Weak acid surface layer pH self-adaptive allocation and online closed-loop calibration method

Abstract

The invention discloses a weak acid surface layer pH self-adaptive allocation and online closed-loop calibration method, and relates to the technical field of material surface treatment and regulation, and the method comprises the following steps: S1, obtaining static process parameters and dynamic behavior response data of a weak acid surface layer material, and carrying out synchronous processing to form a basic data set; s2, calculating a weak acid steady-state response coefficient based on the basic data set; s3, generating a blending instruction according to the deviation between the weak acid steady-state response coefficient and a preset target range; and S4, executing process parameter adjustment based on the allocation instruction, collecting feedback parameters and completing online closed-loop calibration. A weak acid steady-state response coefficient is formed through multi-level fusion processing of multi-source data, dynamic behavior evaluation and intelligent decision making of functions of a weak acid surface layer are achieved, a self-learning online closed-loop calibration system is formed, and therefore the product performance consistency and the production self-adaptive capacity are improved.

Description

Technical Field

[0001] This invention relates to the field of material surface treatment and control technology, specifically a method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer. Background Technology

[0002] In the field of hygiene products, to ensure that products are compatible with the slightly acidic environment of human skin, functionality is typically achieved by adding fixed acidic substances to the surface material or using inherently slightly acidic fibers. These existing technologies have the following limitations: First, their slightly acidic function relies entirely on the raw material formulation and process pre-design before production. Once produced, the product's acid-base properties are fixed, lacking the ability to dynamically respond to actual usage environments. Second, for key indicators such as pH value during production, offline, sampling-based post-production inspection is commonly used, which is typical post-production inspection and cannot capture and compensate for instantaneous fluctuations during production. Furthermore, existing technologies struggle to eliminate the impact of batch-to-batch variations in raw materials and fluctuations in environmental temperature and humidity on the consistency of the final product's slightly acidic performance, leading to challenges in batch-to-batch quality stability.

[0003] Therefore, there is a clear need for improvement in existing technologies in terms of achieving dynamic adaptability of weak acid functions, real-time process quality control, and proactive maintenance of batch consistency. Summary of the Invention

[0004] Based on the shortcomings of the prior art described above, the purpose of this invention is to provide a method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer to solve the aforementioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer, comprising: S1: Obtain the static process parameters and dynamic behavior response data of the weak acid surface layer material and perform synchronous processing to form a basic dataset; S2: Calculate the steady-state response coefficient of weak acid based on the basic dataset; S3: Generate a mixing instruction based on the deviation between the steady-state response coefficient of the weak acid and the preset target range; S4: Based on the adjustment command, process parameters are adjusted, feedback parameters are collected, and online closed-loop calibration is completed.

[0006] The present invention is further configured such that the basic dataset specifically includes: a set of static process parameters and dynamic response waveform data.

[0007] The present invention is further configured such that the specific process of forming the basic dataset in S1 is as follows: By deploying process sensor groups and material information management systems on the production line, a set of static process parameters is collected in real time. The set of static process parameters includes at least: surface material type, substrate weight, moisture content, coating amount, drying temperature, and wind speed. By deploying dynamic behavior testing units at key nodes of the production line, standardized microdroplet interference is applied to the weakly acidic surface layer, and the pH value changes over time are collected simultaneously to generate dynamic response waveform data. The collected static process parameter set and dynamic response waveform data are synchronized and aligned using a unified timestamp. After synchronization and alignment, the data is filtered and outlier detected to remove invalid data points caused by transient interference or equipment noise, thus forming a basic dataset.

[0008] The present invention is further configured such that S2 specifically includes a root gene calculation step, a response factor calculation step, a reserve factor calculation step, and a weak acid steady-state response coefficient calculation step.

[0009] The present invention is further configured such that the root gene calculation step is as follows: Based on the surface material types in the basic dataset, the inherent bonding coefficient corresponding to the current surface material type is obtained by querying a preset material property knowledge base; The inherent binding coefficient and the substrate basis weight and moisture content in the basic dataset are used as inputs and passed to a preset association mapping model for processing. The output of the association mapping model is obtained to obtain the root gene.

[0010] The present invention further specifies that the response factor calculation step is as follows: By performing signal analysis on the dynamic response waveform data, response features are extracted, including: maximum pH shift, time required to recover to a preset threshold, and area under the waveform curve; The response features are passed as input to a pre-trained dynamic behavior quantization model, and the output of the dynamic behavior quantization model is obtained to obtain the response factor.

[0011] The present invention further specifies that the reserve factor calculation step is as follows: The moisture content in the static process parameter set and the corresponding real-time pH value in the dynamic response waveform data are used as input parameters to input the capacity estimation model based on buffer theory. The buffering capacity characterized by real-time pH value is corrected based on the water content using a capacity estimation model, and a reserve factor characterizing the effective buffering capacity is output.

[0012] The present invention further specifies that the calculation step of the steady-state response coefficient of the weak acid is as follows: Based on the preset product function type, the corresponding function-oriented weight configuration strategy is called from the preset target product function category correspondence table; Based on the function-oriented weight configuration strategy, a multi-dimensional weighted fusion operation is performed on the root gene, the response factor, and the reserve factor to generate a weak acid steady-state response coefficient.

[0013] The present invention is further configured such that S3 specifically includes: The steady-state response coefficient of the weak acid is compared with a preset target range to determine whether a deviation has occurred. If a deviation occurs, the dominant factor causing the deviation is determined by numerical comparison and threshold judgment of the root gene, the response factor, and the reserve factor. Based on the category of the dominant factor, the corresponding combination of compensation actions is matched by querying the pre-set root cause-policy mapping relationship library; Based on the aforementioned combination of compensation actions, a scheduling instruction containing specific process parameter adjustment instructions is generated through the instruction synthesis engine.

[0014] The present invention is further configured such that S4 specifically includes: The allocation command is sent to the actuator of the production line to adjust the corresponding production process parameters; After the process parameters are adjusted and the preset process delay time has elapsed, the data acquisition and processing flow is triggered again to obtain new dynamic response waveform data and calculate the new weak acid steady-state response coefficient as a feedback parameter. The new steady-state response coefficient of the weak acid is compared with the preset target range threshold, and logical judgment is made in the preset calibration effect evaluation rule base according to the comparison result to generate the evaluation result. Based on the evaluation results, a calibration is performed. If the calibration is successful, the system status is updated. If the calibration fails to meet the target, a new round of allocation instruction generation process is initiated, and the feedback parameters and evaluation results are entered into the calibration history database.

[0015] This invention provides a method for pH adaptive mixing and online closed-loop calibration of a weakly acid surface layer. The method comprises: S1: acquiring and synchronously processing the static process parameters and dynamic behavior response data of the weakly acid surface layer material to form a basic dataset; S2: calculating the steady-state response coefficient of the weakly acid based on the basic dataset; S3: generating a mixing instruction based on the deviation between the steady-state response coefficient and a preset target range; and S4: adjusting the process parameters based on the mixing instruction, collecting feedback parameters, and completing online closed-loop calibration. The beneficial effects include: By collecting and analyzing "dynamic response waveform data" online, the limitations of traditional methods that rely solely on static indicators such as offline detection of initial pH value after production are overcome. This enables dynamic and real-time quantitative evaluation of the anti-interference and self-recovery capabilities of weakly acidic surface layers in actual use, elevating the quality control dimension from a single chemical indicator to the functional behavior level.

[0016] By integrating static process parameters and dynamic response data, and generating a comprehensive "weak acid steady-state response coefficient" through multi-level processing, an intelligent decision-making mechanism based on the understanding of the material's intrinsic behavior was established. This method overcomes the limitations of traditional single-factor regulation, effectively identifying and distinguishing performance deviations caused by batch differences in raw materials or environmental fluctuations, making control commands more precise and improving product consistency from the source.

[0017] By constructing a closed-loop calibration system encompassing "online sensing, real-time control, effect evaluation, and model optimization," the quality control model is transformed from a passive "open-loop production, offline sampling inspection" to a proactive "online real-time control and continuous optimization." This system not only compensates for production fluctuations in real time but also learns optimization strategies from historical data, significantly improving the adaptability of the production system and the long-term stability and reliability of product performance.

[0018] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 The flowchart illustrates an exemplary embodiment of the present invention, showing a method for pH adaptive adjustment and online closed-loop calibration of a weakly acidic surface layer. Detailed Implementation

[0020] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0021] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0022] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0023] Example: A method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer, such as... Figure 1 As shown, it includes: S1: Obtain the static process parameters and dynamic behavior response data of the weak acid surface layer material and perform synchronous processing to form a basic dataset; S2: Calculate the steady-state response coefficient of weak acid based on the basic dataset; S3: Generate a mixing instruction based on the deviation between the steady-state response coefficient of the weak acid and the preset target range; S4: Based on the adjustment command, process parameters are adjusted, feedback parameters are collected, and online closed-loop calibration is completed.

[0024] The present invention is further configured such that the basic dataset specifically includes: a static process parameter set and dynamic response waveform data. Specifically, the basic dataset is a unified data input source and factual basis, used to provide comprehensive and synchronous raw data support for subsequent calculation of the weak acid steady-state response coefficient and generation of mixing instructions. The basic dataset includes a static process parameter set and dynamic response waveform data. The static process parameter set is a collection of relatively stable or slowly changing parameters used to describe the intrinsic properties of the surface material and the given production process state. Specifically, it includes material information and physical parameters (such as: surface material type, substrate basis weight, moisture content, coating amount, drying temperature, and wind speed). The material information is read from the manufacturing execution system, and the physical parameters are acquired in real time through existing sensors on the production line. The dynamic response waveform data specifically refers to the continuous curve (sequence) of pH value change over time recorded by a high-speed sensor after applying standardized microdroplet interference to the weak acid surface layer. This is obtained by a specially added dynamic behavior testing unit that sprays standard test liquid onto the surface layer and simultaneously records the complete response and recovery process curve using a high-frequency pH electrode.

[0025] The present invention is further configured such that the specific process of forming the basic dataset in S1 is as follows: By deploying process sensor groups and material information management systems on the production line, a set of static process parameters is collected in real time. The set of static process parameters includes at least: surface material type, substrate weight, moisture content, coating amount, drying temperature, and wind speed. By deploying dynamic behavior testing units at key nodes of the production line, standardized microdroplet interference is applied to the weakly acidic surface layer, and the pH value changes over time are collected simultaneously to generate dynamic response waveform data. The collected static process parameter set and dynamic response waveform data are synchronized and aligned using a unified timestamp. The synchronized data is filtered and outlier detected to remove invalid data points caused by transient interference or equipment noise, forming a basic dataset. Specifically, static process parameters are collected in real time through process sensor groups and material information management systems deployed on the production line. The static process parameter set includes at least: surface material type, substrate basis weight, moisture content, coating amount, drying temperature, and air velocity. Among them, the surface material type is read from the manufacturing execution system, the substrate basis weight is obtained through an online quality scanner, the moisture content is obtained through a near-infrared moisture sensor, the coating amount is read from the flow meter of the high-precision metering pump, and the drying temperature and air velocity are read from the temperature controller and anemometer of the drying oven, respectively. Simultaneously, standardized microdroplet interference is applied to the weak acid surface layer using dynamic behavior testing units deployed at key nodes of the production line. These key nodes refer to workstations after the weak acid coating or impregnation process and before lamination with the absorbent layer. This allows for in-situ testing of the surface layer's dynamic performance after its functional formation and before it enters the final product structure. The standardized microdroplet interference consists of a 10-microliter borate buffer solution with a pH of 8.0. Simultaneously, an industrial flat-tipped pH electrode with a response time of less than 100 milliseconds is used to collect the pH value of the weak acid surface layer over time at a sampling frequency of 500 times per second, generating dynamic response waveform data. Subsequently, the collected static process parameter set and dynamic response waveform data are synchronized using a unified timestamp obtained from a network time protocol server to ensure that all data corresponds to the same production moment. Finally, the synchronized data is cleaned. The specific steps are as follows: a moving average filter with a window length of 5 data points is used to smooth the dynamic response waveform data, and the Laida criterion is used to detect outliers. That is, the arithmetic mean and standard deviation of the data sequence are calculated, and data points that deviate from the arithmetic mean by more than three times the standard deviation are identified as invalid data points caused by transient interference or equipment noise and are removed. Finally, the cleaned data constitutes the basic dataset for subsequent calculations. Taking a specific sample as an example, in a specific production batch, the system collected data on the surface material type as "cotton fiber (code MAT-COTTON-01)". Real-time physical parameters included: substrate weight 52 g / m², moisture content 33%, coating weight 4.5 g / m², drying temperature 65℃, and wind speed 2.5 m / s. Simultaneously, the dynamic behavior testing unit recorded a complete pH dynamic response curve: the initial steady-state pH was 5.50, which dropped to a minimum of 5.18 after disturbance, then recovered to 5.48 and stabilized within approximately 1.2 seconds. All data, after synchronization, filtering, and outlier removal, formed the basic dataset for this sample. The present invention is further configured such that S2 specifically includes a root gene calculation step, a response factor calculation step, a reserve factor calculation step, and a weak acid steady-state response coefficient calculation step. Specifically, the root gene calculation step calculates a value characterizing the degree of fixation and binding of the weak acid functional component in the material by processing static parameters such as the surface material type, basis weight, and moisture content, which is used to evaluate the basic static stability of the weak acid surface layer; the response factor calculation step calculates a value characterizing the surface layer's ability to resist external disturbances and quickly recover to steady state by analyzing the pH dynamic response waveform after applying disturbance, which is used to evaluate the dynamic anti-interference and self-recovery performance of the weak acid surface layer; the reserve factor calculation step calculates a value characterizing the surface layer's long-term potential to maintain a preset acidic environment by using a buffer model based on real-time pH value and moisture content, which is used to evaluate the long-term buffering capacity and functional sustainability of the weak acid surface layer; the weak acid steady-state response coefficient calculation step generates a comprehensive index describing the surface layer's health by weighting and integrating the above three factors according to the product functional objectives, which is used to provide the entire system with a unified, quantifiable, and direct reflection of the weak acid functional steady-state level as a core regulatory target and decision-making basis.

[0026] The present invention is further configured such that the root gene calculation step is as follows: Based on the surface material types in the basic dataset, the inherent bonding coefficient corresponding to the current surface material type is obtained by querying a preset material property knowledge base; The inherent bonding coefficient, along with the substrate basis weight and moisture content from the basic dataset, are used as inputs and passed to a preset association mapping model for processing. The output of the association mapping model is then obtained to yield the root gene. Specifically, based on the surface material type in the basic dataset, such as cotton fiber or nonwoven fabric represented by a specific material code, the corresponding inherent bonding coefficient is obtained by querying a preset material property knowledge base. The material property knowledge base is a structured database that stores experimentally measured data; for example, the inherent bonding coefficient of cotton fiber is preset to 0.85, and the inherent bonding coefficient of polyester fiber is preset to 0.60. Subsequently, the retrieved inherent bonding coefficient, along with the substrate basis weight and moisture content collected in real time from the basic dataset, are used as inputs and passed to the preset association mapping model for processing. The association mapping model internally employs a weighted and corrective logic based on explicit rules: firstly, the inherent bonding coefficient... The base score is used as the baseline score. Next, the substrate weight is compared with the preset base weight value of 50 g / m², and a capacity correction factor is generated to adjust the baseline score. Then, the moisture content is compared with the preset optimal moisture content (e.g., setting the default optimal moisture content range of 30% to 50%). The current score is corrected positively or negatively depending on whether the moisture content is within this range. Finally, the processing result is normalized and mapped to the range of 0 to 1. The value between 0 and 1 output by the final correlation mapping model is the root gene. The root gene is used to quantify the fixation and bonding strength of the weak acid functional components in the surface material, that is, the basic static stability of the weak acid surface layer. Using the previous example, the system queries the knowledge base and obtains the inherent binding coefficient of cotton fiber as 0.85. The association mapping model uses this as the basis, and combines the weight of 52 g / m² (higher than the baseline value of 50 g / m², so a positive correction is made) and the moisture content of 33% (within the optimal range of 30%-50%, so a slight positive correction is made). After rule calculation and normalization, the root gene of the current sample is output as Gf = 0.87.

[0027] The present invention further specifies that the response factor calculation step is as follows: By performing signal analysis on the dynamic response waveform data, response features are extracted, including: maximum pH shift, time required to recover to a preset threshold, and area under the waveform curve; The response features are input to a pre-trained dynamic behavior quantization model, and the output of the model is obtained to yield the response factor. Specifically, firstly, the system performs signal analysis on the dynamic response waveform data to extract response features. The dynamic response waveform data is a sequence of pH values ​​changing over time at 500 points per second. The extracted response features include: maximum pH offset, time required to recover to a preset threshold, and area under the waveform curve. The maximum pH offset refers to the maximum absolute value of the pH value deviating from the original steady-state value, specifically identified in the waveform data using a peak-finding algorithm. The time required to recover to the preset threshold refers to the time required for the pH value to recover from the maximum offset point to within ±0.1 of the original steady-state value, specifically calculated precisely in the data sequence of the dynamic response waveform data using linear interpolation. The area under the waveform curve refers to the area enclosed between the pH response curve and the horizontal line representing the original steady state. The area of ​​the region is obtained by numerical integration using the trapezoidal rule. Then, the maximum pH shift, recovery time, and area under the curve are combined to form a feature vector, which is normalized based on the maximum and minimum values ​​of each feature obtained from the historical database. The normalized feature vector is then input into a pre-trained dynamic behavior quantification model, which is a random forest regression model trained using historical production data. This model obtains a value between 0 and 1 by performing a comprehensive nonlinear mapping on the input features; this value is the response factor. The response factor is used to quantify the strength of the weakly acidic surface layer's ability to resist external disturbances and quickly recover to a steady state, i.e., the dynamic anti-interference and self-recovery performance of the weakly acidic surface layer. Using the previous example, the system analyzes the acquired dynamic waveform: the maximum pH offset is calculated to be 0.32, the time required to recover to pH 5.40 (i.e., to enter the range of 5.50±0.1) is 0.98 seconds, and the area under the waveform curve is 0.15; after normalizing these feature values, they are input into the pre-trained random forest regression model, and the model outputs the response factor Rf = 0.72 for the current sample.

[0028] The present invention further specifies that the reserve factor calculation step is as follows: The moisture content in the static process parameter set and the corresponding real-time pH value in the dynamic response waveform data are used as input parameters to input the capacity estimation model based on buffer theory. The capacity estimation model corrects the buffering capacity represented by real-time pH value based on moisture content, and outputs a reserve factor representing the effective buffering capacity. Specifically, the moisture content in the static process parameter set and the corresponding real-time pH value in the dynamic response waveform data are used as input parameters to the capacity estimation model constructed based on buffering theory. The capacity estimation model is based on the pre-defined chemical properties of the buffer system (e.g., using the pH-buffer capacity titration curve of the citric acid-sodium citrate buffer system as built-in data). It converts real-time pH values ​​into theoretical buffer capacity through table lookup or interpolation. The model also pre-defines a baseline moisture content threshold (e.g., an example value of 30%) to indicate whether moisture is sufficient. The model performs a moisture content correction calculation, with the following rules: if the current moisture content is greater than or equal to the baseline threshold, the correction factor is 1; if the current moisture content is less than the baseline threshold, the correction factor is the square root of the quotient of the current moisture content divided by the baseline threshold. For example, when the moisture content is 24% and the baseline threshold is set to 30%, the correction factor is calculated to be 0.894. Subsequently, the model multiplies the theoretical buffer capacity by the moisture content correction factor to obtain the effective buffer capacity. Finally, the model normalizes the effective buffer capacity to the range of 0 to 1, and the output normalized value is the reserve factor. The reserve factor is used to quantify the potential and buffering capacity of the surface layer to maintain a preset acidic environment under the current water content. Following the previous example, the system extracts the real-time pH value of 5.50 and the water content of 33% from the data; the capacity estimation model finds the theoretical buffering capacity based on pH 5.50 and calculates the effective buffering capacity based on the water content of 33% (higher than the baseline threshold of 30%, so the correction factor is 1). After normalization, the reserve factor Sf = 0.78 for the current sample is output.

[0029] The present invention further specifies that the calculation step of the steady-state response coefficient of the weak acid is as follows: Based on the preset product function type, the corresponding function-oriented weight configuration strategy is called from the preset target product function category correspondence table; Based on the aforementioned function-oriented weight configuration strategy, a multi-dimensional weighted fusion operation is performed on the root gene, the response factor, and the reserve factor to generate a weak acid steady-state response coefficient. Specifically, firstly, according to the preset product function type, the corresponding function-oriented weight configuration strategy is retrieved from the preset target product function category correspondence table; the product function type is a preset category based on market positioning, such as "refreshing," "highly absorbent," and "sensitive skin-specific"; the target product function category correspondence table is a structured database that stores the specific weight value set corresponding to different product function types. For example, the corresponding function-oriented weight configuration strategy set for the product function type "sensitive skin-specific" is: root gene weight 0.3, response factor weight 0.5, reserve factor weight 0.2, and the sum of these three weight values ​​is 1. Subsequently, based on the invoked function-oriented weight configuration strategy, a weighted fusion operation is performed on the root gene, response factor, and reserve factor. Specifically, the root gene value is multiplied by its corresponding weight, the response factor value is multiplied by its corresponding weight, and the reserve factor value is multiplied by its corresponding weight. These three products are then summed to obtain an intermediate value between 0 and 1. Finally, this intermediate value is multiplied by 100 to generate a final value within the range of 0 to 100. This final value is the weak acid steady-state response coefficient. The weak acid steady-state response coefficient is a comprehensive health score used to directly quantify the overall functional steady-state level of the current weak acid surface layer under specific product functional objectives. Using the previous example, the current product is "for sensitive skin". The system calls the corresponding weighting strategy (Gf:0.3, Rf:0.5, Sf:0.2). The weighted fusion calculation is: 0.87 * 0.3 + 0.72 * 0.5 + 0.78 * 0.2 = 0.774. Multiplying this intermediate value by 100, we get the weak acid steady-state response coefficient of the current surface layer as 77.4. This value is lower than the preset target range [85, 100], and the system determines that regulation needs to be triggered.

[0030] The present invention is further configured such that S3 specifically includes: The steady-state response coefficient of the weak acid is compared with a preset target range to determine whether a deviation has occurred. If a deviation occurs, the dominant factor causing the deviation is determined by numerical comparison and threshold judgment of the root gene, the response factor, and the reserve factor. Based on the category of the dominant factor, the corresponding combination of compensation actions is matched by querying the pre-set root cause-policy mapping relationship library; Based on the aforementioned compensation action combination, a mixing instruction containing specific process parameter adjustment instructions is generated through an instruction synthesis engine. Specifically, firstly, a comparator compares the steady-state response coefficient of the weak acid with a preset target range to determine if a deviation has occurred. The preset target range is a specific numerical interval, for example, [85, 100]. If the steady-state response coefficient of the weak acid falls within this numerical interval, it is considered to have no deviation; if it is below the lower limit of 85, it is considered to have a negative deviation. When a negative deviation is determined, the dominant factor causing the deviation is identified by comparing the values ​​of the root gene, response factor, and reserve factor with thresholds. The specific implementation method for comparison and threshold determination is as follows: the values ​​of the root gene, response factor, and reserve factor are compared with preset health thresholds (e.g., root gene threshold 0.8, response factor threshold 0.75, reserve factor threshold 0.7); then, from the factors whose values ​​are lower than their own health thresholds, the factor with the lowest value is selected as the dominant factor. Then, based on the determined dominant factor category, a pre-defined root cause-policy mapping database is queried to match the corresponding compensation action combination. This database is a structured database that uses the dominant factor category as the index key and stores predefined compensation action descriptions. For example, compensation actions for the "response factor dominant" category include: linking the actions "increase the spraying ratio of sensitizer B by 3%" and "decrease the initial temperature of the drying zone by 5°C". Finally, based on the matched compensation action combination, a dispensing instruction containing specific process parameter adjustment instructions is generated through an instruction synthesis engine. This engine integrates the production line's equipment driver library and instruction templates, enabling it to convert the parameterized descriptions in the compensation action combination (such as "increase by 3%)" into specific equipment control instructions (such as a flow setting command sent to a designated metering pump) that the production line's programmable logic controller can recognize and execute. Using the previous example, the system's judgment coefficient of 77.4 shows a negative deviation. Comparing the three factor values ​​(Gf=0.87, Rf=0.72, Sf=0.78) with the health thresholds (0.8, 0.75, 0.7), it is found that the response factor Rf (0.72) is the only factor with a value lower than its own threshold (0.75) and the lowest value. Therefore, it is determined that "response factor dominance" is the cause of the deviation. Based on this, the system queries the mapping relationship library, matches the combination of compensating actions (such as increasing the proportion of sensitizer B by 3.5% and reducing the initial drying temperature by 3°C), and generates specific dispensing instructions to be issued to the execution mechanism.

[0031] The present invention is further configured such that S4 specifically includes: The allocation command is sent to the actuator of the production line to adjust the corresponding production process parameters; After the process parameters are adjusted and the preset process delay time has elapsed, the data acquisition and processing flow is triggered again to obtain new dynamic response waveform data and calculate the new weak acid steady-state response coefficient as a feedback parameter. The new steady-state response coefficient of the weak acid is compared with the preset target range threshold, and logical judgment is made in the preset calibration effect evaluation rule base according to the comparison result to generate the evaluation result. Based on the evaluation results, calibration is performed. If calibration is successful, the system status is updated; if calibration fails to meet the target, a new round of allocation instruction generation is initiated, and the feedback parameters and evaluation results are entered into the calibration history database. Specifically, firstly, the allocation instruction is sent to the actuators on the production line via an industrial communication protocol to adjust the corresponding production process parameters, such as adjusting the flow setpoint of the precision metering pump or the temperature setpoint of the temperature controller. Next, after confirming that the process parameter adjustment is complete and after a preset process delay time, the data acquisition and processing process is re-triggered; the process delay time is a preset waiting time to ensure the production status returns to stability, for example, set to 90 seconds. When the data acquisition and processing process is re-triggered, the system acquires the latest dynamic response waveform data at the current moment and calculates the new weak acid steady-state response coefficient as a feedback parameter. Subsequently, the new weak acid steady-state response coefficient is compared with a preset target range threshold, which is the same as the aforementioned threshold, for example, the aforementioned numerical range [85, 100]. Simultaneously, the comparison results and related data are input into a preset calibration effect evaluation rule base for logical judgment. The calibration effect evaluation rule base contains predefined evaluation rules, such as: if the new weak acid steady-state response coefficient is greater than or equal to 85, it is determined to be "entering the target range"; if the new weak acid steady-state response coefficient increases by more than 5% compared to the coefficient before adjustment, it is determined to be "significantly improved". The system integrates all triggered rules to generate an evaluation result including an overall conclusion and specific labels. Finally, calibration is judged based on the evaluation results: if the overall conclusion of the evaluation result determines that the calibration is successful, the system production status is updated and production continues while maintaining the current process parameters; if the calibration fails to meet the target, a new round of dispensing instruction generation is automatically initiated. Regardless of whether the calibration is successful or not, the system records the feedback parameters of this cycle, the executed dispensing instructions, and the generated evaluation results as a complete record in the calibration history database. Following the previous example, after the actuator completes parameter adjustment, the system waits for a 90-second process delay before re-collecting data and calculating a new weak acid steady-state response coefficient of 86.2. The calibration effect evaluation rule base judges that the new coefficient of 86.2 is higher than 85, thus it is determined to be "within the target range"; and compared to the previous coefficient of 77.4, the improvement is 11.4% (>5%), thus it is determined to be "significantly improved". The comprehensive evaluation result is "calibration successful". The system updates the status accordingly, maintains the adjusted parameters for production, and records the entire process data of this successful case into the calibration history database. This exemplary case verifies that the method can effectively improve the health of the weak acid surface layer from 77.4 to 86.2, realizing online closed-loop calibration and performance optimization.

[0032] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer, characterized in that, include: S1: Obtain the static process parameters and dynamic behavior response data of the weak acid surface layer material and perform synchronous processing to form a basic dataset; S2: Calculate the steady-state response coefficient of weak acid based on the basic dataset; S3: Generate a mixing instruction based on the deviation between the steady-state response coefficient of the weak acid and the preset target range; S4: Based on the adjustment command, process parameters are adjusted, feedback parameters are collected, and online closed-loop calibration is completed.

2. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 1, characterized in that, The basic dataset specifically includes: a set of static process parameters and dynamic response waveform data.

3. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 2, characterized in that, The specific process for forming the basic dataset in S1 is as follows: By deploying process sensor groups and material information management systems on the production line, a set of static process parameters is collected in real time. The set of static process parameters includes at least: surface material type, substrate weight, moisture content, coating amount, drying temperature, and wind speed. By deploying dynamic behavior testing units at key nodes of the production line, standardized microdroplet interference is applied to the weakly acidic surface layer, and the pH value changes over time are collected simultaneously to generate dynamic response waveform data. The collected static process parameter set and dynamic response waveform data are synchronized and aligned using a unified timestamp. After synchronization and alignment, the data is filtered and outlier detected to remove invalid data points caused by transient interference or equipment noise, thus forming a basic dataset.

4. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 1, characterized in that, S2 specifically includes the root gene calculation step, the response factor calculation step, the reserve factor calculation step, and the weak acid steady-state response coefficient calculation step.

5. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 4, characterized in that, The root gene calculation steps are as follows: Based on the surface material types in the basic dataset, the inherent bonding coefficient corresponding to the current surface material type is obtained by querying a preset material property knowledge base; The inherent binding coefficient and the substrate basis weight and moisture content in the basic dataset are used as inputs and passed to a preset association mapping model for processing. The output of the association mapping model is obtained to obtain the root gene.

6. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 4, characterized in that, The steps for calculating the response factor are as follows: By performing signal analysis on the dynamic response waveform data, response features are extracted, including: maximum pH shift, time required to recover to a preset threshold, and area under the waveform curve; The response features are passed as input to a pre-trained dynamic behavior quantization model, and the output of the dynamic behavior quantization model is obtained to obtain the response factor.

7. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 4, characterized in that, The steps for calculating the reserve factor are as follows: The moisture content in the static process parameter set and the corresponding real-time pH value in the dynamic response waveform data are used as input parameters to input the capacity estimation model based on buffer theory. The buffering capacity characterized by real-time pH value is corrected based on the water content using a capacity estimation model, and a reserve factor characterizing the effective buffering capacity is output.

8. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 4, characterized in that, The steps for calculating the steady-state response coefficient of the weak acid are as follows: Based on the preset product function type, the corresponding function-oriented weight configuration strategy is called from the preset target product function category correspondence table; Based on the function-oriented weight configuration strategy, a multi-dimensional weighted fusion operation is performed on the root gene, the response factor, and the reserve factor to generate a weak acid steady-state response coefficient.

9. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 1, characterized in that, S3 specifically includes: The steady-state response coefficient of the weak acid is compared with a preset target range to determine whether a deviation has occurred. If a deviation occurs, the dominant factor causing the deviation is determined by numerical comparison and threshold judgment of the root gene, the response factor, and the reserve factor. Based on the category of the dominant factor, the corresponding combination of compensation actions is matched by querying the pre-set root cause-policy mapping relationship library; Based on the aforementioned combination of compensation actions, a scheduling instruction containing specific process parameter adjustment instructions is generated through the instruction synthesis engine.

10. The method for adaptive pH adjustment and online closed-loop calibration of a weakly acidic surface layer according to claim 1, characterized in that, S4 specifically includes: The allocation command is sent to the actuator of the production line to adjust the corresponding production process parameters; After the process parameters are adjusted and the preset process delay time has elapsed, the data acquisition and processing flow is triggered again to obtain new dynamic response waveform data and calculate the new weak acid steady-state response coefficient as a feedback parameter. The new steady-state response coefficient of the weak acid is compared with the preset target range threshold, and logical judgment is made in the preset calibration effect evaluation rule base according to the comparison result to generate the evaluation result. Based on the evaluation results, a calibration is performed. If the calibration is successful, the system status is updated. If the calibration fails to meet the target, a new round of allocation instruction generation process is initiated, and the feedback parameters and evaluation results are entered into the calibration history database.

Images