Intelligent control method and system for wet desulphurization of thermal power plant

Abstract

The invention discloses an intelligent control method and system for wet desulphurization of a thermal power plant, relates to the technical field of wet desulphurization of thermal power plants, and aims to improve the accuracy of limestone slurry supply. The method comprises the following steps: intercepting historical data from a real-time operation data sequence of the desulfurization system to construct a rolling estimation window; establishing a double-process decoupling mechanism model comprising a gas-liquid mass transfer fast dynamic sub-process and a limestone dissolution slow dynamic sub-process; calculating a slurry chemical buffer index in the rolling estimation window; assigning the dynamic lag time in the limestone dissolution slow dynamic sub-process based on a preset nonlinear mapping relation and the slurry chemical buffer index to obtain corrected lag time; in the rolling estimation window, the corrected lag time is used as a known parameter to be substituted into the double-process decoupling mechanism model, and an optimized gas-liquid mass transfer coefficient and an optimized limestone activity coefficient are solved; updating the dual-process decoupling mechanism model according to the optimization coefficient and the correction delay time; a future limestone slurry supply flow command is calculated based on the model.

Description

Technical Field

[0001] This application relates to the field of wet desulfurization technology in thermal power plants, and in particular to an intelligent control method and system for wet desulfurization in thermal power plants. Background Technology

[0002] With increasingly stringent environmental regulations, coal-fired power plants widely adopt limestone-gypsum wet flue gas desulfurization (FGD) technology to reduce sulfur dioxide emissions in flue gas to meet emission standards. In the actual operation of the FGD system, precise control of the limestone slurry flow rate into the absorption tower is necessary to balance environmental compliance with economic efficiency. The ideal control objective is to maintain the pH value of the slurry and the outlet sulfur dioxide concentration within the absorption tower within the optimal range while adapting to fluctuations in coal quality and changes in unit load, thereby reducing limestone waste and equipment power consumption.

[0003] To address the multivariate coupling and large inertial lag issues involved in desulfurization, the control systems of related technologies typically employ a multi-scale dynamic optimization scheme based on model predictive control. The specific implementation of this scheme involves: considering the rapid gas-phase transport and slow liquid-phase reaction characteristics in the desulfurization reaction, the control system establishes a fast dynamic sub-model describing the absorption of gaseous sulfur dioxide, and a slow dynamic sub-model based on the fluid dynamics mechanism of the slurry tank (often referred to as a physical mixing model). For the slow dynamic process, based on physical parameters such as slurry tank volume, fluid velocity, and agitator mixing power, a linearly changing physical lag time constant is set in the aforementioned physical mixing model, and a Smith predictor or time-delay compensation module is introduced into the controller. During operation, the controller predicts the response trajectory of the slurry pH value to the slurry supply flow rate based on this physical mixing model, and uses the lag compensation module to pre-correct control commands to overcome the response delay caused by the large volume, thereby achieving advance adjustment of the slurry supply.

[0004] However, desulfurization slurry is a complex carbonate-sulfite chemical buffer system. Its response resistance (i.e., buffering capacity) to changes in pH does not remain linear as described by the physical mixing model of related technologies. Instead, it exhibits drastic nonlinear fluctuations with different current pH values ​​(for example, there is a strong buffer plateau region between pH 5.0 and 6.0). When the slurry is in the strong buffer plateau region, the pH value's response to the slurry supply becomes sluggish, exhibiting a chemical pseudo-hysteresis. At this time, the linear compensation model of related technologies will identify this chemical buffering phenomenon as insufficient slurry supply or incomplete physical mixing, thus continuously accumulating and outputting excessive slurry supply commands, resulting in inaccurate slurry supply commands for limestone slurry. Summary of the Invention

[0005] This application provides an intelligent control method and system for wet desulfurization in thermal power plants to improve the accuracy of limestone slurry supply.

[0006] Firstly, an intelligent control method for wet desulfurization in thermal power plants is provided, applied to the control system. This method includes: extracting historical data from the real-time operating data sequence of the desulfurization system to construct a rolling estimation window; establishing a dual-process decoupling mechanism model comprising a fast dynamic subprocess of gas-liquid mass transfer and a slow dynamic subprocess of limestone dissolution, wherein the fast dynamic subprocess of gas-liquid mass transfer includes a gas-liquid mass transfer coefficient characterizing the gas-liquid contact mass transfer efficiency, and the slow dynamic subprocess of limestone dissolution includes a limestone activity coefficient characterizing the reactivity of the limestone raw material and a variable dynamic lag time; and calculating the slurry chemical buffer index within the rolling estimation window based on the nonlinear sensitivity characteristics of the slurry pH to reactant concentration, wherein the slurry chemical buffer index characterizes the slurry's resistance to... The magnitude of the inertia of pH change; the dynamic lag time in the slow dynamic subprocess of limestone dissolution is assigned based on a preset nonlinear mapping relationship and slurry chemical buffer index to obtain the corrected lag time; within the rolling estimation window, the corrected lag time is substituted as a known parameter into the dual-process decoupling mechanism model, and the gas-liquid mass transfer coefficient and limestone activity coefficient are used as optimization variables to solve for the optimized gas-liquid mass transfer coefficient and optimized limestone activity coefficient that minimize the deviation between the output trajectory of the dual-process decoupling mechanism model and the historical data trajectory; the dual-process decoupling mechanism model is updated according to the optimized gas-liquid mass transfer coefficient, optimized limestone activity coefficient and corrected lag time; the future limestone slurry supply flow command is calculated based on the updated dual-process decoupling mechanism model.

[0007] By adopting the above technical solution, when the slurry is in the strong buffer plateau zone, the control system recognizes that the sluggish pH response is due to chemical inertia rather than insufficient slurry supply. By increasing the model lag time, the control system informs the controller to wait for the reaction results, thereby avoiding the controller from issuing excessive slurry supply commands due to misjudgment, and thus improving the accuracy of limestone slurry supply.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the step of calculating the slurry chemical buffer index within a rolling estimation window based on the nonlinear sensitivity of slurry pH to reactant concentration specifically includes: constructing an acid-base buffer characteristic curve function based on the ionization equilibrium relationship of the carbonate and sulfite system, showing the change of slurry pH with the amount of added acid and base reagents; substituting the real-time slurry pH within the rolling estimation window into the acid-base buffer characteristic curve function, and calculating the reciprocal of the slope of the tangent line of the acid-base buffer characteristic curve function at each pH point; and normalizing the reciprocal to obtain the slurry chemical buffer index.

[0009] By adopting the above technical solution, the influence of industrial field measurement noise is reduced by using the acid-base buffer characteristic curve function based on the ionization equilibrium relationship of the carbonate and sulfite system, rather than differentiating the high-noise real-time pH data, thereby improving the stability of the buffer index under complex working conditions.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, the step of constructing an acid-base buffering characteristic curve function of the slurry pH value changing with the amount of added acid or base reagents based on the ionization equilibrium relationship of the carbonate and sulfite system specifically includes: constructing an acid-base buffering characteristic curve function based on the buffer capacity definition formula, wherein the acid-base buffering characteristic curve function is configured to characterize the ratio of the differential change in the amount of added acid or base reagents to the differential change in pH that causes the change; the acid-base buffering characteristic curve function includes a mathematical expression with hydrogen ion concentration as the independent variable, and the mathematical expression includes a bisulfite ionization equilibrium constant term and a bicarbonate ionization equilibrium constant term as denominator parameters; when the hydrogen ion concentration corresponding to the slurry pH value makes the denominator parameters approach a minimum value, the slurry chemical buffering index is in a preset peak range, characterizing that the slurry is in a strong buffering state.

[0011] By adopting the above technical solution, when the hydrogen ion concentration causes the denominator parameter to approach its minimum value, the control system can more sensitively detect whether the slurry has entered the sluggish reaction zone, thus improving the sensitivity of identifying strong buffer conditions.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the step of assigning a value to the dynamic lag time in the slow dynamic subprocess of limestone dissolution based on a preset nonlinear mapping relationship and a slurry chemical buffer index to obtain a corrected lag time specifically includes: using the ratio of the liquid level volume of the slurry pool in the absorption tower to the flow rate of the slurry circulation pump as the physical mixing lag reference value; substituting the slurry chemical buffer index as an input variable into the preset nonlinear mapping relationship to calculate the chemical damping coefficient; and determining the product of the physical mixing lag reference value and the chemical damping coefficient as the corrected lag time.

[0013] By adopting the above technical solution, the physical mixing lag reference value affected by fluid dynamics is calculated separately from the chemical damping coefficient affected by chemical equilibrium, and then the product coupling is performed. This allows the corrected lag time to simultaneously adapt to physical disturbances such as fluctuations in absorber level and switching of circulating pumps, as well as chemical disturbances caused by changes in coal sulfur content. This improves the adaptability of the dual-process decoupling mechanism model under wide load and variable operating conditions.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, a preset nonlinear mapping relationship is included, specifically: when the slurry chemical buffer index increases, the chemical damping coefficient obtained according to the preset nonlinear mapping relationship increases monotonically and nonlinearly, and when the slurry chemical buffer index is in the preset peak range, the chemical damping coefficient reaches the maximum threshold.

[0015] By adopting the above technical solution, when the chemical buffer index of the slurry increases, the chemical damping coefficient increases nonlinearly and reaches its maximum in the preset peak range, simulating the nonlinear amplification effect of chemical reaction resistance on apparent lag time, thereby improving the sensitivity of control response in low buffer zones.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, after determining the product of the physical mixing lag reference value and the chemical damping coefficient as the corrected lag time, the method further includes: calculating, within a rolling estimation window, a cross-correlation function of the actual measured slurry pH value trajectory relative to the historical limestone slurry flow trajectory; extracting the actual measured lag time based on the peak position of the cross-correlation function; calculating the time deviation between the corrected lag time and the actual measured lag time; when the actual measured lag time is greater than the corrected lag time and the time deviation is greater than a preset deviation range, increasing the growth rate adjustment parameter in the preset nonlinear mapping relationship by a preset step size to improve the output chemical damping coefficient; when the actual measured lag time is less than the corrected lag time and the time deviation is greater than a preset deviation range, decreasing the growth rate adjustment parameter in the preset nonlinear mapping relationship by a preset step size.

[0017] By adopting the above technical solution, and by comparing the corrected lag time calculated by the model with the actual measured lag time extracted based on the cross-correlation function in real time, the control system can automatically detect whether there is a deviation in the preset mapping relationship. When a deviation occurs, the growth rate adjustment parameter in the nonlinear mapping is automatically adjusted, thereby dynamically compensating for model mismatch problems caused by equipment aging, sensor drift, or slurry impurity accumulation.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, the step of calculating future limestone slurry flow instructions based on the updated dual-process decoupling mechanism model specifically includes: constructing a prediction model including corrected lag time based on the updated dual-process decoupling mechanism model; constructing an objective function for model predictive control based on environmental emission constraints and economic operating cost indicators within a preset future time period, wherein the objective function includes at least a first performance index term for characterizing the deviation of the control target and a second performance index term for characterizing the magnitude of change in limestone slurry flow, the second performance index term being configured with a variable smoothing weight coefficient; obtaining the current slurry chemical buffer index at the optimization solution time of each rolling control cycle; if the current slurry chemical buffer index is in a preset non-high value range, setting the smoothing weight coefficient as a preset benchmark weight value; if the current slurry chemical buffer index is in a preset high value range, calculating the difference between the slurry chemical buffer index and the lower limit of the high value range; calculating the weight increment based on the difference and a preset monotonically increasing function; and determining the current smoothing weight coefficient by the sum of the benchmark weight value and the weight increment.

[0019] By adopting the above technical solution, the penalty weight of the slurry flow rate variation in the objective function is dynamically adjusted according to the buffer state. In the high-risk range of strong buffering, by increasing the smoothing weight coefficient, the control system adopts a more conservative adjustment strategy, suppressing aggressive slurry supply actions, thereby reducing the risk of potential control divergence and improving the safety of the control system.

[0020] In a second aspect, embodiments of this application provide a control system comprising: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, wherein the one or more processors invoke the computer instructions to cause the control system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a control system, cause the control system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a control system, cause the control system to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the control system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. When the slurry is in the strong buffer plateau zone, the control system recognizes that the sluggish pH response is due to chemical inertia rather than insufficient slurry supply. By increasing the model lag time, the control system informs the controller to wait for the reaction results, thereby avoiding the controller from issuing excessive slurry supply commands due to misjudgment, and thus improving the accuracy of limestone slurry supply.

[0026] 2. By calculating the physical mixing hysteresis reference value affected by fluid dynamics and the chemical damping coefficient affected by chemical equilibrium separately, and then performing product coupling, the corrected hysteresis time can simultaneously adapt to physical disturbances such as absorber level fluctuations and circulating pump switching, as well as chemical disturbances caused by changes in coal sulfur content, thereby improving the adaptability of the dual-process decoupling mechanism model under wide load and variable operating conditions.

[0027] 3. By comparing the corrected lag time calculated by the model with the actual measured lag time extracted based on the cross-correlation function in real time, the control system can automatically detect whether there is a deviation in the preset mapping relationship. When a deviation occurs, the growth rate adjustment parameter in the nonlinear mapping is automatically adjusted to dynamically compensate for model mismatch caused by equipment aging, sensor drift, or slurry impurity accumulation. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating an intelligent control method for wet desulfurization in a thermal power plant, as described in an embodiment of this application.

[0029] Figure 2 This is another schematic diagram of a smart control method for wet desulfurization in a thermal power plant, as described in an embodiment of this application.

[0030] Figure 3 This is a schematic diagram of the physical device structure of a control system in an embodiment of this application. Detailed Implementation

[0031] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0033] This application provides an intelligent control method and system for wet desulfurization in thermal power plants to improve the accuracy of limestone slurry supply.

[0034] Please see Figure 1 This is a flowchart illustrating an intelligent control method for wet desulfurization in a thermal power plant, as described in an embodiment of this application.

[0035] S101. Extract historical data from the real-time operation data sequence of the desulfurization system to construct a rolling estimation window.

[0036] The real-time operating data sequence of the desulfurization system refers to the set of process variables continuously collected at a fixed sampling period and arranged by timestamps from the distributed control system or real-time database at the wet desulfurization site of the thermal power plant. These variables include, but are not limited to: flue gas inlet / outlet SO2 concentration, flue gas flow rate, absorber slurry pH value, limestone slurry supply flow rate, and slurry circulation pump flow rate. Historical data refers to the real-time operating data sequence that has occurred and been recorded before the calculation time of the current control cycle. The rolling estimation window is a dynamic data subset selection mechanism that extracts a fixed-time segment (e.g., the past 60 minutes) of the latest data from the historical data sequence. This segment slides forward synchronously with the passage of physical time, always containing the latest information reflecting the current dynamic characteristics of the system.

[0037] Specifically, the control system uses the current time t as a reference and backtracks by a preset window length T. w Thus, a time range of [tT] is extracted. w The data matrix is ​​denoted as t. This matrix contains the measurement sequence of all relevant variables within the time window. The rolling mechanism enables the data used for parameter estimation to capture the non-stationary, time-varying characteristics of the system caused by changes in operating conditions (such as boiler load adjustments and coal quality fluctuations), thus allowing subsequent model updates to adaptively track process changes.

[0038] S102. Establish a dual-process decoupling mechanism model that includes the fast dynamic sub-process of gas-liquid mass transfer and the slow dynamic sub-process of limestone dissolution.

[0039] The fast-dynamic subprocess of gas-liquid mass transfer refers to the physicochemical process in which gaseous SO2 molecules in flue gas are absorbed, dissolved, and ionized by the slurry in the absorption tower to form acidic substances (mainly sulfurous acid), with a time constant typically ranging from seconds to minutes. The slow-dynamic subprocess of limestone dissolution refers to the chemical process in which solid limestone (CaCO3) particles added to the slurry tank slowly dissolve and neutralize with acidic substances, thereby adjusting the pH value of the slurry, with a time constant typically ranging from tens of minutes to several hours. The dual-process decoupling mechanism model is a mathematical model based on the idea of ​​separating the inherent time scale of the system. It decomposes the complex desulfurization reaction into two subprocess models with different but interrelated dynamic characteristics, and provides a mathematical description based on the principles of chemical reaction engineering and fluid dynamics. The gas-liquid mass transfer coefficient ( The activity coefficient (FCO) is a lumped parameter used to characterize the rate at which SO2 is transferred from the gas phase to the liquid phase per unit time and driven by a unit concentration difference. Its magnitude comprehensively reflects the mass transfer efficiency, including the absorber structure, spray layer efficiency, and gas-liquid contact area. The dynamic hysteresis time (ΔH) is a dimensionless parameter used to characterize the chemical reactivity of limestone raw materials, reflecting the influence of differences in purity, particle size, and crystal form between different batches of limestone on their dissolution rate. () is used to characterize the time delay between the addition of limestone slurry to the slurry tank and the time between the dissolution effect producing an observable and significant impact on the pH of the slurry.

[0040] Specifically, fast dynamic subprocess models are usually based on the two-film theory, describing the change of SO2 concentration along the height z of the absorber tower. Their simplified ordinary differential equations can be expressed as:

[0041]

[0042] in, The average SO2 concentration inside the tower. For flue gas flow rate, For gas phase volume, and These represent the partial pressures of SO2 at the gas phase and the gas-liquid interface, respectively. It is the gas-liquid mass transfer coefficient to be identified.

[0043] Slow-dynamic subprocess models are typically built upon material balance in a CSTR (continuous stirred tank reactor) to describe changes in the concentrations of key ions within the slurry tank, such as hydrogen ion concentration. The change can be represented as:

[0044]

[0045] in, The volume of the slurry tank. The rate at which acid is produced by SO2 absorption (determined by the fast process). This represents the rate at which acid is consumed during the dissolution of limestone. This dissolution rate term is related to the limestone slurry supply flow rate. and limestone activity coefficient Related, for example And subject to a dynamic lag time The impact, expressed as .

[0046] By coupling these two sub-processes through intermediate variables (SO2 absorption rate and slurry pH value), a dual-process decoupling mechanism model is formed.

[0047] In some embodiments, the dual-process decoupling mechanism model can be constructed in various ways. Optionally, it can be in the form of a lumped-parameter ordinary differential equation (ODE). As described above, an ODE is established for the fast dynamic subprocess regarding the SO2 removal rate or outlet concentration, containing the parameters to be identified. Establish a framework for slow-dynamic subprocesses within the slurry tank. , , and A set of material balance differential equations for key ion concentrations. Introducing the limestone activity coefficient into the limestone material balance equation. As a multiplier factor for the dissolution rate term, a pure time lag is applied to the limestone slurry flow rate input term. ,in This refers to the dynamic lag time. The equations for the two sub-processes are combined to form a complete nonlinear state-space model, which is used for subsequent simulation and optimization.

[0048] Alternatively, a simplified transfer function model can be used. The fast-dynamic subprocess can be approximated as a first- or second-order transfer function with the slurry pH value as input and the SO2 removal rate as output. The slow-dynamic subprocess can be approximated as a first-order plus-hysteresis model with the limestone slurry supply flow rate as input and the slurry pH value as output: G(s) Among them, the model gain K and the time constant T are related to the limestone activity coefficient. Related, pure lag time This refers to the dynamic lag time. In the control system architecture, two transfer function models are cascaded or combined in a more complex way to form a dual-process decoupling mechanism model. This approach simplifies calculations and is easy to implement in engineering.

[0049] S103. Calculate the slurry chemical buffer index within the rolling estimation window based on the nonlinear sensitivity characteristics of slurry pH to reactant concentration.

[0050] The nonlinear sensitivity of slurry pH to reactant concentration refers to the fact that in desulfurization slurry, a multi-component electrolyte solution, the change in pH is not linearly related to the amount of acid (e.g., dissolved SO2) or alkali (e.g., dissolved limestone) added that causes the change. Instead, it exhibits resistance to pH changes within a specific pH range (e.g., around pH 5.0-6.0), i.e., a chemical buffering effect. The slurry chemical buffering index (…) () is a normalized dimensionless quantitative index used to characterize the magnitude of the inertia of slurry in resisting pH changes caused by external acid or alkali substances under the current slurry chemical composition.

[0051] The acid-base buffer characteristic curve function, at its core, is the buffer capacity in chemistry ( The mathematical expression for buffer capacity describes the functional relationship between buffer capacity and hydrogen ion concentration (i.e., pH value). According to claims 2 and 3, the present invention obtains buffer capacity by calculating the reciprocal of the slope of the tangent line to this function at a specific pH point, which is consistent with the definition of buffer capacity.

[0052] Specifically, the desulfurization slurry contains and Multiple buffer pairs. According to the definition of buffer capacity, It represents the molar concentration of a strong base or acid required to change the pH of a solution by one unit. The theoretical calculation formula is: ,in: and These represent the concentrations of hydrogen ions and hydroxide ions, respectively. ( (where is the ion product constant of water). The total concentration of the i-th buffer system (e.g., total dissolved carbonate concentration) Total dissolved sulfite concentration ). Let be the acid dissociation constant of the i-th buffer system. Summation term This involves traversing all major buffer systems. For desulfurization systems, the carbonic acid and sulfite systems are primarily considered. The formula is a mathematical expression with hydrogen ion concentration as the independent variable, and includes a bisulfite ionization equilibrium constant term (i.e.,...). ) and the equilibrium constant term for bicarbonate ionization (i.e. ) as a denominator parameter (in ( + (Item in the text). When the pH of the slurry is close to the p of a certain buffer pair... Value (i.e.) The denominator of this term reaches a minimum value, making A peak occurs when the slurry is in a strongly buffered state.

[0053] The execution flow of this step is as follows: within the rolling estimation window, calculate the corresponding buffer capacity for each time point of slurry pH measurement. Value, and then the obtained The value sequence is normalized (e.g., divided by a preset or historical maximum). (value), to obtain a time series slurry chemical buffer index in the range [0,1]. .

[0054] In some embodiments, the slurry chemical buffer index can be calculated in several ways: optionally, it can be calculated based on a theoretical buffer capacity formula. The buffer capacity is established as shown above. The theoretical calculation formula. Wherein, each dissociation constant... Given a constant, the total concentration and It can be used as an adjustable parameter or set based on empirical values. Within the rolling estimation window, the real-time slurry pH value at each sampling time is converted to... concentration( ).Will Substituting into the buffer capacity formula, the buffer capacity at each time step can be calculated. Value sequence. (For) The sequence is normalized, for example: ,in and This is a preset buffer capacity range, thus obtaining the slurry chemical buffer index sequence. Optionally, it can also be based on a data-driven local gradient analysis method. This method does not rely on precise chemical parameters. Within the rolling estimation window, the limestone slurry supply flow rate is... (As an alkali input) and slurry pH value (The response output) is considered as an input-output pair. The local slope of the pH response curve with respect to the limestone flow input is calculated using numerical differentiation methods (such as the finite difference method) or local polynomial fitting. Calculate the reciprocal of the slope. The physical meaning of this value is the limestone slurry flow rate required to cause a unit change in pH value; its magnitude directly reflects the buffer strength. By filtering and normalizing the obtained reciprocal sequence, an empirical slurry chemical buffer index sequence can be obtained.

[0055] S104. Based on the preset nonlinear mapping relationship and the slurry chemical buffer index, the dynamic lag time in the slow dynamic subprocess of limestone dissolution is assigned a value to obtain the corrected lag time.

[0056] Among them, the pre-defined nonlinear mapping relationship refers to a mathematical function or lookup table that is pre-established during the controller design phase through mechanism analysis, experimental data, and operational experience. It defines the slurry chemical buffer index as the input. A non-linear correspondence between the lag time and a coefficient used to adjust for the lag time (or the lag time itself). Corrected lag time. It is based on the current chemical buffer state of the slurry (by...) (Quantification) is the effective lag time obtained by dynamically correcting a baseline lag time, which can more accurately reflect the apparent response delay of the system under the current operating conditions.

[0057] Specifically, this step aims to address control deviations caused by the use of fixed or linearly varying hysteresis compensation. When the slurry is in a strongly buffered state ( When the pH value is high, the response of the pH value to the addition of limestone becomes sluggish, macroscopically manifested as an increased response delay, i.e., chemical pseudohysteresis. Therefore, this step utilizes a pre-defined monotonically increasing nonlinear mapping relationship to... The quantified chemical buffer strength is translated into a dynamic adjustment of the model lag time. For example, a mapping relationship of the following form can be constructed: ,in, It is a baseline lag time. It is a monotonically increasing nonlinear function, and .when When it increases, This also increases non-linearly, thus increasing the correction lag time. It gets longer. This is what was obtained. These will be treated as known parameters and used for the next step of model parameter identification.

[0058] In some embodiments, the lag time can be determined in several ways: optionally, a continuous nonlinear mapping can be constructed using the Sigmoid function. The mapping relationship... Designed as a scaled and translated sigmoid function, for example:

[0059]

[0060] in, This is the preset maximum chemical damping coefficient (e.g., 2.5), and k is the growth rate parameter. It is the center point of the buffer exponent. This function can smoothly convert... (range [0, 1]) is mapped to a damping coefficient (range [1, ..., 1]). ]). The calculation of S103 Substitute into this function to calculate the damping coefficient. Then compared with the benchmark lag time Multiply to obtain the corrected lag time. Alternatively, a lookup table can be used. A series of [data points] can be calibrated through offline simulation or historical data analysis. Its corresponding optimal lag time The corresponding points. These data points ( , ), ( , ... are stored in a lookup table. In real-time control, the values ​​calculated based on S103 are... The corrected lag time can be quickly obtained by looking up a table and using linear interpolation. .

[0061] S105. Within the rolling estimation window, the corrected lag time is substituted into the dual-process decoupling mechanism model as a known parameter. The gas-liquid mass transfer coefficient and the limestone activity coefficient are used as optimization variables to solve for the optimized gas-liquid mass transfer coefficient and the optimized limestone activity coefficient that minimize the deviation between the output trajectory of the dual-process decoupling mechanism model and the historical data trajectory.

[0062] The optimization variables are the unknown parameters that need to be solved in this optimization problem, namely the gas-liquid mass transfer coefficient k_La and the limestone activity coefficient. The output trajectory of the dual-process decoupling mechanism model refers to the trajectory of a set of candidate ( , ) and the known information obtained in step S104 After substituting into the model, historical input data (such as flue gas flow rate and limestone slurry supply flow rate) within the rolling estimation window are used as model input. The model predicts the output sequence (such as the predicted SO2 outlet concentration sequence and slurry pH value sequence) through numerical integration simulation. The historical data trajectory refers to the actual value sequence of the corresponding process output variables measured within the rolling estimation window. Minimizing the deviation refers to finding a set of optimal solutions by systematically adjusting the values ​​of the optimization variables. , This minimizes the cumulative error (usually represented by the sum of squared errors) between the model's output trajectory and the historical data trajectory.

[0063] Specifically, in step S104, the time-varying lag time is... After decoupling and determination, the uncertainty of the model mainly focuses on reflecting equipment efficiency. and reflecting the characteristics of raw materials The goal of this step is to use the latest historical data within the scrolling window to reverse-engineer the data that best reproduces this historical process. and The value of . Therefore, construct an objective function J, for example:

[0064]

[0065] Among them, summation Iterate through all sampling points within the rolling estimation window. and These are weighting coefficients used to balance the fitting accuracy of different variables. Then, a numerical optimization algorithm (such as nonlinear least squares method) is used to solve this minimization problem, and the solution found is the optimized gas-liquid mass transfer coefficient and the optimized limestone activity coefficient.

[0066] In some embodiments, the solution can be optimized in various ways: optionally, a gradient-based nonlinear least squares method can be used. Define the error sum of squares objective function J as shown above. Let J be the optimization variable. and Initial guesses based on experience or results from the previous period are provided, and a reasonable physical range for these guesses is set as a constraint. Numerical optimization algorithms, such as the Levenberg-Marquardt (LM) algorithm or the Sequential Quadratic Programming (SQP) algorithm, are used to iteratively solve for the value that minimizes the objective function J. and Alternatively, a gradient-free global optimization algorithm can be used. The objective function J is defined similarly. Particle Swarm Optimization (PSO) or Genetic Algorithm (GA) can be employed. These algorithms maintain a population of solutions and search by simulating natural evolution or social behavior, effectively avoiding local optima and possessing better global search capabilities for complex non-convex optimization problems. When the algorithm finally converges, the optimal individual in the population represents (… , This is the desired optimal solution.

[0067] S106. Update the dual-process decoupling mechanism model based on the optimized gas-liquid mass transfer coefficient, optimized limestone activity coefficient, and corrected lag time.

[0068] Updating the dual-process decoupling mechanism model refers to refining the optimized gas-liquid mass transfer coefficient obtained in step S105. and optimization of limestone activity coefficient and the corrected lag time obtained in step S104 Substitute these values ​​into the dual-process decoupling mechanism model established in step S102, replacing the old parameter values ​​used in the previous cycle, thereby obtaining a new model instance with corrected parameters that can more accurately reflect the current dynamic characteristics of the system.

[0069] Specifically, after optimization in S105, a better combination of parameters was obtained under the current operating conditions. This step then incorporates these improved parameters into the dual-process decoupling mechanism model. Through this operation, the original dual-process decoupling mechanism model with unknown parameters is transformed into a customized model for the current specific operating conditions (specific equipment mass transfer efficiency, specific limestone reactivity, and specific chemical buffer state). This updated model, because its parameters are dynamically calibrated based on the latest data, improves the accuracy of predicting future behavior.

[0070] S107. Calculate future limestone slurry flow instructions based on the updated dual-process decoupling mechanism model.

[0071] Among them, the future limestone slurry supply flow command refers to a series of flow set values ​​that need to be sent to the actuators such as limestone slurry pumps within one or more future control cycles, calculated by the controller.

[0072] Specifically, this step is the decision-making output stage of the entire intelligent control method. First, the updated high-precision model of S106 is used as the prediction model to predict a future time period (referred to as the prediction time domain). Within the control time domain, if a series of different limestone slurry flow rate operations are implemented (control time domain) The system's key outputs (such as slurry pH and outlet SO2 concentration) will evolve. Secondly, based on environmental emission constraints (SO2 concentration must be below the limit) and economic operating cost indicators (pH maintained within the optimal range, limestone consumption minimized, and slurry flow rate stable), a model predictive control objective function is constructed. This objective function includes at least: a first performance indicator term: used to penalize deviations between the future output trajectory and the desired setpoint trajectory. A second performance indicator term: used to penalize the magnitude of changes in the future control quantity (limestone slurry flow rate), configured with a variable smoothing weight coefficient.

[0073] For example, the formula for the model predictive control objective function is as follows: ,in, To predict the time domain, To control the time domain, For the model's predicted output, Set a value as desired. To control the increment, Q and R are weight matrices, and the smoothing weight coefficient R is dynamically adjusted. At each optimization solution time, the current slurry chemical buffer index is obtained. .like If the pH value is within the preset high-buffered range, the value of R is increased. This is because, in the high-buffered range, the pH response is sluggish, and aggressive control actions are not only ineffective but also prone to excessive limestone accumulation. Increasing R makes the controller behave more conservatively, suppressing unnecessary drastic flow changes. Finally, the solution is obtained through online optimization. Minimize the problem to obtain the future Optimal control increment sequence of steps Based on the rolling optimization principle of MPC, only the first control increment is considered. Applying this to the current control variable, we obtain the slurry flow rate command for the current moment. And it was issued to the implementing agency.

[0074] In the above embodiments, when the slurry is in the strong buffer plateau zone, the control system recognizes that the sluggish pH response is due to chemical inertia rather than insufficient slurry supply, and informs the controller to wait for the reaction results by increasing the model lag time, thereby avoiding the controller issuing excessive slurry supply commands due to misjudgment, and thus improving the accuracy of limestone slurry supply.

[0075] However, in the above embodiments, the calculation of the corrected lag time mainly relies on a preset nonlinear mapping relationship, and this relationship directly affects the final lag time. During the long-term operation of a thermal power plant, the physical conditions of the absorber tower (such as the slurry pool level and the number of circulating pumps in operation) will change significantly, causing the basic physical mixing time to fluctuate. At the same time, the accumulation of impurities such as chloride and fluoride ions in the slurry may change the actual buffering characteristics of the slurry, causing the preset mapping relationship to gradually become invalid. This may result in insufficient or excessive correction when physical conditions change drastically or the model ages, thereby affecting the stability of the control system.

[0076] Please see Figure 2 This is another flowchart illustrating an intelligent control method for wet desulfurization in a thermal power plant, as described in this application.

[0077] S201. Extract historical data from the real-time operation data sequence of the desulfurization system to construct a rolling estimation window.

[0078] S202. Establish a dual-process decoupling mechanism model that includes the fast dynamic sub-process of gas-liquid mass transfer and the slow dynamic sub-process of limestone dissolution.

[0079] S203. Calculate the slurry chemical buffer index within the rolling estimation window based on the nonlinear sensitivity characteristics of slurry pH to reactant concentration.

[0080] Step S201 is similar to step S101, step S202 is similar to step S102, and step S203 is similar to step S103, so they will not be described again here.

[0081] S204. The ratio of the liquid level and volume of the slurry pool in the absorption tower to the flow rate of the slurry circulation pump is used as the physical mixing hysteresis reference value.

[0082] Among them, the liquid level volume of the slurry pool in the absorption tower ( The slurry volume ( ) refers to the actual volume of slurry contained in the slurry pool at the bottom of the absorber at the current moment, which can be calculated by combining the real-time measurement value of the level gauge with the geometric parameters of the absorber. Slurry circulation pump flow rate ( The total flow rate of all operating slurry circulation pumps is much greater than the limestone replenishment flow rate and the gypsum discharge flow rate, and is the main factor determining the macroscopic mixing rate of the fluid in the slurry tank. Physical mixing hysteresis reference value ( ) is a time constant calculated based on ideal mixing or macroscopic fluid dynamics. It characterizes the average time required for a newly added tracer (which can be compared to limestone slurry) to achieve macroscopic homogeneous mixing throughout the slurry tank, i.e., the average residence time of the fluid in the reactor.

[0083] Specifically, the overall hysteresis is broken down into two parts: physical mixing and chemical reaction. This step is responsible for quantifying the former. This is done using the formula: This physical mixing hysteresis baseline value is used to estimate the physical mixing hysteresis baseline value. This baseline value represents the inherent delay in material transport and mixing under ideal conditions, without considering any chemical reaction complexity. For example, if the effective volume of the slurry tank is 1200 cubic meters and the total slurry circulation flow rate is 24000 cubic meters per hour, then the physical mixing hysteresis baseline value is approximately: This physical mixing hysteresis baseline will serve as the basis for subsequent multiplication with chemical effects.

[0084] In some embodiments, the physical mixing hysteresis baseline value can be calculated in several ways: optionally, a simplified real-time calculation method can be used. This involves real-time reading of the absorber tower level sensor height value. Calculate the current slurry volume based on the known cross-sectional area A of the absorption tower (if it is a regular cylinder): Real-time reading of the flow rate of all operating slurry circulation pumps. Alternatively, the current signal (converted to flow rate) can be summed to obtain the total circulating flow rate: Finally, the physical hybrid hysteresis baseline value is calculated: Optionally, a correction calculation considering non-ideal flow can be employed. The residence time distribution (RTD) of the absorber slurry pool is analyzed through computational fluid dynamics (CFD) simulations or tracer experiments. The first moment (i.e., the average residence time) is extracted from the residence time distribution curve as... By analyzing the deviation between the actual flow and the ideal CSTR model, a correction coefficient is obtained. (For example, obtained by calculating indices such as short-circuit current and dead zone volume). The final physical hybrid hysteresis baseline value is: .

[0085] S205. Substitute the slurry chemical buffer index as an input variable into the preset nonlinear mapping relationship to calculate the chemical damping coefficient.

[0086] The input variable is the slurry chemical buffer index calculated in step S203. Chemical damping coefficient ( The multiplier factor (F) is a dimensionless multiplier used to quantify the amplification factor of the hysteresis in pH response introduced by the chemical buffering effect of the slurry. This factor is greater than or equal to 1; a value of 1 indicates that the system only has physical mixing hysteresis, with no additional chemical hysteresis. The larger the value, the stronger the apparent hysteresis effect caused by chemical buffering.

[0087] Specifically, this step transforms the buffer strength quantized in step S203 into an amplification factor that can be used to correct the physical lag time. This preset nonlinear mapping is designed as a monotonically increasing nonlinear function, and... When within the preset peak range, The maximum threshold is reached. This simulates the situation when the system enters a strong buffer plateau region. When it increases, the calculated It also increases nonlinearly, thus significantly extending the total lag time in the next step.

[0088] In some embodiments, the chemical damping coefficient can be calculated in a variety of ways: optionally, an exponential function or a sigmoid function with saturation characteristics can be used.

[0089]

[0090] in It is the preset maximum chemical damping coefficient. This is the growth rate adjustment parameter. This function adjusts the growth rate when... When =0, =1. With Increase Nonlinearly increasing and asymptotically approaching The calculation of S203 Substituting into this function yields the chemical damping coefficient. Alternatively, a piecewise function can be used. The range [0, 1] is divided into several sub-intervals, such as [0, 0.3), [0.3, 0.7), and [0.7, 1]. A definition is given for each interval. The calculation method, for example: within [0, 0.3): =1. Within [0.3, 0.7): (Linear growth). Within [0.7, 1]: (Nonlinear growth to) ). Calculated according to S203 The interval to which it belongs is calculated using the corresponding formula. .

[0091] S206. The product of the physical mixing hysteresis reference value and the chemical damping coefficient is determined as the corrected hysteresis time.

[0092] Specifically, this step is the calculation step to form the final effective lag time, and its mathematical expression is: This step couples the effects of the physical and chemical components, which were quantified separately in the previous steps. This is achieved by using a physical mixing hysteresis benchmark representing the basic mixing time. (From S204) and the chemical damping coefficient representing the amplification factor of the chemical hysteresis effect. Multiplying (from S205) yields a corrected lag time that simultaneously reflects the combined effects of physical transport delay and nonlinear damping of chemical reactions. This dynamically adjusted... This enables subsequent models to more accurately predict and adapt to sluggish responses in strong buffers.

[0093] In some embodiments, after determining the correction lag time, an online correction mechanism based on data feedback can be introduced: optionally, an adaptive adjustment method based on cross-correlation mapping parameters can be adopted. Within the rolling estimation window, the historical limestone slurry supply flow trajectory is calculated. The trajectory of the actual measured slurry pH value Cross-correlation function between them: .in For the expected operation, m is the delay step size. Find the delay time corresponding to the peak (or trough, depending on the correlation) of the cross-correlation function. This is used as the actual measurement lag time extracted from the data. Next, calculate the time deviation: .like The absolute value is greater than the preset deviation range (e.g.) Then, the growth rate adjustment parameter of the nonlinear mapping relationship in S205 is... Adjustments need to be made. If... If the actual lag is greater than the model estimate, then increase the step size by the preset step size. This increases the chemical damping coefficient of the output in subsequent calculations, and vice versa. Optionally, the recursive least squares (RLS) method can be used to identify the mapping parameters online. The mapping parameters in S205 (such as...) Treating these as variables to be identified, construct a parameter vector. Build a For measured values, The problem of parameter identification for model predictions, where The RLS algorithm is used to initialize the covariance matrix P(0) and the forgetting factor. The parameters are updated recursively using the following formula. :

[0094]

[0095]

[0096]

[0097] in For the regression vector, It is an identity matrix. This formula uses the time deviation calculated for each period. Recursively update parameters The value of is such that it minimizes the long-term sum of squared prediction errors.

[0098] S207. Within the rolling estimation window, the corrected lag time is substituted into the dual-process decoupling mechanism model as a known parameter. The gas-liquid mass transfer coefficient and the limestone activity coefficient are used as optimization variables to solve for the optimized gas-liquid mass transfer coefficient and the optimized limestone activity coefficient that minimize the deviation between the output trajectory of the dual-process decoupling mechanism model and the historical data trajectory.

[0099] The logic of this step is the same as S105, except that the lag time is corrected. The calculation method is replaced by the coupled calculation results of S204-S206, and the formula of its optimization objective function is as follows:

[0100]

[0101] The specific implementation methods (such as nonlinear least squares method, etc.) are the same as those in S105, and will not be described again here.

[0102] S208. Update the dual-process decoupling mechanism model based on the optimized gas-liquid mass transfer coefficient, optimized limestone activity coefficient, and corrected lag time.

[0103] Step S208 is similar to step S106, and will not be described again here.

[0104] S209. Calculate future limestone slurry flow instructions based on the updated dual-process decoupling mechanism model.

[0105] Step S209 is similar to step S107, and will not be described again here.

[0106] In the above embodiments, the correction logic for the lag time is further refined. A decoupled calculation method is adopted, multiplying the physical mixed lag reference value by the chemical damping coefficient. This allows the control system to independently respond to physical disturbances caused by changes in liquid level / flow rate, as well as chemical disturbances caused by changes in pH. An online feedback correction mechanism based on a cross-correlation function is also introduced. When a deviation occurs between the preset mapping relationship and the actual process, the growth rate adjustment parameter can be automatically adjusted, and the calculation rules for the chemical damping coefficient can be dynamically calibrated. The closed-loop adaptive method in the above embodiments improves the robustness of the control system in complex and variable environments.

[0107] The above describes an intelligent control method for wet desulfurization in a thermal power plant according to an embodiment of this application. The following describes an exemplary control system 300 provided in an embodiment of this application.

[0108] Figure 3This is an exemplary hardware structure diagram of the control system 300 provided in an embodiment of this application. In some embodiments, the control system 300 is a computer device. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements an intelligent control method for wet desulfurization in a thermal power plant according to an embodiment of this application.

[0109] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0110] In some embodiments of this application, a computer-readable storage medium is also provided, including instructions that, when executed on the control system 300, cause the control system 300 to perform an intelligent control method for wet desulfurization in a thermal power plant according to an embodiment of this application.

[0111] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0112] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0113] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0114] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An intelligent control method for wet desulfurization in thermal power plants, characterized in that, Applied to a control system, the method includes: A rolling estimation window is constructed by extracting historical data from the real-time operation data sequence of the desulfurization system; A dual-process decoupling mechanism model is established, which includes a fast dynamic subprocess of gas-liquid mass transfer and a slow dynamic subprocess of limestone dissolution. The fast dynamic subprocess of gas-liquid mass transfer includes a gas-liquid mass transfer coefficient to characterize the gas-liquid contact mass transfer efficiency, and the slow dynamic subprocess of limestone dissolution includes a limestone activity coefficient and a variable dynamic lag time to characterize the reactivity of limestone raw materials. The slurry chemical buffer index within the rolling estimation window is calculated based on the nonlinear sensitivity of slurry pH to reactant concentration. The slurry chemical buffer index characterizes the magnitude of the slurry's inertia in resisting changes in pH. The dynamic lag time in the slow dynamic subprocess of limestone dissolution is assigned a value based on the preset nonlinear mapping relationship and the slurry chemical buffer index to obtain the corrected lag time. Within the rolling estimation window, the corrected lag time is substituted into the dual-process decoupling mechanism model as a known parameter. The gas-liquid mass transfer coefficient and the limestone activity coefficient are used as optimization variables to solve for the optimized gas-liquid mass transfer coefficient and the optimized limestone activity coefficient that minimize the deviation between the output trajectory of the dual-process decoupling mechanism model and the historical data trajectory. The dual-process decoupling mechanism model is updated based on the optimized gas-liquid mass transfer coefficient, the optimized limestone activity coefficient, and the corrected lag time. The future limestone slurry flow command is calculated based on the updated dual-process decoupling mechanism model.

2. The method according to claim 1, characterized in that, The step of calculating the slurry chemical buffer index within the rolling estimation window based on the nonlinear sensitivity of slurry pH to reactant concentration specifically includes: Based on the ionization equilibrium relationship of the carbonate and sulfite system, an acid-base buffering characteristic curve function was constructed to show the change of slurry pH value with the amount of added acid and base reagents. Substitute the real-time slurry pH value within the rolling estimation window into the acid-base buffer characteristic curve function, and calculate the reciprocal of the slope of the tangent line of the acid-base buffer characteristic curve function at each pH value point. The reciprocal is normalized to obtain the slurry chemical buffer index.

3. The method according to claim 2, characterized in that, The step of constructing the acid-base buffering characteristic curve function of the slurry pH value as a function of the amount of added acid and base reagents based on the ionization equilibrium relationship of the carbonate and sulfite system specifically includes: The acid-base buffer characteristic curve function is constructed based on the buffer capacity definition formula. The acid-base buffer characteristic curve function is configured to characterize the ratio of the differential change in the amount of added acid or base reagent to the differential change in pH that causes the change. The acid-base buffer characteristic curve function includes a mathematical expression with hydrogen ion concentration as the independent variable, and the mathematical expression includes bisulfite ionization equilibrium constant term and bicarbonate ionization equilibrium constant term as denominator parameters. When the hydrogen ion concentration corresponding to the pH value of the slurry causes the denominator parameter to approach a minimum value, the chemical buffering index of the slurry is in a preset peak range, indicating that the slurry is in a strong buffering state.

4. The method according to claim 1, characterized in that, The step of assigning a value to the dynamic lag time in the slow dynamic sub-process of limestone dissolution based on a preset nonlinear mapping relationship and the slurry chemical buffer index to obtain the corrected lag time specifically includes: The ratio of the liquid level and volume of the slurry tank in the absorption tower to the flow rate of the slurry circulation pump is used as the physical mixing hysteresis reference value. The chemical buffer index of the slurry is substituted into the preset nonlinear mapping relationship as an input variable to calculate the chemical damping coefficient. The product of the physical mixing hysteresis reference value and the chemical damping coefficient is determined as the corrected hysteresis time.

5. The method according to claim 4, characterized in that, The preset nonlinear mapping relationship specifically includes: When the slurry chemical buffer index increases, the chemical damping coefficient obtained according to the preset nonlinear mapping relationship increases monotonically and nonlinearly, and when the slurry chemical buffer index is in the preset peak range, the chemical damping coefficient reaches the maximum threshold.

6. The method according to claim 5, characterized in that, After determining the corrected hysteresis time by multiplying the physical mixing hysteresis reference value by the chemical damping coefficient, the method further includes: Within the rolling estimation window, the cross-correlation function of the actual measured slurry pH trajectory relative to the historical limestone slurry flow trajectory is calculated; The actual measurement lag time is extracted based on the peak position of the cross-correlation function; Calculate the time deviation between the corrected lag time and the actual measured lag time; When the actual measurement lag time is greater than the corrected lag time and the time deviation is greater than the preset deviation range, the growth rate adjustment parameter in the preset nonlinear mapping relationship is increased by a preset step size to improve the output chemical damping coefficient. When the actual measurement lag time is less than the corrected lag time and the time deviation is greater than the preset deviation range, the growth rate adjustment parameter in the preset nonlinear mapping relationship is reduced by a preset step size.

7. The method according to claim 1, characterized in that, The step of calculating the future limestone slurry supply flow command based on the updated dual-process decoupling mechanism model specifically includes: Based on the updated dual-process decoupling mechanism model, a prediction model incorporating the corrected lag time is constructed. Based on the environmental emission constraints and economic operating cost indicators within a future preset period, a model predictive control objective function is constructed. The objective function includes at least a first performance index term for characterizing the deviation of the control target and a second performance index term for characterizing the change in limestone slurry flow rate. The second performance index term is configured with a variable smoothing weight coefficient. At each optimization solution moment of each rolling control cycle, the current slurry chemical buffer index is obtained; If the current slurry chemical buffer index is in a preset non-high value range, then the smoothing weight coefficient is set to a preset benchmark weight value. If the current slurry chemical buffer index is in a preset high value range, then calculate the difference between the slurry chemical buffer index and the lower limit of the high value range; The weight increment is calculated based on the difference and a preset monotonically increasing function. The sum of the baseline weight value and the weight increment is determined as the current smoothing weight coefficient.

8. An intelligent control system, characterized in that, The intelligent control system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the intelligent control system to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on the intelligent control system, the intelligent control system performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the intelligent control system, the intelligent control system performs the method as described in any one of claims 1-7.

Images