Safety early warning method and system for continuous production of hydrolytic polymaleic anhydride

Abstract

The invention relates to a safety early warning method and system for continuous production of hydrolytic polymaleic anhydride, and the method comprises the steps: L1, in the continuous production process of hydrolytic polymaleic anhydride, obtaining the data information of the temperature of a reaction kettle in real time based on a temperature sensor on the reaction kettle, data information of pressure of the reaction kettle is acquired in real time based on a pressure sensor on the reaction kettle, data information of rotating speed of a stirrer in the reaction kettle is acquired in real time based on a tachometer on the reaction kettle, and data of temperature and pressure of the reaction kettle and data of the rotating speed of the stirrer are fused by adopting an improved multi-modal model level fusion algorithm. And constructing a combined characteristic matrix of the reaction kettle. According to the invention, the operation state of the reaction kettle can be accurately predicted in advance, and safety early warning is carried out, so that the quality of the product is ensured to meet the standard, the whole monitoring process does not need manual participation, the manual error is reduced, and the production efficiency of hydrolytic polymaleic anhydride is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of continuous production of hydrolyzed polymaleic anhydride, and particularly to a safety warning method and system for continuous production of hydrolyzed polymaleic anhydride. Background Art

[0002] Hydrolyzed polymaleic anhydride is a low molecular weight polyelectrolyte, generally with a relative molecular weight of 400-800. It is non-toxic, easily soluble in water, has high chemical and thermal stability, and a decomposition temperature above 330°C. It has an obvious solubility limit effect at high temperature (less than 350°C) and high pH. HPMA is suitable for alkaline water quality or for compounding with other drugs. HPMA still has good scale inhibition and dispersion effects on carbonates below 300°C, and the scale inhibition time can reach 100h. Due to its excellent scale inhibition performance and high temperature resistance, it is widely used in flash evaporation devices for seawater desalination, low-pressure boilers, steam locomotives, crude oil dehydration, water and oil pipelines, and industrial circulating cooling water. In addition, HPMA has a certain corrosion inhibition effect, and has a better compounding effect with zinc salts.

[0003] In the continuous production process of hydrolyzed polymaleic anhydride, if the temperature of the reaction kettle is too high or too low, the pressure of the reaction kettle is too high, or the rotation speed of the stirrer in the reaction kettle does not meet the standard, it will affect the output, quality and production efficiency of the product. Therefore, how to give an early warning of the state of the reaction kettle, so as not to reduce the quality and production efficiency of the product has become an urgent problem for us to solve. Summary of the Invention

[0004] In view of the above problems, the present invention provides a safety warning method and system for continuous production of hydrolyzed maleic anhydride, which can not only accurately predict the operating state of the reaction kettle in advance and give a safety warning, so as to ensure that the quality of the product meets the standard, but also the entire monitoring process does not require manual participation, reduces human error, and improves the production efficiency of hydrolyzed polymaleic anhydride.

[0005] In order to achieve the above object and other related objects, the technical solutions provided by the present invention are as follows:

[0006] A safety warning method for continuous production of hydrolyzed polymaleic anhydride, the method comprising:

[0007] L1. During the continuous production of hydrolyzed polymaleic anhydride, based on the temperature sensor on the reaction kettle, the data information of the temperature of the reaction kettle is obtained in real time, based on the pressure sensor on the reaction kettle, the data information of the pressure of the reaction kettle is obtained in real time, and based on the tachometer on the reaction kettle, the data information of the rotation speed of the stirrer in the reaction kettle is obtained in real time;

[0008] L2. Based on the data information of the temperature of the reactor, the data information of the pressure of the reactor, and the data information of the rotation speed of the agitator in the reactor, the improved multi-modal model-level fusion algorithm is used to fuse the data of the temperature, pressure, and rotation speed of the reactor, and a joint feature matrix of the reactor is constructed, and the data information of the joint feature matrix of the reactor is output;

[0009] L3. Based on the data information of the joint feature matrix of the reactor, the algorithm of the weaver bird optimization BP neural network regression prediction based on adaptive weights is used to predict the operating state of the reactor, and the data information of the predicted operating state of the reactor is obtained;

[0010] L4. Based on the data information of the predicted operating state of the reactor, a safety warning function W for the continuous production of hydrolyzed polymaleic anhydride is established, and the warning value for the safe production of hydrolyzed polymaleic anhydride is calculated, and the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride is output.

[0011] Further, the method further includes:

[0012] L5. Based on the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride, a preset safety threshold is set. If the warning value for the safe production of hydrolyzed polymaleic anhydride is less than the preset safety threshold, the reactor operates normally and continues production. If the warning value for the safe production of hydrolyzed polymaleic anhydride is greater than the preset safety threshold, the reactor operates abnormally and a warning is issued, and temporary maintenance is required.

[0013] Further, the safety warning function W for the continuous production of hydrolyzed polymaleic anhydride is

[0014] where x is the data information of the predicted operating state of the reactor, and α1, α2, and α3 are weight coefficients.

[0015] Further, the constraint conditions of the weight coefficients α1, α2, and α3 are

[0016]

[0017] Further, in step L2, the use of the improved multi-modal model-level fusion algorithm to fuse the data of the temperature, pressure, and rotation speed of the reactor includes:

[0018] L21. Input the data information of the temperature of the reactor into the first layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the first layer of the LSTM model;

[0019] L22. Input the data information of the hidden layer state of each neuron in the first layer of the LSTM model and the data information of the pressure of the reactor into the second layer of the LSTM model for training and learning to obtain the data information of the hidden layer state of each neuron in the second layer of the LSTM model;

[0020] L23. Input the data information of the hidden layer state of each neuron in the second layer of the LSTM model and the data information of the rotation speed of the reactor agitator into the third layer of the LSTM model for training and learning to obtain the data information of the hidden layer state of each neuron in the third layer of the LSTM model;

[0021] L24. Based on the data information of the hidden layer state of each neuron in the third layer of the LSTM model, construct the joint feature function Q of the reactor,

[0022]

[0023] where y is the data information of the hidden layer state of each neuron in the third layer of the LSTM model, and β1, β2, and β3 are the penalty factors of the reactor joint features, which characterize the joint features of the reactor, construct the joint feature matrix of the reactor, and output the data information of the joint feature matrix of the reactor.

[0024] Further, the penalty factors β1, β2, and β3 of the reactor joint features are

[0025]

[0026] where y is the data information of the hidden layer state of each neuron in the third layer of the LSTM model.

[0027] Further, in step L3, the use of the bowerbird optimization BP neural network regression prediction algorithm based on adaptive weights to predict the operating state of the reactor includes:

[0028] L31. Input the data information of the predicted operating state of the reactor into the BP neural network regression prediction model for training and learning, initialize the weights and biases of the model to obtain the data information of the weights and biases of the model;

[0029] L32. Based on the data information of the weights and biases of the model, initialize the bowerbird population, determine the population parameters and the maximum number of iterations M to obtain the data information of the initialized bowerbird population;

[0030] L33. Based on the data information of the initialized bowerbird population, establish the fitness function P of the population individuals,

[0031]

[0032] Among them, z is the data information of the initialized bowerbird population, δ1, δ2 and δ3 are the expansion coefficients of the population individuals, and the fitness values ​​of the population individuals are calculated to obtain the data information of the fitness values ​​of the bowerbird population individuals;

[0033] L34. Based on the data information of the fitness value of the individual gardener bird population, establish the target optimization function S,

[0034]

[0035] Among them, r is the data information of the fitness value of the individual of the gardener bird population, η1, η2 and η3 are adaptive weight coefficients, and the weight and bias of the model are optimized to obtain the data information of the weight and bias of the optimized model;

[0036] L35. Based on the weight and bias data information of the optimized model, an optimized BP neural network regression prediction model is obtained, the data information of the predicted operating status of the reactor is input, the operating status of the reactor is predicted, and the data information of the predicted operating status of the reactor is obtained.

[0037] Furthermore, the adaptive weight coefficients η1, η2 and η3 are,

[0038]

[0039] Among them, r is the data information of the fitness value of the individual bowerbird population.

[0040] In order to achieve the above-mentioned object and other related objects, the present invention also provides a safety early warning system for continuous production of hydrolyzed polymaleic anhydride, comprising a computer device, the computer device being programmed or configured to execute any one of the steps of the safety early warning method for continuous production of hydrolyzed polymaleic anhydride, the system further comprising:

[0041] A data acquisition module is used to acquire data information on the temperature of the reactor in real time based on a temperature sensor on the reactor, acquire data information on the pressure of the reactor in real time based on a pressure sensor on the reactor, and acquire data information on the rotation speed of a stirrer in the reactor in real time based on a tachometer on the reactor;

[0042] A joint feature matrix extraction module of the reactor, connected to the data acquisition module, is used to fuse the data of the temperature, pressure and speed of the stirrer of the reactor using an improved multimodal model-level fusion algorithm, and to construct a joint feature matrix of the reactor, and output data information of the joint feature matrix of the reactor;

[0043] The reactor operation status prediction module is connected to the combined feature matrix extraction module of the reactor, and is used to predict the operation status of the reactor by using the BP neural network regression prediction algorithm optimized by the bowerbird based on adaptive weights, so as to obtain the data information of the operation status of the predicted reactor;

[0044] The safety warning module is connected to the reactor operation status prediction module, and is used to establish the safety warning function W for the continuous production of hydrolyzed polymaleic anhydride, calculate the warning value for the safe production of hydrolyzed polymaleic anhydride, and output the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride;

[0045] The production threshold module of hydrolyzed polymaleic anhydride is connected to the safety warning module, and is used to set a preset safety threshold. If the warning value for the safe production of hydrolyzed polymaleic anhydride is less than the preset safety threshold, the reactor operates normally and continues production. If the warning value for the safe production of hydrolyzed polymaleic anhydride is greater than the preset safety threshold, the reactor operates abnormally and issues a warning, and temporary maintenance is required.

[0046] To achieve the above and other related purposes, the present invention also provides a computer-readable storage medium, on which a computer program is stored that is programmed or configured to execute the safety warning method for the continuous production of hydrolyzed polymaleic anhydride described in any one of the above.

[0047] The present invention has the following positive effects:

[0048] 1. By using the improved multi-modal model-level fusion algorithm to fuse the data of the temperature, pressure of the reactor and the rotation speed of the mixer, constructing the combined feature matrix of the reactor, and combining the BP neural network regression prediction algorithm optimized by the bowerbird based on adaptive weights to predict the operation status of the reactor, the present invention can not only accurately predict the status of the reactor in advance, ensure the quality and production efficiency of hydrolyzed polymaleic anhydride, but also the whole process does not require manual participation, reduces human error, and ensures the safety of the production of hydrolyzed maleic anhydride.

[0049] 2. By establishing the safety warning function W for the continuous production of hydrolyzed polymaleic anhydride and calculating the warning value for the safe production of hydrolyzed polymaleic anhydride, the present invention can not only timely and accurately warn the status of the reactor, ensure the safety of the entire production line, but also further save labor costs and improve the production efficiency of hydrolyzed polymaleic anhydride. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic flow chart of the method of the present invention;

[0051] Figure 2Schematic flow chart of the improved multi-modal model-level fusion algorithm of the present invention;

[0052] Figure 3 Schematic flow chart of the bowerbird optimization BP neural network regression prediction algorithm based on adaptive weights of the present invention;

[0053] Figure 4 Schematic diagram of the system framework of the present invention;

[0054] Figure 5 Schematic diagram of the structure of the reactor of the present invention.

[0055] Explanation of reference numerals in the figure: 1 - temperature sensor, 2 - pressure sensor, 3 - tachometer. Detailed implementation manners

[0056] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to assist understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0057] Embodiment 1: As Figure 1 or Figure 5 shown, a safety warning method for continuous production of hydrolyzed polymaleic anhydride, the method includes:

[0058] L1. During the continuous production of hydrolyzed polymaleic anhydride, based on the temperature sensor 1 on the reactor, the data information of the temperature of the reactor is obtained in real time, based on the pressure sensor 2 on the reactor, the data information of the pressure of the reactor is obtained in real time, and based on the tachometer 3 on the reactor, the data information of the rotation speed of the stirrer in the reactor is obtained in real time;

[0059] L2. Based on the data information of the temperature of the reactor, the data information of the pressure of the reactor, and the data information of the rotation speed of the stirrer in the reactor, an improved multi-modal model-level fusion algorithm is used to fuse the data of the temperature, pressure, and rotation speed of the stirrer of the reactor, and a joint feature matrix of the reactor is constructed, and the data information of the joint feature matrix of the reactor is output;

[0060] L3. Based on the data information of the joint feature matrix of the reactor, a bowerbird optimization BP neural network regression prediction algorithm based on adaptive weights is used to predict the operating state of the reactor, and the data information of the predicted operating state of the reactor is obtained;

[0061] L4. Based on the data information of the predicted operating state of the reactor, establish a safety warning function W for the continuous production of hydrolyzed polymaleic anhydride, calculate the warning value for the safe production of hydrolyzed polymaleic anhydride, and output the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride.

[0062] In this embodiment, the method further includes:

[0063] L5. Based on the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride, set a preset safety threshold. If the warning value for the safe production of hydrolyzed polymaleic anhydride is less than the preset safety threshold, the reactor operates normally and continues production. If the warning value for the safe production of hydrolyzed polymaleic anhydride is greater than the preset safety threshold, the reactor operates abnormally, issues a warning, and requires temporary maintenance.

[0064] In this embodiment, the safety warning function W for the continuous production of hydrolyzed polymaleic anhydride is

[0065]

[0066] where x is the data information of the predicted operating state of the reactor, and α1, α2, and α3 are weight coefficients.

[0067] In this embodiment, the constraint conditions for the weight coefficients α1, α2, and α3 are

[0068]

[0069] In this embodiment, as Figure 2 shown, in step L2, the fusion of the data of the temperature, pressure of the reactor, and the rotation speed of the stirrer by using the improved multi-modal model-level fusion algorithm includes:

[0070] L21. Input the data information of the temperature of the reactor into the first layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the first layer of the LSTM model;

[0071] L22. Input the data information of the hidden layer state of each neuron in the first layer of the LSTM model and the data information of the pressure of the reactor into the second layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the second layer of the LSTM model;

[0072] L23. Input the data information of the hidden layer state of each neuron in the second layer of the LSTM model and the data information of the rotation speed of the reactor stirrer into the third layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the third layer of the LSTM model;

[0073] Based on the data information of the hidden layer state of each neuron in the third layer of the LSTM model, construct the combined feature function Q of the reactor.

[0074]

[0075] Among them, y is the data information of the hidden layer state of each neuron in the third layer of the LSTM model, and β1, β2, and β3 are the penalty factors of the combined features of the reactor, which characterize the combined features of the reactor, construct the combined feature matrix of the reactor, and output the data information of the combined feature matrix of the reactor.

[0076] In this embodiment, the penalty factors β1, β2, and β3 of the combined features of the reactor are

[0077]

[0078] Among them, y is the data information of the hidden layer state of each neuron in the third layer of the LSTM model.

[0079] Reactor R0401A / B

[0080] When the temperature of reactor R0401A / B is too high or too low, the interlock executes actions: close the emergency cut-off valve XZV-1401A / B for concentrated sulfuric acid dropping, close the emergency cut-off valve XZV-1402A / B for hydrogen peroxide dropping, close the emergency cut-off valve XZV-1405A / B for steam, close the drain pipeline valve XZV-1406A / B, open the cooling water inlet valve XZV-1403A / B, and open the cooling water return valve XZV-1404A / B.

[0081] Overpressure interlock protection for reactor R0401A / B

[0082] When the pressure of reactor R0401A / B is too high, the interlock executes actions: close the emergency cut-off valve XZV-1401A / B for concentrated sulfuric acid dropping, close the emergency cut-off valve XZV-1402A / B for hydrogen peroxide dropping, close the emergency cut-off valve XZV-1405A / B for steam, close the drain pipeline valve XZV-1406A / B, open the cooling water inlet valve XZV-1403A / B, and open the cooling water return valve XZV-1404A / B.

[0083] Stirring current fault protection for reactor R0401A / B

[0084] When the stirring current of the reaction kettle R0401A / B fails or stops, the interlock executes actions: closing the emergency cut-off valves XZV-1401A / B for concentrated sulfuric acid dropping, closing the emergency cut-off valves XZV-1402A / B for hydrogen peroxide dropping, closing the emergency cut-off valves XZV-1405A / B for steam, closing the valves of the drain pipe XZV-1406A / B, opening the cooling water inlet valves XZV-1403A / B, and opening the cooling water return valves XZV-1404A / B.

[0085] Example 2: Based on the safety warning method and system for continuous production of hydrolyzed polymaleic anhydride in Example 1, the present invention will be further described and explained below.

[0086] As Figure 1 Or Figure 5 shown, a safety warning method for continuous production of hydrolyzed polymaleic anhydride, the method includes:

[0087] L1. During the continuous production of hydrolyzed polymaleic anhydride, based on the temperature sensor 1 on the reaction kettle, the data information of the temperature of the reaction kettle is obtained in real time, based on the pressure sensor 2 on the reaction kettle, the data information of the pressure of the reaction kettle is obtained in real time, and based on the tachometer 3 on the reaction kettle, the data information of the rotation speed of the stirrer in the reaction kettle is obtained in real time;

[0088] L2. Based on the data information of the temperature of the reaction kettle, the data information of the pressure of the reaction kettle, and the data information of the rotation speed of the stirrer in the reaction kettle, an improved multi-modal model-level fusion algorithm is used to fuse the data of the temperature, pressure, and rotation speed of the stirrer of the reaction kettle, and a joint feature matrix of the reaction kettle is constructed, and the data information of the joint feature matrix of the reaction kettle is output;

[0089] L3. Based on the data information of the joint feature matrix of the reaction kettle, a BP neural network regression prediction algorithm optimized by the bowerbird based on adaptive weights is used to predict the operating state of the reaction kettle, and the data information of the predicted operating state of the reaction kettle is obtained;

[0090] L4. Based on the data information of the predicted operating state of the reaction kettle, a safety warning function W for continuous production of hydrolyzed polymaleic anhydride is established, and the warning value for the safe production of hydrolyzed polymaleic anhydride is calculated, and the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride is output.

[0091] In this embodiment, as Figure 3 shown, in step L3, the using the BP neural network regression prediction algorithm optimized by the bowerbird based on adaptive weights to predict the operating state of the reaction kettle includes:

[0092] L31. Input the data information of the combined feature matrix of the reaction kettle into the BP neural network regression prediction model for training and learning, initialize the weights and biases of the model, and obtain the data information of the weights and biases of the model;

[0093] L32. Based on the data information of the weights and biases of the model, initialize the bowerbird population, determine the population parameters and the maximum number of iterations M, and obtain the data information of the initialized bowerbird population;

[0094] L33. Based on the data information of the initialized bowerbird population, establish the fitness function P of the population individuals,

[0095]

[0096] where z is the data information of the initialized bowerbird population, δ1, δ2, and δ3 are the expansion coefficients of the population individuals, calculate the fitness values of the population individuals, and obtain the data information of the fitness values of the bowerbird population individuals;

[0097] L34. Based on the data information of the fitness values of the bowerbird population individuals, establish the target optimization function S,

[0098]

[0099] where r is the data information of the fitness values of the bowerbird population individuals, η1, η2, and η3 are the adaptive weight coefficients, optimize the weights and biases of the model, and obtain the data information of the optimized weights and biases of the model;

[0100] L35. Based on the data information of the optimized weights and biases of the model, obtain the optimized BP neural network regression prediction model, input the data information of the combined feature matrix of the reaction kettle, predict the operating state of the reaction kettle, and obtain the data information of the predicted operating state of the reaction kettle.

[0101] In this embodiment, the adaptive weight coefficients η1, η2, and η3 are

[0102]

[0103] where r is the data information of the fitness values of the bowerbird population individuals.

[0104] In this embodiment, as Figure 4 shown, the present invention provides a safety warning system for continuous production of hydrolyzed polymaleic anhydride, including a computer device, which is programmed or configured to execute the steps of any one of the safety warning methods for continuous production of hydrolyzed polymaleic anhydride, and the system further includes:

[0105] A data acquisition module, configured to obtain in real time the data information of the temperature of the reactor based on the temperature sensor on the reactor, obtain in real time the data information of the pressure of the reactor based on the pressure sensor on the reactor, and obtain in real time the data information of the rotation speed of the stirrer in the reactor based on the tachometer on the reactor;

[0106] A combined feature matrix extraction module of the reactor, connected to the data acquisition module, configured to fuse the data of the temperature, pressure and rotation speed of the stirrer of the reactor by using an improved multi-modal model-level fusion algorithm, construct a combined feature matrix of the reactor, and output the data information of the combined feature matrix of the reactor;

[0107] A reactor operating state prediction module, connected to the combined feature matrix extraction module of the reactor, configured to predict the operating state of the reactor by using a BP neural network regression prediction algorithm optimized by bowerbirds based on adaptive weights, and obtain the data information of the predicted operating state of the reactor;

[0108] A safety warning module, connected to the reactor operating state prediction module, configured to establish a safety warning function W for the continuous production of hydrolyzed polymaleic anhydride, calculate the warning value for the safe production of hydrolyzed polymaleic anhydride, and output the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride;

[0109] A production threshold module for hydrolyzed polymaleic anhydride, connected to the safety warning module, configured to set a preset safety threshold. If the warning value for the safe production of hydrolyzed polymaleic anhydride is less than the preset safety threshold, the reactor operates normally and continues production. If the warning value for the safe production of hydrolyzed polymaleic anhydride is greater than the preset safety threshold, the reactor operates abnormally and issues a warning, and temporary maintenance is required.

[0110] In this embodiment, the present invention provides a computer-readable storage medium, on which a computer program is stored that is programmed or configured to execute any one of the safety warning methods for the continuous production of hydrolyzed polymaleic anhydride.

[0111] Any reference to memory, storage, database, or other media used in the embodiments provided by this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0112] In summary, the present invention can not only accurately predict the operating state of the reactor in advance and give a safety warning, so as to ensure that the product quality meets the standards, but also the whole monitoring process does not require manual participation, reduces human errors, and improves the production efficiency of hydrolyzed polymaleic anhydride.

[0113] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present disclosure should be included within the protection scope of the present disclosure.

Claims

1. A safety warning method for the continuous production of hydrolyzed polymaleic anhydride, characterized in that, The method includes: L1. During the continuous production of hydrolyzed polymaleic anhydride, based on the temperature sensor on the reactor, the data information of the temperature of the reactor is obtained in real time, based on the pressure sensor on the reactor, the data information of the pressure of the reactor is obtained in real time, and based on the tachometer on the reactor, the data information of the rotation speed of the agitator in the reactor is obtained in real time; L2. Based on the data information of the temperature of the reactor, the data information of the pressure of the reactor, and the data information of the rotation speed of the agitator in the reactor, an improved multi-modal model-level fusion algorithm is used to fuse the data of the temperature, pressure, and rotation speed of the agitator of the reactor, and a joint feature matrix of the reactor is constructed, and the data information of the joint feature matrix of the reactor is output; L3. Based on the data information of the joint feature matrix of the reactor, an adaptive weight-based bowerbird optimization BP neural network regression prediction algorithm is used to predict the operating state of the reactor, and the data information of the predicted operating state of the reactor is obtained; L4. Based on the data information of the predicted operating state of the reactor, a safety warning function W for the continuous production of hydrolyzed polymaleic anhydride is established, and the warning value for the safe production of hydrolyzed polymaleic anhydride is calculated, and the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride is output.

2. The safety warning method for continuous production of hydrolyzed polymaleic anhydride according to claim 1, characterized in that, The method further includes: L5. Based on the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride, a preset safety threshold is set. If the warning value for the safe production of hydrolyzed polymaleic anhydride is less than the preset safety threshold, the reactor operates normally and continues production. If the warning value for the safe production of hydrolyzed polymaleic anhydride is greater than the preset safety threshold, the reactor operates abnormally and a warning is issued, and temporary maintenance is required.

3. The safety warning method for continuous production of hydrolyzed polymaleic anhydride according to claim 1, wherein: The safety warning function W for the continuous production of hydrolyzed polymaleic anhydride is where x is the data information of the predicted operating state of the reactor, and α1, α2, and α3 are weight coefficients.

4. The safety warning method for the continuous production of hydrolyzed polymaleic anhydride according to claim 3, wherein: The constraint conditions for the weight coefficients α1, α2, and α3 are 5. The safety warning method for continuous production of hydrolyzed polymaleic anhydride according to claim 1, wherein, In step L2, the use of the improved multi-modal model-level fusion algorithm to fuse the data of the temperature, pressure, and rotation speed of the agitator of the reactor includes: L21. Input the data information of the temperature of the reactor into the first layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the first layer of the LSTM model; L22. Input the data information of the hidden layer state of each neuron in the first layer of the LSTM model and the data information of the pressure of the reactor into the second layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the second layer of the LSTM model; L23. Input the data information of the hidden layer state of each neuron in the second layer of the LSTM model and the data information of the rotation speed of the reactor agitator into the third layer of the LSTM model for training and learning, and obtain the data information of the hidden layer state of each neuron in the third layer of the LSTM model; L24. Based on the data information of the hidden layer state of each neuron in the third layer of the LSTM model, a joint feature function Q of the reactor is constructed. Among them, y is the data information of the hidden layer state of each neuron in the third layer of the LSTM model. β1, β2, and β3 are the penalty factors of the combined features of the reactor, which characterize the combined features of the reactor, construct the combined feature matrix of the reactor, and output the data information of the combined feature matrix of the reactor.

6. The safety warning method for the continuous production of hydrolyzed polymaleic anhydride according to claim 5, characterized in that: The penalty factors β1, β2, and β3 of the combined features of the reactor are Among them, y is the data information of the hidden layer state of each neuron in the third layer of the LSTM model.

7. The safety warning method for continuous production of hydrolyzed polymaleic anhydride according to claim 1, characterized in that, In step L3, the use of the bowerbird optimization BP neural network regression prediction algorithm based on adaptive weights to predict the operating state of the reactor includes: L31. Input the data information of the predicted operating state of the reactor into the BP neural network regression prediction model for training and learning, initialize the weights and biases of the model, and obtain the data information of the weights and biases of the model. L32. Based on the data information of the weights and biases of the model, initialize the bowerbird population, determine the population parameters and the maximum number of iterations M, and obtain the data information of the initialized bowerbird population. L33. Based on the data information of the initialized bowerbird population, establish the fitness function P of the population individuals. Among them, z is the data information of the initialized bowerbird population, and δ1, δ2, and δ3 are the expansion coefficients of the population individuals, which are used to calculate the fitness values of the population individuals, and obtain the data information of the fitness values of the bowerbird population individuals. L34. Based on the data information of the fitness values of the bowerbird population individuals, establish the target optimization function S. Among them, r is the data information of the fitness values of the bowerbird population individuals, and η1, η2, and η3 are the adaptive weight coefficients, which are used to optimize the weights and biases of the model, and obtain the data information of the optimized weights and biases of the model. L35. Based on the data information of the optimized weights and biases of the model, obtain the optimized BP neural network regression prediction model, input the data information of the predicted operating state of the reactor, predict the operating state of the reactor, and obtain the data information of the predicted operating state of the reactor.

8. The safety warning method for the continuous production of hydrolyzed polymaleic anhydride according to claim 7, characterized in that: The adaptive weight coefficients η1, η2, and η3 are Among them, r is the data information of the fitness values of the bowerbird population individuals.

9. A safety warning system for the continuous production of hydrolyzed polymaleic anhydride, including a computer device, which is programmed or configured to execute the steps of the safety warning method for the continuous production of hydrolyzed polymaleic anhydride according to any one of claims 1 to 8, characterized in that, The system further includes: A data acquisition module, which is used to obtain the data information of the temperature of the reactor in real time based on the temperature sensor on the reactor, obtain the data information of the pressure of the reactor in real time based on the pressure sensor on the reactor, and obtain the data information of the rotation speed of the agitator in the reactor in real time based on the tachometer on the reactor. A combined feature matrix extraction module of the reactor, which is connected to the data acquisition module, and is used to fuse the data of the temperature, pressure, and rotation speed of the agitator of the reactor by using an improved multi-modal model-level fusion algorithm, construct the combined feature matrix of the reactor, and output the data information of the combined feature matrix of the reactor. The reactor operating state prediction module, connected to the combined feature matrix extraction module of the reactor, is used to predict the operating state of the reactor by using the BP neural network regression prediction algorithm optimized by the bowerbird based on adaptive weights, and obtain the data information of the predicted operating state of the reactor; The safety warning module, connected to the reactor operating state prediction module, is used to establish the safety warning function W for the continuous production of hydrolyzed polymaleic anhydride, calculate the warning value for the safe production of hydrolyzed polymaleic anhydride, and output the data information of the warning value for the safe production of hydrolyzed polymaleic anhydride; The hydrolyzed polymaleic anhydride production threshold module, connected to the safety warning module, is used to set a preset safety threshold. If the warning value for the safe production of hydrolyzed polymaleic anhydride is less than the preset safety threshold, the reactor operates normally and continues production. If the warning value for the safe production of hydrolyzed polymaleic anhydride is greater than the preset safety threshold, the reactor operates abnormally, issues a warning, and requires temporary maintenance.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is programmed or configured to execute the safety warning method for the continuous production of hydrolyzed polymaleic anhydride according to any one of claims 1 to 8.

Images