following will describe the above technical solution in detail in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0028] As Figure 1 shown, a cleaning control method based on an electronic cleaning agent is proposed in the first embodiment of the present invention, including the following steps: Step S100: Obtain the micron-level particle distribution data, chemical residue concentration data, and surface roughness parameters of the contaminants on the surface of the electronic workpiece, and use a convolutional neural network algorithm to perform feature extraction and classification on the data to obtain the contaminant distribution characteristics and the material sensitivity judgment result.

[0029] The contaminant distribution characteristics include the contaminant type and the spatial distribution characteristics. Among them, the contaminant type includes chemical composition and physical form. The chemical composition includes organic matter, inorganic matter, and particulate matter. The physical form includes polar contaminants (ionic substances), non-polar contaminants (non-ionic substances and mixed-state contaminants). The spatial distribution characteristics include position concentration and distribution density and uniformity. The position concentration includes key functional areas and edges and holes. The distribution density and uniformity include local high density and large-area low density.

[0030] Material sensitivity refers to the degree and rate of degradation of the physical, chemical, or electrical properties of a material under the action of a specific environment or contaminants. Material sensitivity is a key indicator for measuring the resistance of a material to contaminants (such as chemical residues, particulate matter, moisture, etc.) or environmental conditions (temperature, humidity, mechanical stress), and directly affects the reliability and life of the electronic workpiece.

[0031] Material sensitivity includes chemical sensitivity, physical sensitivity, electrical sensitivity, and environmental interaction sensitivity.

[0032] Step S200: According to the contaminant distribution characteristics and the material sensitivity judgment result, match candidate formulas from the preset cleaning agent formula database, calculate the predicted values of the contaminant removal rate for different concentration and time combinations through a linear regression algorithm, and determine the optimized cleaning agent formula parameters and cleaning intensity control parameters.

[0033] A candidate formula refers to multiple potential material or process combination solutions formed through systematic screening and evaluation based on specific performance goals, environmental conditions, and process requirements during product R & D or manufacturing. Its core purpose is to determine the best solution that meets the application requirements through experimental verification and parameter optimization.

[0034] The predicted pollutant removal rate refers to the removal efficiency (percentage or ratio) of a specific cleaning process or technology for target pollutants pre-estimated through theoretical models, experimental data, or algorithms (such as machine learning). Its core objective is to provide a quantitative basis for process optimization (such as adjusting cleaning parameters) and cost control (such as chemical usage).

[0035] The cleaning agent formulation parameters are the core indicators used to quantify the formulation composition and performance, covering the solvent system, additive ratio, functional characteristics, and process adaptability. Their definition and optimization directly affect cleaning efficiency, safety, and cost control.

[0036] The cleaning intensity control parameters refer to a set of quantitative indicators that achieve efficient pollutant removal and avoid substrate damage by adjusting key variables such as mechanical force, chemical action, thermal action, and time in the cleaning process. Its core role is to balance cleaning efficiency and safety, ensuring process repeatability and economy.

[0037] Step S300: During the cleaning process, the residual amount of pollutants and the degree of surface damage are collected in real-time. If the residual amount of pollutants exceeds the preset threshold T1 or the degree of surface damage exceeds the preset threshold T2, the injection pressure and cleaning time parameters of the cleaning agent are dynamically adjusted to generate updated cleaning execution parameters.

[0038] The injection pressure refers to the dynamic impact force exerted when the cleaning agent passes through the nozzle, which directly affects the intensity of the mechanical action of the cleaning agent on pollutant stripping. The commonly used unit is MPa or Bar. Its core role is to balance cleaning efficiency and substrate protection requirements.

[0039] The cleaning time refers to the continuous action duration of the cleaning agent in contact with pollutants, which needs to be comprehensively set according to the pollutant dissolution kinetics and substrate tolerance. The unit is seconds, minutes, or hours.

[0040] The cleaning execution parameters refer to a set of core control variables set to achieve an efficient, safe, and repeatable cleaning process, covering key elements such as mechanical action, chemical action, thermal action, and time dimension, and need to be comprehensively defined based on pollutant characteristics, substrate tolerance, and environmental protection requirements.

[0041] Step S400: Analyze the chemical composition of the wastewater generated during cleaning, use the support vector machine algorithm to identify the recyclable cleaning agent components, and determine the wastewater recovery parameters and the concentration value of the recyclable cleaning agent.

[0042] The wastewater recovery parameters refer to a set of core control indicators set to achieve the goal of wastewater reuse, covering dimensions such as recovery efficiency, water quality control, process stability, and economy, and need to be comprehensively defined based on wastewater characteristics, treatment processes, and reuse scenarios.

[0043] The concentration value of the recyclable cleaning agent refers to the proportion of the effective active ingredients or the mass of the solute per unit volume in the cleaning agent reused through the recycling system in the cleaning process. It needs to be dynamically adjusted in combination with the cleaning target, substrate compatibility, and environmental protection requirements, and is usually expressed as a percentage (%) or mass concentration (mg / L).

[0044] Step S500: Based on the wastewater recovery parameters and the concentration value of the recyclable cleaning agent, establish an energy consumption optimization model, calculate the optimal power distribution plan for the cleaning system and the wastewater treatment equipment, and generate the system operation parameters with the lowest total energy consumption.

[0045] The energy consumption optimization model refers to a quantitative analysis tool that systematically reduces energy consumption (electricity, heat energy, mechanical energy, etc.) in industrial cleaning or production processes through mathematical modeling and algorithm design while meeting process constraints (such as cleaning efficiency, equipment safety). Its core goal is to maximize energy utilization efficiency and minimize carbon emissions, and it needs to be comprehensively constructed in combination with equipment characteristics, process parameters, and dynamic monitoring data.

[0046] The system operation parameters with the lowest total energy consumption refer to a set of key variables that minimize the total energy consumption (electricity, heat energy, mechanical energy, etc.) of the entire process through multi-dimensional collaborative optimization (such as improving equipment efficiency, adjusting process conditions, dynamic load matching, etc.) while meeting the system function requirements and safety constraints. Its core goal is to reduce the energy cost per unit output while ensuring the stability and compliance of system operation.

[0047] Furthermore, for the cleaning control method based on the electronic cleaning agent proposed in this embodiment, step S100 includes: Step S110: Obtain the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters on the surface of the electronic workpiece through a scanning electron microscope, and integrate the data using a data fusion method to obtain a fusion data set containing particle distribution, chemical residue, and roughness characteristics.

[0048] A scanning electron microscope is a high-resolution imaging tool that can obtain microscopic information on the surface of an electronic workpiece. For example, by scanning the surface of the workpiece with its high-energy electron beam, a two-dimensional image reflecting the particle distribution can be generated, with the particle size accurate to the micron level. At the same time, combined with an energy dispersive spectrometer, the elemental composition and concentration of surface chemical residues can be detected.

[0049] In addition, by analyzing the electron scattering signal, surface roughness parameters such as the average roughness Ra and the peak-to-valley height Rz can also be calculated.

[0050] In a possible implementation, assuming the workpiece is a semiconductor wafer, the scanning microscope can detect particles with a diameter of 1 to 5 micrometers, identify residual elements such as silicon, oxygen, carbon, etc., and measure the Ra value to be 0.2 micrometers. These data provide a diverse characterization basis for subsequent analysis.

[0051] The data fusion method is used to integrate multi-source information to form a unified data set. Specifically, principal component analysis or weighted average method can be adopted to map particle distribution, chemical residue concentration, and roughness data into a unified feature space. Exemplarily, for the wafer surface, data such as 10 particles per square millimeter in particle density, 5% oxygen element concentration, and an Ra value of 0.2 micrometers can be normalized to generate a fused feature vector. The advantage of this method is that it synthesizes multi-dimensional information, reduces redundancy, and improves the processing efficiency of subsequent algorithms.

[0052] Step S120: Use the convolutional neural network algorithm to extract features from the fused data set, generate feature vectors describing the distribution law of pollutants and surface texture characteristics, and obtain a set of feature vectors available for classification.

[0053] The convolutional neural network performs excellently in feature extraction and is particularly suitable for processing surface data with strong spatial correlation. For example, a network containing multiple layers of convolution and pooling can be designed. After inputting the fused data set, vectors reflecting the distribution of pollutants and texture features are extracted.

[0054] In one embodiment, the network can identify the density of particle aggregation regions or the periodic changes in surface texture and output 128-dimensional feature vectors. Such vectors not only capture local features but also retain the global distribution law, which is helpful for subsequent classification tasks.

[0055] Step S130: If the spatial distribution value of the feature vector is greater than a preset threshold, then classify the feature vector through a clustering algorithm to obtain high-density and low-density distribution categories of pollutants.

[0056] If the spatial distribution value of the feature vector exceeds the preset threshold, such as the Euclidean distance being greater than 2, then clustering classification is required.

[0057] It can be understood that the K-means clustering algorithm can classify the feature vectors into high-density and low-density categories.

[0058] Preferably, for the wafer surface, the clustering result may show that the particle density in some areas is 20 per square millimeter, which is classified as the high-density category, while in other areas it is only 5, which is classified as the low-density category. This classification clarifies the regional differences in pollutant distribution and provides a basis for precise cleaning.

[0059] Step S140: According to the high-density and low-density distribution categories of pollutants and the chemical residue concentration value, use the decision tree algorithm to evaluate the data and determine the sensitivity level of the workpiece material.

[0060] The decision tree algorithm is used to evaluate the sensitivity of the workpiece material and make a judgment by combining the pollutant distribution category and the chemical residue concentration value.

[0061] In a possible implementation, the decision tree can set the rule: if the proportion of the high-density area exceeds 30% and the oxygen concentration is greater than 3%, the material sensitivity is high.

[0062] Exemplarily, on the surface of a certain wafer, the high-density area accounts for 40% and the oxygen concentration is 4%, which is rated as high sensitivity, indicating that the cleaning process needs to be optimized. This method has clear logic, is easy to explain, and can directly guide production decisions.

[0063] It should be noted that the advantage of the above method lies in the full-chain analysis from microscopic data to macroscopic evaluation, which significantly improves the accuracy of workpiece quality control. For example, clustering classification can guide local cleaning strategies, and decision tree evaluation optimizes the selection of process parameters, overall improving production efficiency and product reliability.

[0064] Preferably, for the cleaning control method based on electronic cleaning agents proposed in this embodiment, step S200 includes: Step S210: Receive the pollutant distribution characteristic data and the material sensitivity data, and use the clustering algorithm to classify the pollutant distribution characteristic data to obtain the pollutant distribution category and the material sensitivity level.

[0065] Exemplarily, in the scenario of cleaning pollutants on the surface of electronic workpieces, the acquisition of pollutant distribution characteristic data is usually based on high-precision detection equipment. For example, a laser microscope is used to scan the surface of the workpiece to generate a spatial distribution image of pollutant particles, and the data includes particle size, density, and distribution uniformity.

[0066] Suppose it is detected that the particle density on the surface of a certain workpiece is 1000 per square millimeter and the size range is between 1 and 5 microns. These data constitute the core content of the pollutant distribution characteristics.

[0067] It should be noted that the generation of material sensitivity data depends on the analysis of the chemical and physical properties of the workpiece material. For example, the surface chemical composition is detected by a spectrometer to judge the tolerance of the material to acidic or alkaline cleaning agents, and the sensitivity level, such as low, medium, and high, is obtained.

[0068] In a possible implementation, a clustering algorithm is used for classifying pollutant distribution characteristics. For example, based on the K-means clustering algorithm, the particle density and distribution uniformity are used as inputs to classify high-density areas and low-density areas. Suppose the particle density in the high-density area on the surface of a workpiece is 1500 per square millimeter, and that in the low-density area is 500. Such classification results provide a basis for subsequent formulation selection.

[0069] Specifically, the material sensitivity level can be mapped through a decision table. For example, materials with low sensitivity are resistant to strong acid cleaning, while materials with high sensitivity require mild cleaning agents.

[0070] Step S220: According to the pollutant distribution category and the material sensitivity level, query for matching candidate formulations from a preset cleaning agent formulation database. If the number of candidate formulations is greater than a preset threshold, sort the formulations according to the material sensitivity level to determine a priority formulation list.

[0071] The classified pollutant distribution category and the material sensitivity level jointly determine the pertinence of the cleaning strategy. For example, the query process in the cleaning agent formulation database depends on preset rules.

[0072] Suppose the database contains 100 formulations, and the query conditions are high-density pollutants and medium-sensitivity materials, and 10 candidate formulations are screened out. If the preset threshold is 5 formulations, further sorting is required.

[0073] Preferably, sort the formulations according to the material sensitivity, and give priority to selecting formulations with less damage to the material. For example, formulation A is a neutral cleaning agent, and formulation B is a weakly alkaline cleaning agent. Based on medium-sensitivity materials, A is prior to B to generate a priority formulation list. This sorting method ensures the balance between cleaning effect and material protection.

[0074] Step S230: For each formulation in the priority formulation list, calculate the predicted pollutant removal rate values for different concentration and cleaning time combinations through a linear regression algorithm to obtain a pollutant removal rate prediction data set.

[0075] In an embodiment, a linear regression algorithm is used to predict the pollutant removal rate. For example, for formulation A, the input variables are the cleaning agent concentration and the cleaning time, and the output is the predicted removal rate.

[0076] Suppose the test concentration range is 5% to 15%, and the time range is 10 to 30 minutes. The prediction result shows that the removal rate for the combination of 10% concentration and 20 minutes is 95%.

[0077] It can be understood that the pollutant removal rate prediction data set provides data support for optimizing the cleaning parameters.

[0078] The parameter combination corresponding to the highest removal rate is extracted, such as a concentration of 10% and a time of 20 minutes, as the initial optimization plan.

[0079] Step S240: Extract the concentration and time combination corresponding to the highest removal rate from the pollutant removal rate prediction dataset, and use logical judgment to process and optimize the cleaning intensity control parameters to determine the final cleaning agent formulation parameters and cleaning intensity control parameters.

[0080] Specifically, logical judgment is used to optimize the cleaning intensity control parameters.

[0081] For example, based on the removal rate data, it is judged whether the cleaning intensity needs to be adjusted. If the removal rate has reached 95%, but the material sensitivity is high, the concentration is appropriately reduced to 8% to reduce potential damage.

[0082] The finally determined cleaning agent formulation parameters are a neutral cleaning agent with a concentration of 8%, a cleaning time of 20 minutes, and the cleaning intensity control parameter is low-speed stirring.

[0083] This parameter combination achieves a balance between efficient pollutant removal and material protection.

[0084] It should be noted that the flexibility of logical judgment allows dynamic adjustment according to the actual workpiece state to ensure the adaptability and reliability of the cleaning process.

[0085] Furthermore, for the cleaning control method based on electronic cleaning agents proposed in this embodiment, step S300 includes: Step S310: Real-time collect data on the pollutant residue amount and surface damage degree during the cleaning process, and perform denoising and formatting processing on the original data through the data processing module to obtain the cleaning state data.

[0086] Real-time collection of data on the pollutant residue amount and surface damage degree during the cleaning process is a key link to ensure the cleaning effect and material safety.

[0087] Exemplarily, the cleaning equipment is equipped with high-precision sensors, such as laser scanners and chemical residue analyzers, which are used to detect the pollutant residue amount and surface micro-damage respectively. These sensors collect data once per second to form a time series dataset. For example, for cleaning the oil stain on the surface of metal parts, the laser scanner can judge the oil stain coverage area through the change in light reflection intensity, while the chemical residue analyzer detects the grease molecule concentration through spectral analysis.

[0088] The collected original data usually contains noise, such as outliers caused by environmental light interference or sensor jitter.

[0089] In one possible implementation, the data processing module performs denoising and formatting on the raw data. Denoising can use the mean filtering method to replace the abnormal and sudden data points in the time series with the average value of the previous and next data. For example, assuming that the pollutant residue data collected at a certain time is 100, 102, 150, and 103 mg / cm², 150 is obviously too high, and it can be replaced by (102+103) / 2=102.5 mg / cm².

[0090] The formatting process unifies the data from different sensors into a standard JSON format for easy subsequent analysis.

[0091] Step S320: If the amount of pollutant residue in the cleaning status data exceeds the preset threshold value T1 or the degree of surface damage exceeds the preset threshold value T2, the cleaning agent injection pressure and cleaning time parameters are iteratively calculated through a parameter optimization algorithm to obtain adjusted parameter values.

[0092] The processed cleaning status data includes two indicators: the amount of pollutant residue and the degree of surface damage, which are compared with the preset thresholds T1 and T2 respectively. For example, T1 is set to 10mg / cm² and T2 is set to the surface scratch depth of 0.01mm.

[0093] It should be noted that if the amount of pollutant residue exceeds T1 or the degree of surface damage exceeds T2, the parameter optimization algorithm will intervene to adjust the cleaning agent injection pressure and cleaning time.

[0094] Preferably, the algorithm is based on the principle of gradient descent, and searches for the balance between pollutant removal efficiency and material damage through multiple iterations. For example, when the residual amount is detected to be 15mg / cm², exceeding T1, and the surface damage is 0.005mm, not reaching T2, the algorithm may suggest reducing the injection pressure from 5bar to 4bar, and extending the cleaning time from 10 seconds to 12 seconds to reduce the impact on the material.

[0095] The adjusted parameter values are used to generate new cleaning execution parameters through the data processing module, such as a spray pressure of 4 bar and a cleaning time of 12 seconds.

[0096] Step S330: Generate updated cleaning execution parameters through the data processing module according to the adjusted parameter values, and determine the operation instructions of the cleaning equipment.

[0097] Specifically, the cleaning process control module transmits the cleaning execution parameters to the cleaning equipment and obtains the equipment operation status data in real time. For example, the operation status data returned by the equipment may show that the actual injection pressure is 4.1 bar, with a deviation of 0.1 bar.

[0098] The control module will determine whether the deviation is within the allowable range, such as ±0.2 bar. If it meets the requirement, the operation will continue; if not, a warning will be issued and the cleaning will be paused.

[0099] Step S340: Transmit the cleaning execution parameters to the cleaning equipment through the cleaning process control module, obtain the operation status data of the equipment, and determine whether the cleaning process meets the preset conditions.

[0100] In one embodiment, the operation status data further includes the cleaning liquid flow rate and the equipment temperature to ensure the stability of the cleaning process. For example, when the flow rate is maintained at 2 L / min and the temperature is below 50 °C, it is determined that the cleaning process meets the preset conditions.

[0101] It can be understood that the implementation of the above method can dynamically adapt to the complex situations in the cleaning process. For example, for metal parts of different batches, the amount of pollutant residue may vary due to different types of oil stains. The collaborative work of sensors and optimization algorithms can ensure that the expected effect can be achieved for each cleaning.

[0102] In addition, the real-time monitoring of the surface damage degree avoids the material deterioration caused by over-cleaning and extends the service life of the parts.

[0103] Preferably, for the cleaning control method based on electronic cleaning agents proposed in this embodiment, step S400 includes: Step S410: Detect the chemical components of the wastewater generated by cleaning through a liquid chromatograph, obtain the wastewater component data from the detection results, and perform dimensionality reduction processing on the wastewater component data using the principal component analysis method to obtain the wastewater component feature set after feature extraction.

[0104] Detecting the chemical components of the wastewater by a liquid chromatograph is an important part of the post-treatment of the cleaning process, aiming to analyze the chemical substances remaining in the wastewater. For example, in an industrial cleaning scenario, the wastewater may contain cleaning agents, oils or metal ions. These components can be separated and detected by a liquid chromatograph to generate wastewater component data including concentration and type. Exemplarily, a certain cleaning equipment uses an alkaline cleaning agent, and the liquid chromatograph detects 0.5 mg / L of surfactant and 0.2 mg / L of organic solvent in the wastewater. These data provide a basis for subsequent analysis.

[0105] The purpose of performing dimensionality reduction processing on the wastewater component data using the principal component analysis method is to extract key features and reduce the data complexity.

[0106] It can be understood that the principal component analysis transforms the high-dimensional data into a low-dimensional feature space through linear transformation, retaining the main information.

[0107] In a possible implementation, assume that the wastewater composition data contains the concentrations of 10 chemical substances. Principal component analysis may extract 3 main eigenvectors, representing 80% of the information variance. This method can effectively reduce the computational amount and facilitate subsequent classification.

[0108] Step S420: According to the wastewater composition feature set, use the support vector machine classification method to identify the components of the recyclable cleaning agent, determine whether there are components of the recyclable cleaning agent in the wastewater, and obtain the classification result of the recyclable cleaning agent.

[0109] The support vector machine classification method is used to identify the components of the recyclable cleaning agent in the wastewater. The core lies in constructing a hyperplane to distinguish recyclable and non-recyclable substances.

[0110] Specifically, the support vector machine classifies the components of the cleaning agent according to the chemical characteristics in the feature set, such as molecular weight or polarity. For example, if the characteristics of a certain surfactant detected in the wastewater match those of a known recyclable cleaning agent, the classification result indicates that it is recyclable. This classification method has clear logic and can quickly determine the direction of wastewater treatment.

[0111] Step S430: If the classification result indicates that there are components of the recyclable cleaning agent, obtain the wastewater recovery parameters matching the classification result through a preset wastewater recovery parameter database, and use the linear regression method to optimize the wastewater recovery parameters to determine the wastewater recovery parameter set.

[0112] The wastewater recovery parameter database provides matching parameters for the classification result, and the linear regression method further optimizes these parameters.

[0113] Preferably, the database stores the recovery process parameters of different cleaning agents, such as filtration rate or distillation temperature.

[0114] Linear regression adjusts the parameters according to historical data to make the recovery efficiency higher.

[0115] In one embodiment, for a certain recyclable cleaning agent, the database recommends a filtration rate of 1 L / min, and after optimization by linear regression, it is adjusted to 1.2 L / min to ensure a more stable recovery process.

[0116] Step S440: According to the wastewater recovery parameter set, use the mass balance calculation method to calculate the concentration value of the recyclable cleaning agent, and obtain the concentration value of the recyclable cleaning agent by comparing it with a preset concentration threshold.

[0117] The mass balance calculation method is used to determine the concentration value of the recyclable cleaning agent, and it is judged whether it can be recycled by comparing with a preset threshold value. For example, the calculated concentration of the recyclable cleaning agent in the wastewater is 0.3 mg / L, while the threshold value is 0.1 mg / L, indicating that it can be recycled. This method is based on the principle of conservation of matter, and the calculation result is reliable, providing a basis for the reuse of wastewater.

[0118] It should be noted that the concentration value calculation takes into account the total amount of wastewater and the distribution of the cleaning agent to ensure the accuracy of the result.

[0119] In a possible implementation manner, the above technical topics form a complete chain: the liquid chromatograph provides the original data, the principal component analysis streamlines the data, the support vector machine accurately classifies, the database and the linear regression optimize the parameters, and the mass balance calculation finally confirms the recyclable concentration. This progressive logic ensures the efficiency and science of wastewater treatment and avoids waste of resources.

[0120] Furthermore, the cleaning control method based on the electronic cleaning agent proposed in this embodiment, step S500 includes: Step S510, obtain the data of the wastewater flow rate, pollutant concentration, and cleaning agent concentration, and calculate the initial allocation ratio of the power of the cleaning equipment and the wastewater treatment equipment through a preset linear regression algorithm to obtain a preliminary power allocation plan.

[0121] In the industrial cleaning scenario, wastewater treatment and equipment power allocation are key links to ensure recycling.

[0122] Obtaining the data of the wastewater flow rate, pollutant concentration, and cleaning agent concentration is a basic step. For example, a certain cleaning equipment treats 1000 L of wastewater per day, the pollutant concentration is 2 mg / L, and the cleaning agent concentration is 0.4 mg / L. These data are collected in real time through a flow meter and a chemical analyzer, providing a basis for subsequent calculations.

[0123] It can be understood that the data collection needs to consider the volatility of the wastewater flow rate and the uniformity of the cleaning agent distribution to ensure the reliability of the input parameters.

[0124] Calculate the power allocation ratio of the cleaning equipment and the wastewater treatment equipment through a preset linear regression algorithm to form a preliminary plan.

[0125] Specifically, the linear regression analyzes the correlation between the wastewater flow rate and the pollutant concentration on the power demand based on historical data.

[0126] In an embodiment, assume that the initial power of the cleaning equipment is 10 kW and the wastewater treatment equipment is 5 kW. The algorithm calculates that the preliminary allocation ratio is 2:1. This ratio is based on the balance of wastewater characteristics and equipment performance and aims to optimize the energy consumption allocation.

[0127] Step S520: According to the preliminary power distribution scheme, if the circulating water quality is lower than the preset threshold, adjust the cleaning agent replenishment rate, obtain the new cleaning agent concentration, and determine the updated circulating water quality parameters.

[0128] If the circulating water quality is lower than the preset threshold, for example, the turbidity exceeds 50 NTU, it is necessary to adjust the cleaning agent replenishment rate to improve the water quality.

[0129] Exemplarily, the initial cleaning agent concentration is 0.4 mg / L, and after replenishment, it is increased to 0.6 mg / L. Through chemical analysis, it is confirmed that the circulating water quality parameter is updated to a turbidity of 40 NTU.

[0130] It should be noted that the adjustment of the replenishment rate needs to be combined with the wastewater flow rate to avoid cost increase caused by excessive addition.

[0131] Step S530: Using the updated circulating water quality parameters and cleaning agent concentration, calculate the running time of the cleaning equipment and the wastewater treatment equipment through the running time calculation formula to obtain the adjusted power distribution scheme, where the running time calculation formula is T = Q / (P1 + P2), T represents the running time, Q represents the wastewater flow rate, P1 represents the power of the cleaning equipment, and P2 represents the power of the wastewater treatment equipment.

[0132] The new water quality parameters provide a basis for the subsequent running time calculation.

[0133] Using the updated circulating water quality parameters and cleaning agent concentration, determine the equipment running time through the running time calculation formula.

[0134] Preferably, the formula takes into account the relationship between the wastewater flow rate and the total power. In a possible implementation, the wastewater flow rate is 1000 L, the power of the cleaning equipment is 10 kW, and the power of the wastewater treatment equipment is 5 kW. The calculated running time is 66.7 minutes. This method provides a basis for power distribution adjustment by quantifying the running time.

[0135] Step S540: Obtain the adjusted power distribution scheme, calculate the total energy consumption in combination with the wastewater flow rate and the pollutant concentration. If the total energy consumption is higher than the preset threshold, determine the final operating parameters by iteratively optimizing the power distribution ratio.

[0136] After obtaining the adjusted power distribution scheme, calculate the total energy consumption in combination with the wastewater flow rate and the pollutant concentration. For example, after adjustment, the power of the cleaning equipment is 12 kW, and the power of the wastewater treatment equipment is 4 kW, and the total energy consumption is 16 kWh. If the total energy consumption is higher than the preset threshold of 15 kWh, the power ratio is readjusted through iterative optimization. In one embodiment, during the iteration process, the power of the cleaning equipment is reduced to 11 kW, and the power of the wastewater treatment equipment is increased to 5 kW, and the final energy consumption is reduced to 14.5 kWh. Iterative optimization ensures the balance between energy consumption and water quality treatment requirements by simulating different ratios multiple times.

[0137] Throughout the process, each link is closely connected, forming a complete chain from data collection to power distribution, then to water quality adjustment and energy consumption optimization.

[0138] Exemplarily, iterative optimization not only focuses on energy consumption but also takes into account the stability of equipment operation to avoid mechanical losses caused by frequent adjustments.

[0139] This progressive logic provides efficient support for wastewater treatment through multi-dimensional data analysis and dynamic adjustment.

[0140] Another aspect of the present invention relates to a cleaning control system based on an electronic cleaning agent for implementing the above-mentioned cleaning control method based on an electronic cleaning agent. The cleaning control system based on an electronic cleaning agent includes an acquisition module, a first determination module, a first generation module, a second determination module, and a second generation module. Among them, the acquisition module is used to acquire the micron-level particle distribution data, chemical residue concentration data, and surface roughness parameters of contaminants on the surface of the electronic workpiece, and uses a convolutional neural network algorithm to extract and classify the data to obtain the contaminant distribution characteristics and the material sensitivity judgment result; the first determination module is used to match candidate formulas from a preset cleaning agent formula database according to the contaminant distribution characteristics and the material sensitivity judgment result, calculate the predicted values of the contaminant removal rate for different concentration and time combinations through a linear regression algorithm, and determine the optimized cleaning agent formula parameters and cleaning intensity control parameters; the first generation module is used to collect the contaminant residue amount and the surface damage degree in real time during the cleaning process. If the contaminant residue amount exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, the cleaning agent injection pressure and cleaning time parameters are dynamically adjusted to generate updated cleaning execution parameters; the second determination module is used to analyze the chemical composition of the wastewater generated by the cleaning, use a support vector machine algorithm to identify the recyclable cleaning agent components, and determine the wastewater recovery parameters and the concentration value of the recyclable cleaning agent; the second generation module is used to establish an energy consumption optimization model according to the wastewater recovery parameters and the concentration value of the recyclable cleaning agent, calculate the optimal power distribution plan for the cleaning system and the wastewater treatment equipment, and generate the system operation parameters with the lowest total energy consumption.

[0141] Furthermore, for the cleaning control system based on electronic cleaning agents provided in this embodiment, the acquisition module includes a first acquisition unit, a second acquisition unit, a third acquisition unit, and a first judgment unit. Among them, the first acquisition unit is used to obtain the micron-level particle distribution data, chemical residue concentration data, and surface roughness parameters on the surface of the electronic workpiece through a scanning electron microscope, and integrate the data using a data fusion method to obtain a fusion data set containing particle distribution, chemical residue, and roughness characteristics; the second acquisition unit is used to extract features from the fusion data set using a convolutional neural network algorithm to generate feature vectors describing the distribution law of pollutants and surface texture characteristics, and obtain a set of feature vectors available for classification; the third acquisition unit is used to, if the spatial distribution value of the feature vector is greater than a preset threshold, classify the feature vectors through a clustering algorithm to obtain the high-density and low-density distribution categories of pollutants; the first judgment unit is used to evaluate the data using a decision tree algorithm based on the high-density and low-density distribution categories of pollutants and the chemical residue concentration value, and judge the sensitivity level of the workpiece material.

[0142] Furthermore, for the cleaning control system based on electronic cleaning agents provided in this embodiment, the first determination module includes a fourth acquisition unit, a first determination unit, a fifth acquisition unit, and a second determination unit. Among them, the fourth acquisition unit is used to receive the pollutant distribution characteristic data and the material sensitivity data, classify the pollutant distribution characteristic data using a clustering algorithm, and obtain the pollutant distribution category and the material sensitivity level; the first determination unit is used to query the matching candidate formulas from the preset cleaning agent formula database according to the pollutant distribution category and the material sensitivity level. If the number of candidate formulas is greater than the preset threshold, sort the formulas according to the material sensitivity level to determine the priority formula list. The fifth acquisition unit is used to calculate the predicted pollutant removal rate values for different concentration and cleaning time combinations through a linear regression algorithm for each formula in the priority formula list, and obtain a predicted pollutant removal rate data set; the second determination unit is used to extract the concentration and time combination corresponding to the highest removal rate from the predicted pollutant removal rate data set, and optimize the cleaning intensity control parameters through logical judgment processing to determine the final cleaning agent formula parameters and the cleaning intensity control parameters.

[0143] Furthermore, for the cleaning control system based on electronic cleaning agents provided in this embodiment, the first generation module includes a sixth acquisition unit, a seventh acquisition unit, a third determination unit, and a second judgment unit. Among them, the sixth acquisition unit is used to collect data on the residual amount of pollutants and the degree of surface damage in real time during the cleaning process, and through the data processing module, perform denoising and formatting processing on the original data to obtain cleaning status data; the seventh acquisition unit is used to, if the residual amount of pollutants in the cleaning status data exceeds the preset threshold T1 or the degree of surface damage exceeds the preset threshold T2, perform iterative calculations on the cleaning agent injection pressure and cleaning time parameters through a parameter optimization algorithm to obtain the adjusted parameter values; the third determination unit is used to generate updated cleaning execution parameters through the data processing module according to the adjusted parameter values and determine the operation instructions of the cleaning equipment; the second judgment unit is used to transmit the cleaning execution parameters to the cleaning equipment through the cleaning process control module, obtain the operation status data of the equipment, and judge whether the cleaning process meets the preset conditions.

[0144] This embodiment provides a cleaning control method and system based on electronic cleaning agents. Compared with the prior art, first, microscopic data of pollutants on the workpiece surface is collected, and a convolutional neural network is used for feature extraction and classification to obtain pollutant distribution characteristics and material sensitivity. According to these results, a cleaning agent formula is matched from a preset database, and the cleaning parameters are optimized through a linear regression algorithm; during the cleaning process, the residual amount of pollutants and the degree of surface damage are monitored in real time, and the cleaning parameters are dynamically adjusted according to preset thresholds; the composition of the wastewater is analyzed, and a support vector machine algorithm is used to identify recyclable components and determine the wastewater recovery parameters; finally, an energy consumption optimization model is established to calculate the optimal power distribution scheme to achieve the operating parameters with the lowest total energy consumption. The cleaning control method and system based on electronic cleaning agents provided in this embodiment achieve precise and efficient cleaning of the electronic workpiece surface through intelligent algorithms and real-time monitoring, while maximizing resource recycling and reducing energy consumption, with significant economic and environmental benefits.

[0145] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention. Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. A cleaning control method based on an electronic cleaning agent, characterized in that, It includes the following steps: Obtain the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters of the contaminants on the surface of the electronic workpiece. Use the convolutional neural network algorithm to extract features and classify the data to obtain the contaminant distribution characteristics and the material sensitivity judgment result; According to the contaminant distribution characteristics and the material sensitivity judgment result, match candidate formulas from the preset cleaning agent formula database. Calculate the predicted values of the contaminant removal rate for different concentration and time combinations through the linear regression algorithm, and determine the optimized cleaning agent formula parameters and cleaning intensity control parameters; During the cleaning process, real-time collect the contaminant residue amount and the surface damage degree. If the contaminant residue amount exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, dynamically adjust the cleaning agent injection pressure and cleaning time parameters to generate updated cleaning execution parameters; Conduct chemical composition analysis on the wastewater generated by cleaning. Use the support vector machine algorithm to identify the recyclable cleaning agent components, and determine the wastewater recovery parameters and the concentration value of the recyclable cleaning agent; According to the wastewater recovery parameters and the concentration value of the recyclable cleaning agent, establish an energy consumption optimization model, calculate the optimal power distribution scheme for the cleaning system and the wastewater treatment equipment, and generate the system operation parameters with the lowest total energy consumption.

2. The cleaning control method based on an electronic cleaning agent according to claim 1, wherein, The steps of obtaining the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters of the contaminants on the surface of the electronic workpiece, and using the convolutional neural network algorithm to extract features and classify the data to obtain the contaminant distribution characteristics and the material sensitivity judgment result include: Obtain the micron-scale particle distribution data, chemical residue concentration data, and surface roughness parameters of the surface of the electronic workpiece through a scanning electron microscope. Use the data fusion method to integrate the data to obtain a fusion data set containing the particle distribution, chemical residue, and roughness characteristics; Use the convolutional neural network algorithm to extract features from the fusion data set, generate feature vectors describing the contaminant distribution law and surface texture characteristics, and obtain a set of feature vectors available for classification; If the spatial distribution value of the feature vector is greater than the preset threshold, classify the feature vector through the clustering algorithm to obtain the high-density and low-density distribution categories of the contaminants; According to the high-density and low-density distribution categories of the contaminants and the chemical residue concentration value, use the decision tree algorithm to evaluate the data and judge the sensitivity level of the workpiece material.

3. The cleaning control method based on an electronic cleaning agent according to claim 1, wherein, The steps of, according to the contaminant distribution characteristics and the material sensitivity judgment result, matching candidate formulas from the preset cleaning agent formula database, calculating the predicted values of the contaminant removal rate for different concentration and time combinations through the linear regression algorithm, and determining the optimized cleaning agent formula parameters and cleaning intensity control parameters include: Receive the contaminant distribution characteristic data and the material sensitivity data. Use the clustering algorithm to classify the contaminant distribution characteristic data to obtain the contaminant distribution category and the material sensitivity level; According to the pollutant distribution category and material sensitivity level, query the matching candidate formulations from a preset cleaning agent formulation database. If the number of candidate formulations is greater than a preset threshold, sort the formulations according to the material sensitivity level to determine a priority formulation list. For each formulation in the priority formulation list, calculate the predicted pollutant removal rate values for different combinations of concentration and cleaning time through a linear regression algorithm to obtain a pollutant removal rate prediction data set. Extract the concentration and time combination corresponding to the highest removal rate from the pollutant removal rate prediction data set, and use logical judgment processing to optimize the cleaning intensity control parameters to determine the final cleaning agent formulation parameters and cleaning intensity control parameters.

4. The cleaning control method based on an electronic cleaning agent according to claim 1, wherein, During the cleaning process, the steps of dynamically adjusting the cleaning agent injection pressure and cleaning time parameters to generate updated cleaning execution parameters when the pollutant residue amount exceeds a preset threshold T1 or the surface damage degree exceeds a preset threshold T2 include: Collect the pollutant residue amount and surface damage degree data in real time during the cleaning process, and perform denoising and formatting processing on the original data through a data processing module to obtain cleaning status data. If the pollutant residue amount in the cleaning status data exceeds the preset threshold T1 or the surface damage degree exceeds the preset threshold T2, perform iterative calculation on the cleaning agent injection pressure and cleaning time parameters through a parameter optimization algorithm to obtain the adjusted parameter values. According to the adjusted parameter values, generate updated cleaning execution parameters through a data processing module to determine the operation instructions for the cleaning equipment. Transmit the cleaning execution parameters to the cleaning equipment through a cleaning process control module, obtain the operation status data of the equipment, and determine whether the cleaning process meets the preset conditions.

5. The cleaning control method based on an electronic cleaning agent according to claim 1, wherein The steps of analyzing the chemical composition of the wastewater generated by cleaning, using a support vector machine algorithm to identify recyclable cleaning agent components, and determining the wastewater recovery parameters and recyclable cleaning agent concentration values include: Detect the chemical composition of the wastewater generated by cleaning through a liquid chromatograph, obtain the wastewater component data from the detection results, and perform dimensionality reduction processing on the wastewater component data using the principal component analysis method to obtain the wastewater component feature set after feature extraction. According to the wastewater component feature set, use the support vector machine classification method to identify recyclable cleaning agent components, and determine whether there are recyclable cleaning agent components in the wastewater to obtain the classification result of recyclable cleaning agents. If the classification result indicates the existence of recyclable cleaning agent components, obtain the wastewater recovery parameters matching the classification result from a preset wastewater recovery parameter database, and optimize the wastewater recovery parameters using a linear regression method to determine the wastewater recovery parameter set. According to the wastewater recovery parameter set, use the mass balance calculation method to calculate the concentration value of the recyclable cleaning agent, and obtain the recyclable cleaning agent concentration value by comparing it with a preset concentration threshold.

6. The cleaning control method based on an electronic cleaning agent according to claim 1, wherein, According to the waste water recovery parameters and the concentration value of the recyclable cleaning agent, steps for establishing an energy consumption optimization model, calculating the optimal power distribution scheme of the cleaning system and the waste water treatment equipment, and generating the system operation parameters with the lowest total energy consumption include: Obtaining waste water flow rate, pollutant concentration, and cleaning agent concentration data, calculating the initial distribution ratio of the power of the cleaning equipment and the waste water treatment equipment through a preset linear regression algorithm, and obtaining the preliminary power distribution scheme; According to the preliminary power distribution scheme, if the circulating water quality is lower than a preset threshold, adjusting the cleaning agent replenishment rate, obtaining a new cleaning agent concentration, and determining the updated circulating water quality parameters; Using the updated circulating water quality parameters and the cleaning agent concentration, calculating the operation time of the cleaning equipment and the waste water treatment equipment through the operation time calculation formula to obtain the adjusted power distribution scheme, where the operation time calculation formula is T = Q / (P1 + P2), T represents the operation time, Q represents the waste water flow rate, P1 represents the power of the cleaning equipment, and P2 represents the power of the waste water treatment equipment; Obtaining the adjusted power distribution scheme, calculating the total energy consumption in combination with the waste water flow rate and the pollutant concentration, and if the total energy consumption is higher than a preset threshold, determining the final operation parameters through iterative optimization of the power distribution ratio.

7. A cleaning control system based on an electronic cleaning agent, which is used to implement the cleaning control method based on the electronic cleaning agent described in any one of claims 1 to 6, characterized in that, The cleaning control system based on the electronic cleaning agent includes: An acquisition module for obtaining the micron-level particle distribution data, chemical residue concentration data, and surface roughness parameters of the pollutants on the surface of the electronic workpiece, performing feature extraction and classification on the data by using a convolutional neural network algorithm, and obtaining the pollutant distribution characteristics and the material sensitivity judgment result; A first determination module for matching a candidate formula from a preset cleaning agent formula database according to the pollutant distribution characteristics and the material sensitivity judgment result, calculating the predicted pollutant removal rate values for different concentration and time combinations through a linear regression algorithm, and determining the optimized cleaning agent formula parameters and the cleaning intensity control parameters; A first generation module for real-time collecting the pollutant residue amount and the surface damage degree during the cleaning process. If the pollutant residue amount exceeds a preset threshold T1 or the surface damage degree exceeds a preset threshold T2, dynamically adjusting the cleaning agent injection pressure and the cleaning time parameters to generate the updated cleaning execution parameters; A second determination module for analyzing the chemical components of the waste water generated by the cleaning, identifying the recyclable cleaning agent components by using a support vector machine algorithm, and determining the waste water recovery parameters and the concentration value of the recyclable cleaning agent; A second generation module for establishing an energy consumption optimization model according to the waste water recovery parameters and the concentration value of the recyclable cleaning agent, calculating the optimal power distribution scheme of the cleaning system and the waste water treatment equipment, and generating the system operation parameters with the lowest total energy consumption.

8. The cleaning control system based on an electronic cleaning agent according to claim 7, wherein The acquisition module includes: A first acquisition unit for obtaining the micron-level particle distribution data, chemical residue concentration data, and surface roughness parameters of the surface of the electronic workpiece through a scanning electron microscope, integrating the data by using a data fusion method, and obtaining a fusion data set including the particle distribution, chemical residue, and roughness characteristics; A second acquisition unit, configured to extract features from the fused data set by using a convolutional neural network algorithm, generate a feature vector describing the distribution law of pollutants and the surface texture characteristics, and obtain a set of feature vectors available for classification; A third acquisition unit, configured to, if the spatial distribution value of the feature vector is greater than a preset threshold, classify the feature vector by using a clustering algorithm to obtain high-density and low-density distribution categories of pollutants; A first judgment unit, configured to evaluate the data by using a decision tree algorithm according to the high-density and low-density distribution categories of the pollutants and the chemical residue concentration value, and judge the sensitivity level of the workpiece material.

9. The cleaning control system based on an electronic cleaning agent according to claim 7, wherein The first determination module includes: A fourth acquisition unit, configured to receive pollutant distribution feature data and material sensitivity data, classify the pollutant distribution feature data by using a clustering algorithm to obtain a pollutant distribution category and a material sensitivity level; A first determination unit, configured to query a matching candidate formula from a preset cleaning agent formula database according to the pollutant distribution category and the material sensitivity level, and if the number of the candidate formulas is greater than a preset threshold, sort the formulas according to the material sensitivity level to determine a priority formula list. A fifth acquisition unit, configured to calculate predicted pollutant removal rate values for different combinations of concentrations and cleaning times by using a linear regression algorithm for each formula in the priority formula list to obtain a predicted pollutant removal rate data set; A second determination unit, configured to extract the concentration and time combination corresponding to the highest removal rate from the predicted pollutant removal rate data set, and optimize the cleaning intensity control parameters by using a logical judgment process to determine the final cleaning agent formula parameters and the cleaning intensity control parameters.

10. The cleaning control system based on an electronic cleaning agent according to claim 7, characterized in that, The first generation module includes: A sixth acquisition unit, configured to collect data on the residual amount of pollutants and the degree of surface damage in real time during the cleaning process, and perform denoising and formatting processing on the original data through a data processing module to obtain cleaning state data; A seventh acquisition unit, configured to, if the residual amount of pollutants in the cleaning state data exceeds a preset threshold T1 or the degree of surface damage exceeds a preset threshold T2, perform iterative calculation on the cleaning agent injection pressure and the cleaning time parameters by using a parameter optimization algorithm to obtain adjusted parameter values; A third determination unit, configured to generate updated cleaning execution parameters through a data processing module according to the adjusted parameter values, and determine the operation instruction of the cleaning device; A second judgment unit, configured to transmit the cleaning execution parameters to the cleaning device through a cleaning process control module, obtain the operation state data of the device, and judge whether the cleaning process meets the preset conditions.

Images