Method and device for predicting concentration of hydrogen sulfide
By combining neural network model and mechanism model in the supergravity desulfurization system, the problem of insufficient prediction accuracy of gas concentration at the outlet of the supergravity reactor in the prior art is solved, and higher prediction accuracy and reliability are achieved.
Patent Information
- Application Number
- CN202411396312.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-08
- Publication Date
- 2025-02-25
AI Technical Summary
In the prior art, the accuracy of predicting the gas concentration at the outlet of the supergravity reactor is poor, resulting in a large deviation from the actual measurement results.
A method combining gray box model and white box model is used to predict through the ultra-gravity desulfurization neural network model and the ultra-gravity desulfurization mechanism model. The neural network model is used to fit complex data laws in the desulfurization process, while the mechanism model is based on the desulfurization mass transfer reaction model and the mass balance model, and accurately follows physical laws.
The prediction accuracy of concentration at the reactor outlet is significantly improved, and the data fitting ability of the black box model and the physical law constraints of the white box model are combined to ensure the reliability and accuracy of the prediction results.
Smart Images
Figure CN119517212A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of desulfurization technology, in particular to a method for preparing hydrogen sulfide Concentration prediction method and device. Background Art
[0002] In order to calculate the inlet and outlet concentrations of a chemical reactor, it is necessary to assume certain conditions and simplify certain realistic laws in order to establish a mathematical model.
[0003] In related technologies, continuous stirred tank reactors (CSTRs) or tubular reactors (PFRs) are often modeled based on material conservation equations. However, since ultra-gravity reactors involve multiphase flow and uneven mixing, the accuracy of material conservation is poor, resulting in a large deviation between the predicted gas concentration at the outlet and the actual measurement results. Summary of the invention
[0004] The present invention provides a hydrogen sulfide The invention discloses a concentration prediction method and device, which are used to solve the defect of poor accuracy in predicting the gas concentration at the outlet of a supergravity reactor in the prior art.
[0005] The present invention provides a hydrogen sulfide The concentration prediction method comprises the following steps.
[0006] The composition of the desulfurizer and the desulfurization operating conditions are input into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of the concentration; the high gravity desulfurization neural network model is used to predict the concentration at the outlet of the high gravity reactor concentration; According to the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system, we can get The second prediction result of the concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and the chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; According to the first prediction result and the second prediction result, hydrogen sulfide is obtained. Target prediction results of concentration.
[0007] A hydrogen sulfide provided by the present invention Concentration prediction method, the ultra-gravity desulfurization mechanism model is established based on the following method: According to the Hebe-Bess penetration theory, a desulfurization mass transfer reaction model in spherical coordinate system is established; According to the desulfurization mass transfer reaction model, a liquid mass transfer coefficient is determined; the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor; The high-gravity desulfurization mechanism model is established according to the liquid mass transfer coefficient and the desulfurization mass balance model.
[0008] A hydrogen sulfide provided by the present invention The concentration prediction method, the desulfurization mass transfer reaction model in the spherical coordinate system, comprises: = ( )-
[0009] in, represents the solute concentration; D represents the diffusion coefficient of the solute in the liquid phase; r represents the penetration radius; t represents the time; Represents the reaction rate.
[0010] A hydrogen sulfide provided by the present invention The concentration prediction method, the liquid mass transfer coefficient, comprises:
[0011] in, represents the liquid mass transfer coefficient; It represents the mass transfer rate per unit interfacial area; It means that the desulfurization reaction reaches equilibrium concentration; express initial concentration.
[0012] A hydrogen sulfide provided by the present invention The concentration prediction method, the ultra-gravity desulfurization mechanism model, comprises:
[0013] in, The outlet of the reactor concentration; Indicates the gas-liquid interface The mole fraction of represents the liquid mass transfer coefficient; T represents the temperature; express Henry's constant; express The partial pressure of represents the mass transfer interface area; h represents the thickness of the packing layer; Represents the outer diameter of the reactor; represents the inner diameter of the reactor; G represents of traffic.
[0014] A hydrogen sulfide provided by the present invention Concentration prediction method, the high gravity desulfurization neural network model is trained based on the following method: Acquire training samples; the training samples include desulfurization agent samples and desulfurization operation condition samples; The desulfurization agent sample and the desulfurization operation condition sample are input into the high-gravity desulfurization neural network model to obtain the high-gravity reactor outlet Concentration prediction results; According to the outlet of the supergravity reactor The concentration prediction results and the measured concentration at the outlet of the high gravity reactor concentration, training the high-gravity desulfurization neural network model to obtain a trained high-gravity desulfurization neural network model.
[0015] A hydrogen sulfide provided by the present invention The concentration prediction method is to obtain hydrogen sulfide according to the first prediction result and the second prediction result. The target prediction results of concentration include at least one of the following: The first prediction result and the second prediction result are weighted to obtain hydrogen sulfide Target prediction results of concentration; The second prediction result is input into the high-gravity desulfurization neural network model, and the hydrogen sulfide output by the high-gravity desulfurization neural network model is The first prediction result of the concentration is optimized to obtain hydrogen sulfide Target prediction results of concentration.
[0016] The present invention also provides a hydrogen sulfide Concentration prediction, including the following modules: The first prediction module is used to input the composition of the desulfurization agent and the desulfurization operating conditions into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of the concentration; the high gravity desulfurization neural network model is used to predict the concentration at the outlet of the high gravity reactor concentration; The second prediction module is used to obtain the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system. The second prediction result of the concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and the chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; A target prediction module is used to obtain hydrogen sulfide according to the first prediction result and the second prediction result. Target prediction results of concentration.
[0017] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the hydrogen sulfide Concentration prediction method.
[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned hydrogen sulfide Concentration prediction method.
[0019] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the hydrogen sulfide Concentration prediction method.
[0020] Hydrogen sulfide provided by the present invention The concentration prediction method and device can effectively fit the complex data rules in the desulfurization process through the high-gravity desulfurization neural network model, and can accurately follow the physical laws through the high-gravity desulfurization mechanism model, that is, it combines the data fitting ability of the black box model and the physical law constraints of the white box model, thereby greatly improving the concentration at the reactor outlet. The prediction accuracy of concentration. In addition, the high-gravity desulfurization mechanism model in the embodiment of the present application is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model, that is, the high-gravity desulfurization mechanism model is established according to the physical and chemical mechanisms of the high-gravity desulfurization system; wherein, the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor, thereby accurately describing the mass transfer process, chemical reaction and mass relationship before and after desulfurization in the desulfurization process, and then based on the established desulfurization mass transfer reaction model, it is also possible to accurately predict the mass at the outlet. concentration. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0022] Figure 1 The hydrogen sulfide provided by the present invention One of the flow charts of the concentration prediction method.
[0023] Figure 2 The hydrogen sulfide provided by the present invention Schematic diagram of the concentration prediction method (part 2).
[0024] Figure 3 The hydrogen sulfide provided by the present invention Schematic diagram of the structure of the concentration prediction device.
[0025] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0026] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0027] In order to facilitate a clearer understanding of the technical solutions of the embodiments of the present application, some technical contents related to the embodiments of the present application are first introduced.
[0028] (1) Basic mass transfer process
[0029] In the ultra-gravity desulfurization system, the mass transfer process between the gas and liquid phases is the key factor affecting the desulfurization efficiency. The mass transfer enters the liquid phase through the gas-liquid interface and reacts with the alkaline solution in the liquid phase. The mass transfer process can be described by the double membrane theory. The mass transfer rates of the gas phase and the liquid phase are: = ( )
[0030] = ( )
[0031] and are the mass transfer rates of the gas phase and the liquid phase, respectively. and are the mass transfer coefficients of the gas phase and the liquid phase, respectively. and are the concentrations of the gas phase and the liquid phase, respectively. and are the concentrations at the gas-liquid interface, respectively.
[0032] Assuming that the mass transfer between the gas and liquid phases reaches equilibrium, we have: ( )= ( )
[0033] (2) Basic reaction process
[0034] In the liquid phase, It reacts chemically with alkaline solution to generate sulfide and other products. The reaction steps are as follows: Desulfurization stage: Na2CO3 + H2S → NaHS + NaHCO3
[0035] 2NaHS + 4NaVO3+ H2O → Na2V4O9+ 4NaOH + 2S
[0036] 2TQ (quinone state) + Na2V4O9 + 2NaOH + H20 → 2THQ (phenol state) + 4NaVO3
[0037] Regeneration phase: THQ(phenolic state) + O2→TQ(quinone state) + H2O2
[0038] (3) Study of reaction rate
[0039] In order to better understand the reaction process, it is necessary to study the reaction rate. The reaction rate describes the speed at which reactants are converted into products and is one of the important characteristics of chemical reactions. Specifically, the reaction rate can be determined by fitting the reaction kinetics model with experimental data. The following are several common reaction rate models: Zero-order reaction: The reaction rate is independent of the concentration of the reactants and is expressed as: r = K0, where K0 is the zero-order reaction rate constant.
[0040] First-order reaction: The reaction rate is proportional to the concentration of the reactants, expressed as: r= K1C, where K1 is the first-order reaction rate constant and C is the concentration of the reactants.
[0041] Second order reaction: The reaction rate is proportional to the square of the reactant concentration, expressed as: r = K2C 2 , where K2 is the second-order reaction rate constant. In the present embodiment, it is assumed that The reaction with alkaline solution is a pseudo-first-order reaction, and the reaction rate is expressed as: r = K H2S C H2S。
[0042] Combine the following Figure 1-Figure 4 Description of the invention Hydrogen sulfide Concentration prediction method and device.
[0043] Figure 1 The hydrogen sulfide provided by the present invention One of the flow charts of the concentration prediction method is shown in Figure 1. Figure 1 As shown, the method includes the following: Step 101: Input the composition of the desulfurizing agent and the desulfurizing operating conditions into the high-gravity desulfurization neural network model in the gray box model of the high-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of concentration; The high-gravity desulfurization neural network model is used to predict the concentration at the outlet of the high-gravity reactor. concentration.
[0044] Specifically, in order to accurately predict the outlet of the reactor In the embodiment of the present application, the composition of the desulfurizing agent and the desulfurization operation conditions are first input into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of concentration; wherein, the supergravity desulfurization neural network model is a deep neural network with feature extraction and multiple hidden layers containing attention mechanism, which is used to process and learn data collected from the desulfurization system. Optionally, an attention mechanism layer is integrated in the network to give the model a global view and focus on the impact The key factors of concentration change, such as the composition of the desulfurizer, operating conditions, etc. Optionally, the model can be based on LSTM, CNN, and ATTENTION modules to enhance the model's ability to identify important features in time series data; and through the mechanism of attention, it can learn which factors are important under different operating conditions. The concentration has the greatest impact, and the learned high-gravity desulfurization neural network model can accurately predict the concentration at the outlet of the high-gravity reactor. Optionally, the operating conditions of the desulfurization process include the concentration of the desulfurization agent, the contact time and the temperature.
[0045] Optionally, common desulfurization methods include adsorption, absorption, combustion and biological desulfurization. In this application, the desulfurization agent is selected to remove There are different reagents for desulfurization, such as tannin extract, complex iron and persulfate method (PDS). In this application, tannin extract is selected as the desulfurization agent for removal. Optionally, desulfurization and learning and training of the high-gravity desulfurization neural network model can also be performed based on other methods and other desulfurizers, which is not limited in the embodiments of the present application. It should be noted that the high-gravity desulfurization neural network model focuses on capturing and learning complex patterns and dynamic changes that are difficult to directly extract from traditional mechanism models, and pays more attention to global features, so that it can accurately predict the high-gravity reactor outlet. For example, the method of selecting a desulfurization agent is to In the removal process of sulfur dioxide, we can focus on the tannin extract method (TQ) and the complex iron method. The following are the specific implementation details of these two methods: (1) Tannin extract method (TQ): Prepare natural plant materials such as oak fruit, sumac leaves, quench wood and castanopsis bark. Quench these plant materials with water to prepare a water quenching liquid for desulfurization. Set experimental conditions, such as the concentration of the desulfurizer, contact time and temperature, to simulate the actual industrial desulfurization process. Collect information about Data on removal efficiency and reaction rate are collected to analyze the desulfurization effect. (2) Complex iron method: Select a suitable iron-based desulfurizer, such as the iron salt used in the LO-CAT process or the iron ion solution used in the GLT process. Determine the design and operating parameters of the reactor, such as the concentration of iron ions, reaction temperature and pH value, conduct experiments to simulate the process of iron ions forming complexes with sulfides, and monitor The removal effect of The removal efficiency and reaction rate data are used to analyze the desulfurization effect. (3) PDS: After PDS is added to the desulfurization solution, it will decompose to generate free radicals such as hydrogen peroxide (HO2·) and hydroxyl (HO·). These free radicals react with the hydroxide (OH - ) and hydrogen sulfide (HS - ) plasma to react, thereby removing hydrogen sulfide ( ) purpose. Comparing the three systems, the cost of tannin extract and its The removal rate is high, so tannin extract can be selected as the desulfurizing agent, and the data of tannin extract are collected for building the model.
[0046] Step 102: According to the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system, obtain The second prediction result of concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor.
[0047] Specifically, in the embodiment of the present application, not only the high-gravity desulfurization neural network model predicts the high-gravity reactor outlet concentration, and also predicted the concentration at the outlet of the high-gravity reactor based on the high-gravity desulfurization mechanism model. concentration, so that on the one hand, the high-gravity desulfurization neural network model can effectively fit the complex data rules in the desulfurization process, and on the other hand, the high-gravity desulfurization mechanism model can accurately follow the physical laws and chemical reactions in the desulfurization process, that is, it combines the data fitting ability of the black box model with the physical laws and chemical reaction constraints of the white box model, which greatly improves the reactor outlet. The prediction accuracy of the concentration.
[0048] Optionally, in the prior art, common reactor modeling is based on ideal conditions and is often based on material conservation equations. Such import and export constant calculations have poor accuracy, which leads to a large deviation between the predicted gas concentration at the outlet and the actual measurement results. In order to solve the above problems, the ultra-gravity desulfurization mechanism model in the embodiment of the present application is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model, that is, the ultra-gravity desulfurization mechanism model is established according to the physical and chemical mechanisms of the ultra-gravity desulfurization system; wherein, the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the ultra-gravity reactor and the mass at the outlet of the ultra-gravity reactor, thereby accurately describing the mass transfer process, chemical reaction and mass relationship before and after desulfurization in the desulfurization process, and then based on the established desulfurization mass transfer reaction model, the outlet can also be accurately predicted. concentration.
[0049] Step 103: Obtain hydrogen sulfide according to the first prediction result and the second prediction result. Target prediction results of concentration.
[0050] Specifically, the high-gravity desulfurization neural network model obtains the high-gravity reactor outlet The first prediction result of the concentration is obtained based on the high-gravity desulfurization mechanism model at the outlet of the high-gravity reactor. After obtaining the second prediction result of the concentration, the prediction results of the mechanism model and the attention mechanism neural network can be effectively integrated in the embodiment of the present application, so as to not only understand and predict the system behavior based on physical laws, but also adapt to and learn those complex phenomena and patterns that can only be revealed from data. This dual learning mechanism ensures that the model can self-adjust and optimize according to actual operation data while maintaining physical laws, greatly improving the accuracy of the model. This method not only overcomes the limitations of the model, but also improves the prediction accuracy of concentration changes in the high-gravity desulfurization process. Concentrations show higher accuracy and reliability. In general, this application provides a gray box model that can accurately predict the performance of an ultra-gravity desulfurization system. This model is particularly suitable for dealing with systems with complex nonlinear characteristics and high-dimensional dynamics, and can ensure efficient and stable operation throughout the desulfurization process. By implementing the technical solution in this application, the operating efficiency of the chemical process can be significantly improved, providing an innovative and effective solution for industrial applications that meets the requirements for high precision and reliability.
[0051] The method of the above embodiment, on the one hand, can effectively fit the complex data rules in the desulfurization process through the high-gravity desulfurization neural network model, and on the other hand, can accurately follow the physical laws through the high-gravity desulfurization mechanism model, that is, it combines the data fitting ability of the black box model and the physical law constraints of the white box model, thereby greatly improving the reactor outlet. The prediction accuracy of concentration. In addition, the high-gravity desulfurization mechanism model in the embodiment of the present application is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model, that is, the high-gravity desulfurization mechanism model is established according to the physical and chemical mechanisms of the high-gravity desulfurization system; wherein, the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor, thereby accurately describing the mass transfer process, chemical reaction and mass relationship before and after desulfurization in the desulfurization process, and then based on the established desulfurization mass transfer reaction model, it is also possible to accurately predict the mass at the outlet. concentration.
[0052] In one embodiment, the high gravity desulfurization mechanism model is established based on the following method: According to the Hebe-Bess penetration theory, a desulfurization mass transfer reaction model in spherical coordinate system is established; According to the desulfurization mass transfer reaction model, the liquid mass transfer coefficient is determined; the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor; According to the liquid mass transfer coefficient and the desulfurization mass balance model, a high-gravity desulfurization mechanism model is established.
[0053] Specifically, the high-gravity desulfurization mechanism model in the embodiment of the present application is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model, that is, the high-gravity desulfurization mechanism model is established according to the physical and chemical mechanisms of the high-gravity desulfurization system. Optionally, in the embodiment of the present application, firstly, according to the Hebe penetration theory, a desulfurization mass transfer reaction model in a spherical coordinate system is established; wherein, in the high-gravity desulfurization system, the mass transfer process between the gas and liquid phases is a key factor affecting the desulfurization efficiency. In the embodiment of the present application, a desulfurization mass transfer reaction model is accurately established based on the Hebe penetration theory, that is, it is assumed that the mass transfer process of the solute in the liquid phase is driven by the instantaneous renewal event caused by the turbulence at the gas-liquid interface, and the residence time of each liquid element at the interface is limited, and mass transfer occurs during this period of time, thereby accurately describing the mass transfer and chemical reaction process in the desulfurization process.
[0054] Optionally, in the embodiment of the present application, a desulfurization mass transfer reaction model is accurately established based on the Schiff penetration theory, and the main assumptions include: (1) Turbulent renewal of the liquid phase: Solute mass transfer in the liquid phase is driven by instantaneous renewal events caused by turbulence at the gas-liquid interface. Each liquid element stays at the interface for a limited time before being replaced by a new liquid element. Modeled as a droplet, it can be accurately described mass transfer reaction process.
[0055] (2) Residence time of liquid element: The residence time of liquid element at the gas-liquid interface is constant and can be expressed as an average residence time.
[0056] (3) Diffusion control: The mass transfer process is mainly controlled by the diffusion of solute in the liquid phase, that is, it is assumed that the mass transfer resistance in the gas phase can be ignored.
[0057] (3) Steady-state mass transfer: During the residence time of the liquid element, the mass transfer process reaches a quasi-steady state and the mass transfer rate is constant.
[0058] (4) Constant interfacial concentration: During the mass transfer process, the solute concentration at the gas-liquid interface remains constant.
[0059] In addition, it should be noted that the desulfurization mass transfer reaction model in the spherical coordinate system is established according to the Hebbe penetration theory in the embodiment of the present application, which can more accurately describe the desulfurization process, and thus can more accurately predict the reactor outlet. Alternatively, the reasons and functions for establishing the desulfurization mass transfer reaction model in the spherical coordinate system are as follows: (1) Conform to the actual droplet shape: In the embodiment of the present application, during the desulfurization process They often exist in the form of droplets. Spherical coordinates can more accurately describe the geometry and mass transfer behavior of these droplets.
[0060] (2) Simplified mathematical processing: For spherical droplets, spherical coordinates can simplify the processing of mathematical equations. In particular, when describing the mass transfer and reaction process inside the droplet, spherical coordinates can more naturally reflect the concentration gradient and mass transfer path.
[0061] (3) Applicable to multiphase flow systems: In multiphase flow systems, such as the flow in a high-gravity reactor, the motion and renewal process of droplets often match spherical coordinates, which makes spherical coordinates more advantageous in describing these systems.
[0062] Optionally, after establishing the desulfurization mass transfer reaction model in the spherical coordinate system according to the Hebbit penetration theory, the liquid mass transfer coefficient is further determined according to the desulfurization mass transfer reaction model in the embodiment of the present application; wherein the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor, and thus also to predict the liquid mass transfer coefficient at the reactor outlet. In the process of concentration, not only the mass transfer process and chemical reaction process in the desulfurization process are accurately modeled, but also the material transfer rate and reactor performance are considered, which makes the final constructed ultra-gravity desulfurization mechanism model more accurate, and can more accurately predict the reactor outlet concentration.
[0063] Furthermore, after establishing the desulfurization mass transfer reaction model in the spherical coordinate system and determining the liquid mass transfer coefficient, the present application further establishes a supergravity desulfurization mechanism model based on the liquid mass transfer coefficient and the desulfurization mass balance model, thereby accurately describing the mass transfer process, chemical reaction and mass relationship before and after desulfurization in the desulfurization process. Based on the established desulfurization mass transfer reaction model, the outlet temperature can be accurately predicted. concentration, effectively increasing the reactor outlet Prediction accuracy of concentration.
[0064] The method of the above embodiment accurately establishes a desulfurization mass transfer reaction model in a spherical coordinate system based on the Hebbit penetration theory, which accurately describes the mass transfer and chemical reaction process in the desulfurization process; and further determines the liquid mass transfer coefficient based on the desulfurization mass transfer reaction model, and establishes a high-gravity desulfurization mechanism model based on the liquid mass transfer coefficient and the desulfurization mass balance model, that is, further considers the material transfer rate, the performance of the reactor, and the mass relationship before and after desulfurization, so that the high-gravity desulfurization mechanism model finally constructed is more accurate, and the reactor outlet can be more accurately predicted. concentration, effectively increasing the reactor outlet Prediction accuracy of concentration.
[0065] In one embodiment, the desulfurization mass transfer reaction model in a spherical coordinate system includes: = ( )-
[0066] in, represents the solute concentration; D represents the diffusion coefficient of the solute in the liquid phase; r represents the penetration radius; t represents the time; Represents the reaction rate.
[0067] Specifically, in the related art, the mass transfer reaction equation not based on the Schiff-Behringer penetration theory is as follows: = ( )-
[0068] Where: C represents concentration; D represents the diffusion coefficient of the solute in the liquid phase; x represents the spatial coordinate; KC represents the reaction rate.
[0069] In the embodiment of the present application, a desulfurization mass transfer reaction model in a spherical coordinate system is accurately established based on the Higbie penetration theory, which more accurately describes the mass transfer and chemical reaction process in the desulfurization process. Optionally, the mass transfer-reaction model based on the Higbie penetration theory in a spherical coordinate system is expressed as: = ( )-
[0070] in, represents the solute concentration; D represents the diffusion coefficient of the solute in the liquid phase; r represents the penetration radius; t represents the time; Represents the reaction rate.
[0071] The method of the above embodiment accurately establishes a desulfurization mass transfer reaction model in a spherical coordinate system based on the Hebbit penetration theory, accurately describes the mass transfer and chemical reaction process in the desulfurization process, and thus can more accurately predict the reactor outlet. concentration.
[0072] In one embodiment, the liquid mass transfer coefficient comprises:
[0073] in, represents the liquid mass transfer coefficient; It represents the mass transfer rate per unit interfacial area; It means that the desulfurization reaction reaches equilibrium concentration; express initial concentration.
[0074] Specifically, after establishing the desulfurization mass transfer reaction model in the spherical coordinate system according to the Hebbit penetration theory, the liquid mass transfer coefficient is further determined according to the desulfurization mass transfer reaction model in the embodiment of the present application; wherein the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor, and thus also to predict the reactor outlet. In the process of concentration, not only the mass transfer process and chemical reaction process in the desulfurization process are accurately modeled, but also the material transfer rate and reactor performance are considered, which makes the final constructed ultra-gravity desulfurization mechanism model more accurate, and can more accurately predict the reactor outlet concentration.
[0075] Optionally, the liquid mass transfer coefficient comprises:
[0076] in, represents the liquid mass transfer coefficient; It represents the mass transfer rate per unit interfacial area; It means that the desulfurization reaction reaches equilibrium concentration; express initial concentration.
[0077] For example, in the embodiment of the present application, according to the desulfurization mass transfer reaction model, the specific process of determining the liquid mass transfer coefficient is as follows: First determine the liquid holding volume And the average diameter d of the droplets:
[0078] Where: N represents the rotation speed, r represents the penetration radius; represents the acceleration due to gravity, u represents the liquid flow rate, represents the characteristic flow velocity, v represents the kinematic viscosity of the liquid, Represents characteristic kinematic viscosity.
[0079] d=12.84
[0080] Where: σ represents the surface tension of the liquid, and ρ represents the density of the liquid.
[0081] Then, the mass transfer interfacial area a is calculated and quantified as: a=6
[0082] Then calculate the droplet renewal time t, droplet renewal frequency s and mass transfer rate per unit interface area : The update time of the droplet can be obtained by the following formula: =
[0083] in: and Respectively represent the inner diameter and outer diameter of the reactor, L represents the liquid flow rate, represents specific fluid dynamics parameters, N represents the rotation speed, and R represents the radius of the reactor.
[0084] Optionally, the update frequency s of the droplet: S=
[0085] Alternatively, the mass transfer rate per unit interfacial area : = (r= ) dt
[0086] Where D represents The diffusion coefficient, C, represents concentration.
[0087] Finally, the liquid mass transfer coefficient is calculated by:
[0088] in, represents the liquid mass transfer coefficient; It represents the mass transfer rate per unit interfacial area; It means that the desulfurization reaction reaches equilibrium concentration; express That is, by calculating the liquid holding capacity, the average diameter of the droplets, the mass transfer interface area, the droplet renewal time and frequency, and the unit interface area The absorption rate and liquid phase mass transfer coefficient are used to quantify the mass transfer process and calculate the mass transfer coefficient.
[0089] The method of the above embodiment determines the liquid mass transfer coefficient according to the desulfurization mass transfer reaction model; wherein the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor, and thus to predict the liquid mass transfer coefficient at the reactor outlet. In the process of concentration, not only the mass transfer process and chemical reaction process in the desulfurization process are accurately modeled, but also the material transfer rate and reactor performance are considered, which makes the final constructed ultra-gravity desulfurization mechanism model more accurate, and can more accurately predict the reactor outlet concentration.
[0090] In one embodiment, the high gravity desulfurization mechanism model includes:
[0091] in, The outlet of the reactor concentration; Indicates the gas-liquid interface The mole fraction of represents the liquid mass transfer coefficient; T represents the temperature; express Henry's constant; express The partial pressure of represents the mass transfer interface area; h represents the thickness of the packing layer; Represents the outer diameter of the reactor; represents the inner diameter of the reactor; G represents of traffic.
[0092] Specifically, after determining the liquid mass transfer coefficient, the concentration can be determined by combining the mass balance equation, that is, the concentration at the reactor outlet. The specific steps are as follows: Perform mass balance calculations, gas phase The equation at the gas-liquid interface is:
[0093] Where: y represents the gas phase The mole fraction of represents the mass transfer interface area, h represents the thickness of the packing layer, R represents the geometric mean radius, which is the average of the inner and outer diameters of the reactor, and G represents of traffic, It represents the gas-liquid mass transfer coefficient, and the calculation formula is:
[0094] in, express Henry's constant, express The partial pressure.
[0095] Further according to the above gas phase The equation at the gas-liquid interface can be used to calculate the reactor outlet concentration .
[0096]
[0097] The method of the above embodiment, in predicting the reactor outlet In the process of concentration, not only the mass transfer process and chemical reaction process in the desulfurization process are accurately modeled, but also the material transfer rate, reactor performance and the mass relationship before and after desulfurization are further considered, which makes the final constructed ultra-gravity desulfurization mechanism model more accurate, and can more accurately predict the reactor outlet concentration, effectively increasing the reactor outlet Prediction accuracy of concentration.
[0098] In one embodiment, the high-gravity desulfurization neural network model is trained based on the following method: Acquire training samples; the training samples include desulfurization agent samples and desulfurization operation condition samples; The desulfurization agent samples and desulfurization operation condition samples were input into the high-gravity desulfurization neural network model to obtain the high-gravity reactor outlet Concentration prediction results; According to the outlet of the supergravity reactor The concentration prediction results and the measured concentration at the outlet of the high gravity reactor concentration, train the high-gravity desulfurization neural network model, and obtain the trained high-gravity desulfurization neural network model.
[0099] Specifically, in the embodiment of the present application, the high-gravity desulfurization neural network model is a deep neural network including feature extraction of attention mechanism and multiple hidden layers, which is used to process and learn data collected from the desulfurization system. Optionally, the high-gravity desulfurization neural network model focuses on capturing and learning complex patterns and dynamic changes that are difficult to directly extract from traditional mechanism models, and pays more attention to global features, so that it can accurately predict the high-gravity reactor outlet. Optionally, integrate an attention layer into the network to give the model a global view and focus on the factors that affect The key factors of concentration change, such as the composition of the desulfurizer, operating conditions, etc. Optionally, the model can be based on LSTM, CNN, and ATTENTION modules to enhance the model's ability to identify important features in time series data; and through the mechanism of attention, the model can learn which factors are important under different operating conditions. The concentration has the greatest impact, and the learned high-gravity desulfurization neural network model can accurately predict the concentration at the outlet of the high-gravity reactor. concentration.
[0100] Exemplarily, in the embodiment of the present application, the high gravity desulfurization neural network model is trained based on the following method, and the specific process is as follows: (1) Data collection and preprocessing.
[0101] Collect historical operation data, including operating parameters and results. Record the type and concentration of desulfurizer used. Record key control parameters in the process, such as gravity level, pH value, liquid flow, gas flow, and operation time.
[0102] Discretization: Convert continuous running data into discrete data points for model processing. This includes converting continuously recorded parameter values into a series of time series data points. Normalization: Normalize the collected data to eliminate the impact of different dimensions and ranges so that the data is on a unified scale. This step helps improve the efficiency of model training and the accuracy of prediction.
[0103]
[0104] u is the mean of the sample data in column X, is the standard deviation of the X column data. is the result after the new data is normalized.
[0105] That is, in the embodiment of the present application, the historical operation data of the high gravity selective desulfurization system is first collected to record the composition of the desulfurizer: the concentration of tannin persulfate method; and the operating conditions, such as: gravity level, inlet Flow, temperature, etc. These data are discretized and normalized to provide standardized data sets for deep learning models.
[0106] (2) Neural network model based on attention mechanism
[0107] First, a deep neural network with multiple hidden layers based on the Transformer model is designed to process and learn the data collected from the desulfurization system. An attention mechanism layer is integrated into the network to enable the model to focus on the impact Key factors affecting concentration changes, such as the composition of the desulfurizer, operating conditions, etc.
[0108] Integrate attention mechanisms in the network to learn complex patterns and dynamic changes. The specific steps are as follows: Input layer: including control conditions such as gas flow, pH value, concentration, and rotation speed.
[0109] Attention mechanism layer: Use self-attention mechanism to capture the relationship and importance between different features.
[0110] Hidden layer: includes multiple hidden layers, which are used to process and extract higher-level features.
[0111] Output layer: predict the exit concentration.
[0112] (3) Implementation of the attention mechanism: In order to enhance the global features between the features extracted by the model, an attention mechanism suitable for time series data is implemented: self-attention. Through this mechanism, the model can learn which factors are important under different operating conditions. Concentration has the greatest impact.
[0113] Input vector: The data (such as concentration, operating conditions, etc.) is set as a set of input vectors X = {x1, x2, ..., xn}, each input vector xi has d dimensions.
[0114] Linear transformation: Perform a linear transformation on the input vector to obtain query, key, and value vectors. These vectors are transformed through different weight matrices. , , The calculation results are: Q=X , K = X , V = X
[0115] Calculate the attention score: For each feature element of the input, its relative importance to other elements in the sequence is calculated, so an attention score is needed to compare the importance. The attention score is calculated by taking the dot product of the query vector and the key vector, and then scaling the dot product: Attention score =
[0116] The denominator is a scaling factor, which is used to prevent the gradient from disappearing due to the dot product value being too large.
[0117] Apply the Softmax function: The attention scores are normalized by the Softmax function to obtain the weight distribution: =softmax( )=
[0118] in, express Input Vector The attention weights for the input vector.
[0119] The attention scores calculated using the softmax function are weighted and summed over the value vectors to obtain weighted value vectors that capture important information about different parts of the input sequence.
[0120] Weighted sum value vector: Use the calculated attention weights to perform weighted summation on the value vector to get the final attention output vector: Attention(Q,K,V)=
[0121] in, is the jth vector in the value vector V. Summarizing the above formulas, we get: Attention(Q,K,V)=softmax( )V
[0122] The output of the self-attention mechanism is used as a feature for subsequent networks, such as predicting In this way, the model can associate other operating conditions according to the changes in operating conditions, adjust their weight matrices, and pay global attention to Factors affecting concentration.
[0123] (4) Network training and optimization: Use the collected data set to train the neural network. Choose a suitable loss function. Loss function: Introduce a custom loss function (custom_loss) based on the mean square error (MSE), add a penalty term when the output exceeds the predetermined boundary, and enhance the accuracy and reliability of the model prediction. Apply optimization algorithms such as stochastic gradient descent (SGD) and Adam to optimize the network weights and achieve the best prediction effect.
[0124] (5) Training platform: Model training on the CUDA platform significantly shortens the training time (from 35 minutes to 4 minutes) compared to CPU, improving training efficiency. 2000 iterations of training are performed to ensure that the model fully learns the dynamic relationship between data features and time series. After training, the model is saved to facilitate continuous operation and obtain accurate simulation data.
[0125] The method of the above embodiment inputs the desulfurization agent sample and the desulfurization operation condition sample into the high-gravity desulfurization neural network model to obtain the high-gravity reactor outlet Concentration prediction results; according to the The concentration prediction results and the measured concentration at the outlet of the high gravity reactor concentration, train the high-gravity desulfurization neural network model, obtain the trained high-gravity desulfurization neural network model, and then use the high-gravity desulfurization neural network model to accurately predict the concentration at the reactor outlet. concentration.
[0126] In one embodiment, hydrogen sulfide is obtained according to the first prediction result and the second prediction result. The target prediction results of concentration include at least one of the following: The first prediction result and the second prediction result are weighted to obtain hydrogen sulfide Target prediction results of concentration; The second prediction result is input into the high-gravity desulfurization neural network model, and the hydrogen sulfide output by the high-gravity desulfurization neural network model is The first prediction result of the concentration is optimized to obtain hydrogen sulfide Target prediction results of concentration.
[0127] Specifically, in the embodiment of the present application, on the one hand, the high-gravity desulfurization neural network model can effectively fit the complex data rules in the desulfurization process, and on the other hand, the high-gravity desulfurization mechanism model can accurately follow the physical laws, that is, it combines the data fitting ability of the black box model and the physical law constraints of the white box model, thereby greatly improving the prediction of the outlet Alternatively, the composition of the desulfurization agent and the desulfurization operating conditions are input into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of the concentration, and the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system, are obtained. After the second prediction result of the concentration is obtained, in the embodiment of the present application, the first prediction result and the second prediction result can be weighted to obtain the hydrogen sulfide The target prediction result of concentration is to assign different weights to the prediction results of the white box model and the black box model, so as to obtain the final Concentration prediction results. That is, by adjusting the weights, the contributions of the two models can be balanced to ensure that the final output reflects both the accuracy of physical and chemical laws and incorporates data-based deep learning insights. Optionally, the second prediction result, the composition of the desulfurizer, and the desulfurization operating conditions can be input into the ultra-gravity desulfurization neural network model to obtain hydrogen sulfide. The target prediction result of concentration is to use the prediction result of the white box as the training data of the black box, so that the black box can learn more content, thus making the final model more accurate. Concentration prediction effectively improves Accuracy of concentration prediction results.
[0128] For example, Figure 2 As shown, in the present application, the mechanism model and the neural network model are integrated in parallel or in series to form a final gray box model of the supergravity desulfurization system.
[0129] (1) Parallel integration: Specifically, in the process of model integration, different weights are assigned to the mechanism model and the neural network model respectively. The specific formula is as follows:
[0130] Here α is the weight factor. By adjusting the weight, the contribution of the two models can be balanced to ensure that the final output not only reflects the accuracy of physical and chemical laws, but also incorporates data-based deep learning insights.
[0131] (2) Series integration: During the model integration process, the mechanical model can be calculated first, and then the results output by the mechanical model can be As part of the black box model neural network. The specific steps are as follows: Compute mass transfer reaction mechanism model output .
[0132] The output of the mechanism model As one of the inputs to the neural network model.
[0133] The neural network model receives the output of the mass transfer reaction mechanism model and other input conditions (such as gas flow, acid and alkali pH value, concentration, rotation speed, etc.) through multi-head attention mechanism and hidden layer processing, and finally predicts the outlet concentration.
[0134] The method of the above embodiment can assign different weights to the mechanism model and the neural network model respectively. By adjusting the weights, the contributions of the two models can be balanced to ensure that the final output not only reflects the accuracy of physical and chemical laws, but also incorporates deep learning insights based on data, so that the final The concentration prediction results are more accurate. You can also output the results of the mechanism model As part of the black box model neural network, the black box can learn more content, so that the final model can be more accurate. Concentration prediction effectively improves Accuracy of concentration prediction results.
[0135] The hydrogen sulfide provided by the present invention is as follows The concentration prediction device is described below. The concentration prediction device is similar to the hydrogen sulfide described above. The concentration prediction methods can be referenced to each other. Concentration prediction device such as Figure 3 As shown, including: The first prediction module 310 is used to input the composition of the desulfurizing agent and the desulfurization operation conditions into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of concentration; The high-gravity desulfurization neural network model is used to predict the concentration at the outlet of the high-gravity reactor. concentration; The second prediction module 320 is used to obtain the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system. The second prediction result of concentration; The high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; The desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; The desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; The target prediction module 330 is used to obtain hydrogen sulfide according to the first prediction result and the second prediction result. Target prediction results of concentration.
[0136] Optionally, the high gravity desulfurization mechanism model is established based on the following method: According to the Hebe-Bess penetration theory, a desulfurization mass transfer reaction model in spherical coordinate system is established; According to the desulfurization mass transfer reaction model, the liquid mass transfer coefficient is determined; the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor; According to the liquid mass transfer coefficient and the desulfurization mass balance model, a high-gravity desulfurization mechanism model is established.
[0137] Optionally, the desulfurization mass transfer reaction model in a spherical coordinate system includes: = ( )-
[0138] in, represents the solute concentration; D represents the diffusion coefficient of the solute in the liquid phase; r represents the penetration radius; t represents the time; Represents the reaction rate.
[0139] Optionally, the liquid mass transfer coefficient comprises:
[0140] in, represents the liquid mass transfer coefficient; It represents the mass transfer rate per unit interfacial area; It means that the desulfurization reaction reaches equilibrium concentration; express initial concentration.
[0141] Optionally, a high gravity desulfurization mechanism model, including:
[0142] in, The outlet of the reactor concentration; Indicates the gas-liquid interface The mole fraction of represents the liquid mass transfer coefficient; T represents the temperature; express Henry's constant; express The partial pressure of represents the mass transfer interface area; h represents the thickness of the packing layer; Represents the outer diameter of the reactor; represents the inner diameter of the reactor; G represents of traffic.
[0143] Optionally, the high gravity desulfurization neural network model is trained based on the following method: Acquire training samples; the training samples include desulfurization agent samples and desulfurization operation condition samples; The desulfurization agent samples and desulfurization operation condition samples were input into the high-gravity desulfurization neural network model to obtain the high-gravity reactor outlet Concentration prediction results; According to the outlet of the supergravity reactor The concentration prediction results and the measured concentration at the outlet of the high gravity reactor concentration, train the high-gravity desulfurization neural network model, and obtain the trained high-gravity desulfurization neural network model.
[0144] Optionally, the target prediction module 330 is specifically used to perform weighted processing on the first prediction result and the second prediction result to obtain hydrogen sulfide Target prediction results of concentration; The second prediction result is input into the high-gravity desulfurization neural network model, and the hydrogen sulfide output by the high-gravity desulfurization neural network model is The first prediction result of the concentration is optimized to obtain hydrogen sulfide Target prediction results of concentration.
[0145] Figure 4The physical structure diagram of an electronic device is illustrated, and the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the hydrogen sulfide The concentration prediction method comprises: inputting the composition of the desulfurizing agent and the desulfurization operation conditions into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of concentration; The high-gravity desulfurization neural network model is used to predict the concentration at the outlet of the high-gravity reactor. concentration; According to the ultra-gravity desulfurization mechanism model in the gray box model of the ultra-gravity desulfurization system, we can get The second prediction result of the concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; according to the first prediction result and the second prediction result, the hydrogen sulfide Target prediction results of concentration.
[0146] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0147] On the other hand, the present invention also provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer readable storage medium, and when the computer program is executed by a processor, the computer can execute hydrogen sulfide The concentration prediction method comprises: inputting the composition of the desulfurizing agent and the desulfurization operation conditions into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of concentration; The high-gravity desulfurization neural network model is used to predict the concentration at the outlet of the high-gravity reactor. concentration; According to the ultra-gravity desulfurization mechanism model in the gray box model of the ultra-gravity desulfurization system, we can get The second prediction result of the concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; according to the first prediction result and the second prediction result, the hydrogen sulfide Target prediction results of concentration.
[0148] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a method for performing hydrogen sulfide treatment. The concentration prediction method comprises: inputting the composition of the desulfurizing agent and the desulfurization operation conditions into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of concentration; The high-gravity desulfurization neural network model is used to predict the concentration at the outlet of the high-gravity reactor. concentration; According to the ultra-gravity desulfurization mechanism model in the gray box model of the ultra-gravity desulfurization system, we can get The second prediction result of the concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; according to the first prediction result and the second prediction result, the hydrogen sulfide Target prediction results of concentration.
[0149] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0150] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. Hydrogen sulfide The concentration prediction method is characterized in that include: The composition of the desulfurizer and the desulfurization operating conditions are input into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of the concentration; the high gravity desulfurization neural network model is used to predict the concentration at the outlet of the high gravity reactor concentration; According to the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system, we can get The second prediction result of the concentration; the high gravity desulfurization mechanism model is used to predict the concentration of concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; According to the first prediction result and the second prediction result, hydrogen sulfide is obtained. Target prediction results of concentration.
2. Hydrogen sulfide according to claim 1 The concentration prediction method is characterized in that The high gravity desulfurization mechanism model is established based on the following method: According to the Hebe-Bess penetration theory, a desulfurization mass transfer reaction model in spherical coordinate system is established; According to the desulfurization mass transfer reaction model, a liquid mass transfer coefficient is determined; the liquid mass transfer coefficient is used to characterize the material transfer rate and the performance of the reactor; The high-gravity desulfurization mechanism model is established according to the liquid mass transfer coefficient and the desulfurization mass balance model.
3. Hydrogen sulfide according to claim 2 The concentration prediction method is characterized in that The desulfurization mass transfer reaction model in the spherical coordinate system includes: = ( )- in, represents the solute concentration; D represents the diffusion coefficient of the solute in the liquid phase; r represents the penetration radius; t represents the time; Represents the reaction rate.
4. Hydrogen sulfide according to claim 2 The concentration prediction method is characterized in that The liquid mass transfer coefficient includes: in, represents the liquid mass transfer coefficient; It represents the mass transfer rate per unit interfacial area; It means that the desulfurization reaction reaches equilibrium concentration; express initial concentration.
5. Hydrogen sulfide according to claim 2 The concentration prediction method is characterized in that The high gravity desulfurization mechanism model includes: in, The outlet of the reactor concentration; Indicates the gas-liquid interface The mole fraction of represents the liquid mass transfer coefficient; T represents the temperature; express Henry's constant; express The partial pressure of represents the mass transfer interface area; h represents the thickness of the packing layer; Represents the outer diameter of the reactor; represents the inner diameter of the reactor; G represents of traffic.
6. Hydrogen sulfide according to any one of claims 1 to 5 The concentration prediction method is characterized in that The high gravity desulfurization neural network model is trained based on the following method: Acquire training samples; the training samples include desulfurization agent samples and desulfurization operation condition samples; The desulfurization agent sample and the desulfurization operation condition sample are input into the high-gravity desulfurization neural network model to obtain the high-gravity reactor outlet Concentration prediction results; According to the outlet of the supergravity reactor The concentration prediction results and the measured concentration at the outlet of the high gravity reactor concentration, training the high-gravity desulfurization neural network model to obtain a trained high-gravity desulfurization neural network model.
7. Hydrogen sulfide according to any one of claims 1 to 5 The concentration prediction method is characterized in that The method of obtaining hydrogen sulfide according to the first prediction result and the second prediction result The target prediction results of concentration include at least one of the following: The first prediction result and the second prediction result are weighted to obtain hydrogen sulfide Target prediction results of concentration; The second prediction result is input into the high-gravity desulfurization neural network model, and the hydrogen sulfide output by the high-gravity desulfurization neural network model is The first prediction result of the concentration is optimized to obtain hydrogen sulfide Target prediction results of concentration.
8. Hydrogen sulfide The concentration prediction device is characterized in that include: The first prediction module is used to input the composition of the desulfurization agent and the desulfurization operating conditions into the ultra-gravity desulfurization neural network model in the gray box model of the ultra-gravity desulfurization system to obtain hydrogen sulfide The first prediction result of the concentration; the high gravity desulfurization neural network model is used to predict the concentration at the outlet of the high gravity reactor concentration; The second prediction module is used to obtain the high-gravity desulfurization mechanism model in the gray box model of the high-gravity desulfurization system. The second prediction result of the concentration; the high-gravity desulfurization mechanism model is determined based on the desulfurization mass transfer reaction model and the desulfurization mass balance model; the desulfurization mass transfer reaction model is used to characterize the mass transfer process and the chemical reaction process in the desulfurization process; the desulfurization mass balance model is used to characterize the balance relationship between the mass at the inlet of the high-gravity reactor and the mass at the outlet of the high-gravity reactor; A target prediction module is used to obtain hydrogen sulfide according to the first prediction result and the second prediction result. Target prediction results of concentration.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the hydrogen sulfide as described in any one of claims 1 to 7 is realized Steps of the concentration prediction method.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the hydrogen sulfide treatment process described in any one of claims 1 to 7 is realized. Steps of the concentration prediction method.