Grease product production control method and system based on intelligent digitization

Abstract

The invention relates to the technical field of grease production control, in particular to a grease product production control method and system based on intelligent digitization. The method comprises the following steps: monitoring temperature, viscosity and turbidity data in a grease crystallization process in real time, and generating grease crystallization state snapshot data; performing crystallization state prediction according to the grease crystallization state snapshot data, and performing reaction kettle cooling rate adjustment to obtain a target control cooling rate; performing real-time stirring rotating speed adjustment fuzzy reasoning on the grease crystallization state snapshot data to generate a corrected grease stirring speed; and performing control instruction conversion on the target control cooling rate and the corrected grease stirring speed, executing a production control instruction, and generating a grease crystallization process control report. By accurately controlling the cooling rate and the stirring speed, intelligent control over the grease crystallization process is achieved, and the quality and stability of grease products are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of oil production control, and particularly to a production control method and system for oil products based on intelligent digitization. Background Art

[0002] In the production process of oil, crystallization is a crucial step. Oil crystallization refers to the process in which triglyceride molecules change from a disordered state to an ordered arrangement during the cooling of liquid oil, forming crystal nuclei and gradually growing into crystals. This process is affected by various factors, including cooling rate, stirring speed, temperature gradient, impurity content, etc. Among them, the cooling rate and stirring speed are the two most critical factors affecting the oil crystallization process. During the crystallization process, if the cooling rate is too fast, it is easy to cause the oil to form a large number of small and unstable crystals, affecting the taste and stability of the oil; if the cooling rate is too slow, the crystals will be too large and unevenly distributed, which will also affect the quality of the oil. The control of the stirring speed is equally important. Too fast a stirring speed will destroy the formed crystal structure, and too slow a stirring speed will cause the oil to stratify and crystallize unevenly. However, the traditional oil crystallization process mainly relies on manual experience for control. Operators set the cooling rate and stirring speed according to experience, with strong subjectivity, making it difficult to achieve precise control, and ultimately affecting the quality and stability of the oil. Summary of the Invention

[0003] Based on this, the present invention provides a production control method and system for oil products based on intelligent digitization to solve at least one of the above technical problems.

[0004] To achieve the above object, a production control method for oil products based on intelligent digitization includes the following steps:

[0005] Step S1: Real-time monitor the temperature, viscosity, and turbidity data during the oil crystallization process in the oil crystallization reactor, and integrate the oil state to generate oil crystallization state snapshot data;

[0006] Step S2: Combine the oil crystallization characteristics according to the oil crystallization state snapshot data to generate an oil crystallization characteristic vector; construct an oil crystallization kinetics model; use the oil crystallization kinetics model to predict the crystallization state of the oil crystallization characteristic vector to generate oil crystallization prediction state data; adjust the cooling rate of the reactor through the oil crystallization prediction state data to obtain the target control cooling rate;

[0007] Step S3: Set the desired heat transfer coefficient based on the target control cooling rate to obtain the desired heat transfer coefficient; perform fuzzy control variable processing on the oil crystallization state snapshot data through the desired heat transfer coefficient, and perform fuzzy inference on the real-time stirring speed adjustment to generate a corrected oil stirring speed;

[0008] Step S4: Convert the target controlled cooling rate and the corrected grease stirring speed into control instructions, and execute the production control instructions to obtain control feedback verification data; store the production process control records according to the control feedback verification data, and generate a grease crystallization process control report.

[0009] By real-time monitoring of key parameters such as temperature, viscosity, and turbidity in the reaction kettle and integrating them to generate snapshot data of the grease crystallization state, the present invention can comprehensively and dynamically reflect the real-time state of grease crystallization. This provides a reliable data basis for subsequent feature extraction, model prediction, and control adjustment, avoiding the subjectivity and lag of traditional processes relying on manual experience judgment, and making the grease crystallization process more transparent and controllable. A grease crystallization kinetic model is established to realize the prediction and control of the crystallization process. Based on the grease crystallization feature vectors generated from real-time monitoring data and combined with the grease crystallization kinetic model, the future state of grease crystallization can be predicted. This enables the production process to no longer passively follow changes, but can anticipate the crystallization trend and make control adjustments in advance, avoiding problems such as too fast or too slow cooling rates, thereby ensuring the uniformity of the size and distribution of grease crystals and ultimately improving the taste and stability of the grease. For example, if the model predicts that the particle size distribution of grease crystallization will become too dispersed in the future under the current cooling rate, the system will automatically adjust the cooling rate to maintain it within the optimal range to prevent the appearance of a large number of small crystals or overly large crystals. Traditional stirring speed control is often fixed and cannot adapt to the complex dynamic changes during the grease crystallization process. However, this method performs fuzzy control variable processing on the snapshot data of the grease crystallization state through the desired heat transfer coefficient and conducts fuzzy inference for real-time stirring speed adjustment, and can dynamically adjust the stirring speed according to the real-time state of grease crystallization. For example, when it is detected that the grease viscosity increases, the system will automatically increase the stirring speed to prevent grease stratification and uneven crystallization; when the crystallization process tends to be stable, the system will decrease the stirring speed to avoid damaging the formed crystal structure. By converting the target controlled cooling rate and the corrected grease stirring speed into control instructions and applying them to the production control system, the automation control of the grease crystallization process is realized, reducing manual intervention and improving production efficiency. At the same time, the system will automatically record the control feedback verification data and generate a grease crystallization process control report, providing data support for the optimization and improvement of the production process, facilitating the traceability of product quality, and ensuring the stability and consistency of product quality. Therefore, a production control method for grease products based on intelligent digitization according to the present invention realizes the precise control of the grease crystallization process by real-time monitoring the temperature, viscosity, and turbidity data during the grease crystallization process, constructs a grease crystallization kinetic model, significantly improves the quality and stability of grease products, and also improves production efficiency.

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

[0011] Step S11: Fix and install temperature sensors, viscosity sensors, and turbidity sensors on the grease crystallization reactor to construct a grease crystallization monitoring network;

[0012] Step S12: Use the grease crystallization monitoring network to continuously monitor the temperature, viscosity, and turbidity data during the grease crystallization process to obtain the original grease crystallization monitoring data stream;

[0013] Step S13: Perform preprocessing on the monitoring data of the original grease crystallization data stream to obtain preprocessed grease crystallization monitoring data;

[0014] Step S14: Align the time series according to the preprocessed grease crystallization monitoring data and integrate the grease states to generate grease crystallization state snapshot data.

[0015] Through the fixedly installed sensors, the present invention constructs a monitoring network, which can continuously and stably obtain the real-time data of these key parameters, so as to comprehensively and accurately master the dynamic change process of grease crystallization. There are problems such as noise, outliers, or missing values in the original data collected by the sensors. Directly using these data will affect the accuracy of subsequent analysis. Through preprocessing, noise can be removed, missing values can be filled, and outliers can be corrected, so as to obtain clean and reliable preprocessed grease crystallization monitoring data. This ensures the data quality input into the subsequent model and improves the reliability and accuracy of the entire control system. Through time series alignment and grease state integration, grease crystallization state snapshot data is generated, realizing a comprehensive characterization of the grease crystallization state. There are time deviations in the data collected by different sensors, and time series alignment can unify these data to the same time reference.

[0016] Preferably, step S2 includes the following steps:

[0017] Step S21: Extract the grease state in the snapshot reactor according to the grease crystallization state snapshot data to obtain the reactor grease temperature data, reactor grease viscosity data, and reactor grease turbidity data respectively;

[0018] Step S22: Combine the grease crystallization characteristics of the reactor grease temperature data, reactor grease viscosity data, and reactor grease turbidity data to generate a grease crystallization feature vector; construct a crystallization kinetics model based on the grease crystallization feature vector to generate a grease crystallization kinetics model;

[0019] Step S23: Use the grease crystallization kinetics model to predict the crystallization state of the grease crystallization feature vector to generate grease crystallization prediction state data;

[0020] Step S24: Obtain the real-time cooling rate of the reactor;

[0021] Step S25: Adjust the real-time cooling rate of the reaction kettle through the predicted state data of oil crystallization to obtain the target controlled cooling rate.

[0022] The kinetic model constructed based on the characteristic vectors of oil crystallization in the present invention is the core for realizing the prediction and control of the oil crystallization process. This model can relate the characteristic vectors of oil crystallization to the future crystallization state. By analyzing the current state, it can predict the future crystallization trend, such as crystal size, distribution uniformity, etc. Using the oil crystallization kinetic model to predict the crystallization state and combining with the real-time cooling rate of the reaction kettle, the cooling rate can be dynamically adjusted to achieve precise control of the crystallization process. For example, if the model predicts that at the current cooling rate, the oil crystallization speed is too fast and a large number of fine crystals are likely to form, the system will automatically reduce the cooling rate to control the crystallization speed and ensure the crystal size and distribution uniformity. This prediction-based dynamic adjustment avoids the influence of too fast or too slow cooling rate on the quality of the oil, and significantly improves the quality and stability of the oil product.

[0023] Preferably, step S22 includes the following steps:

[0024] Step S221: Obtain the target oil material parameters; extract the oil type according to the target oil material parameters to obtain the target oil type data;

[0025] Step S222: Extract the oil solubility curve according to the target oil material parameters, and calculate the oil supersaturation at the current temperature using the oil temperature data of the reaction kettle to generate real-time oil supersaturation data;

[0026] Step S223: Differentiate the oil viscosity data and the oil turbidity data of the reaction kettle with respect to time, and calculate the change rates respectively to obtain the oil viscosity change rate of the reaction kettle and the oil turbidity change rate of the reaction kettle;

[0027] Step S224: Combine the real-time oil supersaturation data, the oil viscosity change rate of the reaction kettle, and the oil turbidity change rate of the reaction kettle to generate an oil crystallization characteristic vector;

[0028] Step S225: Construct a crystallization kinetic model based on the target oil material parameters through the oil crystallization characteristic vector to obtain the oil crystallization kinetic model.

[0029] The present invention extracts the oil solubility curve using the target oil material parameters, and calculates the real-time oil supersaturation degree in combination with the oil temperature data in the reaction kettle, which can more accurately reflect the driving force of oil crystallization. Differentiating the oil viscosity and turbidity data in the reaction kettle with respect to time and calculating the change rate can more sensitively capture the dynamic changes in the oil crystallization process. The change rates of viscosity and turbidity reflect the changes in the crystallization rate and can be used as important control parameters. For example, when the change rate of viscosity suddenly increases, it indicates that the crystallization rate accelerates. At this time, the cooling rate or stirring speed needs to be adjusted to control the crystallization process. Combining the real-time oil supersaturation degree, the change rate of the oil viscosity in the reaction kettle, and the change rate of turbidity to generate an oil crystallization feature vector can more comprehensively and accurately describe the oil crystallization state.

[0030] Preferably, step S225 includes the following steps:

[0031] Conduct crystallization characteristic analysis based on the target oil type data, and screen the classical crystallization kinetic model according to the classical crystallization kinetics theory to obtain the alternative oil crystallization models;

[0032] Evaluate the applicability of the alternative oil crystallization models to generate the oil crystallization model structure data;

[0033] Determine the undetermined parameters of the oil crystallization model structure data through the oil crystallization feature vector to generate a list of model undetermined parameters;

[0034] Query the crystallization experiment data based on the list of model undetermined parameters using the preset historical crystallization experiment database to obtain the historical oil crystallization experiment data;

[0035] Extract the model training features from the historical oil crystallization experiment data to obtain the model training feature data;

[0036] Divide the model training feature data into training data and validation data to obtain a model training data set and a model validation data set;

[0037] Based on the model training data set, use the preset particle swarm optimization algorithm to optimize the list of model undetermined parameters and train the alternative oil crystallization models to obtain a preliminary trained crystallization model;

[0038] Perform cross-validation on the preliminary trained crystallization model through the model validation data set and perform hyperparameter tuning to obtain the oil crystallization kinetics model.

[0039] The present invention analyzes the crystallization characteristics according to the target oil type data and screens the classical crystallization kinetics model, ensuring the applicability and effectiveness of the selected model. By using the oil crystallization characteristic vectors to determine the undetermined parameters of the model and combining with the historical crystallization experiment data for model training, the accuracy and reliability of the model can be effectively improved. Training the model with a large amount of historical experimental data enables the model to better learn the laws of the oil crystallization process, thereby improving the prediction ability of the model. Dividing the model training characteristic data into training data and validation data and using the particle swarm algorithm for parameter optimization and model training can avoid model overfitting and improve the generalization ability of the model.

[0040] Preferably, step S25 includes the following steps:

[0041] Step S251: Perform predictive crystallization rate processing based on the oil crystallization prediction state data to generate a predictive crystallization rate curve;

[0042] Step S252: Identify the crystallization key points based on the predictive crystallization rate curve, calculate the oil crystallization rate deviation, and generate the crystallization rate deviation sequence data;

[0043] Step S253: Calculate the preliminary cooling rate adjustment amount for the real-time cooling rate of the reaction kettle by using the crystallization rate deviation sequence data to generate the preliminary cooling rate adjustment amount;

[0044] Step S254: Perform safety limiting processing on the preliminary cooling rate adjustment amount and perform temperature compensation through the oil temperature data of the reaction kettle to generate the target controlled cooling rate.

[0045] The present invention identifies the crystallization key points on the predictive crystallization rate curve and calculates the oil crystallization rate deviation, which can quantify the difference between the crystallization process and the expected target, providing more accurate guidance for the cooling rate adjustment. By comparing the actual crystallization rate with the target crystallization rate, the direction and amplitude of the cooling rate adjustment can be determined. Calculating the preliminary cooling rate adjustment amount by using the crystallization rate deviation sequence data can achieve dynamic adjustment of the cooling rate, making the crystallization process closer to the expected target. This adjustment amount is calculated based on the crystallization rate deviation, which can more effectively correct the deviation of the crystallization process and ensure that the crystallization process is always in the best state. Performing safety limiting processing on the preliminary cooling rate adjustment amount can avoid excessive cooling rate adjustment and prevent adverse effects on the oil crystallization process. At the same time, performing temperature compensation through the oil temperature data of the reaction kettle can further improve the accuracy of the cooling rate adjustment, making the cooling rate adjustment more in line with the actual situation.

[0046] Preferably, step S252 includes the following steps:

[0047] Based on the predicted crystallization rate curve, analyze the crystal size-time curve and the crystal number-time curve to obtain the oil crystallization size-time curve and the oil crystallization number-time curve respectively;

[0048] Perform first-order numerical differentiation on the oil crystallization size-time curve and the oil crystallization number-time curve respectively to obtain the particle size growth rate curve and the number growth rate curve;

[0049] Align the particle size growth rate curve and the number growth rate curve in time series and perform extreme point analysis to obtain the key point data of oil crystallization;

[0050] Based on the key point data of oil crystallization, divide the crystallization stage to obtain the crystallization stage division data;

[0051] Query the stage target crystallization rate range according to the crystallization stage division data to generate the stage target crystallization rate range;

[0052] Calculate the crystallization rate deviation for the stage target crystallization rate range through the predicted crystallization rate curve, and perform crystallization stage sequence processing to generate the crystallization rate deviation sequence data.

[0053] Through more in-depth analysis of the predicted crystallization rate curve, such as generating the crystal size-time curve and the crystal number-time curve, and performing first-order numerical differentiation respectively to obtain the growth rate curve, the present invention can more comprehensively understand the dynamic changes in the oil crystallization process. By aligning the particle size and number growth rate curves in time series and performing extreme point analysis, the key points in the oil crystallization process can be accurately identified, such as the stages of crystal nucleation, rapid growth, and slow growth. Based on the key point data of oil crystallization, the complex crystallization process can be decomposed into several relatively simple stages, and different target crystallization rate ranges can be set for different stages. By comparing the predicted crystallization rate curve with the stage target crystallization rate range, calculating the crystallization rate deviation, and performing crystallization stage sequence processing, the difference between the crystallization process and the expected target can be more accurately quantified, and the deviation for different stages can be adjusted. For example, in the crystal nucleation stage, the target crystallization rate range should be smaller to control the number of crystal nuclei; while in the rapid growth stage, the target crystallization rate range can be appropriately increased to promote crystal growth. This stage-based deviation calculation and adjustment can more precisely control the oil crystallization process and ultimately obtain a more ideal crystal structure and product quality.

[0054] Preferably, step S3 includes the following steps:

[0055] Step S31: Divide the oil viscosity grades according to the oil viscosity data in the reaction kettle to obtain the current oil viscosity grade data;

[0056] Step S32: Set the desired heat transfer coefficient based on the target controlled cooling rate to obtain the desired heat transfer coefficient;

[0057] Step S33: Set the target controlled cooling rate and the current grease viscosity grade data as the input variables of the fuzzy controller, and set the grease stirring speed as the output variable of the fuzzy controller;

[0058] Step S34: Define the fuzzy subsets for the target controlled cooling rate, the current grease viscosity grade data, and the grease stirring speed, respectively obtaining the viscosity grade fuzzy set, the cooling rate fuzzy set, and the stirring speed fuzzy set;

[0059] Step S35: Based on the desired heat transfer coefficient, construct a fuzzy control rule model for the viscosity grade fuzzy set, the cooling rate fuzzy set, and the stirring speed fuzzy set to generate a fuzzy control rule model;

[0060] Step S36: Obtain the real-time stirring speed of the reactor;

[0061] Step S37: Use the fuzzy control rule model to perform fuzzy inference for adjusting the real-time stirring speed, and correct the real-time stirring speed of the reactor to generate the corrected grease stirring speed.

[0062] The present invention takes into account the influence of the viscosity of the grease on stirring, enhancing the pertinence of control. The viscosity of the grease is an important factor affecting the stirring effect, and greases with different viscosities require different stirring strategies. By dividing the grease viscosity grades and using the viscosity grades as the input variables of fuzzy control, the adjustment of the stirring speed can adapt to the change of the grease viscosity. Based on the target control cooling rate, the desired heat transfer coefficient is set, linking the cooling rate with the adjustment of the stirring speed. The desired heat transfer coefficient reflects the requirement for the heat transfer effect, and the heat transfer effect is related to both the cooling rate and the stirring speed. By setting the desired heat transfer coefficient, the target of the cooling rate is linked with the adjustment of the stirring speed, realizing the coordinated control of the two. The fuzzy control method is adopted to realize the intelligent adjustment of the stirring speed. Fuzzy control can handle uncertainty and nonlinear problems and is suitable for complex control objects such as grease crystallization. The target control cooling rate, the grease viscosity grade, and the stirring speed are used as the input and output variables of the fuzzy controller. By defining fuzzy subsets and constructing a fuzzy control rule model, the refinement of fuzzy control is realized. The fuzzy subsets divide the input and output variables into different grades, and the fuzzy control rules express the relationship between these grades in the form of rules. This enables the fuzzy controller to select appropriate rules according to the current input variables and calculate the corresponding adjustment amount of the stirring speed. The real-time stirring speed is corrected using the fuzzy control rule model, realizing the dynamic and adaptive adjustment of the stirring speed. The real-time stirring speed of the reaction kettle is obtained and input into the fuzzy controller, and the corrected stirring speed is calculated according to the current cooling rate and the grease viscosity.

[0063] Preferably, step S4 includes the following steps:

[0064] Step S41: Convert the target control cooling rate into a cooling control instruction and send it to the cooling system to obtain cooling execution feedback data;

[0065] Step S42: Convert the corrected grease stirring speed into a stirring control instruction and send it to the stirring system to obtain stirring execution feedback data;

[0066] Step S43: Check the cooling execution feedback data and the stirring execution feedback data to generate control feedback check data;

[0067] Step S44: When the control feedback check data is abnormal, an alarm is given and the abnormal handling process is entered to obtain abnormal handling record data;

[0068] Step S45: When the control feedback verification data is normal, collect the oil state in the crystallization reactor again according to the preset monitoring period, and conduct crystal quality evaluation. If the oil crystallization meets the set crystallization requirements, store the target control cooling rate, the corrected oil stirring speed, and the control feedback verification data in the production process control record, and generate an oil crystallization process control report; otherwise, iteratively optimize the crystallization control instruction.

[0069] In the present invention, the target control cooling rate and the corrected oil stirring speed are respectively converted into a cooling control instruction and a stirring control instruction, and sent to the corresponding execution systems. The cooling execution feedback data and the stirring execution feedback data are verified, and control feedback verification data is generated, which can timely detect abnormal situations occurring in the control process, such as equipment failures, data anomalies, etc., and perform corresponding processing, ensuring the stability and reliability of the control system. When the control feedback verification data is abnormal, the system will automatically alarm and enter the abnormal handling process, recording the abnormal handling data, which helps to timely discover and solve problems and avoid causing greater losses. For example, if the cooling system fails and the cooling rate cannot reach the target value, the system will immediately alarm, start the standby cooling system or take other emergency measures, and record the fault information at the same time. When the control feedback verification data is normal, the system will collect the oil state in the crystallization reactor again according to the preset monitoring period and conduct crystal quality evaluation. This closed-loop control mechanism can dynamically adjust the control strategy according to the actual situation of oil crystallization to ensure that the final product meets the preset quality requirements. If the oil crystallization does not meet the set crystallization requirements, the system will iteratively optimize the crystallization control instruction, readjust the cooling rate and the stirring speed until the expected crystallization effect is achieved. Storing the target control cooling rate, the corrected oil stirring speed, and the control feedback verification data in the production process control record and generating an oil crystallization process control report can realize the tracking and recording of the entire production process.

[0070] Preferably, the present invention further provides an intelligent digital-based oil product production control system, which executes the intelligent digital-based oil product production control method as described above. The intelligent digital-based oil product production control system includes:

[0071] An oil state sensing module, which is used to monitor the temperature, viscosity, and turbidity data in the oil crystallization process of the oil crystallization reactor in real time, and conduct oil state integration to generate oil crystallization state snapshot data;

[0072] A cooling rate optimization module is used to combine the characteristics of oil crystallization based on the snapshot data of the oil crystallization state, generate an oil crystallization feature vector; construct an oil crystallization kinetics model; use the oil crystallization kinetics model to predict the crystallization state of the oil crystallization feature vector, generate predicted oil crystallization state data; adjust the cooling rate of the reaction kettle through the predicted oil crystallization state data to obtain the target controlled cooling rate.

[0073] A stirring rate optimization module is used to set the desired heat transfer coefficient based on the target controlled cooling rate to obtain the desired heat transfer coefficient; perform fuzzy control variable processing on the snapshot data of the oil crystallization state through the desired heat transfer coefficient, and perform real-time fuzzy inference on the stirring speed adjustment to generate a corrected oil stirring speed.

[0074] A crystallization adaptive control module is used to convert the target controlled cooling rate and the corrected oil stirring speed into control instructions, and execute the production control instructions to obtain control feedback verification data; store the production process control records according to the control feedback verification data to generate an oil crystallization process control report. Description of the Drawings

[0075] Figure 1 It is a schematic diagram of the step flow of the production control method for oil-based products based on intelligent digitization of the present invention;

[0076] Figure 2 is Figure 1 a detailed implementation step flow diagram of step S3 in

[0077] Figure 3 is Figure 1 a detailed implementation step flow diagram of step S4 in

[0078] The realization, functional features and advantages of the object of the present invention will be further described in conjunction with the embodiments with reference to the drawings. Detailed Embodiments

[0079] The technical method of the present invention will be clearly and completely described below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art within the scope of the present invention without creative work based on the embodiments of the present invention belong to the protection scope of the present invention.

[0080] In addition, the attached drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the 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 in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0081] It should be understood that although terms such as "first" and "second" may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. 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.

[0082] To achieve the above object, please refer to Figures 1 to 3 , the present invention provides a production control method for oil-based products based on intelligent digitization, including the following steps:

[0083] Step S1: Real-time monitor the temperature, viscosity, and turbidity data during the oil crystallization process in the oil crystallization reactor, and integrate the oil state to generate oil crystallization state snapshot data;

[0084] Step S2: Combine the oil crystallization characteristics according to the oil crystallization state snapshot data to generate an oil crystallization characteristic vector; construct an oil crystallization kinetic model; use the oil crystallization kinetic model to predict the crystallization state of the oil crystallization characteristic vector to generate oil crystallization prediction state data; adjust the cooling rate of the reactor through the oil crystallization prediction state data to obtain the target control cooling rate;

[0085] Step S3: Set the desired heat transfer coefficient based on the target control cooling rate to obtain the desired heat transfer coefficient; perform fuzzy control variable processing on the oil crystallization state snapshot data through the desired heat transfer coefficient, and perform fuzzy inference on the real-time stirring speed adjustment to generate a corrected oil stirring speed;

[0086] Step S4: Convert the target control cooling rate and the corrected oil stirring speed into control instructions, and execute the production control instructions to obtain control feedback verification data; store the production process control records according to the control feedback verification data to generate an oil crystallization process control report.

[0087] In an embodiment of the present invention, the production control method for oil-based products based on intelligent digitization includes the following steps:

[0088] Step S1: Monitor the temperature, viscosity, and turbidity data during the oil crystallization process in the oil crystallization reactor in real time, and integrate the oil state to generate oil crystallization state snapshot data;

[0089] In an embodiment of the present invention, Pt100 type thermal resistance temperature sensors are installed at different positions (such as the top, middle, and bottom) on the inner wall of the oil crystallization reactor, fixed by screw connection, and connected to the data acquisition module. The rotary viscosity sensor is fixed to the side wall of the reactor by flange connection, and the sensor probe is immersed in the oil. A turbidity sensor composed of a near-infrared light source and a photodetector is fixedly installed on the opposite side walls of the reactor in a transmission measurement mode. The data acquisition module collects the data of the temperature, viscosity, and turbidity sensors with a sampling period of 1 second, converts the analog signal into a digital signal, and transmits it to the upper computer industrial control system through the RS485 bus. The upper computer industrial control system uses a server based on the Windows Server operating system and installs data storage and processing software. The upper computer system preprocesses the received data, including smoothing using a moving average filtering algorithm (window size of 5 seconds), removing abnormal data using the 3σ criterion, and complementing missing data using linear interpolation. After preprocessing, the temperature, viscosity, and turbidity data at each time point are integrated into a three-dimensional vector to form oil crystallization state snapshot data.

[0090] Step S2: Combine the oil crystallization characteristics according to the oil crystallization state snapshot data to generate an oil crystallization characteristic vector; construct an oil crystallization kinetics model; use the oil crystallization kinetics model to predict the crystallization state of the oil crystallization characteristic vector to generate oil crystallization prediction state data; adjust the cooling rate of the reactor through the oil crystallization prediction state data to obtain the target control cooling rate;

[0091] In the embodiment of the present invention, temperature, viscosity, and turbidity data are extracted from the snapshot data of the oil crystallization state obtained in step S1, and the viscosity change rate and turbidity change rate are calculated. The real-time oil supersaturation (queried and calculated from a preset database based on the oil type and real-time temperature), viscosity change rate, and turbidity change rate are combined into an oil crystallization characteristic vector, such as [0.2, 0.01, 1]. Using historical crystallization data and classical crystallization kinetic models such as the Avrami equation, a multivariate linear regression method is used to construct an oil crystallization kinetic model. The oil crystallization characteristic vector at the current moment is input into the kinetic model to predict the oil crystallization degree in the next 10 minutes, generating oil crystallization prediction state data. Using the PID control algorithm, the deviation between the predicted crystallization degree and the target crystallization degree is used as the input to calculate the adjustment amount of the cooling rate. The adjustment amount is superimposed on the current cooling rate (obtained through the cooling water control system of the reactor) to obtain the target control cooling rate.

[0092] Step S3: Based on the target control cooling rate, set the desired heat transfer coefficient to obtain the desired heat transfer coefficient; perform fuzzy control variable processing on the snapshot data of the oil crystallization state through the desired heat transfer coefficient, and perform fuzzy inference on the real-time stirring speed adjustment to generate a corrected oil stirring speed;

[0093] In the embodiment of the present invention, according to the target control cooling rate in step S2, the target control cooling rate and the current oil viscosity grade (discretizing the viscosity data according to a preset standard) are used as the input variables of the fuzzy controller, and the stirring speed adjustment amount is used as the output variable. Define fuzzy subsets and membership functions. For example, the viscosity grade fuzzy set is {low, medium, high}, the cooling rate fuzzy set is {slow, moderate, fast}, and the stirring speed adjustment amount fuzzy set is {negative large, negative small, zero, positive small, positive large}. Based on the desired heat transfer coefficient and expert experience, construct fuzzy rules. For example, IF the viscosity grade is high AND the cooling rate is fast THEN the stirring speed adjustment amount is positive large. Using the Mamdani fuzzy inference method, calculate the stirring speed adjustment amount. The adjustment amount is superimposed on the current stirring speed (read through the frequency converter) to generate a corrected oil stirring speed.

[0094] Step S4: Convert the target control cooling rate and the corrected oil stirring speed into control instructions, and execute the production control instructions to obtain control feedback verification data; store the production process control records according to the control feedback verification data to generate an oil crystallization process control report.

[0095] In the embodiments of the present invention, the target controlled cooling rate is converted into a cooling water valve opening instruction, and the corrected grease stirring speed is converted into a frequency converter frequency instruction. The instructions are sent to the cooling system PLC and the stirring system frequency converter through the industrial Ethernet. The PLC and the frequency converter execute the instructions and return the execution feedback data, such as the actual valve opening and the actual motor speed. The upper computer system verifies the feedback data, such as data range and data consistency verification, and generates control feedback verification data. If the verification data is abnormal, an alarm is triggered and the abnormal information is recorded. If the verification data is normal, the grease state data is collected according to a preset period, and the crystal quality is evaluated (for example, the crystal particle size distribution is measured by an on-line particle size analyzer). If the crystallization quality meets the requirements, the target controlled cooling rate, the corrected grease stirring speed, and the control feedback verification data are stored in the database, and a grease crystallization process control report (including information such as production time, process parameters, control results, and product quality) is generated. Otherwise, return to step S1 for iterative optimization. The database uses MySQL, and the report is generated and stored in PDF format.

[0096] Preferably, step S1 includes the following steps:

[0097] Step S11: Fix and install temperature sensors, viscosity sensors, and turbidity sensors on the grease crystallization reactor to construct a grease crystallization monitoring network;

[0098] Step S12: Use the grease crystallization monitoring network to real-time monitor the temperature, viscosity, and turbidity data during the grease crystallization process to obtain the original grease crystallization monitoring data stream;

[0099] Step S13: Perform monitoring data preprocessing on the original grease crystallization monitoring data stream to obtain preprocessed grease crystallization monitoring data;

[0100] Step S14: Perform time series alignment according to the preprocessed grease crystallization monitoring data, and perform grease state integration to generate grease crystallization state snapshot data.

[0101] In the embodiments of the present invention, drill holes in the wall of the grease crystallization reactor, and use threaded connectors to firmly install the PT100 platinum resistance temperature sensor inside the reactor, ensuring that the sensor probe is in full contact with the grease and does not collide with the stirring blades. The measurement range of the temperature sensor is set to -20°C to 150°C, the accuracy is ±0.1°C, and the sampling frequency is set to 1Hz. Install the rotary viscosity sensor at the bottom of the reactor through flange connection, with the sensor probe immersed in the grease and maintaining a certain distance from the bottom of the reactor to prevent sediment from interfering with the measurement. The measurement range of the viscosity sensor is set to 0.1 Pa·s to 100 Pa·s, the accuracy is ±1%, and the sampling frequency is set to 1Hz. Fix the turbidity sensor composed of a near-infrared light source and a photodetector through the observation window on the side wall of the reactor. The light beam passes through the grease, and the detector receives the transmitted light intensity to achieve turbidity measurement. The measurement range of the turbidity sensor is set to 0 - 1000 NTU, the accuracy is ±2 NTU, and the sampling frequency is set to 1Hz. Connect the above three sensors to the data acquisition module respectively. The data acquisition module is connected to the upper computer industrial control system through the industrial Ethernet protocol to form a grease crystallization monitoring network. The upper computer industrial control system uses Siemens S7-1500 PLC and WinCC configuration software, which is responsible for data reception, storage, and display. During the grease crystallization process, the data acquisition module collects the analog signals of the temperature sensor, viscosity sensor, and turbidity sensor respectively according to the preset sampling frequency (1Hz), and converts the analog signals into digital signals. The data acquisition module transmits the converted digital signals to the upper computer industrial control system in real time through the industrial Ethernet. The WinCC configuration software in the upper computer industrial control system receives and stores the temperature, viscosity, and turbidity data from the data acquisition module in real time, forming an original grease crystallization monitoring data stream indexed by time stamps and containing temperature, viscosity, and turbidity values. The data stream is stored in CSV format, generating one line of data per second, including four columns of data: time stamp, temperature value, viscosity value, and turbidity value. The upper computer industrial control system calls the MATLAB program to preprocess the original grease crystallization monitoring data stream. First, use linear interpolation to supplement the missing data points. Secondly, use the moving average filtering method to smooth the temperature, viscosity, and turbidity data. The window size is set to 5, that is, take the average value of 5 consecutive data points as the value at this time point to eliminate the influence of random noise. Finally, perform outlier rejection on the data, set the reasonable fluctuation range of temperature, viscosity, and turbidity, and consider the data points outside the range as outliers and replace them with the previous valid data point. The upper computer industrial control system calls the Python program to perform time series alignment on the preprocessed grease crystallization monitoring data. Since the sampling frequencies of the three sensors are the same, there is no need to adjust the time stamps. Then, arrange the preprocessed temperature, viscosity, and turbidity data in chronological order, and extract the data once every 10 seconds to generate a snapshot data of the grease crystallization state.Each snapshot data contains a timestamp and the corresponding temperature, viscosity, and turbidity values at that time point.

[0102] Preferably, step S2 includes the following steps:

[0103] Step S21: Extract the grease state in the snapshot reactor based on the snapshot data of the grease crystallization state, and obtain the grease temperature data, grease viscosity data, and grease turbidity data in the reactor respectively;

[0104] Step S22: Combine the grease crystallization characteristics of the grease temperature data, grease viscosity data, and grease turbidity data in the reactor to generate a grease crystallization characteristic vector; construct a crystallization kinetics model based on the grease crystallization characteristic vector to generate a grease crystallization kinetics model;

[0105] Step S23: Use the grease crystallization kinetics model to predict the crystallization state of the grease crystallization characteristic vector to generate grease crystallization prediction state data;

[0106] Step S24: Obtain the real-time cooling rate of the reactor;

[0107] Step S25: Adjust the cooling rate of the reactor's real-time cooling rate through the grease crystallization prediction state data to obtain the target controlled cooling rate.

[0108] In the embodiments of the present invention, a Python program is used to extract the required information from the snapshot data of the grease crystallization state in JSON format. The program traverses each snapshot data, and finally obtains three lists, which respectively contain the grease temperature data of the reactor, the grease viscosity data of the reactor, and the grease turbidity data of the reactor. The data in each list corresponds one-to-one with the time stamp. The NumPy library of Python is used to combine the grease temperature data, viscosity data, and turbidity data of the reactor into a grease crystallization feature vector. Each feature vector contains a time stamp and the temperature, viscosity, and turbidity values at the corresponding time point. All feature vectors are stored in a matrix. The Avrami equation is used as the grease crystallization kinetics model, and the nonlinear least squares fitting function (lmfit) in the SciPy library of Python is used to fit the temperature, viscosity, and turbidity data in the feature vector matrix to the Avrami equation to obtain the parameters of the Avrami equation, such as the crystallization rate constant k and the Avrami exponent n. The finally generated grease crystallization kinetics model is the Avrami equation containing the fitting parameters. The latest grease crystallization feature vector is input into the Avrami kinetics model constructed in step S22 to predict the grease crystallization state in the next period of time (for example, the next 10 minutes). Specifically, the future time point is substituted into the Avrami equation to calculate the corresponding crystallinity. Combining the current temperature, viscosity, and turbidity data, and the predicted crystallinity, the grease crystallization prediction state data is generated. The reactor is equipped with a cooling water jacket, and the reactor temperature is controlled by adjusting the cooling water flow rate. The temperature sensor data of the cooling water inlet and outlet is read through the PLC system, and combined with the cooling water flow meter data, the real-time cooling rate of the reactor is calculated, and the unit is °C / min. According to the grease crystallization prediction state data generated in step S23, it is judged whether the grease crystallization process meets the expectations. For example, if the predicted crystallinity is too high, it indicates that the cooling rate is too fast and the cooling rate needs to be reduced; on the contrary, if the predicted crystallinity is too low, the cooling rate needs to be increased. The PID controller is used to calculate the adjustment amount of the cooling rate according to the deviation between the predicted crystallinity and the target crystallinity. Adding the adjustment amount to the real-time cooling rate, the target control cooling rate is obtained.

[0109] Preferably, step S22 includes the following steps:

[0110] Step S221: Obtain the target grease material parameters; extract the grease type according to the target grease material parameters to obtain the target grease type data;

[0111] Step S222: Extract the grease solubility curve according to the target grease material parameters, and calculate the grease supersaturation at the current temperature by using the grease temperature data of the reactor to generate the real-time grease supersaturation data;

[0112] Step S223: Perform time differentiation on the grease viscosity data and grease turbidity data of the reactor, and calculate the change rates respectively to obtain the grease viscosity change rate and grease turbidity change rate of the reactor;

[0113] Step S224: Combine the real-time grease supersaturation data, the grease viscosity change rate of the reactor, and the grease turbidity change rate of the reactor to generate a grease crystallization feature vector;

[0114] Step S225: Construct a crystallization kinetics model based on the target grease material parameters through the grease crystallization feature vector to obtain the grease crystallization kinetics model.

[0115] In the embodiments of the present invention, target grease material parameters are input through the human-machine interface (HMI) of the host computer industrial control system, including grease type, fatty acid composition, iodine value, etc. For example, the input target grease material parameters are: {"grease type": "palm oil", "fatty acid composition": "C16:0 45%, C18:0 5%, C18:1 40%", "iodine value": 55}. The system extracts the target grease type data according to the "grease type" field input. For example, "palm oil". According to the target grease type data obtained in step S221, the corresponding grease solubility curve data is retrieved from the pre-established grease database. The grease database stores the solubility data of different types of grease at different temperatures in tabular form. For example, the solubility of palm oil at 25°C is 0.1 g / mL, and the solubility at 30°C is 0.2 g / mL. Using the interpolation function in the SciPy library of Python, a continuous solubility curve function is constructed based on the discrete data points in the database. According to the reactor grease temperature data obtained in step S21, the current temperature is substituted into the solubility curve function to calculate the grease saturation solubility at the current temperature. Assuming the current grease concentration is 1 g / mL and the current temperature is 25°C, the grease saturation solubility at the current temperature is 0.1 g / mL, and the real-time grease supersaturation is calculated as (1 - 0.1) / 0.1 = 9. The supersaturation calculation results at each time point are stored in the real-time grease supersaturation data list. The time differentiation of the reactor grease viscosity data and turbidity data obtained in step S21 is performed using the difference function in the NumPy library of Python. For example, assuming the viscosity data at two adjacent time points are 0.75 Pa·s and 0.8 Pa·s respectively, and the time interval is 10 seconds, then the viscosity change rate is (0.8 - 0.75) / 10 = 0.005 Pa·s / s. Similar calculations are performed for each time point to obtain the reactor grease viscosity change rate list and the reactor grease turbidity change rate list respectively. The real-time grease supersaturation data calculated in step S222, the reactor grease viscosity change rate calculated in step S223, and the reactor grease turbidity change rate are combined into a three-dimensional vector in chronological order to form the grease crystallization characteristic vector. For example, at a certain time point, the grease supersaturation is 0.2, the viscosity change rate is 0.01 Pa·s / min, and the turbidity change rate is 1 NTU / min, then the grease crystallization characteristic vector at this time point is [0.2, 0.01, 1]. According to the target grease material parameters obtained in step S221, a suitable crystallization kinetics model is selected. For example, for palm oil, the improved Avrami equation is selected, which takes into account the effects of supersaturation, viscosity, and turbidity on the crystallization process. Using the curve fitting function in the SciPy library of Python, the grease crystallization characteristic vector generated in step S224 is fitted into the improved Avrami equation to obtain the parameters of the model. For example, the fitted improved Avrami equation is: where C(t) is the crystallinity, k is the crystallization rate constant, S is the supersaturation, η is the viscosity, η 0 is the initial viscosity, τ is the turbidity, τ 0 is the initial turbidity, t is the time, and m, p, q, and n are fitting parameters. The finally obtained oil crystallization kinetic model is the improved Avrami equation containing the fitting parameters.

[0116] Preferably, step S225 includes the following steps:

[0117] Analyze the crystallization characteristics according to the target oil type data, and screen the classical crystallization kinetic model according to the classical crystallization kinetic theory to obtain the alternative oil crystallization models;

[0118] Evaluate the applicability of the alternative oil crystallization models to generate the oil crystallization model structure data;

[0119] Determine the undetermined parameters of the oil crystallization model structure data through the oil crystallization eigenvectors to generate a list of model undetermined parameters;

[0120] Query the crystallization experiment data based on the list of model undetermined parameters using the preset historical crystallization experiment database to obtain the historical oil crystallization experiment data;

[0121] Extract the model training features from the historical oil crystallization experiment data to obtain the model training feature data;

[0122] Divide the model training feature data into training data and validation data to obtain the model training dataset and the model validation dataset;

[0123] Optimize the parameters of the list of model undetermined parameters based on the model training dataset using the preset particle swarm optimization algorithm, and train the alternative oil crystallization models to obtain the preliminary trained crystallization model;

[0124] Perform cross-validation on the preliminary trained crystallization model through the model validation dataset, and perform hyperparameter tuning to obtain the oil crystallization kinetic model.

[0125] In an embodiment of the present invention, it is assumed that the target oil type data is "palm oil". First, consult literature materials and oil databases to analyze the crystallization characteristics of palm oil, such as crystallization type (polymorphism), crystallization rate, crystallization temperature range, etc. Then, according to the classical crystallization kinetics theory, screen the crystallization kinetics models applicable to palm oil. For example, considering the polymorphic characteristics of palm oil, select the Avrami equation and its improved models as alternative models, and at the same time consider models such as the Ozawa model and the Lauricella model to form a set of alternative oil crystallization models. For each alternative oil crystallization model, evaluate its applicability according to the crystallization characteristics of palm oil. For example, analyze whether the model can describe polymorphic transformation and whether it can reflect the induction period in the oil crystallization process. According to the evaluation results, select the most suitable model structure and determine the specific form of the model to generate oil crystallization model structure data. According to the selected oil crystallization model structure data (such as the improved Avrami equation), determine the undetermined parameters in the model. For example, the undetermined parameters in the improved Avrami equation are [k, m, p, q, n]. According to the target oil type data (palm oil), query relevant experimental data from the preset historical crystallization experiment database. The historical crystallization experiment database contains crystallization experiment data of different types of oils under different conditions, such as temperature, time, crystallinity, viscosity, turbidity, etc. The query result is historical oil crystallization experiment data, such as a series of experimental data points including time, temperature, crystallinity, viscosity, and turbidity. According to the oil crystallization model structure data (such as the improved Avrami equation) determined in step 2, extract the feature data required for model training from the historical oil crystallization experiment data. For example, extract data such as time, supersaturation, viscosity change rate, turbidity change rate, and crystallinity as model training feature data. Divide the model training feature data obtained in step 5 into a model training data set and a model validation data set according to a certain ratio (such as 8:2). For example, use 80% of the data for model training and 20% of the data for model validation. Use the preset particle swarm optimization algorithm, combined with the model training data set, to optimize the list of model undetermined parameters determined in step 3 to find the best parameter values, so that the prediction results of the model best fit the experimental data. Substitute the optimized parameters into the oil crystallization model structure data to obtain a preliminary trained crystallization model. Use the model validation data set to perform cross-validation on the preliminary trained crystallization model to evaluate the generalization ability of the model. According to the validation results, adjust the hyperparameters of the model (such as the parameters of the particle swarm optimization algorithm) to further optimize the model performance. Finally, obtain the oil crystallization kinetics model, that is, the improved Avrami equation containing optimized parameters.

[0126] Preferably, step S25 includes the following steps:

[0127] Step S251: Perform predictive crystallization rate processing based on the predicted state data of oil crystallization to generate a predicted crystallization rate curve;

[0128] Step S252: Identify the key points of crystallization according to the predicted crystallization rate curve, calculate the deviation of the oil crystallization rate, and generate the sequence data of the crystallization rate deviation;

[0129] Step S253: Calculate the preliminary adjustment amount of the cooling rate for the real-time cooling rate of the reaction kettle by using the sequence data of the crystallization rate deviation to generate the preliminary cooling rate adjustment amount;

[0130] Step S254: Perform safety limiting processing on the preliminary cooling rate adjustment amount and perform temperature compensation through the oil temperature data of the reaction kettle to generate the target controlled cooling rate.

[0131] In the embodiment of the present invention, the crystallinity data within a future period of time (e.g., the next 10 minutes) is extracted from the predicted state data of the oil crystallization generated in step S23. The NumPy library of Python is used to perform numerical differentiation on the crystallinity data to calculate the crystallization rate at each time point. The time and the corresponding crystallization rate data points are plotted as a curve to generate a predicted crystallization rate curve. For example, if the crystallinities at the t-th minute and the (t + 1)-th minute are 0.2 and 0.25 respectively, the crystallization rate at the (t + 1)-th minute is (0.25 - 0.2) / 1 = 0.05 / min. According to the preset crystallization rate threshold, the crystallization key points are marked on the predicted crystallization rate curve. For example, if the crystallization rate threshold is set to 0.03 / min, when the predicted crystallization rate exceeds this threshold, this time point is marked as a crystallization key point. The predicted crystallization rate of each crystallization key point is compared with the target crystallization rate to calculate the crystallization rate deviation. The target crystallization rate can be preset according to the product quality requirements, for example, set to 0.04 / min. The crystallization rate deviations of each crystallization key point are arranged in chronological order to generate crystallization rate deviation sequence data. Using a PID controller, the preliminary adjustment amount of the cooling rate is calculated according to the crystallization rate deviation sequence data. The PID controller calculates the proportional, integral, and differential terms according to the magnitude and change trend of the deviation value, and adds the three terms to obtain the preliminary adjustment amount of the cooling rate. For example, if the current crystallization rate deviation is -0.02 / min, and the proportional coefficient, integral coefficient, and differential coefficient of the PID controller are Kp = 0.5, Ki = 0.1, and Kd = 0.01 respectively, the preliminary adjustment amount of the cooling rate is -0.02×0.5 + (cumulative value of historical deviations)×0.1 + (current deviation change rate)×0.01. To ensure safety and prevent over-regulation, a safety limit treatment is performed on the preliminary cooling rate adjustment amount calculated in step S253. The upper and lower limits of the cooling rate adjustment amount are set, for example, the upper limit is 0.1℃ / min and the lower limit is -0.1℃ / min. If the preliminary cooling rate adjustment amount exceeds the safety limit range, it is restricted within the range. Then, temperature compensation is performed according to the oil temperature data of the reaction kettle. For example, if the temperature of the reaction kettle is too low, the cooling rate adjustment amount is appropriately reduced, and vice versa. The cooling rate adjustment amount after the safety limit treatment and temperature compensation is added to the real-time cooling rate of the reaction kettle obtained in step S24 to obtain the final target control cooling rate.

[0132] Preferably, step S252 includes the following steps:

[0133] Based on the predicted crystallization rate curve, the crystal size-time curve and the crystal number-time curve are analyzed to obtain the oil crystallization size-time curve and the oil crystallization number-time curve respectively;

[0134] The first-order numerical differentiation is performed on the oil crystal particle size-time curve and the oil crystal number-time curve respectively to obtain the particle size growth rate curve and the number growth rate curve;

[0135] The particle size growth rate curve and the number growth rate curve are aligned in time series, and the extreme point analysis is carried out to obtain the key point data of oil crystal;

[0136] Based on the key point data of oil crystal, the crystallization stage is divided to obtain the crystallization stage division data;

[0137] According to the crystallization stage division data, the range of the target crystallization rate in each stage is queried to generate the range of the target crystallization rate in each stage;

[0138] The crystallization rate deviation is calculated for the range of the target crystallization rate in each stage through the predicted crystallization rate curve, and the crystallization stage sequence is processed to generate the crystallization rate deviation sequence data.

[0139] In the embodiment of the present invention, the host computer industrial control system calls the data analysis module, and based on the known oil crystallization kinetics model and the predicted crystallization rate curve, deduces the crystal size-time curve and the crystal number-time curve. For example, using the Population Balance Model (PBM), combined with the oil crystallization kinetics parameters (such as nucleation rate, growth rate), the crystal size distribution and crystal number at different times are obtained through numerical simulation. Taking the calculation results with time as the abscissa, and the particle size and number as the ordinates respectively, the oil crystallization particle size-time curve and the oil crystallization number-time curve are plotted. The host computer industrial control system calls the numerical calculation module, and uses the difference method to perform the first-order numerical differentiation on the oil crystallization particle size-time curve and the oil crystallization number-time curve respectively. For example, for the particle size-time curve, the particle size growth rate is calculated using the formula (particle size at t2 - particle size at t1) / (t2 - t1), where t1 and t2 are two adjacent times, and the particle size at t1 and the particle size at t2 are the crystal sizes corresponding to t1 and t2 respectively. Similarly, the numerical differentiation is performed on the number-time curve to obtain the number growth rate curve. The host computer industrial control system calls the data analysis module to ensure that the particle size growth rate curve and the number growth rate curve are aligned in time. Using the peak detection algorithm, for example, finding the local maximum, local minimum or inflection point of the curve, the extreme point analysis is performed on the particle size growth rate curve and the number growth rate curve respectively. The detected extreme points (time and corresponding rate values) are used as the key point data of oil crystallization. The host computer industrial control system calls the expert rule base, and based on the preset rules and the key point data of oil crystallization, divides the oil crystallization process into different stages, such as: induction period, rapid growth qi, slow growth qi. The division rules can be set according to the time when the extreme points appear, the magnitude of the rate value, the rate change trend, etc. For example, it can be defined that the period before the number growth rate reaches the first peak is the induction period, the period from then until the particle size growth rate reaches the peak is the rapid growth qi, and the period after that is the slow growth qi. The start and end times of each stage are recorded as the crystallization stage division data. The host computer industrial control system accesses the preset process parameter database, and according to the crystallization stage division data, queries the target crystallization rate range corresponding to each crystallization stage. The database stores the optimal crystallization rate ranges of different oil types in different crystallization stages. The host computer industrial control system calls the data processing module to compare the predicted crystallization rate curve with the stage target crystallization rate range. For each time point, it is judged whether the predicted crystallization rate falls within the target crystallization rate range of the corresponding crystallization stage. If it exceeds the range, the difference between the predicted crystallization rate and the central value of the target range is calculated as the crystallization rate deviation. The crystallization rate deviation at each time point is associated with the corresponding timestamp and crystallization stage to generate the crystallization rate deviation sequence data.

[0140] As an example of the present invention, refer to Figure 2As shown, Figure 1 is a schematic diagram of the detailed implementation steps of step S3 in Figure 1 . In this example, step S3 includes:

[0141] Step S31: Classify the grease viscosity grades according to the grease viscosity data of the reactor to obtain the current grease viscosity grade data;

[0142] In the embodiment of the present invention, a Python program is used to classify the grease viscosity data obtained in step S21. The viscosity grade classification standard is preset. For example, the viscosity lower than 0.5 Pa·s is classified as the "low viscosity" grade, 0.5 Pa·s to 2 Pa·s is classified as the "medium viscosity" grade, and higher than 2 Pa·s is classified as the "high viscosity" grade. The program traverses the list of grease viscosity data of the reactor, judges the viscosity grade to which each viscosity value belongs according to the viscosity value at each time point, and stores the result in the current grease viscosity grade data list. For example, if the viscosity value at a certain time point is 1.2 Pa·s, it is classified as the "medium viscosity" grade and recorded as "medium viscosity" in the current grease viscosity grade data list.

[0143] Step S32: Set the desired heat transfer coefficient based on the target controlled cooling rate to obtain the desired heat transfer coefficient;

[0144] In the embodiment of the present invention, the desired heat transfer coefficient is set according to the target controlled cooling rate calculated in step S25. The target controlled cooling rate reflects the heat removal requirement during the grease crystallization process, while the desired heat transfer coefficient determines the heat transfer capacity of the cooling system. By establishing a mapping relationship between the target controlled cooling rate and the desired heat transfer coefficient, precise control of the cooling process can be achieved. For example, a functional relationship between the target controlled cooling rate and the desired heat transfer coefficient can be established according to an empirical formula or a heat transfer model. Suppose the target controlled cooling rate is 0.2 °C / min, and according to the pre-established functional relationship, the calculated desired heat transfer coefficient is 150 W / (m 2 ·K).

[0145] Step S33: Set the target controlled cooling rate and the current grease viscosity grade data as the input variables of the fuzzy controller, and set the grease stirring rotation speed as the output variable of the fuzzy controller;

[0146] In the embodiment of the present invention, a fuzzy controller is constructed, with the target controlled cooling rate and the current grease viscosity grade data as the input variables of the fuzzy controller, and the grease stirring rotation speed as the output variable of the fuzzy controller. The function of the fuzzy controller is to automatically adjust the stirring speed according to the viscosity of the grease and the target cooling rate to achieve the best crystallization effect. The Mamdani type fuzzy inference system is selected as the fuzzy controller.

[0147] Step S34: Define fuzzy subsets for the target controlled cooling rate, the current grease viscosity grade data, and the grease stirring speed quantity, respectively obtaining the viscosity grade fuzzy set, the cooling rate fuzzy set, and the stirring speed quantity fuzzy set;

[0148] In the embodiment of the present invention, fuzzy subsets are defined for the input variables and output variables of the fuzzy controller. For the current grease viscosity grade data, three fuzzy subsets are defined: low, medium, and high. For example, the membership function of the "low" fuzzy subset can be defined as a trapezoidal function, with a membership degree of 1 when the viscosity is below 0.5 Pa·s and a membership degree of 0 when the viscosity is above 1 Pa·s. For the target controlled cooling rate, three fuzzy subsets are defined: slow, moderate, and fast. For example, the membership function of the "slow" fuzzy subset can be defined as a trigonometric function with a central value of 0.1 °C / min. For the grease stirring speed quantity, three fuzzy subsets are defined: low, medium, and high. For example, the membership function of the "low" fuzzy subset can be defined as a Gaussian function with a central value of 30 rpm. The membership functions of these fuzzy subsets can be adjusted according to the actual situation. These fuzzy subsets are respectively composed into the viscosity grade fuzzy set, the cooling rate fuzzy set, and the stirring speed quantity fuzzy set for the inference process of the fuzzy controller.

[0149] Step S35: Based on the desired heat transfer coefficient, construct a fuzzy control rule model for the viscosity grade fuzzy set, the cooling rate fuzzy set, and the stirring speed quantity fuzzy set, generating a fuzzy control rule model;

[0150] In the embodiment of the present invention, according to the heat transfer mechanism and control experience of the grease crystallization process, a fuzzy control rule model is established. The fuzzy rules associate the fuzzy subsets of the input variables with the fuzzy subsets of the output variables. The desired heat transfer coefficient affects the contribution of the stirring speed to the heat transfer effect, so the fuzzy rules need to be adjusted according to the desired heat transfer coefficient. For example, when the desired heat transfer coefficient is high, the influence of the stirring speed on the heat transfer effect is greater, and the stirring speed needs to be adjusted more precisely. An example rule can be: If the viscosity grade is "high" and the cooling rate is "fast", then the stirring speed is "high". Another example rule can be: If the viscosity grade is "low" and the cooling rate is "slow", then the stirring speed is "low". According to different desired heat transfer coefficients, these rules can be adjusted. For example, when the desired heat transfer coefficient is very high, the rule can be modified to: If the viscosity grade is "high" and the cooling rate is "fast", then the stirring speed is "extremely high". All the rules are combined to form a fuzzy control rule model, which is stored in the fuzzy controller in the form of if-then statements for subsequent fuzzy inference.

[0151] Step S36: Obtain the real-time stirring speed of the reaction kettle;

[0152] In the embodiment of the present invention, the reaction kettle is equipped with a stirrer driven by a variable-frequency motor, and the stirring speed is controlled by a frequency converter. The output frequency of the frequency converter is read through the PLC system, and the real-time stirring speed of the reaction kettle is calculated according to the structural parameters of the stirrer (such as blade diameter, number of blades, etc.). For example, assuming that the output frequency of the frequency converter is 50 Hz and the reduction ratio of the stirrer is 10:1, the real-time stirring speed is 50×60 / 10 = 300 rpm.

[0153] Step S37: Use the fuzzy control rule model to perform fuzzy inference on the real-time stirring speed adjustment, and correct the real-time stirring speed of the reaction kettle to generate a corrected grease stirring speed.

[0154] In the embodiment of the present invention, the current grease viscosity grade data obtained in step S31 and the target control cooling rate calculated in step S25 are used as inputs and input into the fuzzy control rule model constructed in step S35 for fuzzy inference. The fuzzy inference process includes four steps: fuzzification, rule matching, fuzzy inference, and defuzzification. First, the specific numerical values of the input variables are converted into the membership degrees of fuzzy subsets. Then, according to the rules in the fuzzy rule model, the activation degree of each rule is calculated. Next, the activated rules are subjected to fuzzy inference to obtain the fuzzy subset of the output variable. Finally, the fuzzy subset of the output variable is converted into a specific numerical value, that is, the stirring speed adjustment amount. The adjustment amount is added to the real-time stirring speed of the reaction kettle obtained in step S36 to obtain the corrected grease stirring speed. For example, assuming that the stirring speed adjustment amount obtained by fuzzy inference is +20 rpm and the real-time stirring speed of the reaction kettle is 300 rpm, the corrected grease stirring speed is 300 + 20 = 320 rpm. The corrected stirring speed command is sent to the frequency converter to control the speed of the stirring motor, realizing precise control of the grease stirring speed.

[0155] As an example of the present invention, refer to Figure 3 shown in Figure 1 is a detailed implementation step flow diagram of step S4 in

[0156] Step S41: Convert the target control cooling rate into a cooling control command and send it to the cooling system to obtain cooling execution feedback data;

[0157] In the embodiment of the present invention, the upper computer industrial control system converts the numerical target control cooling rate generated in step S25 into a control command that can be recognized by the cooling system. For example, the target control cooling rate is converted into the opening value of the cooling water valve. The upper computer sends the control command to the PLC controller of the cooling system through the Modbus TCP protocol. After the cooling system executes the control command, data such as the actual opening of the cooling water valve, the cooling water flow rate, and the cooling water temperature are fed back to the upper computer to form cooling execution feedback data.

[0158] Step S42: Convert the stirring control instruction for the corrected grease stirring speed and send it to the stirring system to obtain stirring execution feedback data;

[0159] In the embodiment of the present invention, the upper computer industrial control system converts the numerical corrected grease stirring speed generated in step S37 into a control instruction recognizable by the stirring system. For example, the corrected grease stirring speed is converted into the output frequency value of the frequency converter. The upper computer sends the control instruction to the frequency converter of the stirring system through the Profibus-DP protocol. After the stirring system executes the control instruction, data such as the actual output frequency of the frequency converter and the actual rotation speed of the stirring motor are fed back to the upper computer to form stirring execution feedback data.

[0160] Step S43: Perform data verification on the cooling execution feedback data and the stirring execution feedback data to generate control feedback verification data;

[0161] In the embodiment of the present invention, the upper computer industrial control system verifies the received cooling execution feedback data and stirring execution feedback data to determine whether the data is valid. The verification content includes data range, data consistency, data integrity, etc. For example, check whether the opening of the cooling water valve is between 0% and 100%, check whether the output frequency of the frequency converter is consistent with the set frequency range, and check whether the data is complete without omission. The verification result generates control feedback verification data. For example, if all data meet the requirements, the control feedback verification data is "normal"; if there is abnormal data, the control feedback verification data is "abnormal", and the specific abnormal information is recorded.

[0162] Step S44: When the control feedback verification data is abnormal, an alarm is given and the system enters the abnormal handling process to obtain abnormal handling record data;

[0163] In the embodiment of the present invention, if the control feedback verification data generated in step S43 is "abnormal", the upper computer industrial control system immediately gives an audible and visual alarm to remind the operator to pay attention. At the same time, the system enters the abnormal handling process and is processed according to the pre-set abnormal handling plan. For example, if the opening of the cooling water valve exceeds the range, the cooling water valve is immediately closed, and information such as the time of occurrence of the abnormality, the type of abnormality, and the handling measures is recorded to generate abnormal handling record data.

[0164] Step S45: When the control feedback verification data is normal, re-collect the grease state in the crystallization reactor according to the preset monitoring period and perform crystal quality evaluation. If the grease crystallization meets the set crystallization requirements, store the target control cooling rate, the corrected grease stirring speed, and the control feedback verification data for production process control record to generate a grease crystallization process control report; otherwise, iteratively optimize the crystallization control instruction.

[0165] In the embodiment of the present invention, if the control feedback verification data generated in step S43 is "normal", the host computer industrial control system re-collects the grease state data in the crystallization reactor according to a preset monitoring period (for example, 10 minutes), including temperature, viscosity, turbidity, etc. The quality of the grease crystals is evaluated by using image analysis technology or other methods. For example, indicators such as the size, shape, and distribution of the crystals are evaluated. If the grease crystallization meets the set crystallization requirements, for example, the crystallinity reaches 90% and the crystal particle size distribution is uniform, the target control cooling rate, the corrected grease stirring speed, and the control feedback verification data are stored in the production process control record database, and a grease crystallization process control report is generated. If the grease crystallization does not meet the set crystallization requirements, return to step S2, re-perform the crystallization kinetics model prediction and control instruction calculation based on the latest grease state data, and iteratively optimize the crystallization control instruction until the grease crystallization meets the requirements.

[0166] Preferably, the present invention also provides a production control system for grease products based on intelligent digitization, which executes the production control method for grease products based on intelligent digitization as described above. The production control system for grease products based on intelligent digitization includes:

[0167] A grease state sensing module for real-time monitoring of the temperature, viscosity, and turbidity data during the grease crystallization process in the grease crystallization reactor, and integrating the grease state to generate grease crystallization state snapshot data;

[0168] A cooling rate optimization module for combining the grease crystallization characteristics according to the grease crystallization state snapshot data to generate a grease crystallization characteristic vector; constructing a grease crystallization kinetics model; using the grease crystallization kinetics model to predict the crystallization state of the grease crystallization characteristic vector to generate grease crystallization prediction state data; adjusting the cooling rate of the reactor through the grease crystallization prediction state data to obtain the target control cooling rate;

[0169] A stirring rate optimization module for setting the desired heat transfer coefficient based on the target control cooling rate to obtain the desired heat transfer coefficient; performing fuzzy control variable processing on the grease crystallization state snapshot data through the desired heat transfer coefficient, and performing real-time stirring speed adjustment fuzzy reasoning to generate a corrected grease stirring speed;

[0170] A crystallization adaptive control module for converting the target control cooling rate and the corrected grease stirring speed into control instructions, and executing the production control instructions to obtain control feedback verification data; storing the production process control record according to the control feedback verification data to generate a grease crystallization process control report.

[0171] This application is based on snapshot data of the crystal state of oil and fat, constructs a kinetic model of oil and fat crystallization, and realizes the prediction and control of the crystallization process. By extracting the characteristic vectors of oil and fat crystallization and combining with the classical crystallization kinetics theory, a crystallization kinetics model suitable for the target oil and fat type is screened and constructed. This model can predict the future state of oil and fat crystallization according to real-time monitoring data, such as crystal size, distribution uniformity, etc. Based on the prediction results, the system can dynamically adjust the cooling rate of the reaction kettle to ensure the size and uniform distribution of oil and fat crystals, thereby significantly improving the quality and stability of oil and fat products. For example, in the crystal nucleation stage, the system will automatically adjust the cooling rate to control the number of crystal nuclei; in the rapid growth stage, the system will appropriately increase the cooling rate to promote crystal growth. This prediction-based dynamic adjustment avoids the impact of too fast or too slow cooling rate on the quality of oil and fat. Through the fuzzy control method, the intelligent adjustment of the stirring speed is realized. According to the viscosity grade of the oil and fat and the target control cooling rate, the desired heat transfer coefficient is set, and the stirring speed is adjusted in real time through the fuzzy controller. This fuzzy control method can handle the uncertainties and nonlinear problems in the oil and fat crystallization process, making the adjustment of the stirring speed more adaptable to the changes in oil and fat viscosity and cooling rate. For example, when the oil and fat viscosity is high, the system will automatically increase the stirring speed to prevent oil and fat stratification and uneven crystallization; when the crystallization process tends to be stable, the system will reduce the stirring speed to avoid damaging the formed crystal structure. Based on the target control cooling rate, the desired heat transfer coefficient is set, linking the adjustment of the cooling rate and the stirring speed, and realizing the coordinated control of the two. This coordinated control method can better optimize the oil and fat crystallization process and improve the product quality.

[0172] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be included in the present invention.

[0173] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can 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 these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A production control method for oil and fat products based on intelligent digitalization, characterized in that: The following steps are involved: Step S1: monitoring the temperature, viscosity and turbidity data of the oil crystallization reaction kettle in real time during the oil crystallization process, integrating the oil state, and generating snapshot data of the oil crystallization state; Step S2: combining the oil crystallization characteristics according to the oil crystallization state snapshot data to generate an oil crystallization feature vector; constructing an oil crystallization kinetic model; using the oil crystallization kinetic model to predict the crystallization state of the oil crystallization feature vector to generate oil crystallization prediction state data; adjusting the cooling rate of the reactor according to the oil crystallization prediction state data to obtain a target controlled cooling rate; Step S3: setting the expected heat transfer coefficient based on the target controlled cooling rate to obtain the expected heat transfer coefficient; performing fuzzy control variable processing on the snapshot data of the grease crystallization state through the expected heat transfer coefficient, and performing fuzzy reasoning for real-time stirring speed adjustment to generate a corrected grease stirring speed; Step S4: convert the target control cooling rate and the corrected oil stirring speed into control instructions, and execute the production control instructions to obtain control feedback verification data; store the production process control records according to the control feedback verification data, and generate an oil crystallization process control report.

2. The method for controlling the production of oil and fat products based on intelligent digitization according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Fixing and installing a temperature sensor, a viscosity sensor and a turbidity sensor on the oil crystallization reactor to construct an oil crystallization monitoring network; Step S12: using the oil crystallization monitoring network to monitor the temperature, viscosity and turbidity data in the oil crystallization process in real time, and obtaining the original oil crystallization monitoring data stream; Step S13: preprocessing the original oil crystallization monitoring data stream to obtain preprocessed oil crystallization monitoring data; Step S14: performing time series alignment and oil state integration based on the pre-processed oil crystallization monitoring data to generate oil crystallization state snapshot data.

3. The method for controlling the production of oil and fat products based on intelligent digitalization according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: extracting the snapshot reactor grease state according to the grease crystallization state snapshot data, and obtaining reactor grease temperature data, reactor grease viscosity data, and reactor grease turbidity data respectively; Step S22: combining the reaction kettle grease temperature data, the reaction kettle grease viscosity data, and the reaction kettle grease turbidity data with grease crystallization characteristics to generate a grease crystallization characteristic vector; constructing a crystallization kinetic model based on the grease crystallization characteristic vector to generate a grease crystallization kinetic model; Step S23: using the oil crystallization kinetics model to predict the crystallization state of the oil crystallization feature vector to generate oil crystallization prediction state data; Step S24: obtaining the real-time cooling rate of the reactor; Step S25: adjusting the real-time cooling rate of the reactor according to the predicted state data of oil crystallization to obtain a target controlled cooling rate.

4. The method for controlling the production of oil and fat products based on intelligent digitalization according to claim 3 is characterized in that: Step S22 includes the following steps: Step S221: Obtain target oil material parameters; extract oil types according to the target oil material parameters to obtain target oil type data; Step S222: extracting the oil solubility curve according to the target oil material parameters, and calculating the oil supersaturation at the current temperature using the oil temperature data of the reactor to generate real-time oil supersaturation data; Step S223: performing time differentiation on the reaction kettle grease viscosity data and the reaction kettle grease turbidity data, and respectively performing change rate calculations to obtain the reaction kettle grease viscosity change rate and the reaction kettle grease turbidity change rate respectively; Step S224: performing feature combination on the real-time oil supersaturation data, the oil viscosity change rate of the reactor and the oil turbidity change rate of the reactor to generate an oil crystallization feature vector; Step S225: constructing a crystallization kinetics model based on the target oil material parameters through the oil crystallization characteristic vector to obtain the oil crystallization kinetics model.

5. The method for controlling the production of oil and fat products based on intelligent digitization according to claim 4 is characterized in that: Step S225 includes the following steps: Analyze the crystallization characteristics according to the target oil type data, and screen the classical crystallization kinetics model according to the classical crystallization kinetics theory to obtain the candidate oil crystallization model; Evaluate the applicability of the candidate models for oil crystallization and generate structural data of the oil crystallization model; Determine the undetermined parameters of the oil crystallization model structure data through the oil crystallization characteristic vector, and generate a model undetermined parameter list; Based on the model parameter list, a preset historical crystallization experiment database is used to query the crystallization experiment data to obtain the historical oil crystallization experiment data; Perform model training feature extraction on historical oil crystallization experimental data to obtain model training feature data; Divide the model training feature data into training data and verification data to obtain a model training data set and a model verification data set; Based on the model training data set, the preset particle swarm algorithm is used to optimize the parameters of the model's pending parameter list, and the oil crystallization candidate model is trained to obtain a preliminary training crystallization model; The preliminary training crystallization model was cross-validated using the model validation data set, and hyperparameters were tuned to obtain the oil crystallization kinetics model.

6. The method for controlling the production of oil and fat products based on intelligent digitalization according to claim 3 is characterized in that: Step S25 includes the following steps: Step S251: performing predicted crystallization rate processing according to the predicted state data of oil crystallization to generate a predicted crystallization rate curve; Step S252: marking key crystallization points according to the predicted crystallization rate curve, and calculating the crystallization rate deviation of the oil to generate crystallization rate deviation sequence data; Step S253: using the crystallization rate deviation sequence data to calculate the preliminary cooling rate adjustment amount for the real-time cooling rate of the reactor, and generating a preliminary cooling rate adjustment amount; Step S254: Perform safety limiting processing on the preliminary cooling rate adjustment amount, and perform temperature compensation through the reactor grease temperature data to generate a target controlled cooling rate.

7. The method for controlling the production of oil and fat products based on intelligent digitization according to claim 6 is characterized in that: Step S252 The following steps are involved: Based on the predicted crystallization rate curve, the crystal size-time curve and the crystal number-time curve are analyzed to obtain the oil crystal size-time curve and the oil crystal number-time curve respectively; The first-order numerical differentiation of the oil crystal particle size-time curve and the oil crystal quantity-time curve was performed to obtain the particle size growth rate curve and the quantity growth rate curve; The particle size growth rate curve and the number growth rate curve are aligned in time series, and extreme point analysis is performed to obtain the key point data of oil crystallization; Based on the key point data of oil crystallization, the crystallization stage is divided to obtain the crystallization stage division data; According to the crystallization stage division data, the stage target crystallization rate range is queried to generate the stage target crystallization rate range; The crystallization rate deviation is calculated for the stage target crystallization rate range by predicting the crystallization rate curve, and the crystallization stage sequence is processed to generate crystallization rate deviation sequence data.

8. The method for controlling the production of oil and fat products based on intelligent digitalization according to claim 3 is characterized in that: Step S3 includes the following steps: Step S31: classifying the oil viscosity grade according to the oil viscosity data of the reactor to obtain the current oil viscosity grade data; Step S32: setting a desired heat transfer coefficient based on the target controlled cooling rate to obtain a desired heat transfer coefficient; Step S33: setting the target controlled cooling rate and the current grease viscosity grade data as the fuzzy controller input variables, and setting the grease stirring speed as the fuzzy controller output variable; Step S34: fuzzy subsets are defined for the target controlled cooling rate, the current oil viscosity grade data and the oil stirring speed, and the viscosity grade fuzzy set, the cooling rate fuzzy set and the stirring speed fuzzy set are obtained respectively; Step S35: constructing a fuzzy control rule model for the viscosity grade fuzzy set, the cooling rate fuzzy set and the stirring speed fuzzy set based on the expected heat transfer coefficient to generate a fuzzy control rule model; Step S36: obtaining the real-time stirring speed of the reactor; Step S37: using the fuzzy control rule model to perform fuzzy reasoning for real-time stirring speed adjustment, and correcting the real-time stirring speed of the reactor to generate a corrected oil stirring speed.

9. The method for controlling the production of oil and fat products based on intelligent digitalization according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: converting the target control cooling rate into a cooling control instruction and sending it to the cooling system to obtain cooling execution feedback data; Step S42: converting the corrected oil stirring speed into a stirring control instruction and sending it to the stirring system to obtain stirring execution feedback data; Step S43: performing data verification on the cooling execution feedback data and the stirring execution feedback data to generate control feedback verification data; Step S44: when the control feedback verification data is abnormal, an alarm is issued and the abnormality handling process is entered to obtain abnormality handling record data; Step S45: When the control feedback verification data is normal, the oil state in the crystallization reactor is re-collected according to the preset monitoring cycle, and the crystal quality assessment is performed. If the oil crystallization meets the set crystallization requirements, the target control cooling rate, the corrected oil stirring speed and the control feedback verification data are stored in the production process control record to generate an oil crystallization process control report; otherwise, the crystallization control instructions are iteratively optimized.

10. An intelligent digital-based oil and fat product production control system, characterized in that: Used to execute the oil and fat product production control method based on intelligent digitalization as claimed in claim 1, the oil and fat product production control system based on intelligent digitalization comprises: The oil state sensing module is used to monitor the temperature, viscosity and turbidity data of the oil crystallization reactor in real time during the oil crystallization process, integrate the oil state, and generate snapshot data of the oil crystallization state; The cooling rate optimization module is used to combine the oil crystallization characteristics according to the oil crystallization state snapshot data to generate the oil crystallization feature vector; construct the oil crystallization kinetic model; use the oil crystallization kinetic model to predict the crystallization state of the oil crystallization feature vector and generate the oil crystallization prediction state data; adjust the cooling rate of the reactor according to the oil crystallization prediction state data to obtain the target control cooling rate; The stirring rate optimization module is used to set the expected heat transfer coefficient based on the target control cooling rate to obtain the expected heat transfer coefficient; the fuzzy control variable processing is performed on the snapshot data of the grease crystallization state through the expected heat transfer coefficient, and the fuzzy reasoning of the real-time stirring speed adjustment is performed to generate the corrected grease stirring speed; The crystallization adaptive control module is used to convert the target control cooling rate and the corrected oil stirring speed into control instructions, and execute the production control instructions to obtain control feedback verification data; according to the control feedback verification data, the production process control record is stored and the oil crystallization process control report is generated.

Images