Intelligent lubricating oil production control method and system

Abstract

The present invention relates to the field of lubricating oil production control technology, and in particular to an intelligent lubricating oil production control method and system. The method comprises the following steps: batch sample collection of lubricating oil production line, and obtaining production process benchmark cleaning parameters by mapping different batches of production process benchmark parameters; then simulating the movement of oil molecular chains based on the parameters, and evaluating the process deviation coefficient to obtain the viscosity process deviation coefficient; then using multi-dimensional parameter decision optimization to generate process optimization data, and performing logic learning on it through random forest algorithm to obtain optimized logic data; finally designing automated firmware according to the optimized logic data, and embedding it into the control center; the present invention makes the lubricating oil production control technology more perfect by optimizing the lubricating oil production control technology.

Description

Technical Field

[0001] The present invention relates to the technical field of lubricating oil production control, and in particular to an intelligent lubricating oil production control method and system. Background Art

[0002] There are many challenges in the previous lubricant production process, including complex production process, large fluctuations in process parameters and high difficulty in controlling product consistency. In the production process, different batches of raw materials and process conditions are prone to introduce small deviations. These deviations accumulate and lead to a decline in lubricant performance and increase the risk of equipment wear. In addition, with the popularization of Industry 4.0 and intelligent manufacturing concepts, the market demand for high-quality and high-performance lubricants continues to increase, while also placing higher requirements on precise control of the production process and resource utilization efficiency. The operation of various mechanical equipment and systems has an increasing demand for lubricants, and the performance, quality and stability of lubricants directly affect the operating efficiency, life and energy consumption of the equipment. Therefore, how to improve production efficiency and reduce production costs while ensuring the quality of lubricants has become an important challenge facing the lubricant manufacturing industry. However, a traditional intelligent lubricant production control method has the problems of inaccurate analysis of lubricant status and large errors in process parameter control. Summary of the invention

[0003] Based on this, it is necessary to provide an intelligent lubricant production control method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, an intelligent lubricant production control method is provided, the method comprising the following steps:

[0005] Step S1: collecting batch samples of lubricating oil on a lubricating oil production line to obtain lubricating oil batch samples; mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain production process benchmark cleaning parameters;

[0006] Step S2: simulating the movement of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters to obtain the movement data of oil molecular chains; evaluating the process deviation coefficient of the oil molecular chain movement data to obtain the viscosity process deviation coefficient;

[0007] Step S3: performing multidimensional parameter decision optimization according to the viscosity process deviation coefficient to obtain process multidimensional parameter optimization data; performing logic learning on the process multidimensional parameter optimization data based on the random forest algorithm to obtain process multidimensional parameter optimization logic data;

[0008] Step S4: Design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to execute intelligent lubricant production management and control.

[0009] In the lubricating oil production line of the present invention, by sampling different batches of lubricating oil, multiple groups of batch sample data can be obtained. These samples can be used to analyze the production process differences between different batches, and then map out a set of standardized production process benchmark parameters. These benchmark parameters include key factors such as temperature, pressure, stirring speed, etc., which can provide standard basis for subsequent production process optimization and cleaning process. This step helps to discover the process differences between different batches and provides necessary data support and reference for subsequent quality control. According to the production process benchmark cleaning parameters, the motion simulation of the oil molecule chain is carried out in a computer simulation environment to study the behavior of the oil molecule under different process conditions. These simulation data help to understand the flow characteristics of the oil molecules and the dynamic changes of the molecular chain, and further reveal the viscosity changes in the production process. By evaluating the process deviation coefficient of the oil molecule chain motion data, a coefficient on viscosity deviation can be obtained, which provides a quantitative basis for subsequent production process optimization. This step can accurately evaluate the impact of process deviation on lubricating oil performance and provide a scientific basis for optimization. After obtaining the viscosity process deviation coefficient, a multidimensional parameter decision optimization method is used for in-depth analysis to identify key parameters that have a greater impact on the production process. These optimization data not only help to improve the existing process, but also provide support for the development of new products. At the same time, the logic learning based on the random forest algorithm will conduct in-depth analysis and pattern recognition on these multidimensional optimization parameters, thereby generating a set of optimized process parameter logic data. The random forest algorithm can effectively process a variety of input variables, improve the accuracy and stability of parameter optimization, and thus make the production process more intelligent and efficient. Based on the process multidimensional parameter optimization logic data obtained in step S3, the design of automated firmware is carried out to realize the intelligent control of the lubricant production line. By embedding these optimized firmware into the intelligent lubricant production line control center, the system can automatically execute various optimization measures in the production process, automatically adjust the production parameters, and ensure that the lubricant products always maintain the best quality and performance during the production process. This step realizes the intelligence and automation of the production process, which not only improves production efficiency, but also reduces human intervention, and further improves the stability and consistency of lubricant production. Therefore, the present invention is an optimization process made to a traditional intelligent lubricant production control method, which solves the problems of inaccurate analysis of lubricant state and large process parameter control errors in a traditional intelligent lubricant production control method, improves the accuracy of lubricant state analysis, and reduces process parameter control errors.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: collecting batch samples of lubricating oil on the lubricating oil production line to obtain lubricating oil batch samples;

[0012] Step S12: mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain the production process benchmark parameters;

[0013] Step S13: performing data cleaning on the production process benchmark parameters to obtain the production process benchmark cleaning parameters.

[0014] The present invention can collect representative data of different batches by collecting lubricant samples for each production batch on the lubricant production line. These samples contain various key properties of the lubricant, such as viscosity, density, chemical composition, etc. Batch sample collection provides necessary raw data support for subsequent production process analysis. Through this process, it can be ensured that different batches of lubricant products can be accurately monitored and evaluated, so as to effectively identify potential problems or inconsistencies in the production process and provide a basis for process optimization. The core of this step is to map the production process benchmark parameters between different batches of the collected lubricant batch samples. By comparing the sample data of different batches with the corresponding production process parameters (such as temperature, pressure, stirring speed, additive ratio, etc.), a set of standardized process benchmark parameters can be established. These parameters can clearly reflect the process characteristics of each batch in the production process, providing a basis for subsequent process optimization. This step helps to eliminate the production process differences between different batches, ensure the uniformity of the production process, and provide a scientific basis for accurately controlling the production process. Data cleaning of the production process benchmark parameters is a key step to ensure data quality and accuracy. Since the collected data contains noise, outliers or incomplete information, the data cleaning process can remove these inaccurate interference information to ensure that the final process benchmark cleaning parameters have high reliability. In this process, the process benchmark parameters are made more accurate and consistent by standardizing, deduplicating, and filling missing values. The cleaned data provides a more reliable foundation for subsequent production process improvement and optimization, ensuring that more accurate and operational parameters can be obtained when making process adjustments.

[0015] Preferably, step S2 comprises the following steps:

[0016] Step S21: extracting the density difference of lubricating oil between different batches of lubricating oil batch samples to obtain lubricating oil density difference data;

[0017] Step S22: performing oil molecule chain motion simulation between different production process benchmarks on the lubricating oil density difference data according to the production process benchmark cleaning parameters to obtain oil molecule chain motion data;

[0018] Step S23: performing dynamic viscosity correlation linear analysis on the oil molecular chain motion data to obtain dynamic viscosity correlation linear data;

[0019] Step S24: Evaluate the process deviation coefficient of the production process benchmark cleaning parameters based on the dynamic viscosity correlation linear data to obtain the viscosity process deviation coefficient.

[0020] The present invention extracts density differences from batch samples of lubricating oil, and can effectively analyze the differences in physical properties between different batches of lubricating oil. The density of lubricating oil is an important indicator for measuring its molecular structure and composition, and the density difference between different batches indicates the difference in production process or raw materials. By extracting the density difference data of lubricating oil, an important reference can be provided for the subsequent optimization of production process. This step helps to identify potential problems between batches of lubricating oil, ensure the consistency and quality stability of products, and lay the foundation for further optimization of production processes. The motion simulation of oil molecular chains is performed on the lubricating oil density difference data according to the production process benchmark cleaning parameters, which helps to deeply understand the behavioral characteristics of oil molecules under different process parameters. The motion simulation of oil molecular chains reveals the changes in the fluidity, molecular chain stretching, and interaction force of lubricating oil molecules in different production process environments. Through this process, the performance of lubricating oil, such as fluidity and stability, can be predicted and controlled, thereby providing a scientific basis for process adjustment. This step can accurately simulate the molecular level behavior of lubricating oil, which helps to fundamentally improve the production process and improve product quality. Dynamic viscosity correlation linear analysis is performed on the oil molecular chain motion data to quantify the relationship between oil molecular chain motion and viscosity. This analysis can reveal the intrinsic relationship between the motion characteristics of oil molecules and the change in lubricant viscosity, thereby providing a clear mathematical model for viscosity control. Through this process, the viscosity change trend of lubricants under different process conditions can be more accurately understood and predicted, and the viscosity control in the production process can be optimized. The results of this analysis provide a direct basis for the optimization of the production process, helping to improve the performance of lubricants and ensure that they meet quality standards. Based on the dynamic viscosity correlation linear data, the process deviation coefficient of the production process benchmark cleaning parameters is evaluated to quantify the deviation between the actual production process and the standard process. This evaluation compares the dynamic viscosity change with the process benchmark to obtain the viscosity process deviation coefficient, which reflects the degree of deviation in the production process. By analyzing the deviation coefficient, the production parameters that need to be adjusted can be determined, thereby optimizing the production process and reducing quality fluctuations. This step is of great significance for accurately controlling the production process and improving product consistency, and provides data support for process improvement.

[0021] Preferably, step S22 includes the following steps:

[0022] Step S221: extracting the production temperature and production pressure of the production process benchmark cleaning parameters to obtain process production temperature data and process production pressure data;

[0023] Step S222: performing oil molecule nonlinear gravity analysis between different production process benchmarks on the lubricating oil density difference data according to the process production temperature data and the process production pressure data to obtain oil molecule nonlinear gravity data;

[0024] Step S223: performing discrete evaluation of intermolecular forces on the oil molecule nonlinear gravity data to obtain discrete intermolecular forces data;

[0025] Step S224: performing molecular chain morphology tensor analysis according to the intermolecular force discrete data to obtain molecular chain morphology tensor data;

[0026] Step S225: performing oil molecular chain motion simulation between different production process benchmarks according to the molecular chain morphology tensor data and the intermolecular force discrete data to obtain oil molecular chain motion data.

[0027] The present invention can provide key environmental parameters for the performance analysis of lubricating oil by extracting the production temperature and production pressure data in the production process benchmark cleaning parameters. These parameters directly affect the fluidity of oil molecules, the stability of molecular structure and the overall viscosity of oil in the production process of lubricating oil. The extracted temperature and pressure data provide an accurate process background for the subsequent molecular chain motion simulation, ensuring that the simulation results can reflect the impact of the actual production process. Therefore, this step is crucial to understanding the impact of temperature and pressure on the properties of lubricating oil during the production process, and can provide a reliable basis for subsequent optimization. According to the process production temperature data and production pressure data, the density difference data of the lubricating oil is subjected to nonlinear gravity analysis of oil molecules, which helps to reveal the changes in the interaction force between lubricating oil molecules under different production conditions. The nonlinear gravity between oil molecules is an important factor in determining the viscosity and fluidity of lubricating oil. Through this analysis, the behavior of the molecular chain of lubricating oil under different process conditions can be more accurately understood, and the rheological properties of lubricating oil under different production environments can be predicted. This step lays the foundation for further simulation of oil molecule behavior, so that the physical properties of lubricating oil can be accurately controlled under diverse process conditions. The discrete evaluation of intermolecular forces on the nonlinear gravitational data of oil molecules aims to quantify the distribution of forces between lubricant molecules and their discreteness. This process can reveal how the interaction of oil molecular chains affects the fluidity and viscosity of lubricants under different process conditions. The evaluation of the discrete degree of intermolecular forces is crucial to understanding the microscopic behavior of lubricants, and can help optimize the formulation adjustment or process parameters in the production process to improve the performance of lubricants. This step provides a specific quantitative basis for further analysis of molecular behavior and process optimization, which helps to improve the overall quality of lubricants. By performing molecular chain morphology tensor analysis on the discrete data of intermolecular forces, we can further understand the structural characteristics of oil molecular chains and their morphological changes under different process conditions. The morphology tensor analysis of molecular chains can reveal the stretching, bending of oil molecular chains and their interactions with other molecular chains, thereby predicting the flow properties and durability of lubricants. This analysis provides detailed information on the microstructural changes of lubricant molecular chains, which helps to optimize the molecular structure of lubricants to improve their performance under extreme conditions such as high temperature and high pressure. This step provides an important theoretical basis and technical support for improving the performance of lubricant products. Based on the molecular chain morphology tensor data and the discrete data of intermolecular forces, the motion simulation of oil molecular chains between different production process benchmarks can deeply understand the dynamic behavior of lubricants under the production process. Through this simulation, the motion trajectory, interaction and influence of oil molecular chains on lubricant fluidity, viscosity and other properties under different process conditions can be predicted. This process helps to identify the good and bad performance of lubricants under specific process conditions, provide a quantitative basis for the adjustment of production processes and the optimization of lubricant products, and ensure the efficient performance of lubricant products under different use environments.This step is a key link in lubricant process optimization and provides important data support for subsequent quality control and process improvement.

[0028] Preferably, step S224 includes the following steps:

[0029] The moment distribution matrix is ​​decomposed on the discrete data of intermolecular forces to obtain the moment distribution characteristic matrix;

[0030] According to the moment distribution characteristic matrix data, the discrete data of intermolecular forces are subjected to structure layer-by-layer tensor decomposition to obtain the molecular local tensor data;

[0031] Performing tensor field gradient clustering processing on the local tensor data of the molecule to obtain tensor field gradient clustering data;

[0032] The molecular chain morphology tensor data is analyzed based on the tensor field gradient clustering data to obtain the molecular chain morphology tensor data.

[0033] The present invention performs moment distribution matrix decomposition processing on the discrete data of intermolecular forces, and can extract the overall distribution characteristics of intermolecular forces. Through matrix decomposition, complex force data can be decomposed into an easy-to-understand matrix form, thereby revealing the distribution of intermolecular interaction forces in different directions and positions. This process helps to identify the mechanical behavior of lubricating oil molecules under different conditions, including the degree of fit between molecular chains, the strength of interaction, etc., and provides a clearer physical background and basic data for subsequent molecular chain morphological analysis. According to the moment distribution characteristic matrix data, the discrete data of intermolecular forces are subjected to structural layer-by-layer tensor decomposition processing, aiming to analyze the specific structure of intermolecular forces from multiple levels and dimensions. Layer-by-layer tensor decomposition helps to decompose the multiple structural information in the force data, revealing the multi-level relationship of the molecular structure of lubricating oil and its influencing factors. This process helps to represent complex molecular mechanical properties in the form of tensors, can deeply understand the molecular structural characteristics at different levels, such as the rigidity and flexibility of molecular chains, and provide more refined local information for subsequent analysis. Tensor field gradient clustering processing is performed on the local molecular tensor data, which can reveal the local tension and morphological changes of the lubricating oil molecular chain. During the movement of molecular chains, molecules in different regions are subject to different degrees of tension and shear force. Tensor field gradient clustering helps to divide molecular regions into different groups according to these gradient changes, so as to better understand the local changes in intermolecular forces. This method can effectively identify the molecular behavior patterns in different regions, thereby providing a more scientific basis for optimizing the fluidity, stability and other properties of lubricants. Based on the tensor field gradient clustering data, the molecular chain morphology tensor analysis of the molecular local tensor data can more accurately describe the geometric morphology of the lubricant molecular chain and its changes under different external conditions. Through this analysis, the morphological characteristics of the molecular chain can be quantified, such as the stretching degree and curvature of the molecular chain and its interaction with other molecular chains. Morphological tensor analysis not only helps to understand the behavior of molecular chains under complex process conditions, but also predicts the rheological properties and durability of lubricants, thereby providing a quantitative reference for the process optimization of lubricants. This analysis result helps to achieve efficient lubricant design and improve product stability and performance.

[0034] Preferably, step S24 comprises the following steps:

[0035] Step S241: performing grid segment density change analysis on the dynamic viscosity associated linear data to obtain segment viscosity density change data;

[0036] Step S242: performing segment-by-segment nonlinear deviation interference calculation on the production process benchmark cleaning parameters according to the segment viscosity density change data to obtain benchmark deviation interference data;

[0037] Step S243: performing lubricating oil flash point continuity simulation according to the reference deviation interference data and the segmented viscosity density change data to obtain lubricating oil flash point continuity simulation data;

[0038] Step S244: Based on the continuous simulation data of the lubricating oil flash point and the reference deviation interference data, the process deviation coefficient of the segmented viscosity density change data is evaluated to obtain the viscosity process deviation coefficient.

[0039] The present invention performs grid segment density change analysis on dynamic viscosity associated linear data, the purpose of which is to divide the viscosity data into different grid areas and analyze the viscosity change in each segment area. This process can help identify the viscosity change trend of lubricating oil under different conditions, especially the sensitivity when process parameters such as temperature and pressure change. Through this analysis, the viscosity characteristics of lubricating oil can be more accurately grasped, basic data can be provided for subsequent process adjustments, and the formulation of lubricating oil and its production process can be optimized to ensure the flow performance and stability of the product in practical applications. According to the segment viscosity density change data, the nonlinear deviation interference calculation between segments is performed on the production process benchmark cleaning parameters, the purpose of which is to analyze the nonlinear deviation between different segments and evaluate its influence on the lubricating oil production process. Through this calculation, the deviation between each segment and their mutual interference effect under different process conditions can be revealed. This provides key data support for subsequent lubricating oil performance prediction and quality control. Nonlinear deviation interference calculation can identify potential instability or abnormality in the production process, providing a scientific basis for optimizing the production process and improving product consistency. Based on the reference deviation interference data and the segmented viscosity density change data, the flash point persistence simulation of lubricating oil is carried out to evaluate the flash point change of lubricating oil under different process conditions. Flash point is an important performance indicator of lubricating oil, reflecting its ability to withstand high temperatures. Through this simulation, the flash point persistence of lubricating oil under different production process conditions can be predicted to help determine whether the oil can maintain stable performance for a long time under high temperature environment. Flash point persistence simulation provides a scientific evaluation of the high temperature stability of lubricating oil, and can provide data support for process optimization, product improvement and quality control to ensure the safety and reliability of the final product. Based on the flash point persistence simulation data of lubricating oil and the reference deviation interference data, the process deviation coefficient of the segmented viscosity density change data is evaluated to quantify the impact of process parameter deviations on the viscosity and flash point performance of lubricating oil during production. By evaluating the process deviation coefficient, it can be determined which parameter deviations in the production process will have a significant impact on the performance of the final product. This evaluation helps to optimize the production process, reduce the negative impact of deviations on product performance, and ensure the consistency and stability of lubricating oil between different production batches. The process deviation coefficient evaluation provides systematic feedback and optimization solutions, providing an important reference for quality control in the production process.

[0040] Preferably, step S243 includes the following steps:

[0041] Perform dynamic feature vector extraction on the reference deviation interference data to obtain the reference interference feature vector;

[0042] According to the reference interference characteristic vector, the segmented viscosity density change data is subjected to layered viscosity correlation flash point fitting calculation to obtain layered flash point fitting data;

[0043] Performing dynamic continuity curve fitting processing on the layered flash point fitting data to obtain a dynamic flash point continuity curve;

[0044] The nonlinear flash point deduction is performed according to the dynamic flash point continuity curve and the segmented viscosity density change data to obtain the dynamic flash point nonlinear deduction data;

[0045] The lubricant oil flash point continuity simulation is carried out based on the dynamic flash point nonlinear derivation data to obtain the lubricant oil flash point continuity simulation data.

[0046] The present invention extracts dynamic feature vectors from the reference deviation interference data, with the purpose of extracting vector information that can reflect the data change trend and characteristics from the original data. The dynamic feature vector can capture the important patterns and change characteristics in the reference deviation interference data, and then provide key mathematical descriptions for subsequent analysis. Through this process, it is possible to make quantitative expressions of subtle changes in the production process, providing an accurate basis for process optimization, performance prediction and quality control. In addition, this feature vector can also effectively reduce the complexity of the data and enhance the efficiency and accuracy of subsequent processing. According to the reference interference feature vector, the segmented viscosity density change data is subjected to layered viscosity correlation flash point fitting calculation, aiming to reveal the relationship between the viscosity and flash point of the lubricating oil, and fitting is performed according to different levels (such as temperature, pressure and other factors). This step helps to systematically understand the correlation between the segmented viscosity and the flash point, and provides an important prediction tool for the quality control of the lubricating oil. Through layered fitting, the influence of viscosity changes on the flash point under different conditions can be accurately captured, providing detailed data support for subsequent process optimization and performance adjustment. The layered flash point fitting data is subjected to dynamic continuity curve fitting processing, with the purpose of constructing a continuity curve to represent the change trend of the flash point of the lubricating oil under dynamic conditions. This step can effectively describe the changing rules and stability of the flash point of lubricants under different operating conditions during production and use. Through dynamic continuity curve fitting, the smooth transition of the flash point of lubricants with changes in process conditions such as temperature and pressure can be revealed, which helps engineers predict the performance of lubricants under extreme conditions and provides a basis for optimizing design and process adjustment. Nonlinear flash point derivation is performed based on the dynamic flash point continuity curve and segmented viscosity density change data, aiming to establish a nonlinear relationship between the flash point of lubricants and process parameters. Through nonlinear derivation, it is possible to deeply analyze how the flash point is jointly affected by complex factors in different process steps. This derivation helps to reveal complex nonlinear behaviors that traditional linear models cannot accurately capture, and provides theoretical support for the precise design and customization of lubricants. In addition, nonlinear derivation can better adapt to changing production conditions and provide reliable guarantees for the stability and safety of lubricants in practical applications. The flash point continuity simulation of lubricants is based on the dynamic flash point nonlinear derivation data, with the aim of simulating the flash point change trend of lubricants after long-term exposure to different conditions. This step can predict the flash point performance of lubricants under long-term high temperature, high pressure or extreme working conditions by inputting the previously derived data into the simulation model. Flash point is an important safety indicator of lubricants. Simulating its persistence helps to identify potential safety hazards in advance and optimize the formulation and production process of lubricants. This process is crucial to ensure the high temperature stability and long-term safety of lubricants, and provides a scientific basis for the final quality control of lubricants.

[0047] Preferably, step S3 comprises the following steps:

[0048] Step S31: normalizing the viscosity process deviation coefficient to obtain viscosity process deviation normalized data;

[0049] Step S32: performing multi-dimensional parameter decision optimization according to the viscosity process deviation normalization data to obtain process multi-dimensional parameter optimization data;

[0050] Step S33: Performing logic learning on the process multi-dimensional parameter optimization data based on the random forest algorithm to obtain the process multi-dimensional parameter optimization logic data.

[0051] The present invention normalizes the viscosity process deviation coefficient, with the purpose of converting the deviation coefficients of different ranges and magnitudes into a unified standardized range (such as [0, 1]). This step helps to eliminate the scale differences between different process parameters, so that subsequent data analysis and optimization can be carried out under equal weights. The normalized data can improve the calculation efficiency and ensure that the importance of each parameter in the model is treated fairly. This step lays a solid foundation for the subsequent multidimensional parameter decision optimization and logic learning, so that each deviation coefficient can be comprehensively analyzed and optimized on the same scale. The multidimensional parameter decision optimization is carried out according to the viscosity process deviation normalized data, with the purpose of finding the optimal process parameter combination by comprehensively considering the relationship between multiple process parameters. The multidimensional parameter decision optimization can help determine the optimal values ​​of each parameter under different production processes, so as to achieve the purpose of improving the performance of lubricating oil, reducing production deviations and optimizing product quality. Through this optimization process, the normalized data can be fully utilized, the ratio of different process parameters can be comprehensively analyzed and adjusted, and each production link can be ensured to be in the best state, thereby improving the overall production efficiency and product consistency. The purpose of logically learning the multi-dimensional process parameter optimization data based on the random forest algorithm is to establish a nonlinear relationship model between complex process parameters and final product performance through machine learning algorithms. Random forest is a powerful ensemble learning method that can effectively process high-dimensional data and automatically identify the relationship and importance between parameters. Through this algorithm, key optimization rules and patterns can be extracted from a large amount of process optimization data, thereby providing accurate prediction support for decision-making in the production process. This step can greatly improve the intelligence level of process optimization, reduce human intervention, and improve the stability of the production process and the quality of the final product.

[0052] Preferably, step S32 includes the following steps:

[0053] Step S321: performing multi-scale decomposition processing on the viscosity process deviation normalization data to obtain viscosity multi-scale decomposition parameters;

[0054] Step S322: performing nonlinear constraint processing on the viscosity multi-scale decomposition parameters to obtain viscosity multi-scale decomposition constraint data;

[0055] Step S323: performing multi-objective decision-making and solving processing on the viscosity process deviation normalization data according to the viscosity multi-scale decomposition constraint data to obtain the process multi-objective solution parameters;

[0056] Step S324: performing multi-dimensional parameter decision optimization on the viscosity process deviation normalization data according to the process multi-objective solution parameters to obtain process multi-dimensional parameter optimization data.

[0057] The present invention performs multi-scale decomposition processing on the viscosity process deviation normalization data, aiming to decompose complex process data into components of multiple different scales. Multi-scale decomposition helps to reveal the potential patterns and details of viscosity changes at different time scales or spatial scales. This step can effectively identify the regularity of different levels in the data, including short-term fluctuations and long-term trends, thereby improving the understanding and prediction capabilities of viscosity process changes. Through multi-scale decomposition, the data can be carefully analyzed at different scales to provide more accurate and diverse information support for subsequent optimization processing. Nonlinear constraint processing is performed on the viscosity multi-scale decomposition parameters, the purpose of which is to constrain the decomposed parameters through a nonlinear model to ensure the rationality and physical or process consistency between the parameters. Viscosity processes are often affected by a variety of nonlinear factors, such as environmental variables such as temperature and pressure, so the complex effects of these factors on process parameters can be accurately captured through nonlinear constraints. This step can effectively prevent unreasonable or excessive parameter changes, ensure that the final process optimization results meet the actual operation requirements, and thus improve the stability and reliability of the production process. According to the viscosity multi-scale decomposition constraint data, the viscosity process deviation normalization data is processed by multi-objective decision-making, with the aim of finding the best process parameter combination under multiple optimization objectives. In the lubricant production process, multiple objectives need to be optimized simultaneously, such as efficiency and product quality. Through multi-objective decision-making, the optimal solution can be obtained under the premise of ensuring the balance between various objectives. This step ensures that the process optimization not only meets the needs of a single objective, but also balances between multi-dimensional objectives through a comprehensive analysis of the multi-scale constraint data, thereby improving the comprehensiveness and practicality of the process decision. According to the process multi-objective solution parameters, the viscosity process deviation normalization data is optimized by multi-dimensional parameter decision, with the aim of further optimizing the multiple parameter combinations in the viscosity process based on the multi-objective solution results obtained previously. This step uses a multi-dimensional optimization algorithm to comprehensively consider the interaction of various process parameters and their impact on the final product, ensuring that a comprehensive optimal process solution is obtained. Through multi-dimensional optimization, while ensuring product quality and reducing costs, production efficiency can be improved, so that the lubricant production process reaches the optimal state. Finally, the process multi-dimensional parameter optimization data provides a scientific decision-making basis for actual production and effectively improves the stability and controllability of the production process.

[0058] Preferably, the present invention further provides an intelligent lubricant production control system for executing the intelligent lubricant production control method as described above, the intelligent lubricant production control system comprising:

[0059] The process benchmark parameter mapping module is used to collect batch samples of lubricating oil on the lubricating oil production line to obtain lubricating oil batch samples; map the production process benchmark parameters between different batches of lubricating oil batch samples to obtain production process benchmark cleaning parameters;

[0060] The process deviation coefficient evaluation module is used to simulate the movement of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters to obtain the oil molecular chain movement data; the process deviation coefficient of the oil molecular chain movement data is evaluated to obtain the viscosity process deviation coefficient;

[0061] The logic learning adjustment module is used to optimize the multi-dimensional parameter decision according to the viscosity process deviation coefficient to obtain the process multi-dimensional parameter optimization data; the logic learning of the process multi-dimensional parameter optimization data is performed based on the random forest algorithm to obtain the process multi-dimensional parameter optimization logic data;

[0062] The automation firmware design module is used to design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain the process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to perform intelligent lubricant production management and control.

[0063] The beneficial effect of the present invention is that in the lubricating oil production line, by sampling different batches of lubricating oil, multiple groups of batch sample data can be obtained. These samples can be used to analyze the production process differences between different batches, and then map out a set of standardized production process benchmark parameters. These benchmark parameters include key factors such as temperature, pressure, stirring speed, etc., which can provide standard basis for subsequent production process optimization and cleaning process. This step helps to discover the process differences between different batches and provides necessary data support and reference for subsequent quality control. According to the production process benchmark cleaning parameters, the motion simulation of the oil molecule chain is carried out in a computer simulation environment to study the behavior of the oil molecule under different process conditions. These simulation data help to understand the flow characteristics of the oil molecules and the dynamic changes of the molecular chain, and further reveal the viscosity changes in the production process. By evaluating the process deviation coefficient of the oil molecule chain motion data, a coefficient on viscosity deviation can be obtained, which provides a quantitative basis for subsequent production process optimization. This step can accurately evaluate the impact of process deviation on lubricating oil performance and provide a scientific basis for optimization. After obtaining the viscosity process deviation coefficient, a multidimensional parameter decision optimization method is used for in-depth analysis to identify key parameters that have a greater impact on the production process. These optimization data not only help to improve existing processes, but also provide support for the development of new products. At the same time, the logic learning based on the random forest algorithm will conduct in-depth analysis and pattern recognition on these multidimensional optimization parameters, thereby generating a set of optimized process parameter logic data. The random forest algorithm can effectively process a variety of input variables, improve the accuracy and stability of parameter optimization, and thus make the production process more intelligent and efficient. Based on the process multidimensional parameter optimization logic data obtained in step S3, the design of automated firmware is carried out to realize the intelligent control of the lubricant production line. By embedding these optimized firmware into the intelligent lubricant production line control center, the system can automatically execute various optimization measures in the production process, automatically adjust the production parameters, and ensure that the lubricant products always maintain the best quality and performance during the production process. This step realizes the intelligence and automation of the production process, which not only improves production efficiency, but also reduces human intervention, and further improves the stability and consistency of lubricant production. Therefore, the present invention is an optimization process made to a traditional intelligent lubricant production control method, which solves the problems of inaccurate analysis of lubricant state and large process parameter control errors in a traditional intelligent lubricant production control method, improves the accuracy of lubricant state analysis, and reduces process parameter control errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A schematic diagram of the steps of an intelligent lubricant production control method;

[0065] Figure 2 for Figure 1Detailed implementation steps of step S2 in FIG.

[0066] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0067] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0068] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0069] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0070] To achieve this, please refer to Figure 1 to Figure 2 , an intelligent lubricant production control method, the method comprising the following steps:

[0071] Step S1: collecting batch samples of lubricating oil on a lubricating oil production line to obtain lubricating oil batch samples; mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain production process benchmark cleaning parameters;

[0072] Step S2: simulating the movement of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters to obtain the movement data of oil molecular chains; evaluating the process deviation coefficient of the oil molecular chain movement data to obtain the viscosity process deviation coefficient;

[0073] Step S3: performing multidimensional parameter decision optimization according to the viscosity process deviation coefficient to obtain process multidimensional parameter optimization data; performing logic learning on the process multidimensional parameter optimization data based on the random forest algorithm to obtain process multidimensional parameter optimization logic data;

[0074] Step S4: Design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to execute intelligent lubricant production management and control.

[0075] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a schematic diagram of a process flow of an intelligent lubricant production control method of the present invention. In this example, the intelligent lubricant production control method includes the following steps:

[0076] Step S1: collecting batch samples of lubricating oil on a lubricating oil production line to obtain lubricating oil batch samples; mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain production process benchmark cleaning parameters;

[0077] In an embodiment of the present invention, when batch samples of lubricating oil on a lubricating oil production line are collected, samples of lubricating oil are accurately extracted at key links of the production line through an online sampling device. The sampling process uses a sampling needle with a fixed capacity to perform equal time interval sampling within a period of stable flow to ensure the uniformity and representativeness of the sample. After the sample is collected, the impurities are filtered through a separation device to ensure the purity of the sample. When mapping the production process benchmark parameters between different batches of lubricating oil batch samples, the density, viscosity and spectral characteristics of the sample are first measured by a densitometer, a viscometer and an infrared spectrometer. After these data are grouped, a feature mapping matrix for each batch is established by a hierarchical clustering method. Next, a parameter dimensionality reduction method based on principal component analysis (PCA) is used to project the high-dimensional feature space into a low-dimensional space to enhance computational efficiency. The data cleaning step uses a box plot method to remove outliers, and then an interpolation algorithm based on smooth curve fitting is used to supplement the missing values ​​to ensure the integrity of the production process benchmark parameter data, and finally obtain the production process benchmark cleaning parameters.

[0078] Step S2: simulating the movement of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters to obtain the movement data of oil molecular chains; evaluating the process deviation coefficient of the oil molecular chain movement data to obtain the viscosity process deviation coefficient;

[0079] In an embodiment of the present invention, when simulating the motion of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters, the molecular dynamics (MD) method is used. First, the cleaning parameters are converted into initial conditions, including temperature, pressure and molecular arrangement parameters, which are used as input conditions to define the molecular simulation field. Subsequently, the Newton equations of molecular motion are run in the specified field, and the Verlet integration method is used to calculate the molecular position and velocity of each time step to generate the motion trajectory data of the oil molecular chain. When the process deviation coefficient of the oil molecular chain motion data is evaluated, the diffusion characteristics of the molecular chain motion are measured by calculating the mean square displacement (MSD) between each molecular chain. The standard deviation calculated using the MSD of the benchmark group data is used as the deviation benchmark, and the process deviation coefficient is estimated in combination with dynamic linear regression analysis.

[0080] Step S3: performing multidimensional parameter decision optimization according to the viscosity process deviation coefficient to obtain process multidimensional parameter optimization data; performing logic learning on the process multidimensional parameter optimization data based on the random forest algorithm to obtain process multidimensional parameter optimization logic data;

[0081] In the embodiment of the present invention, when multidimensional parameter decision optimization is performed according to the viscosity process deviation coefficient, the Lagrange multiplier method is used to optimize the objective function. The optimization objective function is set as the weighted sum of various parameters (such as temperature, pressure, and shear rate) in the production process, and the process deviation coefficient is added as a constraint condition. The optimization process is divided into the following steps: first, the sensitivity index of the deviation coefficient to each parameter is calculated; second, the weight of each parameter is updated according to the gradient descent algorithm; finally, iterative optimization is performed under the premise of satisfying the constraints until the objective function converges to obtain the process multidimensional parameter optimization data. When the process multidimensional parameter optimization data is logically learned based on the random forest algorithm, the optimization data is divided into a training set and a validation set. In the training stage, multiple decision trees are constructed, each tree is based on different feature subsets and sample subsets, and the nodes are split by the information gain maximization criterion; in the prediction stage, the optimization logic rules are generated by combining the voting results of all trees. Through repeated cross-validation, the accuracy and robustness of the model are evaluated, and the process multidimensional parameter optimization logic data is finally output.

[0082] Step S4: Design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to execute intelligent lubricant production management and control.

[0083] In an embodiment of the present invention, when the automated firmware design is performed based on the process multi-dimensional parameter optimization logic data, the logic data is first converted into an industrial control instruction sequence, and a finite state machine (FSM) is used to model the state transition process of the production line equipment. Then, the optimization rules are implemented as a programmable logic hardware design through the hardware description language (HDL). After the hardware design is completed, the firmware is functionally verified using a verification tool. The verification process includes logic simulation, delay analysis, and power consumption estimation to ensure that the firmware design is correct. Subsequently, the process optimization automation firmware is embedded in the intelligent lubricant production line control center. The firmware communicates with the interface protocol of each production equipment through the control center, collects and controls the key parameters in the production process in real time, and realizes the efficient execution of intelligent lubricant production control.

[0084] Preferably, step S1 comprises the following steps:

[0085] Step S11: collecting batch samples of lubricating oil on the lubricating oil production line to obtain lubricating oil batch samples;

[0086] Step S12: mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain the production process benchmark parameters;

[0087] Step S13: performing data cleaning on the production process benchmark parameters to obtain the production process benchmark cleaning parameters.

[0088] In the embodiment of the present invention, when batch samples of lubricating oil on the lubricating oil production line are collected, an online sampling device is first installed at the key process nodes of the production line. The sampling device is made of high temperature and corrosion resistant materials and is equipped with a flow control module for stabilizing the sampling flow. The sampling operation uses a time-division control logic to divide the sampling time period into equal intervals to avoid data distortion caused by short-term fluctuations during the sampling process. A fixed volume of lubricating oil is accurately extracted into a closed sampling bottle through a sampling needle, and a constant pressure is maintained in the device during the sampling process to prevent bubbles from mixing into the lubricating oil sample. After sampling, the sample is transmitted to an online filtering device, and a high-precision filter membrane with a pore size of 5 microns is used to remove the trace particulate impurities present to ensure the purity of the sample. The sample is then stored in a low-temperature storage device through an automated transmission system, and the temperature is set to 4 degrees Celsius to suppress changes in the internal components of the sample. During the entire operation, the sampling temperature, flow rate and time parameters are monitored in real time to ensure the consistency of the sampling process for each batch. The sample number generates a unique identification code based on the production time, equipment number and batch information, which serves as the basis for subsequent data processing and analysis. When mapping the production process benchmark parameters between different batches of lubricant batch samples, the basic performance parameters of the samples are first determined by physical and chemical methods. The density of the lubricant samples is measured by an online densitometer, and the temperature compensation value is recorded. The viscosity is measured by a viscometer at a constant shear rate, and the shear rate is controlled within the reciprocal of 100 seconds. The measurement results are output in units of dynamic viscosity. The molecular characteristics of the samples are scanned by a Fourier transform infrared spectrometer to obtain the spectrum of the lubricant and record the characteristic peak intensity. After the measurement is completed, the physical and chemical parameters and spectral characteristics of each batch are combined into a feature matrix. In order to unify the data scale between different batches, the normalization method is used to standardize each characteristic parameter and adjust it to the range of 0 to 1. By using a similarity analysis algorithm based on Euclidean distance, the characteristic matrix of each batch is clustered and the data is grouped into several benchmark categories with similar production characteristics. The central eigenvector is calculated in each benchmark category to generate a production process benchmark parameter table. The mapping relationship is bound by the benchmark category identifier and the sample number to calibrate the production process attribution of each batch of samples. When performing data cleaning on the production process benchmark parameters, the initial data is first checked for integrity to identify records with missing values ​​or outliers. For missing values, a linear interpolation method is used to calculate the estimated results of the missing values ​​based on adjacent data points. For outliers, the box plot method is used to detect outliers that are beyond 1.5 times the upper and lower interquartile range, and these outliers are corrected using an alternative method based on quantile regression. After cleaning, the sliding average method is used to smooth the time series data of the production process benchmark parameters to eliminate the interference of short-term random fluctuations. The smoothed data is trended again to ensure that the changes in the parameters conform to physical laws, such as changes in density and viscosity should be linearly correlated within the temperature control range.Finally, the data cleaning results are verified by the control chart method, and the effective range of the production process benchmark parameters is calibrated using the data control upper and lower limits. After the cleaning is completed, the corrected and verified parameters are stored as the production process benchmark cleaning parameters, providing a reliable data input basis for subsequent analysis.

[0089] Preferably, step S2 comprises the following steps:

[0090] Step S21: extracting the density difference of lubricating oil between different batches of lubricating oil batch samples to obtain lubricating oil density difference data;

[0091] Step S22: performing oil molecule chain motion simulation between different production process benchmarks on the lubricating oil density difference data according to the production process benchmark cleaning parameters to obtain oil molecule chain motion data;

[0092] Step S23: performing dynamic viscosity correlation linear analysis on the oil molecular chain motion data to obtain dynamic viscosity correlation linear data;

[0093] Step S24: Evaluate the process deviation coefficient of the production process benchmark cleaning parameters based on the dynamic viscosity correlation linear data to obtain the viscosity process deviation coefficient.

[0094] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0095] Step S21: extracting the density difference of lubricating oil between different batches of lubricating oil batch samples to obtain lubricating oil density difference data;

[0096] In the embodiment of the present invention, when extracting the density difference of lubricating oil between different batches of lubricating oil batch samples, first use a high-precision online density meter to measure the density of each lubricating oil sample, the measurement temperature is set to 25 degrees Celsius, and the density unit is grams per cubic centimeter. In order to avoid measurement errors, each sample is measured three times, and the average value is taken as the final density data. After the measurement is completed, the obtained density data is marked according to the batch to form an initial data table containing batch information and density values. The initial data table is standardized, and the data is adjusted to a unified dimension by subtracting the average value from all density data and dividing it by the standard deviation. Then, the density difference between any two batches of samples is calculated by a calculation method based on a difference matrix, and the specific formula is: the density difference between the two batches of samples = the density value of sample A-the density value of sample B|. The difference values ​​of all batch combinations are stored in a matrix form, and the rows and columns of the matrix represent different batches of samples, and the element values ​​are the corresponding density difference data. Finally, the difference matrix is ​​saved as lubricating oil density difference data, providing a basis for subsequent analysis.

[0097] Step S22: performing oil molecule chain motion simulation between different production process benchmarks on the lubricating oil density difference data according to the production process benchmark cleaning parameters to obtain oil molecule chain motion data;

[0098] In the embodiment of the present invention, when simulating the motion of the oil molecule chain for the lubricating oil density difference data according to the production process benchmark cleaning parameters, it is first necessary to associate and match the lubricating oil density difference data with the production process benchmark cleaning parameters. A hash mapping algorithm is used in the matching process, and the corresponding cleaning parameters are quickly located by using the sample batch number as the key value. After the matching is completed, the molecular chain motion of the lubricating oil sample is modeled using the molecular dynamics simulation method. During the simulation process, a molecular cluster structure representing the lubricating oil molecular chain is selected, and the initial conditions of the molecular model are set based on the proportion of the main components of the lubricating oil. By incorporating the density difference data, the tightness and distribution of the molecular chain are adjusted; the cleaning parameters are used as input boundary conditions to control the temperature and pressure environment of the molecular chain motion. Subsequently, the Brownian dynamics equation is used to describe the random motion trajectory of the molecular chain, and the iterative method is used to calculate the displacement and velocity of the molecular chain during the motion. The motion simulation results are output as the dynamic position data of the molecular chain and the associated molecular potential energy value. After data cleaning and normalization, the oil molecular chain motion data is formed for subsequent viscosity correlation analysis.

[0099] Step S23: performing dynamic viscosity correlation linear analysis on the oil molecular chain motion data to obtain dynamic viscosity correlation linear data;

[0100] In the embodiment of the present invention, when the dynamic viscosity correlation linear analysis is performed on the oil molecular chain motion data, the key variables such as the average motion speed, acceleration and distribution density of the molecular chain are first extracted from the motion data. Then these variables are correlated with the dynamic viscosity data measured in the laboratory, and a linear relationship model is established by regression analysis. In the specific operation, the regression coefficient of the linear equation is calculated using the least squares method, and the explanatory power of the model is verified by the goodness of fit (R square). In order to further improve the analysis accuracy, the data is stratified by temperature, and the molecular chain motion variables and dynamic viscosity under different temperature conditions are analyzed respectively to ensure the robustness of the linear relationship. In the regression analysis, close attention is paid to the multicollinearity problem between the independent variables, and the variance inflation factor (VIF) is used for diagnosis, and the variables with higher collinearity are eliminated to optimize the linear model. The linear data finally output includes the regression coefficient, significance level and residual analysis results under each set of temperature conditions, forming dynamic viscosity correlation linear data.

[0101] Step S24: Evaluate the process deviation coefficient of the production process benchmark cleaning parameters based on the dynamic viscosity correlation linear data to obtain the viscosity process deviation coefficient.

[0102] In an embodiment of the present invention, when evaluating the process deviation coefficient of the production process benchmark cleaning parameters based on the dynamic viscosity-related linear data, the deviation between the dynamic viscosity value of the actual sample and the viscosity value predicted by the linear data is first calculated, and the calculation formula of the deviation value is defined as: Deviation = |actual viscosity value-predicted viscosity value| / predicted viscosity value. After the deviation values ​​of all samples are calculated, they are grouped and statistically analyzed according to the categories of the process benchmark cleaning parameters, and the average deviation value and standard deviation value of each category are calculated. Then, the deviation values ​​of each group are comprehensively calculated using the weighting factor, and the weighting factor is determined according to the proportion of the benchmark category in the overall sample. Finally, in combination with the boundary conditions of the production process benchmark cleaning parameters, the evaluation range of the deviation coefficient is set, and the deviation value of each category is mapped to the process deviation coefficient to generate the viscosity process deviation coefficient data. Through this process, the quantitative evaluation of the stability and consistency of the production process is completed, providing an important reference basis for the subsequent optimization of process parameters.

[0103] Preferably, step S22 includes the following steps:

[0104] Step S221: extracting the production temperature and production pressure of the production process benchmark cleaning parameters to obtain process production temperature data and process production pressure data;

[0105] Step S222: performing oil molecule nonlinear gravity analysis between different production process benchmarks on the lubricating oil density difference data according to the process production temperature data and the process production pressure data to obtain oil molecule nonlinear gravity data;

[0106] Step S223: performing discrete evaluation of intermolecular forces on the oil molecule nonlinear gravity data to obtain discrete intermolecular forces data;

[0107] Step S224: performing molecular chain morphology tensor analysis according to the intermolecular force discrete data to obtain molecular chain morphology tensor data;

[0108] Step S225: performing oil molecular chain motion simulation between different production process benchmarks according to the molecular chain morphology tensor data and the intermolecular force discrete data to obtain oil molecular chain motion data.

[0109] In the embodiment of the present invention, when extracting the production temperature and production pressure of the production process benchmark cleaning parameters, firstly, the production environment data related to the batch, including the real-time temperature and pressure records of the production line, are screened from the cleaning parameter database. These records are derived from online sensing equipment and are based on continuous data collected at a sampling frequency of seconds. During the extraction process, the sliding window method is used to analyze the mean and extreme values ​​of the temperature and pressure data within the time range, and the window size is set to 10 seconds to avoid data distortion caused by instantaneous fluctuations. The extracted temperature and pressure data are smoothed, and the cubic spline interpolation method is used to eliminate noise to generate a production temperature and pressure curve with good continuity. Subsequently, the curve is calibrated with characteristic points, and the maximum value, minimum value, mean value and stable interval value under the stable state are extracted to form standardized process production temperature data and process production pressure data. These data will be used for subsequent lubricating oil molecular chain motion analysis. When the lubricating oil density difference data is subjected to oil molecule nonlinear gravitational analysis according to the process production temperature data and process production pressure data, the magnitude of the gravitational force between molecules under different production conditions is first determined based on the van der Waals force calculation formula and the molecular distance parameter in the density difference data. Substitute the molecular distance of each group of samples in the density difference data into the formula to calculate the corresponding gravity value. Due to the large density difference between different batches, in order to prevent abnormal values ​​from appearing in the calculation results, the robust regression method is used to fit the gravity value and ensure the stability of the nonlinear relationship. Finally, the nonlinear gravity data of oil molecules under each set of production conditions are output and saved in matrix form, with rows and columns corresponding to the discrete intervals of production temperature and pressure. When evaluating the intermolecular force discreteness of the nonlinear gravity data of oil molecules, the normality of the nonlinear gravity data is first tested, and the normality of the data distribution is judged by the Shapiro-Wilk test method. If the data does not conform to the normal distribution, the data distribution is straightened by logarithmic transformation. The discreteness of the processed gravity data is calculated by batch, and the coefficient of variation (CV) is used as a measure of the degree of discreteness. When analyzing the molecular chain morphology tensor based on the intermolecular force discrete data, the discrete data is first mapped to the three-dimensional structure model of the molecular chain, and the morphology tensor is constructed in combination with the magnitude and direction of the intermolecular gravity. A tensor is defined as a three-dimensional symmetric matrix whose elements are the projection values ​​of the intermolecular forces in three-dimensional space. The tensor matrix of each set of data is obtained by calculation, and the principal component analysis method (PCA) is used to extract the main characteristic dimensions, simplify the data dimensions and improve the calculation efficiency. Finally, the molecular chain morphology tensor data is generated and output in the form of a tensor matrix, which provides morphological input for subsequent molecular chain motion simulation. When simulating the motion of oil molecular chains between different production process benchmarks based on the molecular chain morphology tensor data and the intermolecular force discrete data, the principal axis information is first extracted from the tensor data to define the initial direction and distribution state of the molecular chain.Then, combined with the discrete data of the acting force, the force state of each molecular chain is dynamically adjusted. During the simulation, the implicit Euler method is used to solve the dynamics of the molecular chain motion. The segmented time step calculation method is used to divide the molecular chain motion into multiple time periods for iterative calculation, and the motion trajectory of the molecular chain is output in each time period. After the simulation is completed, all trajectory data are interpolated and fitted to generate continuous motion path information, which is output as oil molecular chain motion data, providing a basic basis for subsequent viscosity analysis.

[0110] Preferably, step S224 includes the following steps:

[0111] The moment distribution matrix is ​​decomposed on the discrete data of intermolecular forces to obtain the moment distribution characteristic matrix;

[0112] According to the moment distribution characteristic matrix data, the discrete data of intermolecular forces are subjected to structure layer-by-layer tensor decomposition to obtain the molecular local tensor data;

[0113] Performing tensor field gradient clustering processing on the local tensor data of the molecule to obtain tensor field gradient clustering data;

[0114] The molecular chain morphology tensor data is analyzed based on the tensor field gradient clustering data to obtain the molecular chain morphology tensor data.

[0115] In an embodiment of the present invention, when processing the discrete data of intermolecular forces, the first step of the moment distribution matrix decomposition process is to construct a moment matrix. Each force data point is determined by the vector information in the discrete data of the force according to its direction of action and arm distance through the moment calculation formula. All data points are calculated using a three-dimensional moment matrix to form a three-dimensional distribution matrix. Then, the matrix is ​​decomposed using singular value decomposition (SVD), the main characteristic components are extracted, and the main contribution of the data is retained in the characteristic matrix. The moment distribution characteristic matrix is ​​combined with the discrete data of intermolecular forces, and tensor decomposition is performed according to the hierarchical structure. First, the discrete data of intermolecular forces are reshaped into a three-dimensional tensor structure, and each dimension corresponds to the distance between molecules, the magnitude of the force and the direction. The tensor is decomposed using a high-order orthogonal iterative algorithm (HOOI). This process is based on the reduction of the tensor rank, and the complexity of the original data is reduced through layer-by-layer iterative optimization, the key tensor characteristic information is retained, and the local tensor data of the molecule is finally output. When performing tensor field gradient clustering on local molecular tensor data, a tensor field gradient distribution map is first constructed. By calculating the gradient change rate between any two points in the tensor field, the density peaks clustering algorithm (DPC) is used to classify the tensor data points in the gradient distribution map. According to the local density of each data point and its distance from the nearest high-density point, the cluster center is determined and the points around it are divided into corresponding categories. Finally, the tensor field gradient clustering data is obtained, which contains the category label corresponding to each gradient point and its cluster center. Based on the tensor field gradient clustering data, the tensor distribution characteristics and molecular chain morphological characteristics of each cluster category are analyzed. First, the local tensor data corresponding to each cluster center is extracted as representative data. Combined with the main direction of the local tensor and the tensor changes between the cluster feature points, a tensor morphological analysis model is constructed. Finally, all morphological feature values ​​are normalized to generate molecular chain morphological tensor data, which are saved in the form of a three-dimensional tensor field to provide reliable basic data for subsequent analysis.

[0116] Preferably, step S24 comprises the following steps:

[0117] Step S241: performing grid segment density change analysis on the dynamic viscosity associated linear data to obtain segment viscosity density change data;

[0118] Step S242: performing segment-by-segment nonlinear deviation interference calculation on the production process benchmark cleaning parameters according to the segment viscosity density change data to obtain benchmark deviation interference data;

[0119] Step S243: performing lubricating oil flash point continuity simulation according to the reference deviation interference data and the segmented viscosity density change data to obtain lubricating oil flash point continuity simulation data;

[0120] Step S244: Based on the continuous simulation data of the lubricating oil flash point and the reference deviation interference data, the process deviation coefficient of the segmented viscosity density change data is evaluated to obtain the viscosity process deviation coefficient.

[0121] In an embodiment of the present invention, the dynamic viscosity-related linear data is subjected to grid segment density change analysis. First, the dynamic viscosity data is segmented according to the time series, and the specific steps are to divide the entire data set into several time periods to ensure that the amount of data in each time period is relatively uniform. Subsequently, the density of each segmented data is calculated using the linear interpolation method to obtain the density change of the lubricating oil in each time period. The segmented viscosity density change data is obtained by calculating the relationship between the viscosity and density in each segment. This process requires accurate recording of the start and end time of each time period for subsequent analysis. Based on the segmented viscosity density change data, the nonlinear deviation interference calculation between segments of the production process benchmark cleaning parameters is performed. First, according to the segmented viscosity density change data, a suitable nonlinear regression model, such as a polynomial regression model, is selected to fit the segmented data to obtain a preliminary estimate of the benchmark deviation. Then, the model parameters are optimized by the least squares method to reduce the fitting error. Then, the fitting results are compared with the production process benchmark cleaning parameters to calculate the nonlinear deviation interference value. In this step, it is necessary to ensure that the goodness of fit of the model meets the predetermined standard to ensure the reliability of the calculation results. The flash point continuity simulation of lubricating oil is carried out based on the reference deviation interference data and the segmented viscosity density change data. First, the reference deviation interference data and the segmented data are integrated to form a complete data set containing time series. Next, the change of the flash point of lubricating oil is simulated by the continuity simulation method and the finite element analysis (FEA) technology. In the specific implementation, the physical properties of the lubricating oil are combined with environmental factors (such as temperature and pressure), and the heat conduction equation and the fluid mechanics equation are used for simulation to obtain the flash point continuity change data. The simulation process requires a detailed analysis of the chemical composition and physical properties of the lubricating oil to ensure the accuracy of the model. Based on the flash point continuity simulation data of lubricating oil and the reference deviation interference data, the process deviation coefficient of the segmented viscosity density change data is evaluated. First, the flash point continuity simulation data of lubricating oil and the reference deviation interference data are comprehensively analyzed to form an evaluation model containing multi-dimensional parameters. Then, the weight of each parameter is calculated by the weighted average method to reflect its impact on the process deviation. Then, the data of different dimensions are standardized by normalization to make them comparable. Finally, the viscosity process deviation coefficient is obtained by calculating the deviation degree of each parameter. This step emphasizes the accuracy and systematicness of data processing to ensure the validity of the evaluation results.

[0122] Preferably, step S243 includes the following steps:

[0123] Perform dynamic feature vector extraction on the reference deviation interference data to obtain the reference interference feature vector;

[0124] According to the reference interference characteristic vector, the segmented viscosity density change data is subjected to layered viscosity correlation flash point fitting calculation to obtain layered flash point fitting data;

[0125] Performing dynamic continuity curve fitting processing on the layered flash point fitting data to obtain a dynamic flash point continuity curve;

[0126] The nonlinear flash point deduction is performed according to the dynamic flash point continuity curve and the segmented viscosity density change data to obtain the dynamic flash point nonlinear deduction data;

[0127] The lubricant oil flash point continuity simulation is carried out based on the dynamic flash point nonlinear derivation data to obtain the lubricant oil flash point continuity simulation data.

[0128] In an embodiment of the present invention, first, dynamic feature vector extraction is performed on the reference deviation interference data to obtain the reference interference feature vector. This process involves time series analysis of the interference data, and the principal component analysis (PCA) technology is used to extract the main features of the data. First, the reference deviation interference data is standardized to eliminate the dimension effect. Then, by calculating the covariance matrix, the eigenvalues ​​and eigenvectors are extracted, and the eigenvectors corresponding to the first few eigenvalues ​​are selected to form the reference interference feature vector. These eigenvectors reflect the main change trend of the reference deviation and can provide important information for subsequent analysis. Next, according to the reference interference feature vector, the segmented viscosity density change data is subjected to layered viscosity associated flash point fitting calculation to obtain layered flash point fitting data. First, the segmented viscosity density change data is divided into multiple levels according to the time series to ensure that the data in each level has similar physical properties. Then, a suitable fitting model is selected for each level, such as a polynomial regression or exponential regression model, to perform data fitting. The model parameters are optimized by the least squares method to reduce the fitting error, and finally the layered flash point fitting data of each level is obtained. This process needs to ensure that the fitting model can accurately reflect the relationship between viscosity and flash point of each layer, thus laying the foundation for subsequent analysis. Then, the layered flash point fitting data is processed by dynamic continuity curve fitting to obtain the dynamic flash point continuity curve. First, the layered flash point fitting data is smoothed to eliminate the influence of noise on the data. Then, the smoothed data is continuously fitted by spline interpolation to generate a dynamic flash point continuity curve. This curve shows the trend of flash point change over time and can reflect the flash point characteristics of lubricating oil under different process conditions. This step requires the interpolation process to consider the rate of change of data to ensure the smoothness and accuracy of the fitting curve. Nonlinear flash point deduction is performed based on the dynamic flash point continuity curve and the segmented viscosity density change data to obtain dynamic flash point nonlinear deduction data. This process uses nonlinear fitting methods, such as Gaussian model or power law model, to fit the relationship between the dynamic flash point continuity curve and the segmented viscosity density change data. First, the two sets of data are merged to form a multidimensional data set containing multiple variables. Then, the least squares method is used for nonlinear regression to optimize the model parameters to obtain the best fitting curve. Finally, the dynamic flash point nonlinear derivation data are extracted, which reflects the flash point change trend of the lubricant under specific conditions. Finally, the lubricant flash point continuity simulation is performed based on the dynamic flash point nonlinear derivation data to obtain the lubricant flash point continuity simulation data. First, a mathematical model containing physical parameters and chemical properties is constructed, considering the influence of environmental factors such as temperature and pressure on the flash point. Then, numerical simulation techniques, such as the finite difference method, are used to solve the model to obtain the flash point change of the lubricant under different process conditions. The entire simulation process requires multiple iterations to ensure the accuracy and stability of the results. Finally, the output lubricant flash point continuity simulation data provides an important basis for subsequent production control.

[0129] Preferably, step S3 comprises the following steps:

[0130] Step S31: normalizing the viscosity process deviation coefficient to obtain viscosity process deviation normalized data;

[0131] Step S32: performing multidimensional parameter decision optimization according to the viscosity process deviation normalization data to obtain process multidimensional parameter optimization data;

[0132] Step S33: Performing logic learning on the process multi-dimensional parameter optimization data based on the random forest algorithm to obtain the process multi-dimensional parameter optimization logic data.

[0133] In an embodiment of the present invention, the viscosity process deviation coefficient is normalized to obtain viscosity process deviation normalized data. First, all viscosity process deviation coefficients are collected to form an array containing multiple data points. Then, a minimum-maximum normalization method is used, which calculates the normalized value by subtracting the minimum value from each data point and dividing it by the difference between the maximum value and the minimum value. Multidimensional parameter decision optimization is performed based on the viscosity process deviation normalized data to obtain process multidimensional parameter optimization data. First, the main parameters affecting the lubricant production process are determined, which may include temperature, pressure, flow rate, etc. Then, a multi-objective optimization model is constructed, which uses a weighted summation method to comprehensively consider the influence of each parameter. To this end, a weight needs to be assigned to each parameter, and the setting of the weight is based on its importance assessment to the production process. Then, a genetic algorithm is used for parameter optimization. The genetic algorithm simulates the process of natural selection and generates a new generation of solutions through operations such as selection, crossover and mutation. First, a population of initial solutions is randomly generated, and then the fitness of each solution is evaluated, that is, its performance under the current process conditions. Next, a new solution is generated by cross-selecting solutions with high fitness, and the new solution is mutated in a small range to increase diversity. After multiple generations of iterations, a set of optimized process multidimensional parameters is finally obtained, which meet the production requirements to the greatest extent. In step S33, the process multidimensional parameter optimization data is logically learned based on the random forest algorithm to obtain the process multidimensional parameter optimization logic data. First, the optimized process parameter data is collected and the corresponding production results are marked. The data set includes multiple characteristic variables (such as temperature, pressure, flow rate, etc.) and their corresponding target variables (such as product quality, production efficiency, etc.). Then, a random forest model is constructed. The random forest consists of multiple decision trees, each of which is constructed by random sampling of training data, thereby improving the generalization ability of the model. In the specific implementation, the data set is first randomly divided into a training set and a test set, the training set is used to build the model, and the test set is used to evaluate the performance of the model. During the training process, the voting mechanism of each tree is used, combined with the prediction results of multiple trees, to obtain the final optimization logic data. Finally, the trained model is verified to ensure that its prediction accuracy on the test set meets the predetermined standard. Through this process, the obtained process multi-dimensional parameter optimization logic data can not only guide the production process of lubricants, but also provide a scientific basis for subsequent production decisions.

[0134] Preferably, step S32 includes the following steps:

[0135] Step S321: performing multi-scale decomposition processing on the viscosity process deviation normalization data to obtain viscosity multi-scale decomposition parameters;

[0136] Step S322: performing nonlinear constraint processing on the viscosity multi-scale decomposition parameters to obtain viscosity multi-scale decomposition constraint data;

[0137] Step S323: performing multi-objective decision-making and solving processing on the viscosity process deviation normalization data according to the viscosity multi-scale decomposition constraint data to obtain the process multi-objective solution parameters;

[0138] Step S324: performing multi-dimensional parameter decision optimization on the viscosity process deviation normalization data according to the process multi-objective solution parameters to obtain process multi-dimensional parameter optimization data.

[0139] In the embodiment of the present invention, the viscosity process deviation normalized data is subjected to multi-scale decomposition processing to obtain viscosity multi-scale decomposition parameters. First, all viscosity process deviation normalized data are collected and time series analysis is performed on them. Wavelet transform is used as a method of multi-scale decomposition. Wavelet transform can decompose the signal into components of different frequencies, thereby extracting features on different time scales. In the specific implementation process, an appropriate wavelet basis function, such as Haer wavelet or Moir wavelet, is selected to better capture the local characteristics of the signal. Next, the collected viscosity process deviation normalized data is subjected to wavelet decomposition, and the decomposition process converts the data into coefficients of multiple scales, representing the signal characteristics at different frequencies. The coefficient of each scale will reflect the trend of change under the scale, and a set of viscosity multi-scale decomposition parameters can be obtained by analyzing the coefficients of different scales. These parameters provide a basis for subsequent nonlinear constraint processing. Nonlinear constraint processing is performed on the viscosity multi-scale decomposition parameters to obtain viscosity multi-scale decomposition constraint data. First, nonlinear constraint conditions are set for the multi-scale decomposition parameters obtained in the previous step. Nonlinear constraints are mainly used to ensure that the parameters are within a reasonable range and conform to the actual physical meaning. In the specific implementation, nonlinear optimization methods, such as the Lagrange multiplier method, are used to transform the constraints into optimization problems. In the process, the objective function is defined as the loss function of the viscosity multiscale decomposition parameters, and the constraints are the range limits set according to the physical properties. By solving the nonlinear optimization problem, the viscosity multiscale decomposition constraint data that meets the constraints is obtained. The key to this step is to ensure that all constraints are accurately transformed and satisfied to ensure the effectiveness of subsequent analysis. According to the viscosity multiscale decomposition constraint data, the viscosity process deviation normalization data is processed by multi-objective decision-making to obtain the process multi-objective solution parameters. First, multiple objectives to be optimized are determined, such as maximizing production efficiency, minimizing energy consumption, and minimizing product defect rate. Then, a multi-objective optimization model is constructed, and multiple objectives are integrated into a comprehensive objective function using a weighted method. Next, the particle swarm optimization (PSO) algorithm is used to solve the model. First, a group of random particles are generated, and the position and velocity of the particles represent the parameters and change rate of the current solution, respectively. In each iteration, the fitness of each particle is evaluated according to the objective function, and the velocity and position of the particle are updated to guide it to a better solution. Through multiple iterations, a set of optimized process multi-objective solution parameters are finally obtained, which will provide a basis for subsequent decision optimization. According to the process multi-objective solution parameters, the viscosity process deviation normalization data is optimized by multi-dimensional parameter decision to obtain the process multi-dimensional parameter optimization data. First, combined with the multi-objective solution parameters obtained in the previous step, the optimization goal is set, such as reducing the viscosity deviation in the production process. A multi-dimensional parameter optimization model is constructed to consider the impact of multiple parameters on the final optimization goal. Next, the simulated annealing algorithm is used for optimization.Simulated annealing is an optimization algorithm based on random search, which simulates the physical annealing process to find the global optimal solution. In the specific implementation, the initial temperature and cooling rate are first set, and the initial solution is randomly generated. By gradually lowering the temperature, the solution is allowed to change within a certain range to avoid falling into the local optimum. Through continuous iteration, the performance of each solution is evaluated, and the better solution is selected as the temperature gradually decreases, and finally the multi-dimensional parameter optimization of the viscosity process deviation from the normalized data is achieved. The key to this step is to balance exploration and utilization to ensure the optimal production parameters.

[0140] Preferably, the present invention further provides an intelligent lubricant production control system for executing the intelligent lubricant production control method as described above, the intelligent lubricant production control system comprising:

[0141] The process benchmark parameter mapping module is used to collect batch samples of lubricating oil on the lubricating oil production line to obtain lubricating oil batch samples; map the production process benchmark parameters between different batches of lubricating oil batch samples to obtain production process benchmark cleaning parameters;

[0142] The process deviation coefficient evaluation module is used to simulate the movement of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters to obtain the oil molecular chain movement data; the process deviation coefficient of the oil molecular chain movement data is evaluated to obtain the viscosity process deviation coefficient;

[0143] The logic learning adjustment module is used to optimize the multi-dimensional parameter decision according to the viscosity process deviation coefficient to obtain the process multi-dimensional parameter optimization data; the logic learning of the process multi-dimensional parameter optimization data is performed based on the random forest algorithm to obtain the process multi-dimensional parameter optimization logic data;

[0144] The automation firmware design module is used to design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain the process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to perform intelligent lubricant production management and control.

[0145] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0146] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent lubricant production control method, characterized in that: The following steps are involved: Step S1: collecting batch samples of lubricating oil on a lubricating oil production line to obtain lubricating oil batch samples; Mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain the production process benchmark cleaning parameters; Step S2: performing oil molecule chain motion simulation between different production process benchmarks according to the production process benchmark cleaning parameters to obtain oil molecule chain motion data; The process deviation coefficient of the oil molecular chain motion data is evaluated to obtain the viscosity process deviation coefficient; Wherein, step S2 comprises the following steps: Step S21: extracting the density difference of lubricating oil between different batches of lubricating oil batch samples to obtain lubricating oil density difference data; Step S22: According to the production process benchmark cleaning parameters, and using the molecular dynamics simulation method, the lubricating oil density difference data is simulated for the oil molecular chain motion between different production process benchmarks to obtain the oil molecular chain motion data; specifically, step S22 includes the following steps: Step S221: extracting the production temperature and production pressure of the production process benchmark cleaning parameters to obtain process production temperature data and process production pressure data; Step S222: performing oil molecule nonlinear gravity analysis between different production process benchmarks on the lubricating oil density difference data according to the process production temperature data and the process production pressure data to obtain oil molecule nonlinear gravity data; Step S223: performing discrete evaluation of intermolecular forces on the oil molecule nonlinear gravity data to obtain discrete intermolecular forces data; Step S224: performing molecular chain morphology tensor analysis according to the intermolecular force discrete data to obtain molecular chain morphology tensor data; wherein step S224 includes the following steps: The moment distribution matrix is ​​decomposed on the discrete data of intermolecular forces to obtain the moment distribution characteristic matrix; According to the moment distribution characteristic matrix data, the discrete data of intermolecular forces are subjected to structure layer-by-layer tensor decomposition to obtain the molecular local tensor data; Performing tensor field gradient clustering processing on the local tensor data of the molecule to obtain tensor field gradient clustering data; Performing molecular chain morphology tensor analysis on molecular local tensor data according to tensor field gradient clustering data to obtain molecular chain morphology tensor data; Step S23: performing dynamic viscosity correlation linear analysis on the oil molecular chain motion data to obtain dynamic viscosity correlation linear data; Step S24: evaluating the process deviation coefficient of the production process benchmark cleaning parameters based on the dynamic viscosity correlation linear data to obtain the viscosity process deviation coefficient; specifically, step S241: performing grid segment density change analysis on the dynamic viscosity correlation linear data to obtain segment viscosity density change data; Step S242: performing segment-by-segment nonlinear deviation interference calculation on the production process benchmark cleaning parameters according to the segment viscosity density change data to obtain benchmark deviation interference data; Step S243: performing lubricating oil flash point continuity simulation according to the reference deviation interference data and the segmented viscosity density change data to obtain lubricating oil flash point continuity simulation data; Step S244: evaluating the process deviation coefficient of the segmented viscosity density change data based on the lubricating oil flash point continuous simulation data and the reference deviation interference data to obtain the viscosity process deviation coefficient; Step S3: performing multidimensional parameter decision optimization according to the viscosity process deviation coefficient to obtain process multidimensional parameter optimization data; performing logic learning on the process multidimensional parameter optimization data based on the random forest algorithm to obtain process multidimensional parameter optimization logic data; Step S4: Design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to execute intelligent lubricant production management and control.

2. The intelligent lubricant production control method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: collecting batch samples of lubricating oil on the lubricating oil production line to obtain lubricating oil batch samples; Step S12: mapping the production process benchmark parameters between different batches of lubricating oil batch samples to obtain the production process benchmark parameters; Step S13: performing data cleaning on the production process benchmark parameters to obtain the production process benchmark cleaning parameters.

3. The intelligent lubricant production control method according to claim 1, characterized in that: Step S243 includes the following steps: Perform dynamic feature vector extraction on the reference deviation interference data to obtain the reference interference feature vector; According to the reference interference characteristic vector, the segmented viscosity density change data is subjected to layered viscosity correlation flash point fitting calculation to obtain layered flash point fitting data; Performing dynamic continuity curve fitting processing on the layered flash point fitting data to obtain a dynamic flash point continuity curve; The nonlinear flash point deduction is performed according to the dynamic flash point continuity curve and the segmented viscosity density change data to obtain the dynamic flash point nonlinear deduction data; The lubricant oil flash point continuity simulation is carried out based on the dynamic flash point nonlinear derivation data to obtain the lubricant oil flash point continuity simulation data.

4. The intelligent lubricant production control method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: normalizing the viscosity process deviation coefficient to obtain viscosity process deviation normalized data; Step S32: performing multidimensional parameter decision optimization according to the viscosity process deviation normalization data to obtain process multidimensional parameter optimization data; Step S33: Performing logic learning on the process multi-dimensional parameter optimization data based on the random forest algorithm to obtain the process multi-dimensional parameter optimization logic data.

5. The intelligent lubricant production control method according to claim 4, characterized in that: Step S32 includes the following steps: Step S321: performing multi-scale decomposition processing on the viscosity process deviation normalization data to obtain viscosity multi-scale decomposition parameters; Step S322: performing nonlinear constraint processing on the viscosity multi-scale decomposition parameters to obtain viscosity multi-scale decomposition constraint data; Step S323: performing multi-objective decision-making and solving processing on the viscosity process deviation normalization data according to the viscosity multi-scale decomposition constraint data to obtain the process multi-objective solution parameters; Step S324: performing multi-dimensional parameter decision optimization on the viscosity process deviation normalization data according to the process multi-objective solution parameters to obtain process multi-dimensional parameter optimization data.

6. An intelligent lubricating oil production control system, characterized in that: Used to execute the intelligent lubricant production control method according to claim 1, the intelligent lubricant production control system comprises: The process benchmark parameter mapping module is used to collect batch samples of lubricating oil on the lubricating oil production line to obtain lubricating oil batch samples; map the production process benchmark parameters between different batches of lubricating oil batch samples to obtain production process benchmark cleaning parameters; The process deviation coefficient evaluation module is used to simulate the movement of oil molecular chains between different production process benchmarks according to the production process benchmark cleaning parameters to obtain the oil molecular chain movement data; the process deviation coefficient of the oil molecular chain movement data is evaluated to obtain the viscosity process deviation coefficient; The logic learning adjustment module is used to optimize the multi-dimensional parameter decision according to the viscosity process deviation coefficient to obtain the process multi-dimensional parameter optimization data; the logic learning of the process multi-dimensional parameter optimization data is performed based on the random forest algorithm to obtain the process multi-dimensional parameter optimization logic data; The automation firmware design module is used to design automation firmware based on the process multi-dimensional parameter optimization logic data to obtain the process optimization automation firmware, and embed the process optimization automation firmware into the intelligent lubricant production line control center to perform intelligent lubricant production management and control.

Images