Methanol recovery optimization management system, method, and fatty acid methyl ester production method

Abstract

This invention discloses a methanol recovery rate optimization management system, method, and fatty acid methyl ester production method, relating to the field of data management. By setting parameters to adjust the maximum number of iterations, this invention not only constrains the operation frequency and avoids disorderly fluctuations in operating conditions, but also upgrades the algorithm's capabilities by expanding the dataset and reconstructing the training model after conventional adjustments fail. This overcomes the limitations of a single model in adapting to operating conditions and ensures stable methanol recovery rate compliance. Simultaneously, relying on predicted data, actual detection data, and operating condition labels accumulated from multiple rounds of production operation, an anomaly identification model is built to achieve intelligent correlation analysis between recovery rate deviation data and specific fault types. This comprehensively strengthens the refined control level of the methanol recovery link in the fatty acid methyl ester production process, enhances the initiative and accuracy of process regulation, and promptly locates various production anomalies, ensuring continuous and stable production operation and effectively controlling material loss and production operating costs.

Description

Technical Field

[0001] This invention belongs to the field of data management, and specifically relates to a methanol recovery rate optimization management system, method, and fatty acid methyl ester production method. Background Technology

[0002] Fatty acid methyl esters, as clean energy raw materials, industrial additives and green chemical intermediates, are widely used in many fields such as biodiesel preparation, fine chemical processing and surfactant synthesis. Industrial production mainly relies on core processes such as esterification, neutralization and distillation separation to complete the preparation. Methanol, as a key reaction raw material and extraction medium, participates in the entire reaction conversion and recycling process. Chinese Patent No. CN115496424A discloses a safety management method and system for methanol-to-hydrogen process, including: acquiring multiple process flows and multiple process indicators of a preset methanol-to-hydrogen process; analyzing the correlation parameters between production safety and production quality to obtain multiple correlation parameters and multiple sensitive process indicators; constructing an anomaly detection model including a parameter change rate anomaly detection module and a parameter anomaly degree anomaly detection module; monitoring and collecting the indicator parameters of multiple sensitive process indicators within a preset time range to obtain multiple real-time indicator parameter sequences, calculating multiple real-time parameter change rates and multiple indicator parameter anomalies of the multiple sensitive process indicators; inputting the parameter change rate anomaly detection module and the parameter anomaly degree anomaly detection module to obtain anomaly monitoring result management; However, traditional fatty acid methyl ester methanol recovery production suffers from problems such as a lack of reasonable constraints on parameter adjustment, limited optimization methods for routine operating conditions, and a single prediction model with limited applicability to various scenarios. In addition, it is difficult to meet the target under special operating conditions and it is difficult to quickly trace the source based on the recovery rate deviation and automatically distinguish and identify various types of abnormal production conditions. Summary of the Invention

[0003] In response to the problems in related technologies, this invention proposes a methanol recovery rate optimization management system, a method, and a fatty acid methyl ester production method to overcome the aforementioned technical problems existing in the existing related technologies.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: This invention provides a method for optimizing methanol recovery rate management, comprising the following steps: S1. Collect corresponding methanol recovery process parameters and methanol recovery rate data from historical production. S2. Construct a primary methanol recovery rate mapping model based on the data collected in S1; S3. Input the current methanol recovery process parameter data into the primary methanol recovery rate mapping model for mapping; S4. Set the methanol recovery rate threshold. If the mapping result in S3 meets the threshold, no operation is required. Otherwise, repeatedly adjust the methanol recovery process parameter data and input it into the methanol recovery rate mapping model again to obtain the corresponding recovery rate data. S5. Set a first maximum number of repetitions. If the recovery rate data obtained in S4 meets the threshold and the number of repetitions does not exceed the upper limit, use it as the current final recovery rate and execute S6. Otherwise, expand the data collected in S1 and execute S2 again to retrain the model. Repeat the adjustment and mapping of methanol recovery process parameters until the mapping result meets the threshold and then execute S6. S6. Repeat S4 and S5 multiple times to obtain recovery rate mapping data and actual methanol recovery rate data, and combine them with the corresponding abnormal operating condition detection results to construct the final abnormal operating condition type coding mapping model. Then, obtain the current methanol recovery rate mapping to be tested and the actual data and input them into the final abnormal operating condition type coding mapping model for mapping. Determine the abnormal operating condition based on the mapping result and process it.

[0005] Preferably, step S1 includes the following steps: S11. Set several process parameter types for process nodes such as methanol distillation with water, esterification reaction, and neutralization reaction to obtain a methanol recovery process parameter type set; the methanol recovery process parameter type set includes esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, methanol loss in waste gas, and distillation tower pressure, etc. S12. Based on the methanol recovery process parameter type set, collect various methanol recovery process parameter data and corresponding methanol recovery rate data from several historical fatty acid methyl ester production processes to obtain historical methanol recovery process parameter dataset and historical methanol recovery rate dataset. By collecting various process parameter data and corresponding methanol recovery rate data from historical production processes, the resulting historical dataset is comprehensive and correlated, providing solid data support for the subsequent construction of a dynamic optimization model for methanol recovery efficiency.

[0006] Preferably, step S2 includes the following steps: S21. Based on the historical methanol recovery process parameter dataset and the historical methanol recovery rate dataset, construct a mapping model with various methanol recovery process parameter data as input and methanol recovery rate data as output, and obtain a primary methanol recovery rate mapping model. This mapping model, through deep learning of different combinations of operating conditions, parameter fluctuations, and changes in recovery rate in historical data, can comprehensively integrate effective information in the production process, avoid the subjectivity and bias of human experience judgment, provide intuitive reference for real-time production, and provide scientific basis for subsequent optimization decisions.

[0007] Preferably, step S3 includes the following steps: S31. Based on the methanol recovery process parameter type set, obtain the data of various methanol recovery process parameters corresponding to the current fatty acid methyl ester production process, and obtain the current methanol recovery process parameter dataset. S32. Input the current methanol recovery process parameter dataset into the primary methanol recovery rate mapping model for mapping to obtain the current initial methanol recovery rate data; Real-time mapping not only allows operators to instantly grasp the methanol recovery effect of the current production process and promptly detect abnormal fluctuations in efficiency, but also provides immediate decision-making basis for subsequent optimization and adjustment of process parameters, helping the production process to shift from passive monitoring to active control and effectively ensuring the stability of methanol recovery efficiency.

[0008] Preferably, step S4 includes the following steps: S41. Set the current methanol recovery rate threshold according to the current production requirements of fatty acid methyl esters to obtain the current methanol recovery rate threshold; if the current initial methanol recovery rate data is greater than or equal to the current methanol recovery rate threshold, no action is required; otherwise, proceed to S42. S42. Repeatedly adjust the esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure parameters in the current methanol recovery process parameter dataset. After each repeated adjustment, an adjusted methanol recovery process parameter dataset is obtained. The adjusted methanol recovery process parameter dataset is then input into the methanol recovery rate mapping model for mapping to obtain the adjusted methanol recovery rate data. If the initial methanol recovery rate meets the target, no additional intervention is required, thus reducing the impact of ineffective operations on production continuity. If the target is not met, a parameter adjustment process is initiated to ensure that resources are focused on efficiency optimization.

[0009] Preferably, step S5 includes the following steps: S51. Set a first maximum number of repetitions. If the methanol recovery rate data after one adjustment as described in S42 is greater than or equal to the current methanol recovery rate threshold and the number of repetitions is less than or equal to the first maximum number of repetitions, use the methanol recovery rate data after one adjustment as the current final methanol recovery rate data and execute S6; otherwise, execute S52. S52. Repeatedly expand the historical methanol recovery process parameter dataset and historical methanol recovery rate dataset mentioned in S21. After expansion, S21 is executed again to obtain the secondary methanol recovery rate mapping model. The esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure parameters in the current methanol recovery process parameter dataset are then repeatedly adjusted. After each adjustment, a second-adjusted methanol recovery process parameter dataset is obtained. The second-adjusted methanol recovery process parameter dataset is then input into the second-stage methanol recovery rate mapping model for mapping to obtain the second-adjusted methanol recovery rate data. This process continues until the second-adjusted methanol recovery rate data is greater than or equal to the current methanol recovery rate threshold. Finally, the second-adjusted methanol recovery rate data is used as the current final methanol recovery rate data, and step S6 is executed. By repeatedly adjusting the core process parameters based on the optimized secondary model, and leveraging the upgraded model's feature learning capabilities, the deep correlation between parameters and recovery rate can be captured more accurately, guiding the adjustment direction toward the optimal range and avoiding blind iteration. This not only solves the problems of insufficient adaptability of the initial model and ineffective parameter adjustment, but also achieves dynamic linkage and upgrading between the model and production data.

[0010] Preferably, step S6 includes the following steps: S61. Several abnormal operating conditions are defined in the production process of fatty acid methyl ester to obtain a set of abnormal operating conditions; the set of abnormal operating conditions includes no fault, equipment fault, instrument abnormality and material fluctuation, etc.; numerical codes are performed on each abnormal operating condition type in the set of abnormal operating conditions to obtain a set of abnormal operating condition codes. S62. Repeat S4 and S5 multiple times to obtain the current final methanol recovery rate data corresponding to the repetition process, and collect the actual methanol recovery rate data after each repetition to obtain the methanol recovery rate mapping dataset and the actual methanol recovery rate dataset. S63. Set the current number of fault detection executions and execute S62 multiple times based on this. After each execution, obtain the corresponding methanol recovery rate mapping dataset and methanol recovery rate actual dataset. Detect the abnormal operating conditions in the production process of the corresponding fatty acid methyl ester according to the abnormal operating condition type set, and obtain the historical methanol recovery rate mapping dataset, the historical methanol recovery rate actual dataset, and the historical abnormal operating condition type encoding dataset. S64. Based on the historical methanol recovery rate mapping data set, the historical methanol recovery rate actual data set, and the historical abnormal operating condition type coding dataset, construct a mapping model with multiple sets of methanol recovery rate mapping data and methanol recovery rate actual data as inputs and abnormal operating condition type coding data as outputs, and obtain the final abnormal operating condition type coding mapping model. S65. Based on the current number of fault detection executions, perform the most recent fatty acid methyl ester production process. Repeat S4 and S5 during each process and record the corresponding current final methanol recovery rate data and actual methanol recovery rate data to obtain the methanol recovery rate mapping data set and the methanol recovery rate actual data set. The set of methanol recovery rate to be tested and the set of actual methanol recovery rate to be tested are input into the final abnormal operating condition type coding mapping model for mapping to obtain the current abnormal operating condition type coding data; and the abnormal operating conditions in the recent fatty acid methyl ester production process are processed according to the current abnormal operating condition type coding data. The constructed abnormal operating condition type coding mapping model achieves accurate conversion from recovery rate data differences to operating condition types, breaking the limitations of traditional reliance on manual experience to judge anomalies. It can quickly identify different types of problems such as equipment failure and instrument distortion. In actual production, by collecting real-time recovery rate mapping data and actual data and inputting them into the model, the corresponding abnormal operating condition codes can be output instantly, providing operators with clear fault indications, significantly shortening the time for anomaly investigation and handling, and effectively avoiding increased methanol loss, product quality fluctuations, or production interruptions caused by abnormal operating conditions. This significantly improves the controllability of operating conditions, fault response speed, and overall operational stability of the fatty acid methyl ester production process.

[0011] The methanol recovery rate optimization management system includes a recovery rate data acquisition module, a recovery rate mapping model construction module, a recovery rate mapping module, a primary adjustment module, a dataset expansion module, and an abnormal operating condition handling module. The recovery rate data acquisition module is used to collect corresponding process parameters and methanol recovery rate data from historical production. The recovery rate mapping model building module is used to build a primary methanol recovery rate mapping model; The recovery rate mapping module is used to map the current initial methanol recovery rate data; The single-adjustment module is used to repeatedly adjust the methanol recovery process parameter data and remap it after adjustment; The dataset expansion module is used to expand the data collected in the recovery rate data acquisition module and retrain the model. The abnormal operating condition handling module is used to construct the final abnormal operating condition type encoding mapping model, then obtain the current methanol recovery rate mapping to be tested and the actual data and map them, and determine and handle abnormal operating conditions based on the mapping results.

[0012] The present invention has the following beneficial effects: 1. This invention utilizes a recovery rate threshold determination mechanism to rationally divide the control process, initiating parameter adjustment only when indicators fail to meet standards, thus reducing ineffective operation costs. Furthermore, by setting the maximum number of iterations for parameter adjustment, it not only constrains the operation frequency and avoids disorderly fluctuations in operating conditions, but also upgrades the algorithm's capabilities by expanding the dataset and reconstructing the training model after conventional adjustments fail, overcoming the limitations of a single model's adaptability to operating conditions and ensuring stable methanol recovery rate compliance. Simultaneously, by combining diverse production anomaly classifications and digital coding methods, and relying on prediction data, actual detection data, and operating condition labels accumulated from multiple rounds of production operation, an anomaly identification model is built. This enables intelligent correlation analysis between recovery rate deviation data and specific fault types, thereby comprehensively strengthening the refined control level of the methanol recovery link in the fatty acid methyl ester production process, improving the initiative and accuracy of process control, and promptly locating various production anomalies, ensuring continuous and stable production operation, and effectively controlling material loss and production operating costs.

[0013] 2. In this invention, by repeatedly expanding the historical dataset, new special working conditions and adjusted experience data are incorporated into the model training, making up for the initial model's lack of working condition coverage. This makes the secondary methanol recovery rate mapping model more in line with complex production realities, improving its adaptability to diverse working conditions and prediction accuracy. Thus, it not only solves the problems of insufficient adaptability of the initial model and ineffective parameter adjustment, but also realizes the dynamic linkage and upgrading of the model and production data.

[0014] 3. In this invention, by collecting real-time recovery rate mapping data and actual data and inputting them into the model, the corresponding abnormal operating condition code can be output in real time, providing operators with clear fault indications, greatly shortening the time for abnormal investigation and handling, and effectively avoiding increased methanol loss, product quality fluctuations or production interruptions caused by abnormal operating conditions.

[0015] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram illustrating the process of repeatedly adjusting the current methanol recovery process parameter dataset according to the present invention. Figure 2 A schematic diagram of the process for constructing a primary methanol recovery rate mapping model for this invention; Figure 3This is a schematic diagram of the process for determining the threshold and number of repetitions of methanol recovery rate data after a single adjustment according to the present invention. Figure 4 This is a schematic diagram of the process for determining the threshold of methanol recovery rate data after secondary adjustment according to the present invention. Figure 5 This is a flowchart illustrating the process of constructing the final abnormal operating condition type encoding mapping model for this invention. Detailed Implementation

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0020] Example 1 Please see Figure 1 This embodiment describes a method for optimizing methanol recovery rate management, which includes the following steps: S1. Collect corresponding methanol recovery process parameters and methanol recovery rate data from historical production. Please see Figure 2 S1 includes the following steps: S11. Set several process parameter types for process nodes such as methanol distillation with water, esterification reaction, and neutralization reaction to obtain a methanol recovery process parameter type set; the methanol recovery process parameter type set includes esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, methanol loss in waste gas, and distillation tower pressure, etc. S12. Based on the methanol recovery process parameter type set, collect various methanol recovery process parameter data and corresponding methanol recovery rate data from several historical fatty acid methyl ester production processes to obtain historical methanol recovery process parameter dataset and historical methanol recovery rate dataset. The esterification reaction temperature is acquired by using dual redundant armored platinum resistance temperature sensors installed in the middle of the esterification reactor and in the jacket heating / cooling medium channel. The sensors directly contact the material and the heat exchange medium. The resistance signal is converted into a 4-20mA standard current signal by a temperature transmitter and transmitted to the distributed control system in real time. At the same time, an additional temperature sensor is added in the gas phase space of the reactor for auxiliary calibration. The data acquisition frequency is set to 1 time / second. The DCS performs data filtering and average value calculation to ensure stable and reliable temperature data. The measurement range covers 60-100℃. The neutralization reaction temperature is collected by installing industrial-grade thermocouple temperature sensors in the feed mixing zone, reaction section, and discharge port of the neutralization reactor. The sensor insertion depth is 1 / 3 of the reactor diameter. The thermoelectric potential signal is converted into a digital signal and connected to the production process data acquisition system. The acquisition frequency is 1 time / second. After removing instantaneous fluctuation abnormal values, a 30-second sliding average value is taken. The measurement range is set to 50-80℃. At the same time, the temperature feedback of the jacket heat exchange system is linked to ensure that the data is consistent with the actual reaction state. The moisture content of methanol with water was collected by installing an online microwave moisture analyzer on the feed pipeline from the esterification reactor to the distillation tower. The analyzer calculates the moisture content by measuring the energy attenuation when microwaves penetrate the material, with a measurement range of 5%-20%. At the same time, several offline samples were taken from each batch and tested and calibrated in the laboratory using a Karl Fischer moisture analyzer. The deviation between online and offline data was controlled within ±0.3% to ensure measurement accuracy. The method for collecting methanol loss in waste gas is to install an online gas chromatograph and a mass flow meter in the tail gas emission pipeline of the methanol distillation tower: the mass flow meter measures the total flow rate of waste gas in real time, and the gas chromatograph detects the volume concentration of methanol in waste gas through a hydrogen flame ionization detector. The data from both are correlated and calculated in real time through DCS: Methanol loss in waste gas = Total flow rate of waste gas × Methanol volume concentration × Methanol density; at the same time, a backup detection point is added at the inlet of the incineration system, and the data is compared regularly to ensure the accuracy of the loss calculation. The pressure of the distillation column is acquired by installing an absolute pressure capacitive pressure transmitter on the vapor outlet pipe at the top of the methanol distillation column. The transmitter range is set to 80-120 kPa. The measured vapor pressure is converted into a 4-20 mA standard signal and transmitted to the DCS. After pressure compensation by the DCS, the average value over one minute is output. At the same time, a pressure sensor is added at the bottom of the column to assist in monitoring the pressure gradient inside the column, ensuring that the data reflects the overall pressure status of the distillation column and avoiding measurement deviations caused by local pressure fluctuations. For example, some data from the historical methanol recovery process parameter dataset and the historical methanol recovery rate dataset are shown in Table 1 below: Table 1. Schematic diagram of partial correlation data for methanol recovery rate 78.5 65.3 12.3 0.085 101.3 85.0 79.2 64.8 11.8 0.078 100.8 86.0 77.8 66.1 13.1 0.092 101.5 84.0 79.5 65.5 11.5 0.072 100.5 87.0 78.2 65.9 12.7 0.081 101.1 84.6 By clearly defining the key parameter types of core process nodes such as methanol distillation with water, esterification reaction, and neutralization reaction, and constructing a complete parameter set, we can systematically sort out the multiple factors affecting methanol recovery efficiency. This avoids blind data collection caused by parameter omissions or ambiguity, providing direction for subsequent data collection work and ensuring that the collection direction is accurately aligned with the core needs of methanol recovery efficiency optimization. Based on this parameter set, we collect various process parameter data and corresponding methanol recovery rate data from historical production processes. The resulting historical dataset is comprehensive and correlated, fully restoring the dynamic changes of process parameters under different production conditions and the intrinsic mapping relationship between them and methanol recovery rates. It covers the influence of operating conditions such as reaction temperature and distillation pressure, as well as the role of material characteristics and loss indicators such as the water content of methanol with water and the amount of methanol loss in waste gas. This provides solid data support for the subsequent construction of a dynamic optimization model for methanol recovery efficiency. It can also help to discover undiscovered parameter optimization space, avoiding the limitations of relying on experience-based operations. This lays the foundation for determining the optimal combination of process parameters through model analysis, improving methanol recovery efficiency, and reducing process losses, thereby ensuring the stability and economy of fatty acid methyl ester production. S2. Construct a primary methanol recovery rate mapping model based on the data collected in S1; S2 includes the following steps: S21. Based on the historical methanol recovery process parameter dataset and the historical methanol recovery rate dataset, construct a mapping model with various methanol recovery process parameter data as input and methanol recovery rate data as output, and obtain a primary methanol recovery rate mapping model. S21 includes the following steps: S211. Construct an initial methanol recovery rate mapping model and set a first training data ratio (e.g., 8:2 or 7:3, which can be adjusted adaptively according to the actual training situation); divide the historical methanol recovery process parameter dataset and the historical methanol recovery rate dataset according to the first training data ratio to obtain the first training dataset and the first test dataset. S212. Set a first training error threshold (10%~15%, which can be adjusted adaptively according to the actual training situation); input the first training dataset into the initial methanol recovery rate mapping model for training; during the training process, if the training error is less than the first training error threshold, stop training and obtain the trained methanol recovery rate mapping model; otherwise, continue training until the training error is less than the first training error threshold. S213. Set a first test accuracy threshold (90%~95%, which can be adjusted adaptively according to the actual test situation); input the first test dataset into the trained methanol recovery rate mapping model for testing; after the test is completed, obtain the first test accuracy data; if the first test accuracy data is greater than or equal to the first test accuracy threshold, use the trained methanol recovery rate mapping model as a methanol recovery rate mapping model; otherwise, return to S212 to continue training the trained methanol recovery rate mapping model and repeat S213 until the first test accuracy data is greater than or equal to the first test accuracy threshold. The initial methanol recovery rate mapping model can employ a fully connected deep neural network model to adapt to the nonlinear mapping relationship between process parameters and recovery rate, offering high fitting accuracy and simple deployment. This model includes an input layer, hidden layers, and an output layer. The number of nodes in the input layer equals the total number of methanol recovery process parameters, directly receiving data on all process parameters such as esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, methanol loss in waste gas, and distillation tower pressure. The hidden layer comprises four fully connected layers. The first fully connected layer has 64 neurons and uses the ReLU activation function to initially extract basic features of the process parameters. The second fully connected layer has 128 neurons and uses the ReLU activation function. The model employs a multi-process node correlation feature fusion structure to enhance parameter coupling extraction. The third fully connected layer has 64 neurons and uses the ReLU activation function to reduce the dimensionality of high-order features and filter out invalid data interference. The fourth fully connected layer has 32 neurons and uses the ReLU activation function to further optimize feature representation and improve model fitting stability. The output layer has one neuron and uses the Sigmoid activation function to directly output the predicted methanol recovery rate. Additionally, the model uses an adaptive moment estimation optimizer (Adam optimizer) with a learning rate of 0.001 and a mean squared error (MSE) loss function to adapt to regression prediction tasks. A mapping model is constructed based on the collected historical methanol recovery process parameter dataset and the corresponding methanol recovery rate dataset. This model can deeply correlate scattered process parameter data with core efficiency indicators, breaking the limitations of single-parameter analysis. It accurately captures the complex nonlinear relationship between methanol recovery rate and various factors such as esterification reaction temperature, neutralization reaction conditions, distillation tower operating parameters, material characteristics, and loss indicators. This transforms the previously difficult-to-quantify parameter influence patterns into a calculable and predictable process. Specifically, through deep learning of different operating condition combinations, parameter fluctuations, and recovery rate changes in historical data, this mapping model can comprehensively integrate effective information from the production process, avoiding the subjectivity and bias of human experience judgment, and forming an objective analysis tool with data-driven characteristics. As a result, the final methanol recovery rate mapping model can not only quickly output accurate methanol recovery rate prediction results based on input process parameters, providing an intuitive reference for real-time production, but also provide a scientific basis for subsequent optimization decisions, help identify the sensitive range of key influencing parameters, provide directional guidance for dynamic adjustment of process parameters, and lay the foundation for early prediction and anomaly warning of methanol recovery efficiency. S3. Input the current methanol recovery process parameter data into the primary methanol recovery rate mapping model for mapping; S3 includes the following steps: S31. Based on the methanol recovery process parameter type set, obtain the data of various methanol recovery process parameters corresponding to the current fatty acid methyl ester production process, and obtain the current methanol recovery process parameter dataset. S32. Input the current methanol recovery process parameter dataset into the primary methanol recovery rate mapping model for mapping to obtain the current initial methanol recovery rate data; By acquiring various process parameter data from the current production process, it is possible to ensure that the collected current dataset and the historical modeling dataset maintain a high degree of consistency in parameter dimensions and types, avoiding model mapping failure due to parameter mismatch. Simultaneously, it accurately focuses on core influencing factors such as esterification reaction temperature and neutralization reaction temperature, ensuring that the current data comprehensively reflects the true state of real-time production conditions, providing a reliable data foundation for subsequent mapping calculations. This current dataset is then input into the constructed primary methanol recovery rate mapping model for mapping analysis. Leveraging the complex correlation logic between process parameters and recovery rate learned by the model, real-time production data can be quickly transformed into quantified initial methanol recovery rate data. This breaks through the efficiency lag problem caused by relying on manual calculation or offline detection in traditional production, enabling real-time prediction of methanol recovery efficiency. This real-time mapping method not only allows operators to instantly grasp the methanol recovery effect of the current production stage and promptly detect abnormal efficiency fluctuations, but also provides immediate decision-making basis for subsequent process parameter optimization and adjustment, helping the production process shift from passive monitoring to proactive control, effectively ensuring the stability of methanol recovery efficiency. S4. Set the methanol recovery rate threshold. If the mapping result in S3 meets the threshold, no operation is required. Otherwise, repeatedly adjust the methanol recovery process parameter data and input it into the methanol recovery rate mapping model again to obtain the corresponding recovery rate data. Please see Figure 3 S4 includes the following steps: S41. Set the current methanol recovery rate threshold according to the current production requirements of fatty acid methyl esters to obtain the current methanol recovery rate threshold; if the current initial methanol recovery rate data is greater than or equal to the current methanol recovery rate threshold, no action is required; otherwise, proceed to S42. S42. Repeatedly adjust the esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure parameters in the current methanol recovery process parameter dataset. After each repeated adjustment, an adjusted methanol recovery process parameter dataset is obtained. The adjusted methanol recovery process parameter dataset is then input into the methanol recovery rate mapping model for mapping to obtain the adjusted methanol recovery rate data. By setting targeted current methanol recovery rate thresholds based on actual production needs, the criteria for judging methanol recovery efficiency are highly aligned with the specific goals of fatty acid methyl ester production, avoiding the inadequacy of adaptability caused by a uniform threshold. Secondly, no additional intervention is required when the initial methanol recovery rate meets the target, thereby reducing the impact of ineffective operations on production continuity. However, when the target is not met, a parameter adjustment process is initiated to ensure that resources are concentrated on efficiency optimization. Specifically, core controllable parameters such as esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure are repeatedly adjusted to accurately identify key variables affecting methanol recovery efficiency, avoiding resource waste and process fluctuations caused by blindly adjusting non-core parameters. After each adjustment, the methanol recovery rate data after adjustment is obtained through model remapping, which can provide real-time feedback on the actual effect of parameter adjustment. This fully utilizes the model's precise control over the relationship between parameters and recovery rate, and gradually approaches the optimal parameter combination through repeated iterations, effectively improving the probability of achieving the target methanol recovery rate. At the same time, it avoids the drastic fluctuations in production conditions that may be caused by a one-time large-scale adjustment of parameters, ensuring the stability of the fatty acid methyl ester production process. S5. Set a first maximum number of repetitions. If the recovery rate data obtained in S4 meets the threshold and the number of repetitions does not exceed the upper limit, use it as the current final recovery rate and execute S6. Otherwise, expand the data collected in S1 and execute S2 again to retrain the model. Repeat the adjustment and mapping of methanol recovery process parameters until the mapping result meets the threshold and then execute S6. Please see Figure 4 S5 includes the following steps: S51. Set a first maximum number of repetitions. If the methanol recovery rate data after one adjustment as described in S42 is greater than or equal to the current methanol recovery rate threshold and the number of repetitions is less than or equal to the first maximum number of repetitions, use the methanol recovery rate data after one adjustment as the current final methanol recovery rate data and execute S6; otherwise, execute S52. S52. Repeatedly expand the historical methanol recovery process parameter dataset and historical methanol recovery rate dataset mentioned in S21. After expansion, S21 is executed again to obtain the secondary methanol recovery rate mapping model. The esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure parameters in the current methanol recovery process parameter dataset are then repeatedly adjusted. After each adjustment, a second-adjusted methanol recovery process parameter dataset is obtained. The second-adjusted methanol recovery process parameter dataset is then input into the second-stage methanol recovery rate mapping model for mapping to obtain the second-adjusted methanol recovery rate data. This process continues until the second-adjusted methanol recovery rate data is greater than or equal to the current methanol recovery rate threshold. Finally, the second-adjusted methanol recovery rate data is used as the current final methanol recovery rate data, and step S6 is executed. The methanol recovery rate must be adjusted to meet the threshold before proceeding with subsequent anomaly prediction and judgment. This is because it is necessary to ensure the effectiveness of the model's operating range and the reliability of the baseline for condition judgment. Only when the methanol recovery rate reaches the preset threshold does it mean that the current key processes such as esterification, neutralization, and distillation are within the reasonable and stable range of the process design, and the coupling relationship of various process parameters and the material change law conform to the normal operating condition characteristics learned by the model. Only then can the anomaly judgment based on the model eliminate the basic interference caused by substandard efficiency. If anomaly prediction is performed before the recovery rate meets the standard, controllable optimization problems such as unreasonable parameters and low efficiency will be misjudged as equipment failures or instrument distortions. In cases of real anomalies such as material deterioration, distorted judgment results and frequent false alarms can occur, misleading on-site handling. Calibrating the recovery rate to above the threshold establishes a unified and qualified operating condition benchmark, eliminating fluctuations within the normal optimization range. This allows subsequent monitoring to focus on uncontrollable anomalies such as sudden drift, equipment hazards, and system leaks, improving the accuracy and relevance of anomaly judgment. At the same time, a qualified recovery rate level ensures that production operations are within a safe and economical range, avoiding the risk of misjudgment caused by anomaly analysis under inefficient operating conditions. This ensures the stability and reliability of the entire prediction and judgment system, providing a solid foundation for timely detection of real production anomalies and ensuring the continuous and stable operation of the equipment. By setting a first maximum number of repetitions, a reasonable boundary is defined for parameter adjustment. This avoids production efficiency losses and operating condition fluctuations caused by unlimited repeated adjustments, while ensuring that the parameter optimization space is fully explored within the effective iteration range. If the recovery rate meets the target after adjustment and does not exceed the maximum number of repetitions, the current optimal result is used to advance the subsequent process, ensuring production continuity. When parameter adjustment fails to meet the target, the historical dataset is repeatedly expanded, and new special operating conditions and adjustment experience data are incorporated into the model training to compensate for the initial model's insufficient coverage of operating conditions. This makes the secondary methanol recovery rate mapping model more closely aligned with complex production realities, improving its adaptability to diverse operating conditions and prediction accuracy. Based on the optimized secondary model, the core process parameters are repeatedly adjusted again. By leveraging the upgraded model's feature learning capabilities, the deep correlation between parameters and recovery rate is captured more accurately, guiding the adjustment direction toward the optimal range and avoiding blind iteration. This solves the problems of insufficient adaptability of the initial model and ineffective parameter adjustment, and achieves dynamic linkage and upgrading between the model and production data. While ensuring that the methanol recovery rate ultimately meets the target, the model's adaptability to production conditions is continuously strengthened, balancing the achievement of production goals with the sustainable optimization of technical solutions. S6. Repeat S4 and S5 multiple times to obtain recovery rate mapping data and actual methanol recovery rate data, and combine them with the corresponding abnormal operating condition detection results to construct the final abnormal operating condition type coding mapping model. Then, obtain the current methanol recovery rate mapping to be tested and the actual data and input them into the final abnormal operating condition type coding mapping model for mapping. Determine the abnormal operating condition and process it based on the mapping result. Please see Figure 5 S6 includes the following steps: S61. Define several abnormal operating conditions in the production process of fatty acid methyl esters to obtain a set of abnormal operating condition types; the set of abnormal operating condition types includes no faults, equipment faults (such as the detachment or blockage of internal trays in the methanol distillation tower or the scaling and perforation of the reboiler heating tubes, leading to a decrease in vapor-liquid contact efficiency; damage to the bearings of the agitator in the esterification reactor, causing agitation failure, resulting in uneven material mixing; internal leakage in the delivery pump or incomplete closure of the outlet valve, causing methanol or water-containing methanol backflow and leakage; fan failure in the tail gas incineration system, resulting in abnormally increased methanol volatilization, etc.), and instrument abnormalities (zero drift of the temperature sensor in the esterification reactor, or measured values ​​that are too high or too low; pressure in the distillation tower). The abnormal operating conditions include: pressure transmitter blockage causing pressure display distortion; methanol moisture content analyzer probe contamination causing continuous deviation of detection values ​​from the true value; liquid accumulation in the exhaust gas flow meter pipeline causing inaccurate flow measurement; poor contact in various sensor circuits causing data jumps, etc.; and material fluctuations (such as imported raw material moisture content and acid value exceeding the normal range; catalyst dampness, failure, or inconsistent activity between batches, affecting reaction conversion rate; low methanol purity and excessive water content, resulting in a significant decrease in distillation load and recovery efficiency; impurities mixed in the material causing changes in system viscosity and boiling point, etc.). Numerical coding is performed on various abnormal operating condition types in the above-mentioned production abnormal operating condition type set to obtain a production abnormal operating condition type coding set. S62. Repeat S4 and S5 multiple times to obtain the current final methanol recovery rate data corresponding to the repetition process, and collect the actual methanol recovery rate data after each repetition to obtain the methanol recovery rate mapping dataset and the actual methanol recovery rate dataset. S63. Set the current number of fault detection executions (adaptive settings can be made according to the current actual generation situation) and execute S62 multiple times based on this. After each execution, obtain the corresponding methanol recovery rate mapping dataset and methanol recovery rate actual dataset. Detect the abnormal operating conditions in the production process of the corresponding fatty acid methyl ester according to the abnormal operating condition type set, and obtain the historical methanol recovery rate mapping dataset, the historical methanol recovery rate actual dataset, and the historical abnormal operating condition type encoding dataset. S64. Based on the historical methanol recovery rate mapping data set, the historical methanol recovery rate actual data set, and the historical abnormal operating condition type coding dataset, construct a mapping model with multiple sets of methanol recovery rate mapping data and methanol recovery rate actual data as inputs and abnormal operating condition type coding data as outputs, and obtain the final abnormal operating condition type coding mapping model. S64 includes the following steps: S641. Construct an initial abnormal working condition type encoding mapping model and set a second training data ratio (such as 8:2 or 7:3, which can be adjusted adaptively according to the actual training situation); divide the historical methanol recovery rate mapping data set, the historical methanol recovery rate actual data set, and the historical abnormal working condition type encoding dataset according to the second training data ratio to obtain the second training dataset and the second test dataset. S642. Set a second training error threshold (10%~15%, which can be adjusted adaptively according to the actual training situation); input the second training dataset into the initial abnormal working condition type encoding mapping model for training; during the training process, if the training error is less than the second training error threshold, stop training and obtain the trained abnormal working condition type encoding mapping model; otherwise, continue training until the training error is less than the second training error threshold. S643. Set a second test accuracy threshold (90%~95%, which can be adjusted adaptively according to the actual test situation); input the second test dataset into the trained abnormal working condition type encoding mapping model for testing; after the test is completed, obtain the second test accuracy data; if the second test accuracy data is greater than or equal to the second test accuracy threshold, use the trained abnormal working condition type encoding mapping model as the final abnormal working condition type encoding mapping model; otherwise, return to S642 to continue training the trained abnormal working condition type encoding mapping model and repeat S643 until the second test accuracy data is greater than or equal to the second test accuracy threshold. The initial abnormal operating condition type encoding mapping model can adopt a multilayer perceptron classification model, which is suitable for scenarios where operating condition types are classified based on multiple sets of recovery rate difference features. It boasts high classification accuracy and strong generalization. The model includes an input layer, three fully connected layers, and an output layer. The number of nodes in the input layer corresponds to the total dimension of multiple sets of methanol recovery rate mapping data and actual recovery rate data, and is used to receive the feature vector composed of model predictions and actual measured values. The first fully connected layer has 32 neurons and uses the ReLU linear rectified activation function to initially extract basic features such as the deviation and fluctuation trend between predicted and measured values. The second fully connected layer has 64 neurons and uses ReLU linear rectified activation function. The ReLU linear rectified activation function is used to deeply fuse the temporal variation and difference distribution features of multiple sets of data. The third fully connected layer has 32 neurons and uses the ReLU linear rectified activation function to filter and reduce the dimensionality of high-dimensional features, highlighting the distinguishing features of no fault, equipment fault, instrument abnormality, and material fluctuation. The output layer has the same number of neurons as the abnormal operating condition categories and uses the Softmax normalized exponential activation function to output the probability distribution of each type of operating condition and determine the final operating condition code. In addition, the model uses an adaptive moment estimation optimizer with a learning rate of 0.001 and a cross-entropy loss function to improve classification accuracy and training stability. S65. Based on the current number of fault detection executions, perform the most recent fatty acid methyl ester production process. Repeat S4 and S5 during each process and record the corresponding current final methanol recovery rate data and actual methanol recovery rate data to obtain the methanol recovery rate mapping data set and the methanol recovery rate actual data set. The set of methanol recovery rate to be tested and the set of actual methanol recovery rate to be tested are input into the final abnormal operating condition type coding mapping model for mapping to obtain the current abnormal operating condition type coding data; and the abnormal operating conditions in the recent fatty acid methyl ester production process are processed according to the current abnormal operating condition type coding data. For example, consider a continuous industrial production line for fatty acid methyl esters; as follows: Real-time process data during the steady-state operation of the production line were collected on-site. The data included the current esterification reaction temperature (77.2℃), neutralization reaction temperature (63.5℃), methanol moisture content (12.6%), methanol loss in waste gas (0.091 kg / h), and distillation tower operating pressure (101.6 kPa). All real-time data were integrated to form a complete dataset of current methanol recovery process parameters. This structured process parameter data was directly imported into a trained methanol recovery rate mapping model for intelligent mapping calculation. The model output an initial methanol recovery rate of 83.2%. This was combined with the production line's monthly energy consumption control, environmental emission constraints, and material recycling... Using actual production standards, the current methanol recovery rate threshold for this production cycle was uniformly set at 85.0%. Numerical comparison showed that the initial methanol recovery rate was lower than the preset control threshold, automatically triggering a dynamic adjustment process for process parameters. Multiple rounds of gradient fine-tuning were performed on the core controllable parameters. After the first round of parameter adjustment, the esterification reaction temperature was set to 78.6℃, the neutralization reaction temperature to 64.2℃, the water content of the methanol with water to 11.9%, and the distillation column pressure to 101.1 kPa, forming a dataset of methanol recovery process parameters after the first adjustment. Substituting these parameters into the first methanol recovery rate mapping model, the calculated methanol recovery rate after the first adjustment was 83.9%. The second round of parameter optimization... After adjustment, the esterification reaction temperature was set to 79.3℃, the neutralization reaction temperature to 64.8℃, the water content of methanol with water to 11.3%, and the distillation column pressure to 100.7 kPa. The remapping yielded a recovery rate of 84.5%. The first maximum number of repeated adjustments was preset to 3. After the third round of limit parameter control, the corresponding process parameters were: esterification reaction temperature 79.8℃, neutralization reaction temperature 65.3℃, water content of methanol with water to 10.8%, and distillation column pressure to 100.2 kPa. This mapping showed a methanol recovery rate of 84.8% after one adjustment, still below the 85.0% threshold and having exhausted the preset maximum number of adjustments. Therefore, dataset expansion was initiated. The model iterative upgrade process involved collecting nearly thirty sets of real data on process parameters and corresponding recovery rates under similar inefficient operating conditions on-site. These data were then batch-added to the original historical methanol recovery process parameter dataset and historical methanol recovery rate dataset. Based on the expanded new sample set, the model training iteration was completed again, resulting in a secondary methanol recovery rate mapping model with higher accuracy and better adaptability to operating conditions. Based on the secondary model, parameter refinement and iterative optimization were carried out again. The first round of secondary parameter adjustments used the following values: esterification reaction temperature 79.5℃, neutralization reaction temperature 65.1℃, water content of methanol with water content 11.0%, and distillation tower pressure 100.4 kPa. This resulted in a methanol recovery rate of 85% after the secondary adjustment.The methanol recovery rate is 3%, which meets the preset recovery rate threshold. This value is then determined as the current final methanol recovery rate and proceeds to the subsequent abnormal operating condition detection stage. The range of abnormal operating conditions corresponding to this production is defined in advance, categorized into four types: fault-free operation, equipment failure, instrument malfunction, and material fluctuation. Numerical codes are defined for each category: fault-free operation corresponds to code 01, equipment failure to code 02, instrument malfunction to code 03, and material fluctuation to code 04. The number of fault detection executions is set to 5 times, with five complete production cycles conducted consecutively. Each production cycle fully replicates the aforementioned parameter adjustment, model mapping, and recovery rate achievement process. The current final methanol recovery rate data for each of the five production batches is recorded sequentially as 85.3%, 85.1%, 85.5%, 85.2%, and 85.4%. Simultaneously, offline laboratory testing and online exhaust gas analysis are used to verify the results. Through methods such as material testing and verification, actual methanol recovery rate data for the corresponding batches were collected simultaneously, which were 85.0%, 84.9%, 85.7%, 85.1%, and 85.6%, respectively. All mapping calculation data and actual detection data were integrated and summarized to construct a set of mapped data for the methanol recovery rate to be tested and a set of actual data for the methanol recovery rate to be tested. These two sets of multi-dimensional datasets were simultaneously input into the final abnormal operating condition type coding mapping model after training for feature matching and intelligent judgment. After comprehensively analyzing the recovery rate deviation amplitude, parameter fluctuation patterns, and data deviation characteristics, the model outputs the current abnormal operating condition type code as 03, indicating an instrument malfunction in the production process. Based on the coding judgment result, maintenance personnel disassembled, calibrated, and cleaned the temperature sensor, pressure transmitter, and moisture detection probe accordingly. After rectifying the instrument malfunction, the production line returned to normal and stable operation. By systematically identifying and numerically encoding typical operating condition types such as no-fault, equipment failure, instrument malfunction, and material fluctuation, a standardized and quantifiable abnormal operating condition identification system was constructed. This provides a clear and unified classification basis for subsequent model training and operating condition judgment, avoiding judgment bias caused by ambiguity in operating condition types. Through repeated execution of core process parameter adjustments and model mapping processes, methanol recovery rate mapping data and actual data were continuously accumulated. Combined with operating condition type annotations during multiple fault detection executions, a historical dataset covering diverse operating conditions and with close data correlation was formed. This ensures that the model can fully learn the differences and mapping patterns of recovery rate data under different operating conditions. Based on this dataset, a... The established abnormal operating condition type coding mapping model achieves accurate conversion from recovery rate data differences to operating condition types, breaking the limitations of traditional reliance on manual experience to judge anomalies. It can quickly identify different types of problems such as equipment failure and instrument distortion. In actual production, by collecting real-time recovery rate mapping data and actual data and inputting them into the model, the corresponding abnormal operating condition codes can be output instantly, providing operators with clear fault indications, significantly shortening the time for anomaly investigation and handling, and effectively avoiding increased methanol loss, product quality fluctuations, or production interruptions caused by abnormal operating conditions. This significantly improves the controllability of operating conditions, fault response speed, and overall operational stability of the fatty acid methyl ester production process.

[0021] Example 2 This embodiment discloses a methanol recovery rate optimization management system, which can implement the methods of the above embodiments, including a recovery rate data acquisition module, a recovery rate mapping model construction module, a recovery rate mapping module, a primary adjustment module, a dataset expansion module, and an abnormal operating condition handling module; The recovery rate data acquisition module is used to collect corresponding process parameters and methanol recovery rate data from historical production. The recovery rate mapping model building module is used to build a primary methanol recovery rate mapping model; The recovery rate mapping module is used to map the current initial methanol recovery rate data; The single-adjustment module is used to repeatedly adjust the methanol recovery process parameter data and remap it after adjustment; The dataset expansion module is used to expand the data collected in the recovery rate data acquisition module and retrain the model. The abnormal operating condition handling module is used to construct the final abnormal operating condition type encoding mapping model, then obtain the current methanol recovery rate mapping to be tested and the actual data and map them, and determine and handle abnormal operating conditions based on the mapping results.

[0022] Example 3 A method for producing fatty acid methyl esters includes managing the methanol recovery rate during the fatty acid methyl ester production process. The management of the methanol recovery rate during the fatty acid methyl ester production process employs the method described in the above embodiments, and includes the following steps: 1. Take 1000 kg of palm oil and put it into a thin-film evaporator. Dehydrate it for 20 min at a temperature of 110℃ and a vacuum of 0.08 MPa, controlling the moisture content to 0.2% and the acid value to 3.5 mg KOH / g. 2. Transfer the dehydrated palm oil into the reactor, add 30 kg of concentrated sulfuric acid catalyst (98% industrial grade concentrated sulfuric acid), add 420 kg of methanol at a methanol-to-oil molar ratio of 4:1, stir at 300 r / min, keep the temperature at 75℃, react for 90 min, and the acid value drops to 0.8 mg KOH / g. 3. After neutralization, add 15 kg of sodium methoxide catalyst, add methanol to make the total alcohol-oil molar ratio 6:1, temperature 65℃, react for 120 min, let stand for 30 min to separate the layers, and separate the lower layer of crude glycerol 120 kg. 4. The crude methyl ester is fed into a vacuum distillation column at a temperature of 95°C and a vacuum degree of 0.09 MPa to remove 85 kg of residual methanol, which is then recovered and reused; wherein the recovery and reuse adopts the method of the above embodiment. 5. Wash the crude methyl ester twice with 150 kg of deionized water, stirring for 15 min each time. After standing and separating into layers, drain the water and then vacuum dry (80℃, 0.09 MPa) for 30 min to obtain the finished fatty acid methyl ester.

[0023] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0024] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

Claims

1. A method for optimizing methanol recovery rate management, characterized in that, Includes the following steps: S1. Collect corresponding methanol recovery process parameters and methanol recovery rate data from historical production. S2. Construct a primary methanol recovery rate mapping model based on the data collected in S1; S3. Input the current methanol recovery process parameter data into the primary methanol recovery rate mapping model for mapping; S4. Set the methanol recovery rate threshold. If the mapping result in S3 meets the threshold, no operation is required. Otherwise, repeatedly adjust the methanol recovery process parameter data and input it into the methanol recovery rate mapping model again to obtain the corresponding recovery rate data. S5. Set a first maximum number of repetitions. If the recovery rate data obtained in S4 meets the threshold and the number of repetitions does not exceed the upper limit, use it as the current final recovery rate and execute S6. Otherwise, expand the data collected in S1 and execute S2 again to retrain the model. Repeat the adjustment and mapping of methanol recovery process parameters until the mapping result meets the threshold and then execute S6. S6. Repeat S4 and S5 multiple times to obtain recovery rate mapping data and actual methanol recovery rate data, and combine them with the corresponding abnormal operating condition detection results to construct the final abnormal operating condition type coding mapping model. Then, obtain the current methanol recovery rate mapping to be tested and the actual data and input them into the final abnormal operating condition type coding mapping model for mapping. Determine the abnormal operating condition based on the mapping result and process it.

2. The methanol recovery rate optimization management method according to claim 1, characterized in that: The methanol recovery process parameters include esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, methanol loss in waste gas, and distillation tower pressure.

3. The methanol recovery rate optimization management method according to claim 2, characterized in that: The input to the primary methanol recovery rate mapping model is various methanol recovery process parameter data, and the output is methanol recovery rate data.

4. The methanol recovery rate optimization management method according to claim 3, characterized in that, S3 includes the following steps: S31. Obtain the data of various methanol recovery process parameters corresponding to the current fatty acid methyl ester production process, and input them into the primary methanol recovery rate mapping model for mapping to obtain the current initial methanol recovery rate data.

5. The methanol recovery rate optimization management method according to claim 4, characterized in that, S4 includes the following steps: S41. Set the current methanol recovery rate threshold according to the current production requirements of fatty acid methyl esters to obtain the current methanol recovery rate threshold; if the current initial methanol recovery rate data is greater than or equal to the current methanol recovery rate threshold, no action is required; otherwise, proceed to S42. S42. Repeatedly adjust the esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure parameters in the current methanol recovery process parameter dataset. After each repeated adjustment, an adjusted methanol recovery process parameter dataset is obtained. The adjusted methanol recovery process parameter dataset is then input into the methanol recovery rate mapping model for mapping to obtain the adjusted methanol recovery rate data.

6. The methanol recovery rate optimization management method according to claim 5, characterized in that, S5 includes the following steps: S51. Set a first maximum number of repetitions. If the methanol recovery rate data after one adjustment as described in S42 is greater than or equal to the current methanol recovery rate threshold and the number of repetitions is less than or equal to the first maximum number of repetitions, use the methanol recovery rate data after one adjustment as the current final methanol recovery rate data and execute S6; otherwise, execute S52. S52. Repeatedly expand the historical methanol recovery process parameter dataset and historical methanol recovery rate dataset mentioned in S21. After expansion, S21 is executed again to obtain the secondary methanol recovery rate mapping model. The esterification reaction temperature, neutralization reaction temperature, water content of methanol with water, and distillation tower pressure parameters in the current methanol recovery process parameter dataset are repeatedly adjusted. After each adjustment, a second-adjusted methanol recovery process parameter dataset is obtained. The second-adjusted methanol recovery process parameter dataset is input into the second-stage methanol recovery rate mapping model for mapping to obtain the second-adjusted methanol recovery rate data. This process continues until the second-adjusted methanol recovery rate data is greater than or equal to the current methanol recovery rate threshold. Finally, the second-adjusted methanol recovery rate data is used as the current final methanol recovery rate data, and step S6 is executed.

7. The methanol recovery rate optimization management method according to claim 6, characterized in that, S6 includes the following steps: S61. Several abnormal operating conditions are defined in the production process of fatty acid methyl ester to obtain a set of abnormal operating conditions; the set of abnormal operating conditions includes no fault, equipment fault, instrument abnormality and material fluctuation; numerical codes are performed on each abnormal operating condition type in the set of abnormal operating conditions to obtain a set of abnormal operating condition codes. S62. Repeat S4 and S5 multiple times to obtain the current final methanol recovery rate data corresponding to the repetition process, and collect the actual methanol recovery rate data after each repetition to obtain the methanol recovery rate mapping dataset and the actual methanol recovery rate dataset. S63. Set the current number of fault detection executions and execute S62 multiple times based on this. After each execution, obtain the corresponding methanol recovery rate mapping dataset and the actual methanol recovery rate dataset. Detect the abnormal operating condition types in the production process of the corresponding fatty acid methyl ester according to the abnormal operating condition type set. Then, construct a mapping model with multiple sets of methanol recovery rate mapping data and actual methanol recovery rate data as inputs and abnormal operating condition type encoded data as outputs to obtain the final abnormal operating condition type encoded mapping model.

8. The methanol recovery rate optimization management method according to claim 7, characterized in that, S6 further includes the following steps: S64. Based on the current number of fault detection executions, perform the most recent fatty acid methyl ester production process. Repeat S4 and S5 each time and record the corresponding current final methanol recovery rate data and actual methanol recovery rate data. Input the data into the final abnormal operating condition type coding mapping model for mapping to obtain the current abnormal operating condition type coding data. Process the abnormal operating conditions in the most recent fatty acid methyl ester production process according to the current abnormal operating condition type coding data.

9. A system for implementing the methanol recovery rate optimization management method as described in any one of claims 1-8, characterized in that: It includes a recovery rate data acquisition module, a recovery rate mapping model construction module, a recovery rate mapping module, a primary adjustment module, a dataset expansion module, and an abnormal operating condition handling module; The recovery rate data acquisition module is used to collect corresponding process parameters and methanol recovery rate data from historical production. The recovery rate mapping model building module is used to build a primary methanol recovery rate mapping model; The recovery rate mapping module is used to map the current initial methanol recovery rate data; The single-adjustment module is used to repeatedly adjust the methanol recovery process parameter data and remap it after adjustment; The dataset expansion module is used to expand the data collected in the recovery rate data acquisition module and retrain the model. The abnormal operating condition handling module is used to construct the final abnormal operating condition type encoding mapping model, then obtain the current methanol recovery rate mapping to be tested and the actual data and map them, and determine and handle abnormal operating conditions based on the mapping results.

10. A method for producing fatty acid methyl esters, comprising managing the methanol recovery rate during the fatty acid methyl ester production process, characterized in that: The management of methanol recovery rate in the fatty acid methyl ester production process adopts the methanol recovery rate optimization management method as described in any one of claims 1-8.

Images