Corn oil deodorization temperature accurate control method and system based on data feedback

Abstract

The invention discloses a corn oil deodorization temperature accurate control method and system based on data feedback, and relates to the technical field of temperature control, and the method comprises the following steps: deploying an SCADA system, collecting corn oil deodorization data, and building a self-adaptive temperature target model, a cross-process sub-control data feedback model and a corn oil deodorization temperature prediction model. Key parameters such as a corn oil optimal deodorization temperature target value, a cross-process temperature sub-control feedback coefficient and a corn oil deodorization temperature accurate coefficient are obtained, corn oil deodorization temperature adaptation grades are divided and visually output, and corn oil deodorization temperature accurate control based on data feedback is achieved. A traditional corn oil deodorization temperature accurate control technology has the problems of low corn oil deodorization temperature control accuracy and poor product quality stability caused by data islanding, control parameter immobilization, data acquisition lagging and one-sided and lack of process data feedback, and the intelligent degree in the corn oil deodorization temperature accurate control process is remarkably enhanced.

Description

Technical Field

[0001] This invention relates to the field of temperature control technology, specifically to a method and system for precise temperature control of corn oil deodorization based on data feedback. Background Technology

[0002] Deodorization of corn oil is a core step in the refining process. The deodorization temperature directly determines the quality of the oil and processing efficiency. If the temperature is too low, off-odor substances cannot be removed, affecting the flavor and shelf life of the oil. If the temperature is too high, it can easily cause oil oxidation and the formation of trans fatty acids, reducing the nutritional value of the product and increasing energy consumption and production costs. Traditional corn oil deodorization temperature control mostly adopts an open-loop control mode, relying on preset process parameters and manual experience for adjustment. This mode cannot respond in real time to interference factors such as fluctuations in raw material quality, changes in equipment operating status, and differences in environmental conditions. As a result, the deodorization process is unstable and the product consistency is low, making it difficult to meet the needs of large-scale, high-quality corn oil production. With the development of industrial automation and data acquisition technology, closed-loop control technology based on data feedback is gradually being applied to the oil processing field, achieving dynamic and precise control of deodorization temperature. This method breaks through the limitations of traditional control modes, provides technical support for improving the quality of corn oil refining and reducing energy consumption, and also promotes the transformation and upgrading of the oil processing industry towards intelligence and refinement.

[0003] Traditional corn oil deodorization temperature precision control technology suffers from problems such as data silos, fixed control parameters, lagging and fragmented data acquisition, and lack of process data feedback. This results in low temperature control accuracy, poor product quality stability, and high manual intervention costs. Therefore, it is necessary to explore how to use a SCADA system as the core, combined with machine learning algorithms for modeling, to ultimately achieve visualized supervision, anomaly warning, and model iterative optimization in the corn oil deodorization temperature precision control process based on cross-process data feedback, and to connect the data linkage between upstream and downstream processes to achieve high-precision, adaptive, and full-process collaborative control of deodorization temperature. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for precise control of corn oil deodorization temperature based on data feedback, so as to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: Firstly, a method for precise control of corn oil deodorization temperature based on data feedback, comprising the following steps: Step 1: Deploy the SCADA system to collect corn oil deodorization data, which includes cross-process collaborative data and multi-dimensional data of the deodorization process, and preprocess the collected data. Step 2: Based on the corn oil deodorization data, construct an adaptive temperature target model, output the optimal deodorization temperature target value for corn oil, and calculate the deodorization temperature difference for corn oil. Step 3: Using a neural network algorithm and combining the temperature difference in corn oil deodorization, construct a cross-process temperature control data feedback model to obtain the cross-process temperature control feedback coefficient. Step 4: Through the SCADA system and cross-process temperature control feedback coefficient, the deodorization process is adjusted in a coordinated manner to obtain the real-time degumming temperature of the previous process after adjustment and the real-time degreasing feed temperature of the corn oil deodorization process after adjustment. Combined with convolutional neural network, a corn oil deodorization temperature prediction model is constructed to obtain the predicted deodorization temperature of corn oil. Step 5: Combine the target value of the optimal deodorization temperature of corn oil with the predicted deodorization temperature of corn oil, calculate the accuracy coefficient of corn oil deodorization temperature, classify and visualize the corn oil deodorization temperature adaptation level. Step 6: Based on the corn oil deodorization temperature adaptation level, precisely control the corn oil deodorization temperature.

[0006] A further improvement to the technical solution of this invention is that, in step one, the process of deploying a SCADA system and collecting corn oil deodorization data includes: The SCADA system architecture comprises an equipment layer, a communication layer, and a monitoring layer. The equipment layer deploys online potentiometric titrators, online near-infrared moisture analyzers, armored thermocouples, temperature transmitters, Coriolis mass flow meters, platinum resistance temperature sensors, distributed fiber optic temperature sensors, steam mass flow meters, ultrasonic flow meters, infrared imaging sensors, ultrasonic thickness gauges, and high-precision temperature and humidity sensors to collect corn oil deodorization data. The communication layer uses the OPC UA protocol to enable cross-process collaborative data exchange and multi-dimensional data exchange during the deodorization process, ensuring the stability of data acquisition by transmitting real-time data to the SCADA system. The monitoring layer installs industrial monitoring software, which includes an automatic data entry module responsible for receiving and storing the collected corn oil deodorization data. Cross-process collaborative data includes real-time feed acid value, real-time moisture content, and real-time degumming temperature for the pre-deodorization process of corn oil; real-time discharge flow rate, real-time degreasing feed temperature, and real-time capacity demand for the post-deodorization process of corn oil; and multi-dimensional data for the deodorization process, including real-time deodorization temperature of corn oil, real-time oil temperature in the preheating section, deodorization section, and cooling section of the deodorization tower, real-time steam enthalpy value and real-time oil flow rate, real-time infrared images of the equipment, real-time metal wall thickness, real-time inner and outer temperatures of the metal wall, and real-time temperature and humidity of the corn oil deodorization environment. Through the OPC UA protocol of the communication layer, it automatically connects to the device layer and, in conjunction with the automatic data entry module of the monitoring layer, records the collected corn oil deodorization data into the real-time database of the SCADA system.

[0007] A further improvement to the technical solution of the present invention is that, in step one, the process of preprocessing the collected data includes: A combination of median filtering and Gaussian filtering algorithms is used to remove noise from the real-time infrared image of the equipment. Combined with the mask region convolutional neural network semantic segmentation model (referred to as Mask R-CNN semantic segmentation model), the binarized mask image of the scaled area in the real-time infrared image of the equipment is output, and the pixel mask of the scaled area in the real-time infrared image of the equipment is obtained. The number of pixel masks of the scaled area in the real-time infrared image of the equipment is counted. Combined with the pixel resolution of the infrared imaging sensor, the real-time scaled area of ​​the equipment is calculated. Infrared thermal imaging temperature measurement technology is used to extract the real-time temperature of the outer side of the scale layer of the equipment from the real-time infrared image of the equipment. Combined with the thermal conduction model thickness inversion technology and the one-dimensional steady-state thermal conduction equation, the real-time scale thickness of the equipment is calculated. The real-time scaling area and real-time scaling thickness of the equipment are integrated into the multi-dimensional data of the deodorization process. Data cleaning and standardization are performed on the cross-process collaborative data and the multi-dimensional data of the deodorization process. The standardization process is used to eliminate dimensional issues in the subsequent model building process.

[0008] A further improvement to the technical solution of this invention lies in the fact that, in step two, the process of constructing an adaptive temperature target model based on corn oil deodorization data, outputting the optimal deodorization temperature target value for corn oil, and calculating the temperature difference for corn oil deodorization includes: The weights are allocated to the real-time feed acid value of the corn oil deodorization process before deodorization, the real-time steam enthalpy of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization process after deodorization. The real-time feed acid value of the corn oil deodorization process before deodorization, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization process after deodorization are integrated into a linear weight training set and a linear weight test set in a 7:3 ratio. By combining the weighted summation method and the multiple linear regression algorithm, the linear weight training set data is used as input, and the target value of the optimal deodorization temperature of corn oil is used as output. The linear relationship between the real-time feed acid value of the corn oil deodorization front process, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, the real-time capacity demand of the corn oil deodorization back process and the target value of the optimal deodorization temperature of corn oil is learned by combining the weights, and thus a trained adaptive temperature target model is obtained. The linear weighted test set data is input into the trained adaptive temperature target model. The weights and intercept terms of the adaptive temperature target model are adjusted to optimize the performance of the adaptive temperature target model and obtain the final adaptive temperature target model. The real-time feed acid value of the current corn oil deodorization front process, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization back process are input into the adaptive temperature target model to obtain the optimal deodorization temperature target value for corn oil. The deodorization temperature difference of corn oil is calculated by the absolute value of the difference between the target value of the optimal deodorization temperature and the real-time deodorization temperature of corn oil.

[0009] A further improvement to the technical solution of this invention lies in the fact that, in step three, a neural network algorithm is used to construct a cross-process temperature control feedback model, combined with the temperature difference in corn oil deodorization, to obtain the cross-process temperature control feedback coefficients. This process includes: The temperature difference of corn oil deodorization, the real-time moisture content of the corn oil deodorization process before deodorization, and the real-time discharge flow rate of the corn oil deodorization process after deodorization were integrated into a data feedback training set and a data feedback test set in a ratio of 8:2. A neural network model architecture is built using neural network algorithms. The neural network model architecture includes an input layer, hidden layers, and an output layer. The number of neurons in the input layer is set to 13, corresponding to the 13 types of data in the data feedback training set. The data feedback training set data is used as the input variable. A single neuron is set in the output layer, and the output variable is set as the cross-process temperature control feedback coefficient. A fully connected layer is set in the hidden layer. The coupling relationship is captured by the activation function to achieve nonlinear fitting between the input variable and the output variable, and the trained cross-process temperature control data feedback model is obtained. Select a loss function, set the overfitting threshold of the cross-process control data feedback model, use the data feedback test set to test the overfitting loss value of the cross-process control data feedback model, adjust the learning rate and regularization parameters of the cross-process control data feedback model until the overfitting loss value of the cross-process control data feedback model converges to the overfitting threshold, and embed the current cross-process control data feedback model into the SCADA system. By combining the current temperature difference in corn oil deodorization, the real-time moisture content of the corn oil deodorization process before deodorization, and the real-time discharge flow rate of the corn oil deodorization process after deodorization, a cross-process temperature control feedback coefficient is output.

[0010] A further improvement to the technical solution of this invention lies in the fact that, in step four, the process of collaboratively adjusting the deodorization process through the SCADA system and the cross-process temperature control feedback coefficient includes: The deodorization process is a core step in corn oil refining. The real-time degumming temperature in the pre-deodorization process determines the removal rate of gums in the oil. Excessive gum content increases the deodorization load. The real-time defatting feed temperature in the post-deodorization process determines the crystallization efficiency of fatty acids in the deodorized oil and is an important reference for precise control of the deodorization temperature. Therefore, based on the cross-process temperature control feedback coefficient, the deodorization process is adjusted in a coordinated manner, mainly by adjusting the real-time degumming temperature in the pre-deodorization process and the real-time defatting feed temperature in the post-deodorization process. This incorporates the corn oil deodorization process into the scope of precise deodorization temperature control, providing technical support for subsequent precise control of the corn oil deodorization temperature based on data feedback from the pre- and post-processes, and providing data support for improving the accuracy of corn oil deodorization temperature control. The deodorization temperature compensation value of corn oil is calculated by combining the target value of the optimal deodorization temperature of corn oil, the real-time deodorization temperature of corn oil and the feedback coefficient of cross-process temperature control. Using the SCADA system, weights are assigned to the real-time degumming temperature of the corn oil deodorization pre-process and the real-time degreasing feed temperature of the corn oil deodorization post-process. Based on the comprehensive corn oil deodorization temperature compensation value, the degumming temperature control value and the degreasing feed temperature control value are calculated respectively. Then, the real-time degumming temperature of the corn oil deodorization pre-process and the real-time degreasing feed temperature of the corn oil deodorization post-process are adjusted respectively to obtain the adjusted real-time degumming temperature of the pre-process and the adjusted real-time degreasing feed temperature of the corn oil deodorization post-process.

[0011] A further improvement to the technical solution of this invention lies in the fact that, in step four, a corn oil deodorization temperature prediction model is constructed using a convolutional neural network, and the process of predicting the corn oil deodorization temperature includes: The real-time oil temperature and flow rate of the preheating section, deodorization section and cooling section of the deodorization tower, the real-time scaling area and scaling thickness of the equipment, the real-time temperature and humidity of the corn oil deodorization environment, the real-time degumming temperature of the adjusted preceding process and the real-time degreasing feed temperature of the adjusted corn oil deodorization process are integrated into a convolutional neural network training set and a convolutional neural network test set in a 6:4 ratio. Convolutional neural networks are used to construct convolutional neural network models, which include input layers, convolutional layers, pooling layers, fully connected layers, and output layers. Based on the dimensions of the training and test sets of the convolutional neural network (CNN), the parameters of the input layer neurons are designed. A sliding window is configured in the convolutional layer, and convolution operations are performed on the training and test sets of the CNN using the sliding window to generate feature maps. Pooling layers are then used to reduce the dimensionality of the feature maps, thereby reducing computation and enhancing the robustness of the model. A fully connected layer is configured after the convolutional and pooling layers, and a single neuron is set for the output layer to output the predicted deodorization temperature of corn oil. Choose a loss function, set an optimizer, train a convolutional neural network model using the convolutional neural network training set data, and update the parameters of the convolutional neural network model through the backpropagation algorithm to obtain a trained corn oil deodorization temperature prediction model. The trained corn oil deodorization temperature prediction model was evaluated and optimized using convolutional neural network test set data to obtain the final corn oil deodorization temperature prediction model. Based on the real-time oil temperature and flow rate of the preheating, deodorizing and cooling sections of the deodorizing tower, the real-time scaling area and thickness of the equipment, the real-time temperature and humidity of the corn oil deodorization environment, the real-time degumming temperature of the preceding process after adjustment, and the real-time degreasing feed temperature of the following process after adjustment, the predicted deodorization temperature of corn oil is output.

[0012] A further improvement to the technical solution of this invention lies in the fact that, in step five, the process of calculating the accuracy coefficient of corn oil deodorization temperature by combining the target value of the optimal deodorization temperature of corn oil and the predicted deodorization temperature of corn oil, and classifying and visually outputting the corn oil deodorization temperature adaptation level includes: The accuracy coefficient of corn oil deodorization temperature is calculated by predicting the proportion of the deodorization temperature in the target value of the optimal deodorization temperature of corn oil. Based on the accuracy coefficient of corn oil deodorization temperature, the corn oil deodorization temperature adaptation levels are divided into three levels: Level I, Level II, and Level III. The system integrates the current corn oil deodorization temperature accuracy coefficient and its corresponding corn oil deodorization temperature adaptation level, and outputs the corn oil deodorization temperature adaptation level in the form of data reports. This is a visualization of the corn oil deodorization temperature accuracy control adaptation level optimized based on process data feedback.

[0013] A further improvement to the technical solution of this invention lies in the fact that, in step six, the process of precisely controlling the deodorization temperature of corn oil according to the corn oil deodorization temperature adaptation level includes: When the corn oil deodorization temperature adaptation level is level I, it indicates that the precise control effect of corn oil deodorization temperature based on data feedback is not good. To address this, we need to implement continuous monitoring of corn oil deodorization data during the deodorization process, widen the precise control range of corn oil deodorization temperature, optimize data acquisition accuracy, optimize the algorithm, introduce PID control parameters, and add anti-interference measures for precise corn oil deodorization temperature control. When the corn oil deodorization temperature adaptation level is Level II, it indicates that the precise control effect of corn oil deodorization temperature based on data feedback is generally poor. It is necessary to maintain routine monitoring of corn oil deodorization temperature, analyze the fluctuation and trend of corn oil deodorization temperature, fine-tune the cross-process temperature control feedback coefficient, and optimize the current corn oil deodorization temperature adaptation level. When the corn oil deodorization temperature adaptation level is Level III, it indicates that the precise control of the corn oil deodorization temperature based on data feedback is effective. This is achieved by regularly monitoring and recording the corn oil deodorization data and maintaining the current corn oil deodorization temperature adaptation level.

[0014] Secondly, a data feedback-based corn oil deodorization temperature precision control system is provided to implement the aforementioned data feedback-based corn oil deodorization temperature precision control method. The system includes a corn oil deodorization data acquisition module, an adaptive optimal deodorization temperature evaluation module, a data feedback module, a deodorization temperature prediction module, a deodorization temperature adaptation module, and a temperature precision control module, wherein the modules are interconnected. The corn oil deodorization data acquisition module deploys a SCADA system to collect corn oil deodorization data, which includes cross-process collaborative data and multi-dimensional data of the deodorization process. The collected data is preprocessed to solidify the data foundation for precise control of corn oil deodorization temperature. The adaptive optimal deodorization temperature assessment module constructs an adaptive temperature target model based on corn oil deodorization data, outputs the target value of the optimal deodorization temperature for corn oil, and calculates the temperature difference for corn oil deodorization. The data feedback module uses a neural network algorithm and combines the temperature difference of corn oil deodorization to construct a cross-process sub-control data feedback model and obtain the cross-process temperature sub-control feedback coefficient. This is a key application of data feedback technology and provides a basis for subsequent coordinated adjustment of the upstream and downstream processes of corn oil deodorization to obtain the predicted deodorization temperature of corn oil. The deodorization temperature prediction module, through the SCADA system and cross-process temperature control feedback coefficient, coordinates the pre-process and post-process deodorization of corn oil, and combines convolutional neural network to construct a corn oil deodorization temperature prediction model to obtain the predicted deodorization temperature of corn oil. The deodorization temperature adaptation module combines the target value of the optimal deodorization temperature of corn oil with the predicted deodorization temperature of corn oil to calculate the accuracy coefficient of corn oil deodorization temperature, classify and visualize the corn oil deodorization temperature adaptation level, and realize the evaluation and visualization output of the accuracy of corn oil deodorization temperature control after the coordinated adjustment of the deodorization process. The precise temperature control module performs final precise control of the corn oil deodorization temperature based on the corn oil deodorization temperature adaptation level.

[0015] The beneficial effects of this invention are as follows: Compared with traditional methods and systems for precise control of corn oil deodorization temperature, the data feedback-based method and system for precise control of corn oil deodorization temperature in this invention closely integrates SCADA system data acquisition technology, multi-model fusion modeling technology, and modern information technology. This allows for the precise capture of cross-process collaborative data and multi-dimensional data from the deodorization process, obtaining key parameters such as the optimal target value for corn oil deodorization temperature, cross-process temperature control feedback coefficients, and the precision coefficient of corn oil deodorization temperature. This solves the problems of data silos, fixed control parameters, delayed and fragmented data acquisition, and lack of process data feedback inherent in traditional corn oil deodorization temperature control technologies. The problems of low precision in corn oil deodorization temperature control, poor product quality stability, and high cost of manual intervention are addressed by the method in this invention. This invention provides a more precise dynamic monitoring standard for a data-feedback-based corn oil deodorization temperature precision control method and system, making the monitored data more accurate indicators under the same conditions. The development and application of this method significantly enhances the intelligence level of the data-feedback-based corn oil deodorization temperature precision control process, achieving adaptive deodorization temperature control and precise collaborative control between upstream and downstream processes. This improves the consistency of corn oil product quality, reduces energy consumption and production costs, and promotes the transformation and upgrading of the oil processing industry towards intelligence and refinement. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a flowchart of a method for precise control of corn oil deodorization temperature based on data feedback according to the present invention; Figure 2 This is a block diagram of a precise temperature control system for corn oil deodorization based on data feedback, according to the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, as Figure 1 As shown, this invention provides a method for precise control of corn oil deodorization temperature based on data feedback, comprising the following steps: S101. Deploy the SCADA system to collect corn oil deodorization data, which includes cross-process collaborative data and multi-dimensional data of the deodorization process, and preprocess the collected data. S102. Based on corn oil deodorization data, construct an adaptive temperature target model, output the optimal deodorization temperature target value for corn oil, and calculate the deodorization temperature difference for corn oil. S103. Using a neural network algorithm and combining the temperature difference of corn oil deodorization, a cross-process temperature control data feedback model is constructed to obtain the cross-process temperature control feedback coefficient. S104. By using the SCADA system and cross-process temperature control feedback coefficient, the pre-process and post-process of corn oil deodorization are adjusted in a coordinated manner. The real-time degumming temperature of the pre-process after adjustment and the real-time degreasing feed temperature of the post-process of corn oil deodorization after adjustment are obtained. Combined with convolutional neural network, a corn oil deodorization temperature prediction model is constructed to obtain the predicted deodorization temperature of corn oil. S105. Combining the target value of the optimal deodorization temperature of corn oil and the predicted deodorization temperature of corn oil, calculate the accuracy coefficient of corn oil deodorization temperature, classify and visualize the corn oil deodorization temperature adaptation level. S106. Based on the corn oil deodorization temperature adaptation level, the corn oil deodorization temperature is precisely controlled.

[0020] In some embodiments, S101, the process of deploying a SCADA system and collecting corn oil deodorization data includes: The SCADA system architecture comprises an equipment layer, a communication layer, and a monitoring layer. The equipment layer deploys online potentiometric titrators, online near-infrared moisture analyzers, armored thermocouples, temperature transmitters, Coriolis mass flow meters, platinum resistance temperature sensors, distributed fiber optic temperature sensors, steam mass flow meters, ultrasonic flow meters, infrared imaging sensors, ultrasonic thickness gauges, and high-precision temperature and humidity sensors to collect corn oil deodorization data. The communication layer uses the OPC UA protocol to enable cross-process collaborative data exchange and multi-dimensional data exchange during the deodorization process, ensuring the stability of data acquisition by transmitting real-time data to the SCADA system. The monitoring layer installs industrial monitoring software, which includes an automatic data entry module responsible for receiving and storing the collected corn oil deodorization data. Cross-process collaborative data includes real-time feed acid value, real-time moisture content, and real-time degumming temperature for the pre-deodorization process of corn oil; real-time discharge flow rate, real-time degreasing feed temperature, and real-time capacity demand for the post-deodorization process of corn oil; and multi-dimensional data for the deodorization process, including real-time deodorization temperature of corn oil, real-time oil temperature in the preheating section, deodorization section, and cooling section of the deodorization tower, real-time steam enthalpy value and real-time oil flow rate, real-time infrared images of the equipment, real-time metal wall thickness, real-time inner and outer temperatures of the metal wall, and real-time temperature and humidity of the corn oil deodorization environment. Specifically, an online potentiometric titrator, an online near-infrared moisture analyzer, a Coriolis mass flow meter, and a platinum resistance temperature sensor are used to collect real-time feed acid value and moisture content for the corn oil deodorization process before deodorization, and real-time discharge flow rate and degreasing feed temperature for the corn oil deodorization process after deodorization. Combined with armored thermocouples and temperature transmitters, real-time degumming temperature for the corn oil deodorization process before deodorization is collected. Distributed fiber optic temperature sensors are used to collect real-time oil temperatures and corn oil deodorization temperatures in the preheating, deodorization, and cooling sections of the deodorization tower. Steam mass flow meters and ultrasonic flow meters are used to collect real-time steam enthalpy and oil flow rates in the preheating, deodorization, and cooling sections of the deodorization tower. Infrared imaging sensors are used to collect real-time infrared images of the equipment. Combined with an ultrasonic thickness gauge and armored thermocouples, real-time metal wall thickness, real-time inner metal wall temperature, and real-time outer metal wall temperature are collected. High-precision temperature and humidity sensors are used to collect real-time temperature and humidity of the corn oil deodorization environment. Using the data entry technology of the automatic data entry module, the real-time capacity demand of the corn oil deodorization process is obtained from the real-time database of the production execution system of the corn oil deodorization process. The real-time capacity demand is a numerical value in t / h, which can be used for subsequent standardization and as an input variable for model building. Through the OPC UA protocol of the communication layer, it automatically connects to the equipment layer and, in conjunction with the automatic data entry module of the monitoring layer, enters the collected corn oil deodorization data into the real-time database of the SCADA system in real time.

[0021] In some embodiments, S101 includes the process of preprocessing the collected data: A combination of median filtering and Gaussian filtering algorithms is used to remove noise from the real-time infrared image of the equipment. Combined with a mask region convolutional neural network semantic segmentation model (referred to as Mask R-CNN semantic segmentation model), a binary mask image of the scaled area in the real-time infrared image of the equipment is output, obtaining the pixel mask of the scaled area in the real-time infrared image of the equipment. The number of pixel masks of the scaled area in the real-time infrared image of the equipment is counted. Combined with the pixel resolution of the infrared imaging sensor, the real-time scaled area of ​​the equipment is calculated. The calculation formula includes:

[0022] in, This represents the real-time scaling area of ​​the equipment. The number of pixels masking the scaled area in the real-time infrared image of the equipment; The pixel resolution of an infrared imaging sensor is the actual area corresponding to a single pixel mask. This formula is a common practice for converting image pixels to actual size. Infrared thermal imaging temperature measurement technology is used to extract the real-time temperature of the outer surface of the scale layer from the real-time infrared image of the equipment. Combined with the thickness inversion technology of the heat conduction model and the one-dimensional steady-state heat conduction equation, the real-time scale thickness of the equipment is calculated. The calculation formula is as follows:

[0023]

[0024] in, and These refer to the real-time metal wall thickness and real-time scale thickness of the equipment, respectively, in meters, abbreviated as... ; Heat flux density, its unit is ; and The thermal conductivity values ​​for the equipment's metal wall and structural layers are respectively taken as values. and ; , and These represent the real-time inner temperature of the metal wall, the real-time outer temperature of the metal wall, and the real-time outer temperature of the scale layer, respectively, in Kelvin (abbreviated as Kelvin). The formula here is derived from the one-dimensional steady-state heat conduction equation and its inverse derivation; The real-time scaling area and real-time scaling thickness of the equipment are integrated into the multi-dimensional data of the deodorization process. Data cleaning and standardization are performed on the cross-process collaborative data and the multi-dimensional data of the deodorization process. The standardization process is used to eliminate dimensional issues in the subsequent model building process.

[0025] In some embodiments, S102, the process of constructing an adaptive temperature target model based on corn oil deodorization data, outputting the optimal deodorization temperature target value for corn oil, and calculating the deodorization temperature difference for corn oil includes: The weights are allocated to the real-time feed acid value of the corn oil deodorization process before deodorization, the real-time steam enthalpy of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization process after deodorization. The real-time feed acid value of the corn oil deodorization process before deodorization, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization process after deodorization are integrated into a linear weight training set and a linear weight test set in a 7:3 ratio. By combining the weighted summation method and the multiple linear regression algorithm, the linear weight training set data is used as input, and the target value of the optimal deodorization temperature of corn oil is used as output. The linear relationship between the real-time feed acid value of the corn oil deodorization front process, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, the real-time capacity demand of the corn oil deodorization back process and the target value of the optimal deodorization temperature of corn oil is learned by combining the weights, and thus a trained adaptive temperature target model is obtained. The linear weighted test set data is input into the trained adaptive temperature target model. The weights and intercept terms of the adaptive temperature target model are adjusted to optimize the performance of the adaptive temperature target model and obtain the final adaptive temperature target model. The real-time feed acid value of the current corn oil deodorization front process, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization back process are input into the adaptive temperature target model to obtain the optimal deodorization temperature target value for corn oil. The deodorization temperature difference of corn oil is calculated by the absolute value of the difference between the target value of the optimal deodorization temperature and the real-time deodorization temperature of corn oil. The adaptive temperature target model, constructed by combining the multiple linear regression algorithm and the weighted summation method, is expressed as follows:

[0026] in, The target value for the optimal deodorization temperature of corn oil. This is the expression for the adaptive temperature target model constructed by combining the multiple linear regression algorithm and the weighted summation method. , , , and These correspond to the real-time feed acid value of the pre-deodorization process in the corn oil deodorization tower, the real-time steam enthalpy value of the preheating section, deodorization section, and cooling section of the deodorization tower, and the real-time capacity demand of the post-deodorization process in the corn oil deodorization tower, respectively. , , , and The weights are as follows: real-time feed acid value of the pre-deodorization process in corn oil, real-time steam enthalpy of the preheating, deodorization, and cooling sections of the deodorization tower, and real-time capacity demand of the post-deodorization process in corn oil. and These are the intercept and error terms of the adaptive temperature target model, respectively. This expression is a combination of multiple linear regression algorithm and weighted summation.

[0027] In some embodiments, in S103, the process of using a neural network algorithm, combined with the temperature difference in corn oil deodorization, to construct a cross-process temperature control data feedback model and obtain the cross-process temperature control feedback coefficient includes: The temperature difference of corn oil deodorization, the real-time moisture content of the corn oil deodorization process before deodorization, and the real-time discharge flow rate of the corn oil deodorization process after deodorization were integrated into a data feedback training set and a data feedback test set in a ratio of 8:2. A neural network model architecture is built using neural network algorithms. The neural network model architecture includes an input layer, hidden layers, and an output layer. The number of neurons in the input layer is set to 13, corresponding to the 13 types of data in the data feedback training set. The data feedback training set data is used as the input variable. A single neuron is set in the output layer, and the output variable is set as the cross-process temperature control feedback coefficient. A fully connected layer is set in the hidden layer. The coupling relationship is captured by the activation function to achieve nonlinear fitting between the input variable and the output variable, and the trained cross-process temperature control data feedback model is obtained. Select a loss function, set the overfitting threshold of the cross-process control data feedback model, use the data feedback test set to test the overfitting loss value of the cross-process control data feedback model, adjust the learning rate and regularization parameters of the cross-process control data feedback model until the overfitting loss value of the cross-process control data feedback model converges to the overfitting threshold, and embed the current cross-process control data feedback model into the SCADA system. By combining the current temperature difference in corn oil deodorization, the real-time moisture content of the corn oil deodorization process before deodorization, and the real-time discharge flow rate of the corn oil deodorization process after deodorization, a cross-process temperature control feedback coefficient is output.

[0028] In some embodiments, S104, the process of collaboratively adjusting the deodorization process through the SCADA system and the cross-process temperature control feedback coefficient includes: The deodorization process is a core step in corn oil refining. The real-time degumming temperature in the pre-deodorization process determines the removal rate of gums in the oil. Excessive gum content increases the deodorization load. The real-time defatting feed temperature in the post-deodorization process determines the crystallization efficiency of fatty acids in the deodorized oil and is an important reference for precise control of the deodorization temperature. Therefore, based on the cross-process temperature control feedback coefficient, the deodorization process is adjusted in a coordinated manner, mainly by adjusting the real-time degumming temperature in the pre-deodorization process and the real-time defatting feed temperature in the post-deodorization process. This incorporates the corn oil deodorization process into the scope of precise deodorization temperature control, providing technical support for subsequent precise control of the corn oil deodorization temperature based on data feedback from the pre- and post-processes, and providing data support for improving the accuracy of corn oil deodorization temperature control. Combining the target value of optimal deodorization temperature for corn oil, the real-time deodorization temperature of corn oil, and the feedback coefficient of cross-process temperature control, the deodorization temperature compensation value for corn oil is calculated. The calculation process includes:

[0029] in, This is the temperature compensation value for deodorizing corn oil. This refers to the cross-process temperature control feedback coefficient. The target value for the optimal deodorization temperature of corn oil. The formula for the real-time deodorization temperature of corn oil is mainly derived from process optimization and automatic control practices in the field of oil refining. Using the SCADA system, weights are assigned to the real-time degumming temperature before the corn oil deodorization process and the real-time degreasing feed temperature after the corn oil deodorization process. Based on the comprehensive corn oil deodorization temperature compensation value, the degumming temperature control value and the degreasing feed temperature control value are calculated separately. Then, the current real-time degumming temperature before the corn oil deodorization process and the real-time degreasing feed temperature after the corn oil deodorization process are adjusted respectively, resulting in the adjusted real-time degumming temperature before the corn oil deodorization process and the adjusted real-time degreasing feed temperature after the corn oil deodorization process. The calculation formulas involved are as follows:

[0030]

[0031]

[0032]

[0033] in, and These are the degumming temperature control values ​​and the degreasing feed temperature control values, respectively. and The weights are the real-time degumming temperature before the corn oil deodorization process and the real-time degreasing feed temperature after the corn oil deodorization process, respectively. This is the temperature compensation value for deodorizing corn oil; and These are the real-time degumming temperature of the preceding process after adjustment and the real-time degreasing feed temperature of the following process after corn oil deodorization. and These are the real-time degumming temperature before the corn oil deodorization process and the real-time degreasing feed temperature after the corn oil deodorization process.

[0034] In some embodiments, S104, the process of constructing a corn oil deodorization temperature prediction model using a convolutional neural network to derive the predicted corn oil deodorization temperature includes: The real-time oil temperature and flow rate of the preheating section, deodorization section and cooling section of the deodorization tower, the real-time scaling area and scaling thickness of the equipment, the real-time temperature and humidity of the corn oil deodorization environment, the real-time degumming temperature of the adjusted preceding process and the real-time degreasing feed temperature of the adjusted corn oil deodorization process are integrated into a convolutional neural network training set and a convolutional neural network test set in a 6:4 ratio. Convolutional neural networks are used to construct convolutional neural network models, which include input layers, convolutional layers, pooling layers, fully connected layers, and output layers. Based on the dimensions of the training and test sets of the convolutional neural network (CNN), the parameters of the input layer neurons are designed. A sliding window is configured in the convolutional layer, and convolution operations are performed on the training and test sets of the CNN using the sliding window to generate feature maps. Pooling layers are then used to reduce the dimensionality of the feature maps, thereby reducing computation and enhancing the robustness of the model. A fully connected layer is configured after the convolutional and pooling layers, and a single neuron is set for the output layer to output the predicted deodorization temperature of corn oil. Select a loss function, set an optimizer, train a convolutional neural network model using the training set data, and update the parameters of the convolutional neural network model through the backpropagation algorithm. Learn the real-time oil temperature and real-time oil flow rate in the preheating section, deodorization section and cooling section of the deodorization tower, the real-time scaling area and real-time scaling thickness of the equipment, the real-time temperature and real-time humidity of the corn oil deodorization environment, the real-time degumming temperature of the adjusted preceding process, the real-time degreasing feed temperature of the adjusted corn oil deodorization downstream process and the nonlinear relationship between them and the predicted deodorization temperature of corn oil, and obtain a trained corn oil deodorization temperature prediction model. The trained corn oil deodorization temperature prediction model was evaluated and optimized using convolutional neural network test set data to obtain the final corn oil deodorization temperature prediction model. Based on the real-time oil temperature and flow rate of the preheating, deodorizing and cooling sections of the deodorizing tower, the real-time scaling area and thickness of the equipment, the real-time temperature and humidity of the corn oil deodorization environment, the real-time degumming temperature of the preceding process after adjustment, and the real-time degreasing feed temperature of the following process after adjustment, the predicted deodorization temperature of corn oil is output.

[0035] In some embodiments, S105, the process of calculating the corn oil deodorization temperature accuracy coefficient by combining the target value of the optimal deodorization temperature of corn oil and the predicted deodorization temperature of corn oil, and classifying and visually outputting the corn oil deodorization temperature adaptation level includes: The accuracy coefficient of corn oil deodorization temperature is calculated by predicting the proportion of the deodorization temperature in the target value of the optimal deodorization temperature of corn oil. Based on the accuracy coefficient of corn oil deodorization temperature, corn oil deodorization temperature adaptation levels are classified, including Level I, Level II, and Level III. Specifically, the accuracy coefficient of corn oil deodorization temperature is between 0 and 1. When the accuracy coefficient is less than or equal to 0.5, it corresponds to Level I; when the accuracy coefficient is between 0.5 and 0.8, it corresponds to Level II; and when the accuracy coefficient is greater than or equal to 0.8, it corresponds to Level III. The system integrates the current corn oil deodorization temperature accuracy coefficient and its corresponding corn oil deodorization temperature adaptation level, and outputs the corn oil deodorization temperature adaptation level in the form of data reports. This is a visualization of the corn oil deodorization temperature accuracy control adaptation level optimized based on process data feedback.

[0036] In some embodiments, S106, the process of precisely controlling the deodorization temperature of corn oil according to the corn oil deodorization temperature adaptation level includes: When the corn oil deodorization temperature adaptation level is level I, it indicates that the precise control effect of corn oil deodorization temperature based on data feedback is not good. To address this, we need to implement continuous monitoring of corn oil deodorization data during the deodorization process, widen the precise control range of corn oil deodorization temperature, optimize data acquisition accuracy, optimize the algorithm, introduce PID control parameters, and add anti-interference measures for precise corn oil deodorization temperature control. When the corn oil deodorization temperature adaptation level is Level II, it indicates that the precise control effect of corn oil deodorization temperature based on data feedback is generally poor. It is necessary to maintain routine monitoring of corn oil deodorization temperature, analyze the fluctuation and trend of corn oil deodorization temperature, fine-tune the cross-process temperature control feedback coefficient, and optimize the current corn oil deodorization temperature adaptation level. When the corn oil deodorization temperature adaptation level is Level III, it indicates that the precise control of the corn oil deodorization temperature based on data feedback is effective. This is achieved by regularly monitoring and recording the corn oil deodorization data and maintaining the current corn oil deodorization temperature adaptation level.

[0037] Example 2, as Figure 2 As shown, based on Embodiment 1, the present invention provides a technical solution: a precise temperature control system for corn oil deodorization based on data feedback, used to implement the above-mentioned precise temperature control method for corn oil deodorization based on data feedback, including a corn oil deodorization data acquisition module, an adaptive optimal deodorization temperature evaluation module, a data feedback module, a deodorization temperature prediction module, a deodorization temperature adaptation module, and a precise temperature control module, wherein each module is communicatively connected; The corn oil deodorization data acquisition module deploys a SCADA system to collect corn oil deodorization data, which includes cross-process collaborative data and multi-dimensional data of the deodorization process. The collected data is preprocessed to solidify the data foundation for precise control of corn oil deodorization temperature and ensure the accuracy and completeness of the data. The adaptive optimal deodorization temperature assessment module, based on corn oil deodorization data, constructs an adaptive temperature target model, outputs the target value of the optimal deodorization temperature for corn oil, and calculates the temperature difference of corn oil deodorization, accurately anchoring the optimal temperature control benchmark. The data feedback module uses a neural network algorithm and combines the temperature difference of corn oil deodorization to construct a cross-process sub-control data feedback model and obtain the cross-process temperature sub-control feedback coefficient. This is a key application of data feedback technology and provides a basis for subsequent coordinated adjustment of the upstream and downstream processes of corn oil deodorization. The deodorization temperature prediction module, through the SCADA system and cross-process temperature control feedback coefficient, coordinates the pre-process and post-process deodorization of corn oil, and combines convolutional neural network to construct a corn oil deodorization temperature prediction model to obtain the predicted deodorization temperature of corn oil. The deodorization temperature adaptation module combines the target value of the optimal deodorization temperature for corn oil with the predicted deodorization temperature for corn oil to calculate the accuracy coefficient of the deodorization temperature for corn oil, classify and visualize the deodorization temperature adaptation level of corn oil, realize the evaluation and visualization output of the accuracy of corn oil deodorization temperature control after the coordinated adjustment of the deodorization process, and intuitively present the matching degree and optimization space of corn oil deodorization temperature control. The precise temperature control module, based on the corn oil deodorization temperature adaptation level, performs final precise control of the corn oil deodorization temperature, ensuring the stability of the corn oil deodorization process and product quality.

[0038] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for precise control of corn oil deodorization temperature based on data feedback, characterized in that, Includes the following steps: Deploy a SCADA system to collect corn oil deodorization data, which includes cross-process collaborative data and multi-dimensional data of the deodorization process, and preprocess the collected corn oil deodorization data. Based on the corn oil deodorization data, an adaptive temperature target model is constructed to output the optimal deodorization temperature target value of corn oil, and the deodorization temperature difference of corn oil is calculated in combination with the real-time deodorization temperature of corn oil. A neural network algorithm is used to construct a cross-process temperature control data feedback model based on the temperature difference of corn oil deodorization, so as to obtain the cross-process temperature control feedback coefficient. By using the SCADA system and the cross-process temperature control feedback coefficient, the deodorization process is adjusted in a coordinated manner. Combined with a convolutional neural network, a corn oil deodorization temperature prediction model is constructed to output the predicted deodorization temperature of corn oil. By combining the target value of the optimal deodorization temperature of corn oil and the predicted deodorization temperature of corn oil, the accuracy coefficient of corn oil deodorization temperature is calculated, and the corn oil deodorization temperature adaptation level is classified and visualized. Based on the corn oil deodorization temperature adaptation level, the corn oil deodorization temperature is precisely controlled.

2. The method for precise temperature control of corn oil deodorization based on data feedback according to claim 1, characterized in that, The process of deploying the SCADA system and collecting corn oil deodorization data includes: The SCADA system comprises a device layer, a communication layer, and a monitoring layer. The device layer deploys online potentiometric titrators, online near-infrared moisture analyzers, armored thermocouples, temperature transmitters, Coriolis mass flow meters, platinum resistance temperature sensors, distributed fiber optic temperature sensors, steam mass flow meters, ultrasonic flow meters, infrared imaging sensors, ultrasonic thickness gauges, and high-precision temperature and humidity sensors to collect data on corn oil deodorization. The communication layer schedules the OPC UA protocol, and the monitoring layer installs industrial monitoring software, including an automatic data entry module. The cross-process collaborative data includes real-time feed acid value, real-time moisture content, and real-time degumming temperature for the pre-deodorization process of corn oil, and real-time discharge flow rate, real-time degreasing feed temperature, and real-time capacity demand for the post-deodorization process of corn oil; the multi-dimensional data of the deodorization process includes real-time deodorization temperature of corn oil, real-time oil temperature in the preheating section, deodorization section, and cooling section of the deodorization tower, real-time steam enthalpy value and real-time oil flow rate, real-time infrared image of the equipment, real-time metal wall thickness, real-time inner and outer temperatures of the metal wall, and real-time temperature and humidity of the corn oil deodorization environment; Through the OPC UA protocol of the communication layer, the system automatically connects to the device layer and, in conjunction with the automatic data entry module of the monitoring layer, records the collected corn oil deodorization data into the real-time database of the SCADA system in real time.

3. The method for precise temperature control of corn oil deodorization based on data feedback according to claim 2, characterized in that, The preprocessing of the collected corn oil deodorization data includes: A combination algorithm of median filtering and Gaussian filtering is used to remove noise from the real-time infrared image of the device. Combined with the semantic segmentation model of the mask region convolutional neural network, a binary mask image of the scaled area in the real-time infrared image of the device is output to obtain the pixel mask of the scaled area in the real-time infrared image of the device. The number of pixel masks of the scaled area in the real-time infrared image of the device is counted. Combined with the pixel resolution of the infrared imaging sensor, the real-time scaled area of ​​the device is calculated. Infrared thermal imaging temperature measurement technology is used to extract the real-time temperature of the outer side of the scale layer of the equipment from the real-time infrared image of the equipment. Combined with the thermal conduction model thickness inversion technology and the one-dimensional steady-state thermal conduction equation, the real-time scale thickness of the equipment is calculated. The real-time scaling area and real-time scaling thickness of the equipment are integrated into the multi-dimensional data of the deodorization process. The cross-process collaborative data and the multi-dimensional data of the deodorization process are cleaned and standardized. The standardization process is used to eliminate dimensional issues in the subsequent model building process.

4. The method for precise control of corn oil deodorization temperature based on data feedback according to claim 3, characterized in that, The process of constructing an adaptive temperature target model based on corn oil deodorization data, outputting the optimal deodorization temperature target value for corn oil, and calculating the deodorization temperature difference of corn oil in combination with the real-time deodorization temperature of corn oil includes: The weights are allocated to the real-time feed acid value of the corn oil deodorization pre-process, the real-time steam enthalpy of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization post-process. The real-time feed acid value of the corn oil deodorization pre-process, the real-time steam enthalpy of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity demand of the corn oil deodorization post-process are integrated into a linear weight training set and a linear weight test set. By combining the weighted summation method and the multiple linear regression algorithm, the linear weight training set data is used as input, and the target value of the optimal deodorization temperature of corn oil is used as output. The linear relationship between the real-time feed acid value of the corn oil deodorization front process, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, the real-time capacity demand of the corn oil deodorization back process and the target value of the optimal deodorization temperature of corn oil is learned by combining the weights, and thus a trained adaptive temperature target model is obtained. The linear weight test set data is input into the trained adaptive temperature target model, the weights and intercept terms of the adaptive temperature target model are adjusted, the performance of the adaptive temperature target model is optimized, and the final adaptive temperature target model is obtained. The real-time feed acid value of the current corn oil deodorization front process, the real-time steam enthalpy value of the preheating section, deodorization section and cooling section of the deodorization tower, and the real-time capacity requirement of the corn oil deodorization back process are input into the adaptive temperature target model to obtain the optimal deodorization temperature target value of the corn oil. The deodorization temperature difference of corn oil is calculated by the absolute value of the difference between the target value of the optimal deodorization temperature and the real-time deodorization temperature of corn oil.

5. The method for precise control of corn oil deodorization temperature based on data feedback according to claim 4, characterized in that, The process of using a neural network algorithm, combined with the temperature difference in corn oil deodorization, to construct a cross-process temperature control feedback model to obtain the cross-process temperature control feedback coefficients includes: The temperature difference of corn oil deodorization, the real-time moisture content of the corn oil deodorization pre-process, and the real-time discharge flow rate of the corn oil deodorization post-process are integrated into a data feedback training set and a data feedback test set. Using the aforementioned neural network algorithm, a neural network model architecture is constructed, which includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is set, the data feedback training set data is used as the input variable, a single neuron is set in the output layer, the output variable is set as the cross-process temperature control feedback coefficient, a fully connected layer is set in the hidden layer, and the coupling relationship is captured by the activation function to make the input variable and the output variable nonlinearly fit, so as to obtain the trained cross-process temperature control data feedback model. Select a loss function, set the overfitting threshold of the cross-process control data feedback model, use the data feedback test set to test the overfitting loss value of the cross-process control data feedback model, adjust the learning rate and regularization parameters of the cross-process control data feedback model until the overfitting loss value of the cross-process control data feedback model converges to the overfitting threshold, and embed the tested cross-process control data feedback model into the SCADA system; Based on the current temperature difference of corn oil deodorization, the real-time moisture content of the corn oil deodorization pre-process, and the real-time discharge flow rate of the corn oil deodorization post-process, the cross-process temperature control feedback coefficient is output.

6. The method for precise temperature control of corn oil deodorization based on data feedback according to claim 5, characterized in that, The process of coordinating the adjustment of the deodorization process through the SCADA system and the cross-process temperature control feedback coefficient includes: Based on the target value of the optimal deodorization temperature of corn oil, the real-time deodorization temperature of corn oil and the cross-process temperature control feedback coefficient, the deodorization temperature compensation value of corn oil is calculated. The SCADA system assigns weights to the real-time degumming temperature of the corn oil deodorization process before deodorization and the real-time degreasing feed temperature of the corn oil deodorization process after deodorization. Based on the corn oil deodorization temperature compensation value, the degumming temperature control value and the degreasing feed temperature control value are calculated respectively. Based on the degumming temperature control value and the degreasing feed temperature control value, the real-time degumming temperature of the current corn oil deodorization pre-process and the real-time degreasing feed temperature of the corn oil deodorization post-process are adjusted respectively to obtain the adjusted real-time degumming temperature of the pre-process and the adjusted real-time degreasing feed temperature of the corn oil deodorization post-process.

7. The method for precise control of corn oil deodorization temperature based on data feedback according to claim 6, characterized in that, The process of constructing a corn oil deodorization temperature prediction model using a convolutional neural network to output the predicted deodorization temperature of corn oil includes: The real-time oil temperature and flow rate of the preheating section, deodorization section and cooling section of the deodorization tower, the real-time scaling area and scaling thickness of the equipment, the real-time temperature and humidity of the corn oil deodorization environment, the real-time degumming temperature of the adjusted preceding process and the real-time degreasing feed temperature of the adjusted corn oil deodorization subsequent process are integrated into a convolutional neural network training set and a convolutional neural network test set. Using the aforementioned convolutional neural network, a convolutional neural network model is constructed, which includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. Based on the dimensions of the training set data and the test set data of the convolutional neural network, the parameters of the input layer neurons are designed. A sliding window is configured in the convolutional layer, and the training set data and the test set data of the convolutional neural network are convolved by the sliding window to generate a feature map. The dimension of the feature map is reduced by combining the pooling layer. A fully connected layer is configured after the convolutional layer and the pooling layer, and a single neuron is set for the output layer to output the predicted deodorization temperature of corn oil. Select a loss function, set an optimizer, train the convolutional neural network model using the training set data, and update the parameters of the convolutional neural network model through the backpropagation algorithm to obtain the trained corn oil deodorization temperature prediction model. The trained corn oil deodorization temperature prediction model is evaluated and optimized using the test set data of the convolutional neural network to obtain the final corn oil deodorization temperature prediction model. Based on the real-time oil temperature and flow rate of the preheating, deodorizing, and cooling sections of the deodorizing tower, the real-time scaling area and thickness of the equipment, the real-time temperature and humidity of the corn oil deodorization environment, the real-time degumming temperature of the adjusted preceding process, and the real-time degreasing feed temperature of the adjusted corn oil deodorization following process, the predicted deodorization temperature of the corn oil is output.

8. The method for precise control of corn oil deodorization temperature based on data feedback according to claim 7, characterized in that, The process of combining the target value of the optimal deodorization temperature for corn oil and the predicted deodorization temperature for corn oil to calculate the accuracy coefficient of the deodorization temperature for corn oil, and then classifying and visually outputting the corn oil deodorization temperature adaptation level, includes: The accuracy coefficient of corn oil deodorization temperature is calculated by the proportion of the predicted deodorization temperature in the target value of the optimal deodorization temperature of corn oil. Based on the accuracy coefficient of corn oil deodorization temperature, corn oil deodorization temperature adaptation levels are divided, including corn oil deodorization temperature level I, corn oil deodorization temperature level II, and corn oil deodorization temperature level III. The current corn oil deodorization temperature accuracy coefficient and its corresponding corn oil deodorization temperature adaptation level are integrated and output in the form of a data report, visually representing the corn oil deodorization temperature adaptation level.

9. A method for precise control of corn oil deodorization temperature based on data feedback according to claim 8, characterized in that, The process of precisely controlling the deodorization temperature of corn oil based on the corn oil deodorization temperature adaptation level includes: When the corn oil deodorization temperature adaptation level is the corn oil deodorization temperature level I, it indicates that the precise control effect of corn oil deodorization temperature based on data feedback is not good. To address this, we need to implement continuous monitoring of corn oil deodorization data during the corn oil deodorization process, widen the precise control range of corn oil deodorization temperature, optimize data acquisition accuracy, optimize the algorithm, introduce PID control parameters, and add anti-interference measures for precise corn oil deodorization temperature control. When the corn oil deodorization temperature adaptation level is the corn oil deodorization temperature level II, it indicates that the precise control effect of corn oil deodorization temperature based on data feedback is generally poor. Regular monitoring of corn oil deodorization temperature should be maintained, the fluctuation and trend of corn oil deodorization temperature should be analyzed, the cross-process temperature control feedback coefficient should be fine-tuned, and the current corn oil deodorization temperature adaptation level should be optimized. When the corn oil deodorization temperature adaptation level is level III, it indicates that the precise control of corn oil deodorization temperature based on data feedback is effective. This is achieved by regularly monitoring and recording corn oil deodorization data and maintaining the current corn oil deodorization temperature adaptation level.

10. A precise temperature control system for corn oil deodorization based on data feedback, used to implement the precise temperature control method for corn oil deodorization based on data feedback as described in any one of claims 1-9, comprising a corn oil deodorization data acquisition module, an adaptive optimal deodorization temperature evaluation module, a data feedback module, a deodorization temperature prediction module, a deodorization temperature adaptation module, and a precise temperature control module, wherein, The various modules are connected for communication, characterized in that, The corn oil deodorization data acquisition module is used to deploy a SCADA system and collect corn oil deodorization data, which includes cross-process collaborative data and multi-dimensional data of the deodorization process, and preprocesses the collected corn oil deodorization data. The adaptive optimal deodorization temperature evaluation module is used to construct an adaptive temperature target model based on the corn oil deodorization data, output the target value of the optimal deodorization temperature of corn oil, and calculate the deodorization temperature difference of corn oil in combination with the real-time deodorization temperature of corn oil. The data feedback module is used to construct a cross-process temperature control data feedback model by using a neural network algorithm and combining the temperature difference of corn oil deodorization, so as to obtain the cross-process temperature control feedback coefficient. The deodorization temperature prediction module is used to coordinate the deodorization process through the SCADA system and the cross-process temperature control feedback coefficient, and to construct a corn oil deodorization temperature prediction model by combining a convolutional neural network to output the predicted deodorization temperature of corn oil. The deodorization temperature adaptation module is used to combine the target value of the optimal deodorization temperature of corn oil and the predicted deodorization temperature of corn oil to calculate the accuracy coefficient of corn oil deodorization temperature, classify and visually output the corn oil deodorization temperature adaptation level. The precise temperature control module is used to precisely control the deodorization temperature of corn oil according to the corn oil deodorization temperature adaptation level.

Images