Supervision system and method for industrial oleic acid production

Abstract

The invention discloses a supervision system and method for industrial oleic acid production, and the system comprises a raw material reaction regulation and control module, a real-time parameter collection module, a phase equilibrium analysis module, a data intelligent reasoning module, a cloud centralized management module and a process dynamic adjustment module which are connected through an industrial data transmission protocol, an encrypted data channel and the like to form a complete closed-loop link. The raw material reaction regulation and control module sets reaction initial parameters, the real-time parameter acquisition module continuously captures data such as temperature and oleic acid concentration in a reaction kettle, the phase equilibrium analysis module combines a thermodynamic model to calculate and obtain phase component parameters, and the data intelligent reasoning module complements the parameters and deduces an abnormal trend based on an omission-Bayesian reasoning model. The cloud centralized management module synchronizes data with a related platform, and the process dynamic adjustment module performs stepped correction on the stirring rate and the like. The system realizes coherent operation of parameter acquisition, analysis, reasoning, management and adjustment through cooperation of all the modules, and improves the stability of oleic acid production and the product quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial oleic acid production supervision, and particularly relates to a supervision system and method for industrial oleic acid production. BACKGROUND

[0002] Industrial oleic acid is a basic raw material widely used in chemical industry, daily chemical industry and other industries. Its production process includes complex catalytic reaction and multiphase equilibrium system. The slight fluctuation of reaction temperature, pressure, raw material ratio and other parameters can affect the product quality. The current production mode in the industry relies on traditional control means, which can realize the basic reaction process operation, but lacks systematic control for the dynamic changes of phase equilibrium state in oleic acid production, real-time correlation between reaction parameters and early prediction of abnormal trends. In the production process, the collection, analysis, reasoning and adjustment of various process parameters are often in a scattered state, and a complete closed-loop supervision link has not been formed, which leads to the fact that process adjustment often lags behind the actual changes of the reaction system, and it is difficult to ensure the continuous and stable production of high-quality oleic acid.

[0003] The existing technology has two obvious shortcomings: on the one hand, the phase equilibrium analysis and parameter reasoning are independent of each other. Although the phase component related parameters are calculated by using the phase equilibrium and thermodynamic supervision model, the missing parameters in the collected data cannot be accurately completed by effectively combining the missing-Bayesian reasoning model, which makes the process adjustment scheme based on incomplete data deviate, and further affects the purity and yield of oleic acid products; on the other hand, the data interaction and process adjustment execution lack cooperation. The transmission of various data generated in the production process and intelligent reasoning results is delayed, and the response of the process dynamic adjustment module to the related instructions does not form a stepwise correction logic, which leads to unreasonable adjustment range of reaction conditions and difficulty in maintaining the stable state of the oleic acid production process. SUMMARY

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present application provides a supervision system and method for industrial oleic acid production.

[0005] The technical scheme adopted by the present application is a supervision system for industrial oleic acid production, comprising a raw material reaction control module, a real-time parameter collection module, a phase equilibrium analysis module, a data intelligent reasoning module, a cloud centralized management module and a process dynamic adjustment module. The raw material reaction regulation module sets the starting and running parameters of the reaction process according to the catalytic reaction conditions, raw material ratio parameters and reaction kettle pressure value in the industrial oleic acid production process; the real-time parameter acquisition module continuously captures and converts the temperature change curve, oleic acid concentration gradient distribution, reaction flow rate change and reaction time sequence data in the reaction kettle; the phase equilibrium analysis module receives the oleic acid-water-catalyst three-phase system parameters transmitted by the real-time parameter acquisition module, and obtains the mole fraction distribution and interfacial tension value of each phase component through thermodynamic equation operation; the data intelligent reasoning module, based on the parameter deviation value output by the phase equilibrium analysis module, combines the abnormal parameter occurrence probability in the historical production data to perform parameter omission completion and abnormal trend deduction; the cloud centralized management module classifies and stores the processing results transmitted by the data intelligent reasoning module, and establishes a data synchronization link with the UniformancePHD production platform and K / 3Cloud generation management cloud to realize cross-platform data sharing; the process dynamic adjustment module receives the adjustment instructions issued by the cloud centralized management module to perform stepwise correction on the stirring rate, heating power, raw material feed quantity and catalyst addition frequency of the reaction kettle.

[0006] Further, when the real-time parameter acquisition module and the phase equilibrium analysis module work cooperatively, the correlation between the fatty acid double bond retention rate and the reaction temperature in the industrial oleic acid production process is calculated by the following model formula:

[0007] wherein, is the fatty acid double bond retention rate, is the initial double bond concentration, is the temperature influence coefficient, is the real-time reaction temperature, is the reference reaction temperature, is the reaction duration, is the initial total fatty acid concentration, is the pressure influence factor, is the pressure in the reaction kettle; When the cloud centralized management module interacts with the UniformancePHD production platform, the following model formula is used to correct the acid value prediction value of the oleic acid product:

[0008] wherein, is the corrected oleic acid product acid value, is the measured oleic acid product acid value, is the feed flow rate correction coefficient, is the real-time raw material feed flow rate, is the standard feed flow rate, is the enthalpy change correction coefficient, is the real-time reaction enthalpy change, is the standard reaction enthalpy change.

[0009] Further, the phase equilibrium analysis module and the data intelligent inference module interact with data, and the distribution coefficient of oleic acid in oil-water two-phase is calculated by the following model formula:

[0010] wherein, is the distribution coefficient of oleic acid in oil phase and water phase, is the mole fraction of oleic acid in oil phase, is the activity coefficient of oleic acid in oil phase, is the mole fraction of oleic acid in water phase, is the activity coefficient of oleic acid in water phase, is the change of molar volume of oleic acid from water phase to oil phase, is the system pressure, is the standard atmospheric pressure, is the gas constant, is the absolute temperature.

[0011] Further, when the data intelligent inference module processes abnormal data output by the phase equilibrium analysis module using the missing-Bayesian inference model, the posterior probability of abnormal parameters per unit time is calculated by the following model formula:

[0012] wherein, is the posterior probability of the event occurs under the observation data , is the likelihood probability of the observation data occurs when the event occurs, is the prior probability of the event occurs, is the weight factor of the th abnormal parameter, is the deviation degree of the th parameter, is the total number of abnormal parameters, is the marginal probability of the observation data occurs, is the likelihood probability of the observation data occurs when the event occurs, is the prior probability of the event occurs, is the total number of possible events.

[0013] ​Further, the cloud centralized management module calculates the theoretical yield of oleic acid of each production batch when analyzing the production plan data generated and managed by K / 3Cloud through the following model formula:

[0014] wherein, is the theoretical yield of oleic acid, is the density of the reaction mixture, is the effective volume of the reaction kettle, is the mass fraction of fatty acid in the raw material, is the reaction conversion rate, is the mass fraction of impurities in the product.

[0015] Further, when the process dynamic adjustment module adjusts the reaction conditions according to the instructions issued by the cloud centralized management module, the correction value of the stirring rate is calculated through the following model formula:

[0016] wherein, is the corrected stirring rate, is the reference stirring rate, is the correction coefficient of the oleic acid concentration, is the real-time oleic acid concentration, is the target oleic acid concentration, is the viscosity correction coefficient, is the dynamic viscosity of the reaction mixture, is the reference dynamic viscosity.

[0017] Further, the phase equilibrium analysis module comprises a component mole fraction calculation unit, a phase interface parameter analysis unit, a phase equilibrium stability evaluation unit, and an abnormal data marking unit. The component mole fraction calculation unit receives the temperature, pressure, and concentration data of each component of the oleic acid-water-catalyst three-phase system transmitted by the real-time parameter acquisition module, substitutes these data into the phase equilibrium thermodynamic equation, and obtains the mole fraction and chemical potential of each phase through iterative calculation. The calculation results are stored in the cache area of the unit and marked with a time stamp. The phase interface parameter analysis unit extracts data from the cache area of the component mole fraction calculation unit, calculates the tension value of the three-phase interface, determines the moving speed and thickness change of the phase interface by analyzing the diffusion coefficient of each phase component at different temperatures, converts the calibration data into a standard format, and transmits them to the unit output end. The phase equilibrium stability evaluation unit receives the phase interface parameters transmitted by the phase interface parameter analysis unit, combines the stirring intensity data in the reaction kettle, calculates the phase mixing uniformity index, generates phase equilibrium stability evaluation data by comparing the phase equilibrium state parameters at different reaction stages, and stores them in the database of the unit. The abnormal data marking unit retrieves the phase equilibrium stability evaluation data from the database of the phase equilibrium stability evaluation unit, compares them with the preset phase equilibrium threshold, marks the data that exceeds the threshold range, and transmits the marked data together with the normal data to the data intelligent reasoning module.

[0018] Further, the data intelligent reasoning module comprises a parameter omission completion unit, a parameter trend analysis unit, and an abnormal pattern recognition unit. The parameter omission completion unit receives each parameter data transmitted by the phase equilibrium analysis module, identifies the missing value position in the data sequence, estimates and fills the missing parameters based on the missing-Bayesian inference model and the parameter distribution characteristics under similar reaction conditions in the same period, temporarily stores the filled complete data sequence in the temporary storage area of the unit, and transmits the trend analysis results to the result cache area of the unit. The abnormal pattern recognition unit retrieves the trend analysis results from the result cache area of the parameter trend analysis unit, identifies and classifies the abnormal patterns that may occur in the current parameter sequence by combining the probability distribution of abnormal parameter combinations in historical production data, integrates the classified abnormal recognition results and normal trend data, and transmits them to the cloud centralized management module.

[0019] Further, the cloud centralized management module comprises a data format standardization unit, a cross-platform data interaction unit, a production data correlation index unit, and a panoramic view generation unit. The data format standardization unit receives integrated data transmitted by the data intelligent reasoning module, checks and converts the format of the data to make it comply with the data interaction standards of the UniformancePHD production platform and the K / 3Cloud production management cloud, and stores the converted standardized data in the distributed database of the unit. The cross-platform data interaction unit retrieves standardized data from the distributed database of the data format standardization unit, transmits the data to the UniformancePHD production platform and the K / 3Cloud production management cloud through the established data synchronization link, and receives production state data fed back by the two platforms and stores the feedback data in the feedback data area of the unit. The production data correlation index unit extracts production state data from the feedback data area of the cross-platform data interaction unit, classifies and indexes the data, establishes the correlation between the data and the production batch, the reaction kettle number, and the time sequence, and stores the correlation index results in the index database of the unit. The panoramic view generation unit retrieves the correlation index results from the index database of the production data correlation index unit, combines the historical adjustment records of the process dynamic adjustment module, generates a panoramic data view of the production process, transmits the panoramic data view to the view display cache area of the unit, and simultaneously transmits a preliminary draft of the adjustment instruction to the process dynamic adjustment module.

[0020] A supervision method for industrial oleic acid production, which is applied to a supervision system for industrial oleic acid production, comprising the following steps: In the first step, the raw material reaction control module receives the execution parameters returned by the process dynamic adjustment module, combines the preset reaction initial conditions, sets the feed valve opening degree of the reaction kettle, the working frequency of the catalyst adding pump, and the starting power of the heating device, so that the raw materials enter the reaction kettle at a set ratio and start the reaction. In the second step, the real-time parameter acquisition module continuously acquires the temperature, pressure, oleic acid concentration, reaction flow rate, and reaction time parameters in the reaction process through the sensors deployed at different positions of the reaction kettle, converts the acquired analog signals into digital signals, and transmits the digital signals to the phase equilibrium analysis module after filtering processing. In the third step, the phase equilibrium analysis module receives the digital signals transmitted by the real-time parameter acquisition module, extracts the oleic acid-water-catalyst three-phase system parameters included in the signals, substitutes them into the phase equilibrium and thermodynamic supervision model for calculation, obtains the molar fraction, chemical potential, and phase interface tension parameters of each phase component, and transmits the calculation results to the data intelligent reasoning module. In the fourth step, the data intelligent inference module receives the operation result transmitted by the phase equilibrium analysis module, uses the missing-Bayesian inference model to complete the missing parameters in the result, performs abnormal trend deduction on the completed parameter sequence, and transmits the processing result obtained by the deduction to the cloud centralized management module. In the fifth step, the cloud centralized management module receives the processing result transmitted by the data intelligent inference module, stores the result in a classified manner, synchronizes data with the Uniformance PHD production platform and the K / 3 Cloud generation management cloud, generates a process adjustment instruction according to the production plan and state data obtained by the synchronization, and transmits the instruction to the process dynamic adjustment module. In the sixth step, the process dynamic adjustment module receives the process adjustment instruction transmitted by the cloud centralized management module, analyzes the parameters in the instruction, performs stepwise correction on the stirring rate, heating power, raw material feeding amount and catalyst adding frequency of the reaction kettle according to the analysis result, and returns the corrected execution parameters to the raw material reaction control module.

[0021] Beneficial effects: the present application proposes a supervision system and method for industrial oleic acid production, through the close connection of the phase equilibrium analysis module and the data intelligent inference module, the phase component parameters calculated by the phase equilibrium and thermodynamic supervision model are directly input into the data intelligent inference module combined with the missing-Bayesian inference model, the missing parameters are accurately completed, the adjustment deviation problem caused by the separation of phase equilibrium analysis and parameter inference is solved, the purity and yield of oleic acid product are improved; with the cooperation of the cloud centralized management module and the process dynamic adjustment module, the production data, phase equilibrium analysis and intelligent inference result are transmitted in real time through a special link, the data interaction delay is eliminated, at the same time, the process dynamic adjustment module responds to the instruction according to the stepwise correction logic, the unreasonable adjustment range problem is solved, and the production stability is maintained. In addition, the modules form a closed loop supervision link, realize the seamless connection of parameter acquisition, analysis, inference, management and adjustment, improve the production continuity and stability, and ensure the continuous output of high-quality oleic acid. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The system module composition diagram of the present application is shown in the figure; Figure 2 The method step flow chart of the present application is shown in the figure. DETAILED DESCRIPTION

[0023] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict, and the present application will be further described in detail in combination with the drawings and specific embodiments.

[0024] As Figure 1As shown, a regulatory system for industrial oleic acid production includes a raw material reaction control module, a real-time parameter acquisition module, a phase equilibrium analysis module, a data intelligent reasoning module, a cloud centralized management module, and a process dynamic adjustment module. The raw material reaction control module and the real-time parameter acquisition module establish a bidirectional data interaction link through an industrial data transmission protocol. The output end of the real-time parameter acquisition module is connected to the input end of the phase equilibrium analysis module through an encrypted data channel. The operation result of the phase equilibrium analysis module is transmitted to the data intelligent reasoning module through a special data interface. The processing result of the data intelligent reasoning module is sent to the cloud centralized management module through a verification link. The feedback signal of the cloud centralized management module acts on the process dynamic adjustment module through an instruction transmission channel. The execution parameters of the process dynamic adjustment module are fed back to the raw material reaction control module through a closed-loop feedback link. The raw material reaction control module sets the start-up and operation parameters of the reaction process according to the catalytic reaction conditions, raw material ratio parameters, and reaction kettle pressure values in the industrial oleic acid production process. Specifically, the raw material reaction control module is the starting control core of the industrial oleic acid production process and the regulatory system. Its technical parameters include catalytic reaction conditions, raw material ratio, reaction kettle pressure, and other key elements. The accurate setting of these parameters directly affects the direction and efficiency of the subsequent reaction. Through the closed-loop feedback link with the process dynamic adjustment module, the module realizes integrated management of the initial reaction parameters and process adjustment parameters, provides a stable initial reaction environment for the entire production process, and is the basis for ensuring that the oleic acid production proceeds according to the preset process path. In the specific implementation process, the raw material reaction control module first receives the execution parameters of the previous period returned by the process dynamic adjustment module, including the stirring rate correction value, the heating power adjustment value, and the actual value of the raw material feed quantity. Combined with the preset initial reaction conditions, the module calculates the ratio of fatty acids in the raw material to the catalyst, with the fatty acid mass fraction in the raw material controlled at 65%-75% and the catalyst addition amount determined as 2%-3% of the total mass of the raw material. According to the calculation result, the module sets the initial temperature of the reaction kettle at 180℃-220℃, the initial pressure at 0.3MPa-0.5MPa, and the initial stirring rate at 200r / min-300r / min. At the same time, the module calculates and sets the opening of the feed valve to 30%-50% according to the conveying rate of the raw material, so that the raw material enters the reaction kettle at a flow rate of 5m³-8m³ per hour. The working frequency of the catalyst addition pump is set to start once every 30 minutes, with each duration being 10 seconds-15 seconds. The starting power of the heating device is set at 40%-60% of the total power to ensure that the temperature in the reaction kettle rises to the target initial temperature at a rate of 2℃-3℃ per minute. After these parameter settings are completed, the module transmits instructions to the corresponding execution mechanism and continuously receives feedback signals from the execution mechanism to accumulate initial data for subsequent parameter adjustment.

[0025] The real-time parameter acquisition module continuously captures and converts the temperature change curve in the reactor, the oleic acid concentration gradient distribution, the reactant flow rate change, and the reaction time series data; Specifically, the real-time parameter acquisition module is responsible for continuously capturing and converting key reaction parameters during the industrial oleic acid production process. These parameters include the temperature curve within the reactor, the oleic acid concentration gradient, reactant flow rate variations, and reaction time series. These data form the basis for the calculations performed by the phase equilibrium analysis module and are directly related to the accuracy of subsequent phase equilibrium determination and process adjustments. This module connects to the phase equilibrium analysis module via an encrypted data channel, ensuring the timeliness and integrity of data transmission and providing data support for dynamic monitoring of the entire system.

[0026] During implementation, the real-time parameter acquisition module collects data using sensors deployed at various locations within the reactor. Three to five temperature sensors are installed at the top, middle, and bottom of the reactor. These sensors have a measurement range of 0°C to 300°C and an accuracy of ±0.5°C. The data acquisition frequency is set to once per second, capturing real-time temperature changes within different areas of the reactor. A pressure sensor is installed at the top of the reactor with a measurement range of 0 MPa to 1 MPa and an accuracy of ±0.01 MPa. Pressure data is collected every two seconds. A concentration sensor and a flow rate sensor are installed at the reactor outlet. The concentration sensor measures oleic acid concentration with a measurement range of 0% to 100% and an accuracy of ±1%, collecting data every five seconds. The flow rate sensor has a measurement range of 0 m³ / h to 20 m³ / h and an accuracy of ±0.1 m³ / h, collecting data once per second. Time series data is recorded to the nearest second using a built-in timer. The collected analog signals are converted to digital signals using an analog-to-digital converter with a 16-bit resolution to ensure data accuracy. The converted digital signals are filtered to remove high-frequency interference. The temperature signal is filtered using a sliding average filter with a window size of 5 data points, and the pressure signal is filtered using a median filter with a window size of 3 data points. The processed temperature, pressure, oleic acid concentration, reactant flow rate, and time series data are transmitted once per minute via an encrypted data channel to the phase equilibrium analysis module. The module stores both the raw data and the processed data on a local hard drive with a storage capacity of 100-200GB for 30 days.

[0027] The phase equilibrium analysis module receives the parameters of the oleic acid-water-catalyst three-phase system transmitted by the real-time parameter acquisition module, and obtains the mole fraction distribution of each phase component and the interfacial tension value through thermodynamic equation calculation; Specifically, the phase equilibrium analysis module is the key link connecting real-time parameter acquisition and data intelligent reasoning. Based on the received parameters of the oleic acid-water-catalyst three-phase system, the module uses thermodynamic equations to calculate the mole fraction distribution of each phase component and the interfacial tension value. These parameters are important basis for judging whether the reaction system is in a stable state. The module transmits the calculation results to the data intelligent reasoning module through a special data interface, providing basic data for parameter completion and abnormal trend deduction, and playing an important role in maintaining the phase equilibrium state of the reaction system.

[0028] In the specific implementation process, the phase equilibrium analysis module receives the data of the three-phase system temperature, pressure, and component concentration transmitted by the real-time parameter acquisition module. First, the integrity of the data is checked. If the data is missing or abnormal, a retransmission request is sent to the real-time parameter acquisition module. After the check is passed, the module converts the temperature data into absolute temperature, the pressure data into Pascal units, and the concentration data into molar concentration. Then, the module calls the built-in phase equilibrium thermodynamic equation and substitutes the converted parameters for iterative calculation. The number of iterations is set to 50-100, and the convergence error of the calculation results is less than 0.001. The mole fraction of the oleic acid phase, the water phase, and the catalyst phase is calculated, with the mole fraction of the oleic acid phase controlled in the range of 0.6-0.8, the mole fraction of the water phase controlled in the range of 0.1-0.2, and the mole fraction of the catalyst phase controlled in the range of 0.05-0.15. At the same time, the module calculates the interfacial tension value based on the density and viscosity data of each phase. The interfacial tension value between the oleic acid phase and the water phase ranges from 20 mN / m to 30 mN / m, and the interfacial tension value between the water phase and the catalyst phase ranges from 15 mN / m to 25 mN / m. After the calculation is completed, the module formats the results, adds identification information such as data acquisition time and reactor number, and transmits them to the data intelligent reasoning module through a special data interface. The module stores the calculation results in the internal database each time. The database uses a relational structure, and each data table contains fields such as parameter name, value, calculation time, and error range, which facilitates subsequent data query and traceability.

[0029] The data intelligent reasoning module performs parameter completion and abnormal trend deduction based on the parameter deviation value output by the phase equilibrium analysis module and the probability of abnormal parameters in historical production data. Specifically, the data intelligent reasoning module performs parameter completion and abnormal trend deduction based on the parameter deviation value output by the phase equilibrium analysis module and the probability of abnormal parameters in historical production data. The core is to use the missing-Bayesian reasoning model to process incomplete data and identify abnormal patterns. The module is connected to the cloud centralized management module through a verification link to ensure the accuracy and reliability of the processing results, providing decision support for the cloud centralized management module to generate adjustment instructions, which is of great significance to improve the system's ability to predict abnormal situations.

[0030] In the implementation process, the data intelligent inference module receives the operation results transmitted by the phase balance analysis module, first cleans the data to remove outliers that are obviously beyond the reasonable range, wherein the reasonable range of the mole fraction is set to 0-1, and the reasonable range of the interfacial tension is set to 0 mN / m-50 mN / m. After cleaning, the module uses the missing-Bayesian inference model to identify missing values in the data sequence. The missing value identification uses a sliding window method, and the window size is 10 data points. When three consecutive data points are null, it is determined that the data is missing. For the identified missing values, the module retrieves historical data under the same reaction stage and similar raw material ratio conditions in the past six months, calculates the mean and standard deviation of the corresponding parameters in the historical data, wherein the mean is used to fill in the missing values, and the standard deviation is used to evaluate the filling error. After filling, the module calculates the deviation rate of each parameter from the historical normal parameter value, and the deviation rate calculation formula is (current value - historical mean) / historical mean. When the absolute value of the deviation rate is greater than 10%, it is marked as an abnormal parameter. The module performs trend analysis on the marked abnormal parameters, calculates the change slope of the abnormal parameters in the past 30 minutes, and when the absolute value of the slope is greater than 0.05 / minute, it is determined that the abnormal trend is intensifying. The module integrates the completed parameter sequence, abnormal parameter mark and trend analysis result into a processing report, transmits it to the cloud centralized management module through a verification link, and uses a data checksum algorithm to ensure that there is no loss or tampering during data transmission. At the same time, the module stores the historical data, model parameters and inference results used in the inference process in the local server, with a storage period of 90 days, which facilitates subsequent model optimization and result tracing.

[0031] The cloud centralized management module classifies and stores the processing results transmitted by the data intelligent inference module, and establishes a data synchronization link with the Uniformance PHD production platform and the K / 3 Cloud generation management cloud to share data across platforms. Specifically, the cloud centralized management module is responsible for classifying and storing the processing results transmitted by the data intelligent inference module, and realizing data synchronization and sharing with related production platforms and management clouds, and is the data hub of the entire system. By establishing a cross-platform data synchronization link, it integrates various data in the production process to provide comprehensive production status information for the process dynamic adjustment module, and plays an important role in realizing centralized and intelligent management of industrial oleic acid production.

[0032] In the implementation process, the cloud centralized management module receives the processing report transmitted by the data intelligent reasoning module, first analyzes the format of the report, extracts the parameter name, value, timestamp, abnormal flag and other information. After analysis, the module classifies and stores the data according to the production batch, reaction kettle number, parameter type, uses a distributed database architecture, and the data storage capacity is 1TB-2TB, supporting 100-200 read-write operations per second. The module connects with the relevant production platform and management cloud through the preset data synchronization protocol, and the synchronization frequency is set to every 5 minutes, and the synchronized data includes real-time parameters, phase balance analysis results, and abnormal reasoning results. During data synchronization, the module encrypts the transmitted data using AES-256 encryption algorithm to ensure data transmission security. At the same time, the module receives production plan data and equipment state data from the production platform and management cloud, including daily production batch, target output per batch, and equipment state data including reaction kettle running time, sensor calibration time, etc. The module associates the received feedback data with the locally stored processing results to establish a correspondence between production data and equipment state, and identifies key equipment parameters affecting oleic acid yield through data mining algorithms. The module generates a visual report based on the associated comprehensive data, including hourly oleic acid yield, parameter abnormality frequency, and equipment operating status, and stores it in the cloud database. Authorized users can access and view the report through a dedicated client. At the same time, the module generates preliminary process adjustment suggestions based on the comprehensive data and transmits them to the process dynamic adjustment module, including adjustment parameter name, recommended adjustment range, and adjustment timing.

[0033] The process dynamic adjustment module receives the adjustment instructions issued by the cloud centralized management module to make step-by-step corrections to the reaction kettle stirring rate, heating power, raw material feed quantity, and catalyst addition frequency.

[0034] Specifically, the process dynamic adjustment module, as the execution terminal of the system, receives the adjustment instructions issued by the cloud centralized management module to make step-by-step corrections to the key operating parameters of the reaction kettle, including stirring rate, heating power, raw material feed quantity, and catalyst addition frequency. The module transmits the execution parameters back to the raw material reaction control module through a closed-loop feedback link, forming a closed-loop adjustment mechanism for the production process, which directly affects the stability of the reaction system and improves the quality of the oleic acid product.

[0035] During implementation, the dynamic process adjustment module receives adjustment commands from the cloud-based centralized management module. These commands include the parameter name, target value, adjustment step, and adjustment interval. The module first verifies the validity of the command, checking whether the target value is within the equipment's permitted operating range. The permitted range for stirring rate is 100-500 rpm, for heating power is 0%-100%, for raw material feed rate is 3-10 m³ / h, and for catalyst addition frequency is every 10-60 minutes. Once these commands pass verification, the module divides the target value into multiple adjustment stages based on the adjustment step. For example, if the difference between the target stirring rate and the current value is 100 rpm and the adjustment step is set to 20 rpm, this results in five adjustment stages, each with a five-minute interval. During each adjustment stage, the module sends control signals to the corresponding actuators. The stirring rate is achieved by adjusting the motor frequency, the heating power by adjusting the thyristor conduction angle, the raw material feed rate by adjusting the valve opening, and the catalyst addition frequency by adjusting the timer setting. During the adjustment process, the module collects feedback parameters of the actuator in real time, such as the actual stirring rate, actual heating power, etc. When the deviation between the actual value and the stage target value exceeds 5%, the adjustment time of the stage is extended until the deviation is less than 5%. After all adjustment stages are completed, the module transmits the final execution parameters, including the stabilized stirring rate, heating power, raw material feed amount, catalyst addition frequency and adjustment completion time, to the raw material reaction control module through a closed-loop feedback link. At the same time, the module stores the instruction content, execution process data, and final results of this adjustment in a local log. The log is kept for 180 days to facilitate subsequent process optimization analysis.

[0036] Preferably, when the real-time parameter acquisition module and the phase equilibrium analysis module work together, the correlation between the fatty acid double bond retention rate and the reaction temperature in the industrial oleic acid production process is calculated by the following model formula:

[0037] in, is the fatty acid double bond retention rate (dimensionless), is the initial double bond concentration ( ), is the temperature influence coefficient , is the real-time reaction temperature (°C), is the reference reaction temperature (°C), Reaction duration , is the initial total fatty acid concentration ( ), is the pressure influencing factor ( ), The pressure in the reactor is ; The cloud centralized management module corrects the acid value prediction value of the oleic acid product when interacting with the Uniformance PHD production platform using the following model formula:

[0038] Wherein, The corrected acid value of the oleic acid product (K) is The measured acid value of the oleic acid product (K) is The feed flow rate correction coefficient (dimensionless) is The real-time raw material feed flow rate (L / h) is The standard feed flow rate (L / h) is The enthalpy correction coefficient (dimensionless) is The real-time reaction enthalpy change (kJ / mol) is The standard reaction enthalpy change (kJ / mol) is . .

[0039] ​​​Specifically, the real-time parameter acquisition module and the phase equilibrium analysis module work together around the correlation between the fatty acid double bond retention rate and the reaction temperature, which is crucial for the quality of oleic acid product and directly affects the chemical stability and application performance of oleic acid. The real-time parameter acquisition module continuously captures data such as temperature change curve in the reaction kettle, initial double bond concentration, initial total fatty acid concentration, pressure in the reaction kettle, and reaction duration, etc. The real-time acquisition range of reaction temperature is 180℃-220℃, updated every second; the initial double bond concentration is controlled at 0.8mol / L-1.2mol / L, and the initial total fatty acid concentration is maintained at 1.5mol / L-2.0mol / L; the pressure in the reaction kettle is stabilized at 0.3MPa-0.5MPa, and the pressure value is recorded every 2 seconds; the reaction duration is accurate to seconds, continuously accumulated from the start of the reaction. After receiving these data, the phase equilibrium analysis module combines the preset temperature influence coefficient (value range: 0.002L / (mol・℃・min)-0.005L / (mol・℃・min)), the reference reaction temperature (set at 200℃), and the pressure influence factor (value range: 0.001L / (mol・Pa^0.5)-0.003L / (mol・Pa^0.5)) to perform calculations. During the calculation process, the module first verifies the temperature data to ensure that it is within the reasonable reaction interval, and marks and corrects the data outside the range of 180℃-220℃ using the mean value of adjacent valid data; the pressure data need to be converted to Pascal units before participating in the calculation to ensure unit consistency. Through such collaborative operation, the obtained fatty acid double bond retention rate can accurately reflect the retention status of double bonds under the current reaction conditions, providing key basis for subsequent process adjustment. When the retention rate is less than 60%, the system will trigger further parameter optimization process to ensure that the double bond retention of the oleic acid product meets the production requirements. At the same time, the cloud centralized management module is also rigorous in the correction process of the predicted value of the oleic acid product acid value when interacting with the Uniformance PHD production platform data. The real-time raw material feed flow rate is controlled at 5m³ / h-8m³ / h, and the data is collected every 5 seconds; the real-time reaction enthalpy is obtained through a heat sensor, ranging from -2000J / mol to -1500J / mol; the feed flow rate correction coefficient is set to 0.05-0.1, and the enthalpy correction coefficient is set to 0.03-0.07, with the standard feed flow rate being 6m³ / h and the standard reaction enthalpy being -1800J / mol. After correction, the measured acid value of the oleic acid product can be controlled within an error range of ±0.2mgKOH / g, effectively improving the reliability of the acid value data and providing accurate indicators for production quality evaluation.

[0040] Preferably, during the data interaction between the phase equilibrium analysis module and the data intelligent reasoning module, the distribution coefficient of oleic acid in oil-water two-phase is calculated by the following model formula:

[0041] in, is the partition coefficient of oleic acid between oil phase and water phase (dimensionless), is the mole fraction of oleic acid in the oil phase (dimensionless), is the activity coefficient of oleic acid in the oil phase (dimensionless), is the mole fraction of oleic acid in the aqueous phase (dimensionless), is the activity coefficient of oleic acid in water (dimensionless), is the molar volume change of oleic acid transferred from the aqueous phase to the oil phase ( ), is the system pressure (Pa), is standard atmospheric pressure (Pa), is the gas constant , is the absolute temperature (K).

[0042] Specifically, the data interaction core of the phase equilibrium analysis module and the data intelligent inference module is the calculation of the distribution coefficient of oleic acid in the oil-water two-phase system, which is a key indicator for evaluating the distribution of oleic acid in the two-phase system and is directly related to the reaction efficiency and product separation effect. The phase equilibrium analysis module receives the parameters of the oleic acid-water-catalyst three-phase system transmitted by the real-time parameter acquisition module, including the mole fraction, activity coefficient of oleic acid in the oil phase and water phase, as well as the system pressure, temperature and other data. Among them, the mole fraction of oleic acid in the oil phase is controlled in the range of 0.6-0.8, and the mole fraction in the water phase is controlled in the range of 0.05-0.15; the activity coefficient of oleic acid in the oil phase is 1.0-1.2, and the activity coefficient in the water phase is 1.5-2.0. These activity coefficients are determined by fitting the experimental data in the early stage and stored in the module database, which can be fine-tuned according to the different batches of raw materials. The system pressure is collected by the pressure sensor in real time, ranging from 0.3 MPa to 0.5 MPa, which is converted to pascal units for calculation; the absolute temperature is converted from the temperature in the reaction kettle, ranging from 453 K to 493 K; the change of the molar volume of oleic acid from the water phase to the oil phase is set to -0.00002 m³ / mol to -0.00001 m³ / mol, the gas constant is the standard value of 8.314 J / (mol·K), and the standard atmospheric pressure is set to 101325 Pa. During the data interaction process, the phase equilibrium analysis module first checks the received mole fraction data to ensure that the mole fraction of the oil phase, water phase and catalyst phase is within the range of 0.99-1.01, and normalizes the data that exceeds this range; temperature and pressure data need to be filtered to remove high-frequency interference signals, temperature uses sliding average filtering (window size 5 data points), and pressure uses median filtering (window size 3 data points). The processed data is transmitted to the data intelligent inference module, which verifies the current calculation results combined with historical distribution coefficient data (records of the last 6 months). When the deviation exceeds 10%, the data backtracking mechanism is started to check whether there is an abnormality in the original data acquisition and conversion process. Through such interaction and calculation, the obtained distribution coefficient can accurately reflect the distribution of oleic acid in the two-phase system, providing data support for optimizing reaction conditions and improving the extraction efficiency of oleic acid. When the distribution coefficient is less than 0.5, the system will prompt to adjust the stirring rate or temperature to improve the phase distribution effect.

[0043] Preferably, when the data intelligent inference module uses the missing-Bayesian inference model to process abnormal data output by the phase equilibrium analysis module, the posterior probability of parameter abnormality per unit time is calculated by the following model formula:

[0044] wherein, is the observation data the event Posterior probability of occurrence (dimensionless), for event when observed data occurs, for event Prior probability of occurrence (dimensionless), for the weight factor for the deviation of the parameter (dimensionless), total number of for observed data Marginal probability of occurrence (dimensionless), for event when observed data occurs, for event Prior probability of occurrence (dimensionless), total number of possible events (dimensionless).

[0045] Specifically, when the data intelligent inference module processes the abnormal data output by the phase equilibrium analysis module using the missing-Bayesian inference model, the entire process is centered around the calculation of the posterior probability of a parameter abnormality at a certain time. This probability value is the core basis for judging the possibility of parameter abnormality and plays an important role in timely identifying reaction abnormalities and avoiding production accidents. The module first receives various parameter data transmitted by the phase equilibrium analysis module, including phase component mole fraction, interfacial tension, reaction temperature, pressure, etc. After integrity check, these data enter the processing flow. In the calculation process, event A represents a specific parameter abnormality (such as temperature exceeding the set range), and observation data D is the value sequence of the parameter and associated parameters currently collected. The likelihood probability P(D|A) is determined by analyzing the frequency of occurrence of the corresponding observation data in the historical data when the parameter is abnormal. For example, in the past 1000 temperature abnormality events, a certain temperature change sequence appeared 300 times, and its likelihood probability is 0.3; the prior probability P(A) is calculated based on the ratio of the total number of parameter abnormality occurrences to the total production time in the past 6 months, usually taking a value range of 0.01-0.05. The weight factor λ_i of the abnormal parameter is set according to the degree of influence of the parameter on the reaction, and the weight of key parameters such as temperature and pressure is 0.8-1.0, and the weight of secondary parameters such as interfacial tension is 0.3-0.5; the deviation degree δ_i of the parameter is calculated by (current value-normal range mean value) / normal range standard deviation, and the normal range is determined according to the historical data, for example, the temperature normal range is 180-220℃, the mean value is 200℃, and the standard deviation is 10℃, so the deviation degree of a temperature value of 230℃ is 3. The total number of abnormal parameters n is determined according to the actual number of abnormal parameters detected, and the total number of possible events m covers all possible parameter abnormality combinations (such as temperature abnormality, pressure abnormality, temperature and pressure abnormality at the same time, etc.), usually set to 10-20 kinds. The marginal probability P(D) is obtained by summing the product of the likelihood probability and the prior probability of all possible events. In the calculation process, the module calculates the posterior probability of each parameter abnormality event separately, and when the posterior probability of an event exceeds 0.7, it is determined as a high-probability abnormal event, which is immediately marked and transmitted to the cloud centralized management module. At the same time, the module saves all parameter values and intermediate results used in each calculation, with a saving period of 90 days, so as to analyze and optimize the accuracy of subsequent abnormal judgment. In addition, in order to ensure the calculation efficiency, the model adopts a step-by-step calculation strategy, first calculates the key parameters with greater impact, and then processes the secondary parameters. The entire calculation process is completed within 10 seconds to meet the real-time requirements.

[0046] Preferably, when the cloud centralized management module analyzes the production plan data generated and managed by K / 3Cloud, it calculates the theoretical yield of oleic acid for each production batch by the following model formula:

[0047] wherein, is the theoretical yield of oleic acid , is the density of the reaction mixture , is the effective volume of the reactor , is the mass fraction of fatty acid in the raw material (dimensionless), is the reaction conversion rate (dimensionless), is the mass fraction of impurities in the product (dimensionless).

[0048] Specifically, when the cloud centralized management module parses the production plan data generated and managed by K / 3Cloud, the calculation process of the theoretical yield of oleic acid for each production batch is rigorous and involves multiple key parameters. This yield data is an important basis for formulating production plans and evaluating production efficiency, and is directly related to the rational allocation of production resources. The production plan data received by the module includes production batch number, planned production time, total amount of raw material supply, and other information. Combined with the reaction mixture density (range 850 kg / m³-950 kg / m³, collected every 10 minutes), the effective volume of the reactor (fixed at 10 m³-20 m³ according to the reactor model), the mass fraction of fatty acid in the raw material (65%-75%, determined by testing when the raw material of each batch enters the factory), the reaction conversion rate (calculated based on the collected oleic acid concentration data, range 70%-90%), and the mass fraction of impurities in the product (5%-10%, measured by online detection equipment every hour), the calculation is performed. Before calculation, the module first verifies the reaction mixture density data, removes data that deviates significantly from the range of 850 kg / m³-950 kg / m³, and replaces it with the average value of the adjacent 10 valid data; the mass fraction of fatty acid in the raw material needs to be compared with the quality inspection report of the raw material of this batch, and when the deviation exceeds 2%, manual confirmation is required. The calculation of the reaction conversion rate is based on the change of the oleic acid concentration, i.e. (current oleic acid concentration - initial oleic acid concentration) / (theoretical maximum oleic acid concentration - initial oleic acid concentration), where the theoretical maximum oleic acid concentration is calculated according to the composition of the raw material. The mass fraction of impurities in the product data needs to be filtered and processed using the sliding average method (window size 5 data points) to reduce measurement fluctuations. The calculated theoretical yield of oleic acid is compared with the historical yield data of the same batch (records saved for the last 3 months), and when the deviation exceeds 15%, the parameter review process is started to check whether the collection and calculation of key parameters such as density and conversion rate are accurate. Through such parsing and calculation, the theoretical yield obtained can provide accurate guidance for the execution of the production plan, and when the actual yield deviates from the theoretical yield by more than 10%, the system will prompt that the reaction conditions or raw material quality need to be checked to ensure that the production is carried out according to the plan.

[0049] Preferably, when the process dynamic adjustment module adjusts the reaction conditions according to the instructions issued by the cloud centralized management module, the correction value of the stirring rate is calculated by the following model formula:

[0050] wherein, is the corrected stirring rate (V), is the reference stirring rate (V0), is the correction coefficient of the stirring rate (dimensionless), is the real-time oleic acid concentration (C), is the target oleic acid concentration (C0), is the correction coefficient of the oleic acid concentration (dimensionless), is the viscosity correction coefficient (dimensionless), is the dynamic viscosity of the reaction mixture (μ), is the reference dynamic viscosity (μ0).

[0051] ​​​​​Specifically, when the process dynamic adjustment module adjusts the reaction conditions according to the instructions issued by the cloud centralized management module, the correction process of the stirring rate is the key link to ensure the uniform mixing of the reaction system and improve the reaction efficiency. The corrected stirring rate directly affects the phase equilibrium state and the reaction rate. The adjustment instructions received by the module include the target oleic acid concentration (set to 1.0 mol / L-1.5 mol / L), the reference dynamic viscosity (set to 0.05 Pa・s-0.1 Pa・s according to the type of raw materials), and other parameters. At the same time, the current oleic acid concentration (measured every 5 seconds), the dynamic viscosity of the reaction mixture (range 0.04 Pa・s-0.12 Pa・s, obtained through a viscosity sensor), and the reference stirring rate (initially set to 200 r / min-300 r / min, which can be adjusted according to the reaction stage) are collected in real time. In the correction calculation process, the oleic acid concentration correction coefficient κ is set according to the reaction stage, with a value of 0.3-0.5 for the early reaction (0-2 hours), 0.5-0.7 for the middle reaction (2-4 hours), and 0.7-0.9 for the late reaction (more than 4 hours); the viscosity correction coefficient μ is fixed at 0.2-0.4 and determined through historical data fitting. The module first checks whether the current oleic acid concentration is within the reasonable range of 0.5 mol / L-2.0 mol / L, and if it is outside the range, the data is marked and corrected using the linear interpolation method; the dynamic viscosity data of the reaction mixture needs to be converted to Pa・s units to ensure consistency with the reference dynamic viscosity unit. In the correction calculation, the oleic acid concentration deviation ratio (current value-target value) / target value is first calculated. When the ratio is positive, it means that the concentration is too high, and the stirring rate may need to be reduced to reduce the reaction contact; when it is negative, the stirring rate needs to be increased. At the same time, the viscosity deviation ratio (current mixture viscosity-reference viscosity) / reference viscosity is calculated. The higher the viscosity, the greater the required stirring rate. Multiply these two deviation ratios by the corresponding correction coefficients, add 1, and then multiply by the reference stirring rate to get the corrected stirring rate. During the correction process, the module monitors the current and speed feedback of the stirring motor in real time (every second). When the actual speed deviates from the corrected value by more than 5 r / min, the motor output frequency is adjusted to reduce the deviation. The corrected stirring rate needs to be controlled within the device allowable range of 100 r / min-500 r / min. If the calculation result exceeds the range, the nearest boundary value is taken as the final execution value. Through this correction process, the stirring rate can accurately adapt to the current reaction conditions. When the viscosity of the reaction mixture increases by 20%, the stirring rate will increase by 5%-10% to maintain good mixing effect and ensure uniform reaction.

[0052] Preferably, the phase equilibrium analysis module comprises a component mole fraction calculation unit, a phase interface parameter analysis unit, a phase equilibrium stability evaluation unit, and an abnormal data marking unit; the component mole fraction calculation unit receives the temperature, pressure, and concentration data of each component of the oleic acid-water-catalyst three-phase system transmitted by the real-time parameter acquisition module, substitutes these data into the phase equilibrium thermodynamic equation, obtains the mole fraction and chemical potential of each phase through iterative calculation, and stores the calculation results in the cache area of the unit and marks the time stamp; the phase interface parameter analysis unit extracts data from the cache area of the component mole fraction calculation unit, calculates the tension value of the three-phase interface, determines the moving rate and thickness change of the phase interface by analyzing the diffusion coefficient of each phase component at different temperatures, converts the calibration data into a standard format, and transmits the converted data to the output end of the unit; the phase equilibrium stability evaluation unit receives the phase interface parameters transmitted by the phase interface parameter analysis unit, combines the stirring intensity data in the reaction kettle, calculates the phase mixing uniformity index, generates phase equilibrium stability evaluation data by comparing the phase equilibrium state parameters at different reaction stages, and stores the data in the database of the unit; the abnormal data marking unit retrieves the phase equilibrium stability evaluation data from the database of the phase equilibrium stability evaluation unit, compares the data with the preset phase equilibrium threshold value, marks the data that exceeds the threshold range, and transmits the marked data and normal data to the data intelligent reasoning module.

[0053] Specifically, in the phase equilibrium analysis module, the component mole fraction calculation unit, the phase interface parameter analysis unit, the phase equilibrium stability evaluation unit, and the abnormal data marking unit work together to comprehensively analyze and evaluate the phase equilibrium state in the oleic acid production process, and provide accurate phase equilibrium data support for subsequent process adjustment. The component mole fraction calculation unit receives the temperature (range 180-220°C, updated every 10 seconds), pressure (0.3-0.5 MPa, collected every 5 seconds), and concentration of each component (oleic acid concentration 0.5-2.0 mol / L, water concentration 1.0-3.0 mol / L, catalyst concentration 0.1-0.3 mol / L, obtained every 15 seconds) of the oleic acid-water-catalyst three-phase system transmitted by the real-time parameter acquisition module, and substitutes these data into the phase equilibrium thermodynamic equation to obtain the mole fraction (oleic acid phase 0.6-0.8, water phase 0.1-0.2, catalyst phase 0.05-0.15) and chemical potential of each phase through iterative calculation (iteration number set to 30-50 times until the calculation error is less than 0.001), and stores the calculation results in the built-in cache area and marks the time stamp to the millisecond. The phase interface parameter analysis unit extracts data from the cache area of the component mole fraction calculation unit, calculates the tension value of the three-phase interface (20-30 mN / m for the oleic acid phase and the water phase, 15-25 mN / m for the water phase and the catalyst phase), analyzes the diffusion coefficient of each phase component at different temperatures (1.0×10-9-3.0×10-9 m² / s for the diffusion coefficient of oleic acid in the oil phase), determines the moving rate (0.01-0.05 mm / min) and thickness change (0.1-0.5 mm) of the phase interface, and transmits the related data to the unit output end after converting them into standard JSON format. The phase equilibrium stability evaluation unit receives the phase interface parameters transmitted by the phase interface parameter analysis unit, combines the stirring intensity data (200-300 r / min, real-time collection) in the reaction kettle, calculates the phase mixing uniformity index (range 0-1, the larger the value, the more uniform the mixing), compares the phase equilibrium state parameters at different reaction stages (0-2 hours for the early stage, 2-4 hours for the middle stage, and more than 4 hours for the late stage), generates phase equilibrium stability evaluation data (including stability index 0.6-0.9, the higher the index, the better the stability), and stores them in the relational database of the unit (database capacity 100 GB, supporting data writing operation 50 times per second). The abnormal data marking unit retrieves the phase equilibrium stability evaluation data from the database of the phase equilibrium stability evaluation unit, compares them with the preset phase equilibrium threshold (lower limit of stability index 0.6, upper limit of interface tension 30 mN / m, etc.), marks the data beyond the threshold range in red, and transmits the marked data together with the normal data to the data intelligent reasoning module in a package (data package size controlled within 1 MB).

[0054] Preferably, the data intelligent inference module comprises a parameter missing completion unit, a parameter trend analysis unit, and an abnormal pattern recognition unit. The parameter missing completion unit receives each parameter data transmitted by the phase balance analysis module, identifies the missing value position in the data sequence, estimates and fills the missing parameter based on the missing-Bayes inference model and in combination with the parameter distribution characteristics under the similar reaction conditions in the same period of history, temporarily stores the filled complete data sequence in the temporary storage area of the unit, the parameter trend analysis unit extracts the complete data sequence from the temporary storage area of the parameter missing completion unit, analyzes the change trend of the parameters in the sequence, determines the inflection point and slope change of the trend by calculating the first and second derivatives of the parameters, transmits the trend analysis result to the result cache area of the unit, the abnormal pattern recognition unit retrieves the trend analysis result from the result cache area of the parameter trend analysis unit, identifies and classifies the abnormal patterns that may appear in the current parameter sequence in combination with the probability distribution of the abnormal parameter combination in the historical production data, and transmits the integrated abnormal recognition result and normal trend data to the cloud centralized management module.

[0055] Specifically, in the data intelligent inference module, the parameter missing completion unit, the parameter trend analysis unit, and the abnormal pattern recognition unit cooperate in sequence to realize deep processing of the phase equilibrium analysis data, provide complete, reliable and trend-judgment data support for the cloud centralized management module, and help precise process adjustment decision. The parameter missing completion unit receives various parameter data (including phase component mole fraction, interfacial tension, stability index, etc.) transmitted by the phase equilibrium analysis module, scans the data sequence point by point to identify the missing value position (the missing value determination standard is that 3 or more consecutive data points are null), calls the parameter distribution characteristic data (stored in a historical database with a capacity of 500 GB, indexed according to reaction conditions) under similar reaction conditions (temperature deviation ± 5℃, pressure deviation ± 0.05MPa) in the same period (in the past 6 months) based on the missing-Bayesian inference model, estimates (the estimation error is controlled within 5%) and fills the missing parameters, and temporarily stores the filled complete data sequence (the data point interval remains consistent with the original sequence, which is 10 seconds per point) in the temporary storage area (uses RAID5 array to ensure data redundancy and fast access) of the unit. The parameter trend analysis unit extracts the complete data sequence from the temporary storage area of the parameter missing completion unit, calculates the first derivative (reflecting the change rate, unit: unit / min) and the second derivative (reflecting the change acceleration, unit: unit / min²) of the parameter using numerical differentiation algorithm, determines the inflection point (the inflection point determination standard is that the first derivative changes from positive to negative or from negative to positive) and the slope change (the slope change is more than 0.1 unit / min is considered as significant change) of the trend through the sign change of the derivative, and transmits the trend analysis results (including inflection point time, slope value, trend direction, etc.) to the result cache area (cache area capacity 10 GB, data retention 24 hours) of the unit. The abnormal pattern recognition unit retrieves the trend analysis results from the result cache area of the parameter trend analysis unit, combines the probability distribution of abnormal parameter combinations in the historical production data (such as the probability of high temperature and low stability index appearing at the same time, etc., the probability data is updated every hour), uses pattern matching algorithm (matching accuracy is set to more than 80%) to identify and classify the possible abnormal patterns (such as stability index continuously decreasing and slope absolute value greater than 0.05 / min, etc. Preset patterns) in the current parameter sequence (into three categories: slight abnormality, moderate abnormality, and severe abnormality), integrates the classified abnormal recognition results (including abnormal type, start time, associated parameters) and normal trend data into a structured report (report format conforms to XML standard), and transmits it to the cloud centralized management module (transmission rate is not less than 10 Mbps to ensure real-time performance).

[0056] Preferably, the cloud centralized management module comprises a data format standardization unit, a cross-platform data interaction unit, a production data correlation index unit, and a panoramic view generation unit. The data format standardization unit receives integrated data transmitted by the data intelligent reasoning module, checks and converts the format of the data to make it comply with the data interaction standards of the UniformancePHD production platform and the K / 3Cloud production management cloud, and stores the converted standardized data in the distributed database of the unit. The cross-platform data interaction unit retrieves standardized data from the distributed database of the data format standardization unit, transmits the data to the UniformancePHD production platform and the K / 3Cloud production management cloud through the established data synchronization link, respectively, receives production status data fed back by the two platforms, and stores the feedback data in the feedback data area of the unit. The production data correlation index unit extracts production status data from the feedback data area of the cross-platform data interaction unit, classifies and indexes the data, establishes the correlation between the data and the production batch, the reaction kettle number, and the time sequence, and stores the correlation index result in the index database of the unit. The panoramic view generation unit retrieves the correlation index result from the index database of the production data correlation index unit, combines the historical adjustment records of the process dynamic adjustment module, generates a panoramic data view of the production process, transmits the panoramic data view to the view display cache area of the unit, and simultaneously transmits a preliminary draft of the adjustment instruction to the process dynamic adjustment module.

[0057] Specifically, in the cloud centralized management module, the data format standardization unit, the cross-platform data interaction unit, the production data correlation index unit, and the panoramic view generation unit work together to realize the standardization processing, cross-platform sharing, correlation analysis, and visualization presentation of production data, providing comprehensive and systematic information support for process dynamic adjustment. The data format standardization unit receives integrated data (including completed parameters, trend analysis, and abnormal identification) transmitted by the data intelligent reasoning module, checks the field format (such as date format YYYY-MM-DDHH:MM:SS, numerical value retains two decimal places, etc.) and data type (such as temperature is floating point type, batch number is string type, etc.) of the data, and performs format conversion (conversion accuracy reaches 100%) on the data that does not meet the requirements, so that it meets the data interaction standard (follows OPCUA protocol specification) of UniformancePHD production platform and K / 3Cloud generation management cloud, and stores the converted standardized data in the distributed database of the unit (uses Hadoop distributed file system, storage capacity 1TB, supports 100 concurrent accesses per second). The cross-platform data interaction unit retrieves standardized data from the distributed database of the first management unit at a frequency of every 5 minutes through the established special data synchronization link (uses optical fiber transmission, bandwidth 100Mbps, delay control within 100ms), and transmits it to UniformancePHD production platform (used for production process monitoring) and K / 3Cloud generation management cloud (used for production plan scheduling), respectively. At the same time, it receives production state data (such as device running state, plan completion progress, etc., data update frequency is every minute) feedback from the two platforms, and stores the feedback data in the feedback data area of the unit (uses partition table structure, partitioned by date) according to the platform source. The production data correlation index unit extracts production state data from the feedback data area of the cross-platform data interaction unit, classifies and indexes the data (establishes B+ tree index to improve query speed) according to production batch number, reaction kettle unique identifier, and time stamp, etc. key fields (classified into device type, plan type, quality type, etc.), establishes the multi-dimensional correlation relationship between data and production batch (every 8 hours as a batch), reaction kettle number (5 reaction kettles in total, numbered as R1-R5), and time sequence (accurate to minute) through key field matching, and stores the correlation index results (stored in the form of correlation table, including each key field and corresponding data address) in the index database of the unit (uses MySQL database, supports transaction processing).The panoramic view generating unit calls the associated index result from the index database of the production data associated index unit, combines the historical adjustment record (including adjustment time, parameter name, value before and after adjustment, etc., saving the record of the last 3 months) of the process dynamic adjustment module, uses data visualization technology (using D3.js visualization library) to generate panoramic data view of the production process (including temperature and pressure trend chart, phase equilibrium state distribution chart, yield curve chart, etc.), transmits the panoramic data view (picture format is PNG, resolution is 1920x1080) to the view display cache area (supports real-time preview on the web side) of the unit, simultaneously generates the preliminary draft of the adjustment instruction (including the suggested adjustment parameter, adjustment amplitude, execution time, etc.) according to the associated data and the historical adjustment effect (such as the time of the adjusted parameter returning to the normal range, etc.), and transmits the preliminary draft to the process dynamic adjustment module after priority sorting.

[0058] In the present application, the UniformancePHD production platform is an industrial process historical data management platform, which can collect, store and monitor real-time parameters and historical data in the industrial oleic acid production process; the K / 3Cloud production management cloud is a cloud production management platform, which is used for production plan scheduling, resource allocation and production state tracking. In the present application, the UniformancePHD production platform receives the standardized data transmitted by the cloud centralized management module, realizes real-time monitoring of the oleic acid production process, and simultaneously feeds back production data such as equipment operation state to the cloud centralized management module; the K / 3Cloud production management cloud also receives the standardized data, which is used for formulating and scheduling the production plan, and feeds back information such as production plan completion progress, and the two realize data interaction through the data synchronization link of the cloud centralized management module, cooperate with other modules, so that the production data can be integrated and shared, provide comprehensive production state and plan information for process dynamic adjustment, and support stable and efficient operation of the entire industrial oleic acid production process and supervision system.

[0059] The industrial oleic acid phase equilibrium and thermodynamic supervision model in the present invention is a model for analyzing the equilibrium state of the oleic acid-water-catalyst three-phase system in oleic acid production. By receiving parameters such as temperature, pressure, and concentration of each component, it calculates the mole fraction, chemical potential, interfacial tension, etc. of each phase to reflect the phase equilibrium state and change characteristics; the omission-Bayesian reasoning model is a model based on historical data and probability statistics, which is used to identify missing values ​​in the data sequence and perform estimation and completion, while analyzing parameter abnormal trends and patterns. In the present invention, the industrial oleic acid phase equilibrium and thermodynamic supervision model provides an operational basis for the phase equilibrium analysis module, so that it can accurately obtain phase equilibrium related data, laying the foundation for subsequent processing; the omission-Bayesian reasoning model supports the operation of the data intelligent reasoning module, the parameter omission completion unit uses it to complete the missing parameters, and the abnormal pattern recognition unit relies on it to identify abnormal patterns. The two work together to make the data transmitted to the cloud centralized management module complete and reliable, provide accurate data support for process adjustment, and ensure stable and efficient production.

[0060] like Figure 2 As shown, a supervision method for industrial oleic acid production is applied to a supervision system for industrial oleic acid production, comprising the following steps: In the first step, the raw material reaction control module receives the execution parameters returned by the process dynamic adjustment module. Based on the preset initial reaction conditions, it sets the reactor feed valve opening, the operating frequency of the catalyst addition pump, and the starting power of the heating device, so that the raw materials enter the reactor at the set ratio and the reaction begins. In the second step, the real-time parameter acquisition module continuously collects temperature, pressure, oleic acid concentration, reactant flow rate, and reaction time parameters during the reaction process through sensors deployed at different locations in the reactor. The collected analog signals are converted into digital signals, filtered, and transmitted to the phase equilibrium analysis module. In the third step, the phase equilibrium analysis module receives the digital signal transmitted by the real-time parameter acquisition module, extracts the parameters of the oleic acid-water-catalyst three-phase system included in the signal, substitutes them into the phase equilibrium and thermodynamic supervision model for calculation, and obtains the mole fraction, chemical potential, and interfacial tension parameters of each phase component. The calculation results are transmitted to the data intelligent reasoning module; In the fourth step, the data intelligent reasoning module receives the calculation results transmitted by the phase equilibrium analysis module, uses the omission-Bayesian inference model to complete any missing parameters in the results, deduces abnormal trends in the completed parameter sequence, and transmits the deduced processing results to the cloud-based centralized management module; In the fifth step, the cloud centralized management module receives the processing result transmitted by the data intelligent inference module, classifies and stores the result, synchronizes data with the Uniformance PHD production platform and the K / 3 Cloud generation management cloud, generates a process adjustment instruction according to the production plan and state data obtained through synchronization, and transmits the instruction to the process dynamic adjustment module. In the sixth step, the process dynamic adjustment module receives the process adjustment instruction transmitted by the cloud centralized management module, analyzes the parameters in the instruction, and performs stepwise correction on the stirring rate, heating power, raw material feeding amount, and catalyst addition frequency of the reaction kettle according to the analysis result, and returns the corrected execution parameters to the raw material reaction control module.

[0061] A monitoring system and method for industrial oleic acid production, the system establishes a close data transmission link between the phase equilibrium analysis module and the data intelligent inference module, directly inputs the phase component parameters calculated by the phase equilibrium and thermodynamic monitoring model into the data intelligent inference module, and the latter relies on the missing-Bayesian inference model to accurately complete the missing parameters, solves the problem of incomplete data caused by the separation of phase equilibrium analysis and parameter inference, makes the process adjustment scheme generated based on complete data more suitable for the actual reaction, and effectively improves the purity and yield of oleic acid product.

[0062] The cloud centralized management module and the process dynamic adjustment module build an efficient collaborative mechanism, the former realizes real-time transmission of production data, phase equilibrium analysis results and intelligent inference conclusions through a special data synchronization link, eliminating data interaction delay; the latter adjusts the reaction conditions according to the stepwise correction logic after receiving the cloud instruction, solves the problem of too large or too small process adjustment range, ensures that the temperature, pressure and other parameters in the reaction kettle are stable in a reasonable range, and maintains the continuous stability of the oleic acid production process.

[0063] The system forms a closed-loop monitoring link through six modules, realizes seamless connection of the whole process from raw material reaction control, real-time parameter acquisition, phase equilibrium analysis, data intelligent inference, cloud centralized management to process dynamic adjustment. The data interaction and instruction transmission between the modules form a complete feedback loop, making the collection, analysis, inference and adjustment of parameters in the production process form an organic whole, greatly improving the continuity and stability of industrial oleic acid production, and providing strong support for large-scale production of high-quality oleic acid.

[0064] In the description of the application, it should be noted that unless otherwise explicitly specified and limited, the terms "arranged", "mounted", "connected", "linked", "fixed" should be understood broadly, for example, can be fixedly connected, can be detachably connected, or integrally connected; can be mechanically connected, can be electrically connected; can be directly connected, can be indirectly connected through an intermediate medium, or can be internal communication of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0065] Although embodiments of the present application have been shown and described, it would be appreciated by those of ordinary skill in the art that various equivalents, modifications, replacements and variations of these embodiments can be made without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalent scope.

Claims

1. A supervisory system for industrial oleic acid production, characterized in that: include: Raw material reaction control module, real-time parameter acquisition module, phase equilibrium analysis module, data intelligent reasoning module, cloud-based centralized management module, and process dynamic adjustment module; The raw material reaction control module implements the startup and operation parameter settings of the reaction process based on the catalytic reaction conditions, raw material ratio parameters, and reactor pressure value in the industrial oleic acid production process; The real-time parameter acquisition module continuously captures and converts the temperature change curve, oleic acid concentration gradient distribution, reactant flow rate changes, and reaction time series data in the reactor. The phase equilibrium analysis module receives the oleic acid-water-catalyst three-phase system parameters transmitted by the real-time parameter acquisition module and calculates the mole fraction distribution of each phase component and the interfacial tension value through thermodynamic equation calculation. The data intelligent reasoning module uses the parameter deviation values ​​output by the phase equilibrium analysis module and the probability of abnormal parameters in historical production data to complete parameter omissions and deduce abnormal trends. The cloud-based centralized management module categorizes and stores the processing results transmitted by the data intelligent reasoning module, and establishes data synchronization links with the Uniformance PHD production platform and the K / 3Cloud generation management cloud for cross-platform data sharing. The process dynamic adjustment module receives adjustment instructions from the cloud-based centralized management module and makes step-by-step corrections to the reactor stirring rate, heating power, raw material feed amount, and catalyst addition frequency.

2. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: When the real-time parameter acquisition module and the phase equilibrium analysis module work together, the correlation between the fatty acid double bond retention rate and the reaction temperature in the industrial oleic acid production process is calculated using the following model formula: , in, is the fatty acid double bond retention rate, is the initial double bond concentration, is the temperature influence coefficient, is the real-time reaction temperature, is the base reaction temperature, is the reaction duration, is the initial total fatty acid concentration, is the pressure influencing factor, is the pressure inside the reactor; When the cloud-based centralized management module interacts with the UniformancePHD production platform for data, the following model formula is used to correct the predicted acid value of the oleic acid product: , in, is the corrected acid value of oleic acid product, The acid value of the oleic acid product is measured. is the feed flow rate correction coefficient, is the real-time raw material feed flow rate, is the standard feed flow rate, is the enthalpy change correction factor, is the real-time reaction enthalpy change, is the standard reaction enthalpy change.

3. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: During the data interaction between the phase equilibrium analysis module and the data intelligent reasoning module, the distribution coefficient of oleic acid in the oil-water phase is calculated using the following model formula: , in, is the partition coefficient of oleic acid between oil phase and water phase, is the mole fraction of oleic acid in the oil phase, is the activity coefficient of oleic acid in the oil phase, is the mole fraction of oleic acid in the aqueous phase, is the activity coefficient of oleic acid in water phase, is the molar volume change of oleic acid transferred from the aqueous phase to the oil phase, is the system pressure, is standard atmospheric pressure, is the gas constant, is the absolute temperature.

4. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: When the data intelligent reasoning module uses the omission-Bayesian reasoning model to process the abnormal data output by the phase equilibrium analysis module, the posterior probability of the abnormality of the unit time parameter is calculated by the following model formula: , in, For observation data Next event The posterior probability of occurrence, For events Observation data when it occurs The likelihood of occurrence, For events The prior probability of occurrence, For the The weight factor of the abnormal parameter, For the The deviation of the parameters, is the total number of abnormal parameters, For observation data The marginal probability of occurrence, For events Observation data when it occurs The likelihood of occurrence, For events The prior probability of occurrence, is the total number of possible events.

5. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: When the cloud-based centralized management module analyzes the production plan data transmitted by the K / 3Cloud generation management cloud, it calculates the theoretical yield of oleic acid for each production batch using the following model formula: , in, is the theoretical yield of oleic acid, is the density of the reaction mixture, is the effective volume of the reactor, is the mass fraction of fatty acids in the raw materials, is the reaction conversion rate, is the mass fraction of impurities in the product.

6. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: When the process dynamic adjustment module adjusts the reaction conditions according to the instructions issued by the cloud centralized management module, the correction value of the stirring rate is calculated by the following model formula: , in, is the corrected stirring rate, is the base stirring rate, is the correction factor for oleic acid concentration, is the real-time oleic acid concentration, is the target oleic acid concentration, is the viscosity correction factor, is the dynamic viscosity of the reaction mixture, is the reference dynamic viscosity.

7. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: The phase equilibrium analysis module includes a component mole fraction calculation unit, a phase interface parameter analysis unit, a phase equilibrium stability assessment unit, and an abnormal data marking unit; the component mole fraction calculation unit receives the temperature, pressure, and component concentration data of the oleic acid-water-catalyst three-phase system transmitted by the real-time parameter acquisition module, substitutes these data into the phase equilibrium thermodynamic equation, obtains the mole fraction and chemical potential of each phase through iterative calculation, and then stores the calculation results in the unit's built-in buffer area and marks them with a timestamp; The phase interface parameter analysis unit extracts data from the buffer area of ​​the component mole fraction calculation unit, calculates the tension value of the three-phase interface, and determines the movement rate and thickness change of the phase interface by analyzing the diffusion coefficient of each phase component at different temperatures. The calibration data is converted into a standard format and transmitted to the unit output terminal; The phase equilibrium stability evaluation unit receives the phase interface parameters transmitted by the phase interface parameter analysis unit, combines the stirring intensity data in the reactor, calculates the phase mixing uniformity index, and generates phase equilibrium stability evaluation data by comparing the phase equilibrium state parameters in different reaction stages. The data is stored in the unit's database; the abnormal data marking unit retrieves the phase equilibrium stability evaluation data from the database of the phase equilibrium stability evaluation unit, compares it with the preset phase equilibrium threshold, marks the data that exceeds the threshold range, and packages the marked data together with the normal data and transmits it to the data intelligent reasoning module.

8. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: The data intelligent reasoning module includes a parameter omission completion unit, a parameter trend analysis unit, and an abnormal pattern recognition unit; the parameter omission completion unit receives the parameter data transmitted by the phase equilibrium analysis module, identifies the position of the missing value in the data sequence, estimates and fills the missing parameters based on the omission-Bayesian inference model, and combines the parameter distribution characteristics under similar reaction conditions in the same period of history, and temporarily stores the filled complete data sequence in the temporary storage area of ​​the unit; the parameter trend analysis unit extracts the complete data sequence from the temporary storage area of ​​the parameter omission completion unit, analyzes the change trend of the parameters in the sequence, determines the inflection point and slope change of the trend by calculating the first-order derivative and second-order derivative of the parameters, and transmits the trend analysis results to the result buffer area of ​​the unit; abnormal The pattern recognition unit retrieves the trend analysis results from the result buffer of the parameter trend analysis unit, and combines the probability distribution of abnormal parameter combinations in historical production data to identify and classify abnormal patterns that may appear in the current parameter sequence. The classified abnormal recognition results are integrated with the normal trend data and transmitted to the cloud-based centralized management module.

9. A supervisory system for industrial oleic acid production according to claim 1, characterized in that: The cloud-based centralized management module includes a data format standardization unit, a cross-platform data interaction unit, a production data association index unit, and a panoramic view generation unit; the data format standardization unit receives the integrated data transmitted by the data intelligent reasoning module, verifies and converts the data format to make it conform to the data interaction standard between the UniformancePHD production platform and the K / 3Cloud generation management cloud, and stores the converted standardized data in the unit's distributed database; the cross-platform data interaction unit retrieves standardized data from the distributed database of the data format standardization unit, transmits the data to the UniformancePHD production platform and the K / 3Cloud generation management cloud respectively through the established data synchronization link, and receives production status data fed back by the two platforms, and stores the feedback data in the unit's feedback data area; The production data association index unit extracts production status data from the feedback data area of ​​the cross-platform data interaction unit, classifies and indexes the data, establishes an association relationship between the data and the production batch, reactor number, and time series, and stores the association index results in the unit's index database; the panoramic view generation unit retrieves the association index results from the index database of the production data association index unit, combines the historical adjustment records of the process dynamic adjustment module, generates a panoramic data view of the production process, transmits the panoramic data view to the unit's view display cache, and at the same time generates a preliminary draft of the adjustment instruction and transmits it to the process dynamic adjustment module.

10. A method for supervising industrial oleic acid production, characterized in that: The method is applied to a monitoring system for industrial oleic acid production as described in claim 1, The following steps are involved: In the first step, the raw material reaction control module receives the execution parameters returned by the process dynamic adjustment module. Based on the preset initial reaction conditions, it sets the reactor feed valve opening, the operating frequency of the catalyst addition pump, and the starting power of the heating device, so that the raw materials enter the reactor at the set ratio and the reaction begins. In the second step, the real-time parameter acquisition module continuously collects temperature, pressure, oleic acid concentration, reactant flow rate, and reaction time parameters during the reaction process through sensors deployed at different locations in the reactor. The collected analog signals are converted into digital signals, filtered, and transmitted to the phase equilibrium analysis module. In the third step, the phase equilibrium analysis module receives the digital signal transmitted by the real-time parameter acquisition module, extracts the parameters of the oleic acid-water-catalyst three-phase system included in the signal, substitutes them into the phase equilibrium and thermodynamic supervision model for calculation, and obtains the mole fraction, chemical potential, and interfacial tension parameters of each phase component. The calculation results are transmitted to the data intelligent reasoning module; In the fourth step, the data intelligent reasoning module receives the calculation results transmitted by the phase equilibrium analysis module, uses the omission-Bayesian inference model to complete any missing parameters in the results, deduces abnormal trends in the completed parameter sequence, and transmits the deduced processing results to the cloud-based centralized management module; In the fifth step, the cloud-based centralized management module receives the processing results transmitted by the data intelligent reasoning module, classifies and stores the results, and synchronizes data with the Uniformance PHD production platform and K / 3Cloud production management cloud. It generates process adjustment instructions based on the synchronously acquired production plan and status data and transmits the instructions to the process dynamic adjustment module. In the sixth step, the process dynamic adjustment module receives the process adjustment instructions transmitted by the cloud-based centralized management module, parses the parameters in the instructions, and makes step-by-step corrections to the reactor's stirring rate, heating power, raw material feed amount, and catalyst addition frequency based on the analysis results. The corrected execution parameters are then transmitted back to the raw material reaction control module.

Images