Chemical filling production line whole-process automatic control method based on artificial intelligence

Abstract

The invention relates to a chemical filling production line full-process automatic control method based on artificial intelligence, which breaks through the limitation of traditional fixed logic control through multi-module cooperation and closed-loop control, and triggers graded early warning and self-adaptive adjustment according to a deviation range set by dynamically comparing theoretical filling parameters with actual parameters. The chemical property fluctuation is effectively dealt with, the filling precision and the product percent of pass are greatly improved, the raw material loss is reduced, and precise upgrading from experience control to data driving is realized; on the basis of real-time data acquisition and a millisecond-level response mechanism, the system can actively identify parameter deviation and equipment abnormity, quickly trigger early warning and generate a control strategy, so that the response capability to emergency situations is remarkably enhanced, non-planned shutdown is reduced, and the safety management and control level of a production line is improved; by executing result feedback and model iterative optimization, the system has continuous self-learning and process parameter adaptive adjustment capabilities, the flexible adaptation capability of a production line to multi-variety production is remarkably improved, the manual intervention cost is reduced, and the production efficiency and the automation level are comprehensively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to an artificial intelligence-based chemical filling production line full-process automatic control method, belonging to the field of chemical filling. BACKGROUND

[0002] With the global manufacturing industry transforming towards intelligence and digitization, chemical filling as a key process in the chemical, pharmaceutical, electronic and other industries, its automation control precision and production safety are increasingly important. The traditional chemical filling production line mainly relies on manual parameter setting or fixed logic-based automatic devices to perform operations, which not only cannot adapt to the differences in viscosity, density, volatility and other characteristics of various chemicals, but also has low filling precision, delayed equipment failure warning, and relies on manual experience for process parameter adjustment.

[0003] In recent years, artificial intelligence technology has gradually penetrated into the field of industrial control, but there are still significant technical bottlenecks in the chemical filling scenario: most existing solutions only achieve local optimization of a single link, lack of fusion analysis of raw material characteristics, equipment status, environmental parameters and other data; real-time data processing capability is insufficient, unable to dynamically respond to the instantaneous changes in the rheological properties of chemicals; the parameter deviation evaluation system relies on preset thresholds, which is difficult to adapt to the process differences of different batches of chemicals, and lacks a self-learning optimization mechanism based on historical data.

[0004] At the same time, production line data security and cross-device collaborative control are also current industry pain points: fragmented device status data caused by decentralized data storage makes it difficult to build a full-process quality traceability system; the communication protocols of different brands of filling equipment are not unified, causing data interaction delay and affecting real-time execution of control strategies. Therefore, developing an artificial intelligence-based chemical filling production line full-process automatic control method to realize intelligent perception, dynamic optimization and safety control of the chemical filling process has become an urgent need to improve the level of lean manufacturing. SUMMARY

[0005] The purpose of the present application is to provide an artificial intelligence-based chemical filling production line full-process automatic control method to solve the above problems.

[0006] To achieve the above purpose, the present application provides the following technical solution: an artificial intelligence-based chemical filling production line full-process automatic control method, the chemical filling production line full-process automatic control method comprising: S1, a data acquisition module acquires initial data of a production line to be filled with chemicals, uploads to a control center, and marks as a production line initial data set; S2, a data analysis module extracts key parameters from the production line initial data set, and calculates a theoretical filling parameter set; S3, the real-time data acquisition module collects real-time data of the production line filled with chemicals, uploads to the control center, and is marked as a production line real-time data set; S4, the control center processes the real-time data to obtain an actual filling parameter set; S5, the control center calls the theoretical filling parameter set and the actual filling parameter set, constructs a parameter deviation evaluation system, calculates a deviation range set, and determines whether to trigger a hierarchical early warning; S6, the control center generates a control strategy according to the deviation range set, and evaluates the filling state and the quality prediction result; S7, the production line execution module adjusts the equipment operating parameters according to the control strategy, executes the filling closed-loop control, and feeds back the execution result to the control center for model iteration and optimization.

[0007] Further, the production line initial data set includes the target filling amount Vt, the current residual amount Vr, the chemical density p, the maximum filling rate Qm and the deviation proportion a allowed by production before the filling operation is performed.

[0008] Further, the calculation formula of the theoretical filling parameter set includes: required filling amount : ; theoretical filling time : ; target quality : ; allowed quality deviation : .

[0009] Further, the actual filling parameter set includes the actual filling amount Vr', the actual filling rate Qr and the filled quality mr collected after the filling operation is performed.

[0010] Further, the deviation range set includes: flow deviation: ; volume progress deviation: ; quality deviation: .

[0011] Further, the hierarchical early warning includes: when AQ<10% and AVp<5% and A mp<3%, it is determined as normal state; when AQ>10% or AVp>5% or A mp>3%, it is determined as warning state; When Δmp> 2a or ΔQ> 30%, it is determined as an emergency suspension state.

[0012] Further, the control strategy includes valve opening adjustment, and an adjustment formula of the valve opening adjustment amount ΔK is as follows: .

[0013] Further, the control strategy further includes mass prediction adjustment, and a mass prediction formula is as follows: , The mass determination formula is as follows: If yes, it is determined as qualified, otherwise, it is determined as unqualified.

[0014] Further, the control center is further used for pre-processing real-time data, and the pre-processing includes outlier elimination, data smoothing, time alignment and format standardization.

[0015] Further, the production line execution module feeds back an execution result to the control center after executing the control strategy, for iterative optimization of an artificial intelligence model and self-adaptive adjustment of process parameters.

[0016] The application has the following beneficial effects: the application breaks through the limitation of traditional fixed logic control through multi-module cooperation and closed-loop control, fills parameters and actual parameters through dynamic comparison theory, triggers graded early warning and self-adaptive adjustment according to a deviation range set, effectively deals with chemical property fluctuation, greatly improves filling precision and product qualification rate, reduces raw material loss, realizes precise upgrading from experience control to data driving; relying on real-time data acquisition and millisecond-level response mechanism, the system can actively identify parameter deviation and equipment abnormality, quickly triggers early warning and generates a control strategy, significantly enhances response ability to sudden conditions, reduces unplanned downtime, and improves production line safety control level; through execution result feedback and model iterative optimization, the system has continuous self-learning and process parameter self-adaptive adjustment ability, significantly improves the flexible adaptation ability of the production line to multi-species production, reduces manual intervention cost, and comprehensively improves production efficiency and automation level.

[0017] The above description is only a summary of the technical scheme of the application, in order to more clearly understand the technical means of the application, and the content of the specification can be implemented, the following preferred embodiments of the application are described in detail with reference to the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flowchart of a chemical filling production line full-process automatic control method based on artificial intelligence according to an embodiment of the application; Figure 2A structural block diagram of a control system shown in an embodiment of the present application. DETAILED DESCRIPTION

[0019] The specific embodiments of the present application will be further described in conjunction with the drawings and examples. The following examples are used to illustrate the present application, but are not used to limit the scope of the present application.

[0020] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", "third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0021] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0022] Please refer to Figure 1 , a chemical filling production line full-process automatic control method based on artificial intelligence shown in an embodiment of the present application, the chemical filling production line full-process automatic control method comprises: S1, the data acquisition module collects the initial data of the production line to be filled with chemicals, uploads to the control center, and marks as the production line initial data set; S2, the data analysis module extracts key parameters from the production line initial data set, and calculates to obtain a theoretical filling parameter set; S3, the real-time data acquisition module collects real-time data of the production line filled with chemicals, uploads to the control center, and marks as the production line real-time data set; S4, the control center processes the real-time data to obtain an actual filling parameter set; S5, the control center calls the theoretical filling parameter set and the actual filling parameter set, constructs a parameter deviation evaluation system, calculates a deviation range set, and determines whether to trigger a hierarchical early warning; S6, the control center generates a control strategy according to the deviation range set, and evaluates the filling state and the quality prediction result; S7, the production line execution module adjusts the equipment operating parameters according to the control strategy, executes the filling closed-loop control, and feeds back the execution result to the control center for model iteration and optimization.

[0023] It needs to be specifically explained that the real-time data acquisition module dynamically acquires physical parameters closely related to the filling process through the sensor network deployed at each key node of the filling production line. In the material conveying link, the flow sensor monitors the flow rate of the chemical in real time to provide real-time feedback for the filling progress; the pressure sensor is installed at the connection between the pipeline and the valve to acquire pressure fluctuation data in the conveying process for judging whether there is a risk of pipeline blockage or leakage; the weight sensor is integrated at the bottom of the filling container to directly measure the real-time weight change of the chemical in the container; and the liquid level sensor feeds back the liquid level height in the container in real time through non-contact detection technology. All sensors acquire data according to the system preset synchronous clock period, convert the analog signals into digital signals, and then directly transmit them to the control center through the industrial Ethernet protocol.

[0024] It needs to be explained that after the control center receives the real-time data stream, it first performs preprocessing operations on the data. The system automatically identifies and eliminates abnormal jump values caused by sensor failure or electromagnetic interference, and uses a sliding window filtering algorithm to smooth the data; at the same time, in view of the differences in sampling frequency of different types of sensors, the data synchronization alignment operation is performed to ensure that all parameters have comparability in the time dimension. After completing the preprocessing, the system performs format standardization conversion on the data, unifies the data of different protocol formats into an internal standard data structure, and assigns a time stamp and a device identifier to each data point. Subsequently, the control center classifies and stores the data according to the parameter type, establishes an index relationship based on time series, and forms a structured production line real-time data set. This data set is not only used for real-time comparison of theoretical parameters and actual filling state, but also participates in subsequent deviation analysis and control decision as process quality data, ensuring that the entire filling process is in an intelligent management state of being controllable and traceable.

[0025] In an embodiment, the initial data set of the production line includes the target filling volume Vt, the current residual volume Vr, the chemical density p, the maximum filling rate Qm, and the deviation ratio a allowed by production before the filling operation is performed.

[0026] It needs to be further explained that the data acquisition module extracts the production line initial data from the raw material storage tank, production work order system and other data sources according to the preset acquisition strategy. Among them, the target filling volume Vt is obtained by analyzing the process instructions of the production work order, which is used to define the theoretical target value of single filling; the current residual volume Vr is measured by the liquid level sensor installed at the bottom of the filling tank in real time, which is used to compensate the residual medium of the pipeline and the tank; the chemical density p is detected by the online density meter and synchronized to the data acquisition module, which is the core parameter for calculating the conversion between filling weight and volume; the maximum filling rate Qm is retrieved from the equipment parameter library, which is limited by the pipe diameter and pump performance; the allowed deviation ratio a is pre-set by the process standard, which is used to define the qualified range of filling accuracy.

[0027] It needs to be explained that the collected data is uploaded to the control center through a safe network channel after preliminary processing. During the uploading process, data encryption technology is used to encrypt the data to ensure the security of the data during transmission and prevent data leakage.

[0028] In an embodiment, the calculation formula of the theoretical filling parameter set includes: : : ; Vn is the filling volume, Vt is the target filling volume, and Vr is the current residual volume.

[0029] Theoretical filling time : ; Tf is the theoretical filling time, Vn is the filling volume, and Qm is the maximum filling rate.

[0030] Target mass : ; mt is the target mass, Vt is the target filling volume, and p is the chemical density.

[0031] Allowed mass deviation : . is the allowed mass deviation, which is the mass fluctuation range allowed by production, is the target mass, is the allowed error ratio, which is the percentage of the mass deviation allowed by the production process.

[0032] It needs to be further explained that after the above core parameter calculation is completed, the data analysis module will systematically integrate and verify the theoretical filling parameter set. First, the calculated filling amount Vn is compared with the device safety volume. If Vn exceeds the maximum carrying capacity of the device, a warning is automatically triggered and the process parameter is prompted to be corrected; second, according to the adaptability analysis of the theoretical filling time Tf and the production rhythm requirement, if Tf exceeds the preset production cycle limit, the maximum filling rate Qm is adjusted in proportion to recalculate, and the theoretical parameters are ensured to meet the actual production efficiency requirements.

[0033] At the same time, the data analysis module will establish the correlation constraint relationship between the theoretical filling parameters, combine the target mass mt with the allowable mass deviation Am to generate the mass control interval [mt-Am, mt+Am], and use it as a reference to check other parameters. If there is a logical conflict between parameters, such as the calculated target mass exceeding the device bearing limit, the system will start the parameter optimization algorithm to automatically adjust the parameter combination under the premise of meeting the process standard, and generate a feasible theoretical filling parameter set.

[0034] It needs to be explained that the finally generated theoretical filling parameter set will be assigned a unique version number and stored in association with the initial data set of the production line. This parameter set not only serves as a theoretical reference for subsequent real-time control, but also is synchronized to the parameter database of the control center, providing basic data support for dynamic deviation analysis and control strategy generation, ensuring that the parameter calculation of the entire filling process is scientific, accurate and traceable.

[0035] In an embodiment, the actual filling parameter set includes the actual filling amount Vr', the actual filling rate Qr and the filled mass mr collected after the filling operation has been performed.

[0036] It needs to be explained that after the control center obtains real-time liquid level sensor data, the current liquid level height is subtracted from the initial liquid level height of the container, and the actual filling amount change value is obtained by combining the cross-sectional area of the container. If there are multiple liquid level sensors, the system automatically selects valid data for weighted average calculation to avoid data deviation caused by single sensor error. At the same time, the control center continuously monitors the liquid level change trend, and when the liquid level data appears abnormal fluctuation, the adjacent time period data is immediately retrieved for comparison and verification to ensure the accuracy of the actual filling amount calculation; For the actual filling rate Qr, the control center analyzes the actual filling amount change in a plurality of consecutive sampling periods. By calculating the filling amount increment per unit time, the real-time filling rate is obtained. If the rate change of the current sampling period and the last period exceeds the preset threshold, the system will enable the moving average algorithm to smooth the rate data of the last 5 periods, eliminating the influence of accidental fluctuations. In addition, the control center will also combine the ideal filling rate curve preset by the production work order to make trend prediction on the actual filling rate, and identify potential rate abnormalities in advance. For the filled mass mr, the control center multiplies the real-time collected chemical density data with the actual filling amount to obtain the filled mass. To deal with the density fluctuation problem, the system uses a dynamic weighted average method to give different weights to the last 10 density measurements, with the recent data having a higher weight. After calculating the weighted average density, it is used to calculate the mass.

[0037] After the control center completes the above parameter calculation, it will perform a secondary verification on the actual filling parameter set. By comparing the logical correlation between parameters, such as whether the change trend of the actual filling rate and the filled mass matches, and combining historical data rules for reasonableness verification. If parameter abnormalities are found, the system will immediately trigger the data review mechanism to trace back the original sensor data and transmission records in the corresponding time period, and locate the problem source. The final generated actual filling parameter set will be time-stamped and synchronized to the real-time database of the control center, and will be used as the core basis for real-time parameter comparison, deviation analysis, and control strategy generation, ensuring that the parameter calculation of the entire filling process is reliable and effective.

[0038] In an embodiment, the deviation range set includes: Flow deviation: ; AQ is the flow deviation, i.e., the deviation ratio of the actual filling rate and the maximum rate, Qr is the actual filling rate, i.e., the actual filling speed of the production line, Qm is the maximum filling rate.

[0039] Volume progress deviation: ; AVp is the volume progress deviation, i.e., the deviation ratio of the actual filling volume progress and the time progress, Vr' is the actual filled volume, Vn is the required filling amount, t is the actual time used, Tf is the theoretical filling time.

[0040] Mass deviation: . Dmp is the mass deviation, i.e., the deviation ratio of the actual mass and the target mass, mr is the actual filled mass, mt is the target mass.

[0041] In an embodiment, the hierarchical early warning includes: When AQ < 10% and AVp < 5% and Dmp < 3%, it is determined to be a normal state; that is, only when all three deviations do not exceed the threshold, it is a normal condition.

[0042] When AQ>10% or AVp>5% or Ampi>3%, it is determined as a warning state; if one of the flow, volume progress, and mass deviation is greater than the normal threshold, the control center will remind the production line filling module to adjust the filling strategy; When Ampi>2a or AQ>30%, it is determined as an emergency pause state. The mass deviation exceeds 2 times the allowed error, or the flow deviation exceeds 30%, at which time the control center will remind the production line filling module to pause the production line Further explanation is needed. After completing the deviation calculation and early warning determination, the control center will execute a differentiated response mechanism according to the hierarchical early warning results. When in the normal state, the system will continuously collect and analyze data, dynamically update the deviation evaluation model, and compare the current parameter state with the historical high-quality batch data to provide reference for process optimization; if a warning level early warning is triggered, the control center will immediately start the adaptive adjustment program, generate compensation instructions for the out-of-limit indicators based on the pre-set parameter adjustment strategy library; When the flow deviation AQ is out of limit, the system automatically reduces the filling pump speed and adjusts the valve opening to match the theoretical filling rate; if the volume progress deviation AVp is abnormal, the remaining filling amount and time allocation are recalculated, and the subsequent filling rhythm is dynamically corrected; For the case of pausing the production line, the control center will immediately issue an alarm to the on-site operator through the sound and light alarm system and push the abnormal information to the management personnel terminal, detailing the fault type, occurrence time, and possible causes; at the same time, the system will automatically save all running parameters and sensor data snapshots to form a complete abnormal event log for subsequent fault tracing and root cause analysis. In addition, the control center will start the backup data verification mechanism to cross-verify sensor data and calculation results to avoid misjudgment of shutdown due to single-point failure.

[0043] It needs to be explained again that the entire parameter deviation evaluation system has self-learning ability. The control center will continuously record the parameter fluctuation characteristics and early warning response effect under different working conditions, analyze historical early warning data based on machine learning algorithm, and dynamically optimize each deviation threshold and response strategy; by analyzing the parameter evolution law in multiple emergency shutdown events, the system can automatically adjust the trigger thresholds of Ampi and AQ to make them more suitable for the actual production safety boundary; at the same time, according to the response effect of the warning level, the adjustment amplitude of the parameter compensation instruction is adaptively calibrated, so as to realize the continuous optimization and precision of the early warning mechanism, and continuously improve the stability and safety of the production line operation.

[0044] In an embodiment, the control strategy includes valve opening adjustment, and the adjustment formula of the valve opening adjustment amount AK is: .

[0045] AK is the valve opening adjustment, that is, how much the valve opening needs to change, Qr is the actual filling rate; Qm is the maximum filling rate.

[0046] If the actual rate Qr is less than 90% of the maximum rate, the valve opening is increased by 5% to increase the flow rate; if the actual rate Qr is greater than 110% of the maximum rate, the valve opening is decreased by 5% to decrease the flow rate; if the actual rate Qr is between 90%-110%, the valve opening remains unchanged.

[0047] In an embodiment, the control strategy further includes a mass prediction adjustment, and the mass prediction formula is: , M is the predicted mass, that is, the predicted total mass after the final filling, mr is the actual filled mass, Vn is the required filling mass, Vr' is the actual filled volume, and p is the chemical filling density.

[0048] The mass determination formula is: If it is true, it is determined to be qualified, otherwise it is not qualified.

[0049] M is the predicted mass, mt is the target mass, and Am is the allowed mass deviation. If the deviation between the predicted mass and the target mass is less than or equal to the allowed mass deviation Am, it is determined to be qualified; otherwise, it is not qualified.

[0050] It needs to be explained that in terms of mass prediction, the control center receives a new sampling period data every time, that is, the predicted mass M is recalculated and compared with the historical predicted value. If the deviation between the adjacent two predicted results exceeds 10% of the allowed mass deviation, the system will immediately start the data traceability process, check the state of the density sensor, verify the continuity of the volume measurement data, and perform error compensation on the real-time flow integral.

[0051] In an embodiment, the control center is also used for pre-processing of real-time data, including outlier rejection, data smoothing, time alignment and format standardization. Since raw industrial data collected directly from sensors is usually noisy, incomplete, out-of-sync or heterogeneous, if these "dirty data" are directly used for calculating deviation, triggering early warning or training AI models, it will lead to false alarms, misoperations, model failures and other serious consequences, therefore, it is necessary to transform the raw data into clean and reliable data for analysis. The above-mentioned pre-processing techniques, such as outlier rejection, can set reasonable physical upper and lower limits through the control center, the system will automatically scan the real-time data stream, any point exceeding these thresholds will be marked as an outlier; data smoothing uses digital filtering algorithms, that is, instead of using a single instantaneous value, an average value of data in a short period of time is calculated as the current valid value; time alignment, the system maintains a high-precision synchronous clock, all data are time-stamped when collected, when processing, the control center will align the data of different sensors to the same time axis according to the time stamp, for sensors with different frequencies, interpolation and other methods may be used to generate data sequences with uniform time intervals; and format standardization is to identify multiple industrial protocols through the built-in protocol parser in the control center, which are all prior art and will not be described here.

[0052] In an embodiment, the production line execution module feeds back the execution result to the control center after executing the control strategy, for iterative optimization of the artificial intelligence model and adaptive adjustment of the process parameters. That is, the system compares the predicted value of the model before with the real value fed back. If there is a persistent error, the system will retrain or fine-tune the AI model using these new experience data, so that the next prediction or decision of the model is more accurate.

[0053] Please refer to Figure 2 The application also provides a control system applying the above control method, which comprises a data collection module 10, a real-time data collection module 20, a data analysis module 30, a control center 40 and a production line execution module 50.

[0054] The application has the beneficial effects that: the application breaks through the limitation of traditional fixed logic control through multi-module cooperation and closed-loop control, fills the parameters with actual parameters through dynamic comparison theory, triggers graded early warning and self-adaptive adjustment according to the deviation range set, effectively deals with the fluctuation of chemical characteristics, greatly improves the filling precision and product qualification rate, reduces the raw material loss, realizes the precise upgrading from experience control to data driving, actively identifies the parameter deviation and equipment abnormality relying on real-time data acquisition and millisecond level response mechanism, quickly triggers early warning and generates control strategy, significantly enhances the response ability to sudden conditions, reduces the unplanned downtime, improves the production line safety control level, through the execution result feedback and model iteration optimization, the system has the ability of continuous self-learning and self-adaptive adjustment of process parameters, significantly improves the flexible adaptation ability of the production line to multi-species production, reduces the cost of manual intervention, and comprehensively improves the production efficiency and automation level.

[0055] The technical features of the above-described embodiments can be combined arbitrarily, and to make the description concise, all possible combinations of the technical features in the above-described embodiments are not described, however, as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present disclosure.

[0056] The above-described embodiments only express several implementation manners of the present application, the description is more specific and detailed, but it should not be understood as the limitation of the scope of the patent. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. An artificial intelligence-based full-process automatic control method for a chemical filling production line, characterized in that, The chemical filling production line full-process automatic control method comprises: S1, a data acquisition module acquires initial data of a chemical to be filled in a production line, uploads the initial data to a control center, and marks the initial data as a production line initial data set; S2, a data analysis module extracts key parameters from the production line initial data set, and calculates a theoretical filling parameter set; S3, a real-time data acquisition module acquires real-time data of a chemical filled in a production line, uploads the real-time data to a control center, and marks the real-time data as a production line real-time data set; S4, the control center processes the real-time data to obtain an actual filling parameter set; S5, the control center calls the theoretical filling parameter set and the actual filling parameter set, constructs a parameter deviation evaluation system, calculates a deviation range set, and determines whether to trigger a hierarchical early warning; S6, the control center generates a control strategy according to the deviation range set, and evaluates the filling state and the quality prediction result; S7, a production line execution module adjusts equipment operating parameters according to the control strategy, executes filling closed-loop control, and feeds back execution results to the control center for model iteration and optimization.

2. The artificial intelligence-based chemical filling production line full-process automation control method of claim 1, wherein, The production line initial data set includes a target filling amount Vt, a current residual amount Vr, a chemical density ρ, a maximum filling rate Qm, and a production allowed deviation ratio a of the production line before filling operation.

3. The artificial intelligence-based chemical filling production line full-process automation control method of claim 2, wherein, The calculation formula of the theoretical filling parameter set comprises: amount to be filled : ; theoretical filling time : ; Target mass : ; allowing quality deviations : .

4. The artificial intelligence-based chemical filling production line full-process automation control method of claim 1, wherein, The actual filling parameter set includes an actual filling amount Vr', an actual filling rate Qr, and a filled mass mr of the production line after filling operation.

5. The artificial intelligence-based chemical filling production line full-process automation control method of claim 4, wherein, The deviation range set comprises: Flow deviation: ; Volume progress deviation: ; Quality deviation: .

6. The artificial intelligence-based chemical filling production line full-process automation control method of claim 5, wherein, The hierarchical early warning comprises: When ΔQ < 10% and ΔVp < 5% and Δmp < 3%, it is determined as a normal state; When ΔQ > 10% or ΔVp > 5% or Δmp > 3%, it is determined as a warning state; When Δmp > 2a or ΔQ > 30%, it is determined as an emergency pause state.

7. The artificial intelligence-based chemical filling production line full-process automation control method of claim 6, wherein, The control strategy comprises valve opening degree adjustment, and an adjustment formula of the valve opening degree adjustment amount ΔK is: 。 8. The artificial intelligence-based chemical filling production line full-process automation control method of claim 2, wherein, The control strategy further comprises quality prediction adjustment, and a quality prediction formula is: ， The quality determination formula is: , if true, then the result is "pass", otherwise "fail".

9. The artificial intelligence-based chemical filling production line full-process automation control method of claim 2, wherein, The control center is further used for preprocessing real-time data, and the preprocessing comprises outlier rejection, data smoothing, time alignment, and format standardization.

10. The artificial intelligence-based chemical filling production line full-process automation control method of claim 8, wherein, After executing the control strategy, the production line execution module feeds back execution results to the control center for iterative optimization of an artificial intelligence model and adaptive adjustment of process parameters.

Images