Boiling furnace desulfurization process multivariable coordination energy intensity optimization method

Abstract

This invention relates to a multi-variable collaborative energy intensity optimization method for fluidized bed desulfurization processes, and to the field of fluidized bed desulfurization process optimization. This invention defines a comprehensive desulfurization energy intensity index that uniformly converts the heat penalty, electricity consumption, and limestone embodied energy in the desulfurization process into standard energy consumption; establishes a fluidized bed desulfurization process model coupling microscopic reaction kinetics and macroscopic fluid dynamics, and uses its output data combined with actual operating data to train the energy consumption model; with the goal of minimizing comprehensive desulfurization energy intensity and with emission compliance and safe operation as constraints, a multi-objective collaborative optimization model is established, and a multi-objective optimization algorithm is used to collaboratively optimize bed temperature, calcium-sulfur ratio, primary air rate, and limestone particle size to obtain the optimal combination of operating parameters. This invention can optimize the energy utilization efficiency of the desulfurization process, significantly reduce the comprehensive energy consumption of the fluidized bed system while ensuring emission compliance, and has significant economic and social benefits.

Description

Technical Field

[0001] This invention relates to the field of fluidized bed desulfurization optimization technology, and in particular to a multi-variable synergistic energy intensity optimization method for fluidized bed desulfurization processes. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] Fluidized bed desulfurization is currently recognized as one of the most efficient and economical coal-fired desulfurization technologies, especially in circulating fluidized bed boilers. Fluidized bed desulfurization primarily employs in-furnace calcium injection. Its basic principle is to directly introduce the desulfurizing agent (usually limestone or dolomite) into the combustion chamber. The high temperature inside the furnace (typically controlled at 850℃-950℃) causes the limestone to decompose, generating porous calcium oxide (CaO) and carbon dioxide. The sulfur dioxide produced by coal combustion comes into full contact with the calcium oxide under the conditions of the porous structure and fluidized motion, undergoing a gas-solid reaction to generate calcium sulfite, which is further oxidized to stable calcium sulfate. Fluidized bed desulfurization mainly includes the following key steps: desulfurizing agent preparation and feeding; crushing and grinding the limestone into powder of a certain fineness; injecting the limestone powder into the combustion chamber via a pneumatic conveying device or directly mixing it with the coal according to a certain calcium-to-sulfur ratio, based on the sulfur content in the coal. The fuel and desulfurizing agent violently tumble and mix in the fluidized bed. By adjusting the primary air rate and external heat exchanger, the bed temperature is strictly controlled within the optimal desulfurization temperature window (850℃-950℃). Excessive temperature will cause calcium oxide to "die-burn" and lose its activity, while excessively low temperature will result in a too slow reaction rate. The calcium sulfate produced in the reaction, along with unreacted calcium oxide, is discharged from the combustion chamber along with the fly ash. For circulating fluidized bed boilers, most solid particles are separated by a cyclone separator and returned to the combustion chamber to continue participating in the reaction (circulating combustion), which greatly improves the utilization efficiency of the desulfurizing agent. Finally, the desulfurization products (gypsum mixed with ash) are discharged through a dry ash discharge or pneumatic conveying system.

[0004] To achieve stringent emission targets, conservative operating strategies are often adopted in practice: to ensure SO2 emissions do not exceed limits, a high calcium-to-sulfur ratio (Ca / S molar ratio) is tended to be set, typically exceeding 2.5 or even reaching 3.0 or higher. To ensure complete combustion and the oxygen required for desulfurization, the oxygen supply in the exhaust gas is often too high, leading to increased heat loss in the flue gas. To avoid temperature fluctuations causing a decrease in desulfurization efficiency, the bed temperature control range is too narrow, sacrificing some combustion economy. This "each doing their own thing" single-variable conservative adjustment method, while ensuring emission compliance in the short term, brings serious energy consumption problems.

[0005] Operating a desulfurization system incurs energy costs. These costs are not only reflected in the consumption of limestone, but also in the associated changes in combustion efficiency, increased power consumption of fans, and other aspects. This cost can be quantified as desulfurization energy intensity—that is, the total amount of energy consumed to remove a unit mass of SO2.

[0006] An excessively high calcium-to-sulfur ratio leads to significant limestone waste. The mining, crushing, transportation, and calcination processes of limestone are themselves energy-intensive. The comprehensive energy consumption for producing 1 ton of limestone powder (250 mesh) is approximately 30-40 kg of standard coal. When the Ca / S ratio increases from 2.0 to 3.0, limestone consumption increases by 50%, and the resulting direct material and energy consumption, as well as indirect calcination heat absorption (approximately 178 kJ / mol), will significantly reduce the overall thermal efficiency of the boiler.

[0007] To ensure fluidization quality and transport excess limestone, the primary air fan, secondary air fan, and limestone conveying fan need to provide higher head and flow rate. For large CFB boilers, the shaft power of the fans can reach several thousand kilowatts, accounting for 20%-30% of the plant's power consumption. For every 10% increase in excess air distribution, the fan energy consumption will increase by about 15-20% (because the fan power consumption has a non-linear relationship with air pressure and flow rate).

[0008] The desulfurization process is deeply coupled with the combustion process. Excessive limestone entering the combustion chamber causes a strongly endothermic decomposition reaction, which lowers the bed temperature. To maintain a stable bed temperature, the coal feed rate often needs to be increased, which further increases the total amount of SO2 produced by combustion, creating a vicious cycle. Furthermore, to ensure the desulfurization reaction occurs within the optimal temperature window (850-950℃), a more efficient combustion temperature must be sacrificed, leading to increased heat loss from incomplete combustion.

[0009] Therefore, how to minimize the overall energy consumption intensity of the desulfurization system through refined and coordinated multi-variable control while meeting ultra-low emissions has become a core technical challenge for energy saving and carbon reduction in fluidized bed boilers. Summary of the Invention

[0010] To solve the above-mentioned technical problems, or at least partially solve them, the present invention provides a multi-variable synergistic energy intensity optimization method for fluidized bed desulfurization process.

[0011] This invention provides a multi-variable synergistic energy intensity optimization method for fluidized bed desulfurization processes, comprising: A fluidized bed desulfurization process model was established, which includes a limestone calcination model, a sulfidation reaction model, a fluidized bed gas-solid two-phase fluid dynamics model, and a fluidized bed heat balance model, in order to model the mechanism of action of decision variables in the fluidized bed process. The decision variables include: bed temperature, calcium-sulfur ratio, primary air rate, and average limestone particle size. Virtual training data covering a wide range of operating conditions is generated using a desulfurization model. Combined with actual fluidized bed furnace DCS operation data, training data for an energy consumption model based on a deep neural network is obtained. The energy consumption model predicts the comprehensive energy consumption intensity index and desulfurization efficiency of desulfurization based on decision variables. A multivariate collaborative optimization model was constructed with the objective function of minimizing the comprehensive energy consumption intensity of desulfurization, and the constraints of sulfur dioxide emission concentration, bed temperature safety range and equipment constraints. The multi-objective optimization algorithm is used to solve the multivariate collaborative optimization model to obtain the optimal combination of operating parameters, and the optimal combination of operating parameters is applied to the fluidized bed furnace control system.

[0012] Furthermore, the description of the limestone calcination process includes the following steps: heat is transferred from the gas flow to the particle surface; chemical decomposition occurs at the reaction interface; carbon dioxide diffuses through the product layer to the particle surface; carbon dioxide diffuses from the particle surface to the bulk gas flow; assuming the particles are spheres with an initial radius... initial density The limestone calcination reaction begins on the outer surface and gradually progresses inward, forming unreacted nuclei. Based on the random pore model, the limestone conversion rate... The change over time can be described by the following equation: ; in: The surface reaction rate constant during the calcination process depends on the bed temperature; This represents the initial specific surface area of ​​the limestone. The pore structure parameters are determined by the initial porosity and pore length, and are obtained by fitting the pore structure data of limestone after calcination through experiments such as mercury intrusion porosimetry. Z1 is a parameter related to the diffusion resistance of the product layer during calcination; Z2 is the molar volume ratio of the product to the reactants during the calcination reaction.

[0013] Furthermore, in the vulcanization reaction model, a stochastic pore model considering pore blockage introduces the variation of structural parameters over time, relating the reaction rate to pore surface area and diffusion resistance; wherein, the variation of structural parameters over time includes: When considering pore blockage, due to product layer accumulation, the effective pore radius r decreases with reaction time, expressed as: ,in, It refers to the thickness of the product layer when the pore radius shrinks to a critical value. At that time, the throat of the orifice is blocked, and the orifice no longer participates in the reaction; Considering congestion, using the congestion probability function Describe the effective surface area : ,in It is related to the local conversion rate or the thickness of the product layer. When the thickness of the local product layer reaches the pore throat radius, =1; The surface area is not considered when clogging occurs; When considering blockage, effective porosity is described by effective surface area: ; in, is the pore shrinkage rate constant, which controls the rate at which porosity decreases; f is the molar volume effect function, which reflects the effect of the ratio of the molar volume of solid product to reactant on pore shrinkage. When the molar volume of product is greater than that of reactant, f is positive, leading to a decrease in porosity. Correlating reaction rate with pore surface area and diffusion resistance includes: The reaction rate R is directly proportional to the effective surface area: ,in, is the reaction rate constant, C is the reactant concentration, and n is the order; The effective diffusion coefficient of sulfur dioxide is determined by introducing diffusion resistance to sulfur dioxide through porosity. Represented as: ; in The intrinsic diffusion coefficient is the diffusion coefficient of a gas in a completely open channel or free space. The initial porosity, It is an empirical, structurally sensitive parameter. This reflects the degree to which a decrease in porosity hinders the diffusion path. Used to describe the increased tortuosity caused by congestion; The differential equation for the reaction rate of gas-phase sulfur dioxide is obtained by using the law of mass conservation, the effective diffusion coefficient, and the reaction rate: .

[0014] Furthermore, the fluidized bed gas-solid two-phase fluid dynamics model includes: a two-phase sub-model that divides the bed into a bubble phase and an emulsion phase and describes the mass exchange between the two phases; a particle residence time distribution sub-model that describes the residence time distribution of particles in the fluidized bed furnace; and a fluidized bed fluid dynamics simulation sub-model.

[0015] Furthermore, the combustion chamber, cyclone separator, and return feeder constitute a closed-loop circulation system; the particle age distribution function is obtained by solving the mass balance of the closed-loop circulation system, including: assuming the average time for a particle to pass through the combustion chamber once is... The cyclone separator efficiency is The average number of cycles for the particles is Total average stay The axial concentration distribution of particles is described by an axial diffusion model, thereby obtaining the residence time distribution of particles of different sizes.

[0016] Furthermore, the fluid dynamics sub-model of the fluidized bed furnace is simplified into a one-dimensional axial model. The combustion chamber is divided into several micro-elements along the height. In each micro-element, it is assumed that the gas and solid are uniformly mixed. A set of ordinary differential equations is established through mass and energy balance. With the combustion chamber height z as the independent variable, the axial distribution of sulfur dioxide concentration and bed temperature is established.

[0017] Furthermore, the solution strategy for the fluidized bed desulfurization process model includes: dividing the combustion chamber into several control volumes along the axial direction; assuming uniform gas phase concentration, solid phase conversion rate, and temperature within each control volume; establishing a set of ordinary differential equations for mass balance and energy balance for each control volume; given boundary conditions, including inlet coal quantity, limestone quantity, air volume, and return material quantity, and using an iterative method to solve the steady-state solution; for dynamic simulation, using a numerical integration method to solve the differential equations, and using the desulfurization efficiency, bed temperature, and flue gas temperature calculated by the fluidized bed desulfurization process model as sample data for subsequent energy consumption models.

[0018] Furthermore, the process of training the energy consumption surrogate model includes: using a multi-layer feedforward neural network as the surrogate model architecture, with input layer nodes including bed temperature, calcium-sulfur ratio, primary air rate, average limestone particle size, boiler load, and coal sulfur content; output layer nodes including desulfurization efficiency and comprehensive desulfurization energy intensity index; using the fluidized bed desulfurization process model to generate virtual samples in the feasible region through Latin hypercube sampling, which are then mixed with actual DCS historical data to form a training set; and incorporating physical constraints into the loss function of the neural network to ensure that the prediction results conform to physical common sense.

[0019] Furthermore, the formula for calculating the comprehensive energy consumption intensity index of desulfurization is as follows: ; in, This refers to the amount of sulfur dioxide removed. To reduce heat loss due to combustion efficiency; Power consumption for auxiliary equipment, including primary air fans, secondary air fans, induced draft fans, limestone grinding and conveying equipment; The standard coal equivalent coefficient for electricity; This refers to the amount of limestone consumed. This represents the unit implicit energy of limestone.

[0020] Secondly, the present invention provides an energy intensity optimization device for multi-variable coordinated operation of fluidized bed desulfurization process, comprising: at least one processing unit, the processing unit being connected to a storage unit via a bus unit, the storage unit storing a computer program that can run on a processor, and the processing unit implementing the energy intensity optimization method for multi-variable coordinated operation of fluidized bed desulfurization process by running the computer program stored in the storage unit.

[0021] The technical solutions provided in the embodiments of the present invention have the following advantages compared with the prior art: This invention constructs a fluidized bed desulfurization process model that couples microscopic reaction mechanisms with macroscopic flow field characteristics. Based on this model, a comprehensive energy consumption intensity index encompassing heat, electricity, and materials is proposed, enabling a fair comparison of energy utilization efficiency under different operating conditions and avoiding the one-sidedness of optimizing a single index. To accommodate real-time control requirements, this application utilizes virtual training data generated from the fluidized bed desulfurization process model and DCS operating data to obtain an energy consumption model based on a deep neural network. The model predicts the comprehensive desulfurization energy consumption intensity index and desulfurization efficiency based on decision variables, ensuring physical rationality while achieving millisecond-level rapid prediction, meeting real-time optimization needs, and solving the problems of slow computation of mechanistic models and poor generalization of data models. Furthermore, a multi-objective intelligent optimization algorithm is used to explore a multi-variable collaborative optimization strategy for the fluidized bed desulfurization process under different loads. Through multi-variable collaborative optimization, the calcium-sulfur ratio, fan power consumption, and combustion efficiency loss can be significantly reduced while ensuring emission compliance, achieving maximum energy efficiency in the desulfurization process. This invention is applicable to different loads and coal quality conditions, and can establish a mapping relationship between load and the optimal set of operating parameters, achieving full-condition optimization. While reducing energy consumption, it ensured that sulfur dioxide emissions met standards, and reduced greenhouse gas emissions (due to reduced coal consumption) and solid waste emissions (due to reduced limestone consumption). Attached Figure Description

[0022] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart of a multi-variable coordinated energy intensity optimization method for fluidized bed desulfurization process provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a fluidized bed desulfurization system provided in an embodiment of the present invention; Figure 3 This is a flowchart illustrating the multi-objective optimization algorithm solution provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a multi-variable synergistic energy intensity optimization device for a fluidized bed desulfurization process provided in an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0027] Example 1 like Figure 1 As shown, the present invention provides a multi-variable synergistic energy intensity optimization method for fluidized bed desulfurization processes, comprising: S100 is defined as a comprehensive energy consumption intensity index for desulfurization that can uniformly convert the heat consumption penalty, power consumption and limestone embodied energy in the desulfurization process into standard energy consumption.

[0028] To scientifically evaluate the energy efficiency of desulfurization processes, this application utilizes energy quality balance to identify energy losses at each stage and proposes the comprehensive energy consumption intensity index for desulfurization. This comprehensive energy consumption intensity index uniformly converts different types of energy consumption (heat, electricity, material consumption, and implicit energy) related to the desulfurization process onto the same benchmark, enabling fair comparison and optimization.

[0029] like Figure 2 As shown, the fluidized bed desulfurization system includes the following subsystems: The reaction occurs in the combustion chamber. Fuel (coal), desulfurizing agent (limestone), and circulating ash are in a fluidized state within the combustion chamber, where they mix and burn vigorously.

[0030] The coal feeding subsystem includes: raw coal bunker, coal feeder, and coal chute.

[0031] The desulfurizing agent preparation and feeding subsystem includes: crushing and grinding equipment for crushing and grinding limestone to the required particle size, such as a hammer crusher, Raymond mill, or vertical mill; a pneumatic conveying device, such as a Roots blower, for conveying limestone powder through pipelines and injecting it into a set location in the combustion chamber; such as a return pipe, secondary air nozzle, or lower part of the combustion chamber.

[0032] The flue gas subsystem includes: a primary air fan that provides high-pressure air for fluidization at the bottom of the combustion chamber; a secondary air fan that provides air for fuel combustion at the top of the combustion chamber and enables staged combustion; an induced draft fan that maintains negative pressure in the combustion chamber, extracts the flue gas, and sends it to the tail flue; and an air preheater that uses waste heat from the flue gas to heat the primary and secondary air to improve the efficiency of the fluidized bed furnace.

[0033] The material recycling subsystem includes: a cyclone separator located at the combustion chamber outlet to separate a large number of solid particles (unburned carbon, limestone, ash) carried in the flue gas; and a return feeder that seals the high-temperature material separated by the cyclone separator through mechanical or non-mechanical means and sends it back to the combustion chamber to form a circulating combustion.

[0034] The tail-end flue gas treatment subsystem includes: a dust collector, such as an electrostatic precipitator or a bag filter, to further remove fine dust from the flue gas; and an induced draft fan and chimney to discharge the purified flue gas into the atmosphere.

[0035] To analyze energy consumption, the material and energy flows of the fluidized bed desulfurization system were determined. The material flow during the desulfurization process is as follows: coal, limestone, and air enter the fluidized bed desulfurization furnace. The output contains nitrogen, carbon dioxide, water vapor, oxygen, sulfur dioxide (trace amounts, meeting emission standards), and flue gas containing nitrogen oxides. Fine ash not captured by the cyclone separator is also output, containing unreacted CaO, calcium sulfate, and inert ash. Coarse slag discharged from the bottom of the combustion chamber, containing large ash particles and a small amount of unreacted limestone nuclei, is also output. Gypsum is also discharged mixed with the ash.

[0036] Energy flow analysis during desulfurization: The energy provided includes: fuel chemical energy; physical heat of materials introduced into the furnace by coal, limestone, and air; and electrical energy consumed by fans, crushers, feeders, etc., which is ultimately converted into heat or dissipated. Output and loss of energy include: effectively utilized heat: the heat load for generating steam or hot water; heat carried away by flue gas; heat loss due to unburned combustible gases in the flue gas; heat loss due to unburned carbon in ash (fly ash, bottom ash); heat dissipated to the environment by furnace walls, pipes, etc.; heat carried away by high-temperature ash discharged from the furnace; and heat absorption during limestone calcination, a significant endothermic process that lowers the combustion chamber temperature and increases coal consumption. While the exothermic sulfidation reaction exists, its heat output is far less than the heat absorbed during calcination. The concentration is low, so the net heat effect is endothermic.

[0037] Based on the above material and energy analysis, the comprehensive energy consumption intensity of desulfurization is defined as the energy consumption intensity at which desulfurization is guaranteed. Under the premise of meeting emission standards, the fluidized bed furnace system removes a unit mass of The formula for calculating the total amount of standard coal consumed is as follows: ; In the formula: for Removal amount, ,in For coal feed rate, For basic sulfur content, For desulfurization efficiency, K represents the conversion of sulfur to sulfur. The coefficient. The energy consumption penalty for combustion efficiency refers to the standard coal consumption converted from the decrease in fluidized bed furnace thermal efficiency caused by desulfurization (such as controlling bed temperature and excessive air distribution), calculated using the boiler efficiency under actual operating conditions. Compared with theoretical optimal efficiency The calculation yielded: ; The total calorific value of the coal fed into the furnace is 29.3076 MJ / kgce, which is the standard coal calorific value.

[0038] The power consumption of auxiliary equipment, including primary air fans, secondary air fans, induced draft fans, limestone grinding and conveying equipment, can be calculated based on the rated power and operating current of the equipment, or obtained through DCS data. For auxiliary equipment power consumption calculation, This is the standard coal equivalent coefficient for electricity. This refers to the amount of limestone consumed. This represents the embodied energy per unit of limestone. As a material, limestone itself consumes energy during its mining, crushing, and transportation processes. The embodied energy generated during limestone processing is included in the total energy consumption for desulfurization.

[0039] The comprehensive energy consumption intensity index for desulfurization reflects three types of energy consumption in the desulfurization process of the fluidized bed furnace system: the interference and loss of desulfurization on the combustion process itself; the power consumption of auxiliary equipment such as conveying and crushing; and the front-end energy consumption of resources.

[0040] The optimization objective is to minimize the comprehensive energy consumption intensity index of desulfurization while meeting emission constraints, and to maximize the energy efficiency of the desulfurization process.

[0041] S200 establishes a fluidized bed desulfurization process model that includes a limestone calcination model, a sulfidation reaction model, a fluidized bed gas-solid two-phase fluid dynamics model, and a fluidized bed heat balance model, in order to model the mechanism of the decision variables of the fluidized bed process.

[0042] The decision variables include: bed temperature, calcium-to-sulfur ratio, primary air ratio, and limestone particle size. These variables interact and collectively determine the final desulfurization efficiency and overall energy consumption. The optimization process requires defining the value ranges for the decision variables: bed temperature (optimization range: 850℃ to 950℃); calcium-to-sulfur ratio (optimization range: 1.5 to 3.0); primary air ratio (the proportion of primary air volume to total air volume) (optimization range: 0.4 to 0.7); and average limestone particle size (optimization range: 0.1 mm to 1.0 mm).

[0043] Bed temperature affects desulfurization efficiency, and the sulfidation reaction in desulfurization... It is a gas-solid catalytic reaction. According to the Arrhenius equation, the reaction rate constant has an exponential relationship with the bed temperature. As the temperature increases, the reaction rate accelerates. However, at the same time... At high temperatures (>1000℃), CaO decomposes, and the CaO particles sinter, reducing porosity and specific surface area. Therefore, there is an optimal temperature window for the bed. Below this window, the kinetics of the desulfurization reaction are limited; above this window, the desulfurizing agent deactivates. Bed temperature affects the efficiency of the fluidized bed furnace. Excessively high bed temperatures may lead to coking inside the furnace, requiring adjustment by reducing the bed thickness or increasing excess air, which reduces combustion efficiency or increases flue gas losses. Conversely, excessively low bed temperatures directly increase heat loss due to unburned carbon in the ash (fly ash, bottom ash). Experience shows that for every 50℃ decrease in bed temperature, the carbon content of fly ash may increase by 0.5-1 percentage point.

[0044] The construction of a hybrid-driven energy consumption proxy model involves generating virtual training data covering a wide range of operating conditions using a desulfurization model, combining this data with actual DCS operation data to obtain training data for training the energy consumption model. This allows for the training of a high-precision, high-generalization-ability deep neural network-based proxy model, enabling rapid prediction of energy intensity and desulfurization efficiency.

[0045] Research on Multivariate Collaborative Optimization Strategies: With energy intensity minimization and emission compliance as constraints, a multi-objective genetic algorithm (NSGA-II) is employed to globally optimize key operational variables. The optimal operating domain under different load conditions is analyzed, and collaborative control rules applicable to engineering are extracted.

[0046] The calcium-to-sulfur ratio (CTR) refers to the ratio of the number of moles of calcium in the limestone added to the number of moles of sulfur in the coal. Under given conditions (bed temperature, limestone particle size, residence time), the CTR increases with increasing CTR, but exhibits a clear law of diminishing marginal returns, which can be described by the Langmuir equation. The CTR also affects heat consumption penalty, as limestone calcination is endothermic. An increase in the CTR means an increase in the amount of calcium carbonate entering the fluidized bed furnace, leading to increased heat absorption during calcination. This heat absorption requires additional coal to compensate. According to heat balance, each additional 1 mol of calcium carbonate requires approximately 178 kJ of heat, equivalent to about 0.00607 kgce. If this heat is provided by coal, it will correspondingly increase carbon dioxide emissions and coal consumption, further increasing carbon dioxide generation, forming a small positive feedback loop. Finally, the CTR affects power consumption penalty. An increase in the CTR leads to an increase in limestone feed, which in turn increases the load on the crusher and conveyor fans, resulting in higher power consumption. The calcium-to-sulfur ratio affects the energy consumption of materials, and the direct energy consumption of materials is directly proportional to the calcium-to-sulfur ratio.

[0047] The primary air in a fluidized bed desulfurization system is responsible for fluidizing the bed material and providing the oxygen required for initial fuel combustion. Secondary air is introduced in stages, responsible for fuel burnout and forming a reduction zone to lower nitrogen oxide content. The primary air ratio affects desulfurization efficiency, impacting fluidization quality and particle separation. A high primary air ratio results in a large fluidization number, leading to more fine particles (including fine limestone) being rapidly entrained from the fluidized bed, reducing their residence time in the dense phase zone and decreasing desulfurizer utilization. A low primary air ratio results in poor fluidization, potentially causing bed stratification or coking. The primary air ratio also affects oxygen concentration distribution: the penetration depth and mixing effect of the secondary air determine the oxygen concentration distribution in the upper part of the combustion chamber. Sulfidation requires oxygen; if the secondary air stage is too strong, it can lead to localized oxygen deficiency in the upper part of the combustion chamber (dilute phase zone), inhibiting the desulfurization reaction in that area. The primary air ratio affects energy consumption: primary and secondary air fans consume electricity to supply primary and secondary air, with fan power consumption accounting for a large proportion of total electricity consumption. When the total air volume is fixed, changing the primary air ratio will affect the fan's operating point because the primary air fan requires a higher pressure head, while the secondary air fan requires a lower pressure head. The air volume ratio directly affects the excess air coefficient and staged combustion effect, influencing flue gas heat loss and chemical / mechanical incomplete combustion losses. The primary air ratio affects the oxygen concentration distribution and particle residence time in the furnace, which in turn affects the required effective calcium-sulfur ratio and the selection of the optimal bed temperature.

[0048] Limestone particle size affects desulfurization efficiency. According to the core shrinkage model, the CaO inside large limestone particles is not fully utilized, and the reaction time is proportional to the square of the particle size. Reducing the particle size can significantly improve the reaction rate and final conversion rate. However, if the particle size is too small, it becomes precipitated ash, which has a very short residence time in the furnace and is carried away by the flue gas before it can react. Therefore, there exists an optimal particle size range that matches the reaction time with the residence time. Limestone particle size also affects energy consumption. The smaller the particle size requirement, the more crushing stages are required, increasing the power consumption of grinding. Particle size affects the efficiency of cyclone separators. Coarser particles are more easily captured and returned by cyclone separators, gaining more reaction time; extremely fine particles pass through in one go, resulting in a short reaction time and wasting limestone, which indirectly increases the calcium-sulfur ratio required to maintain the same desulfurization efficiency, thus increasing material and energy consumption. For smaller particle sizes, the reaction is more kinetically controlled, and the optimal reaction temperature may be slightly lower than that for larger particle sizes. At the same time, particle size distribution also affects the fluidization characteristics of the bed.

[0049] Modeling of fluidized bed desulfurization process: The desulfurization process in a fluidized bed furnace involves complex gas-solid reactions and multiphase flows. To accurately describe this process, this application establishes a fluidized bed furnace desulfurization process model that couples microscopic reaction kinetics with macroscopic fluid dynamics.

[0050] Limestone calcination model: After entering the combustion chamber, the limestone undergoes a calcination reaction. The limestone calcination reaction is a strongly endothermic reaction, and the reaction rate is affected by temperature and carbon dioxide partial pressure. The calcination product, CaO, has a porous structure and serves as an adsorbent for the subsequent sulfidation reaction. The limestone calcination process is described by considering the following steps: heat is transferred from the gas flow to the particle surface; heat is conducted within the particles; chemical decomposition occurs at the reaction interface; carbon dioxide diffuses through the product layer (CaO layer) to the particle surface; and carbon dioxide diffuses from the particle surface into the bulk gas flow. Due to the small size of the limestone particles and the strong convective heat transfer within the furnace, the internal temperature gradient of the particles is ignored, and the limestone calcination process is modeled using isothermal particles. The particles are assumed to be spheres with an initial radius of... initial density The limestone calcination reaction begins on the outer surface and gradually progresses inward, forming unreacted nuclei. Based on the random pore model, the limestone conversion rate... The change over time can be described by the following equation: ; in: The surface reaction rate constant (m / s) during the calcination process depends on the bed temperature and is given by the Arrhenius equation. Give the activation energy of the calcination reaction. Between 170 and 210 kJ / mol, the pre-exponential factor It is related to the type of limestone. This represents the initial specific surface area of ​​the limestone. The pore structure parameters are determined by the initial porosity and pore length, and are obtained by fitting the pore structure data of limestone after calcination through experiments such as mercury intrusion porosimetry. Z1 is a parameter related to the diffusion resistance of the product layer during calcination; Z2 is the molar volume ratio of the product to the reactants during the calcination reaction.

[0051] The sulfidation reaction model describes the process by which CaO produced during calcination undergoes a sulfidation reaction with sulfur dioxide and oxygen in the flue gas. Sulfidation reaction: ; Sulfidation is an exothermic reaction, but the amount of heat released is much less than the amount of heat absorbed during calcination. Sulfidation is also a gas-solid reaction, but the dense CaSO4 product layer severely hinders the diffusion of sulfur dioxide into the interior. Therefore, sulfidation typically follows the behavior of unreacted core contraction, but is significantly controlled by the diffusion of the product layer.

[0052] The vulcanization reaction process is described using a stochastic pore model that considers pore blockage. This model introduces variations in structural parameters over time, relating the reaction rate to pore surface area and diffusion resistance.

[0053] Establish pore structure variation models, including effective pore size variation models, effective surface area variation models, and porosity variation models.

[0054] When considering pore blockage, due to product layer accumulation, the effective pore radius r decreases with reaction time, expressed as: ,in, It refers to the thickness of the product layer. When the pore radius shrinks to a critical value... When the pore throat is blocked, that part of the pore no longer participates in the reaction. This means the effective porosity... and effective surface area It decays faster.

[0055] Effective surface area When considering clogging, the conversion rate is a function of the orifice connectivity; a clogging probability function is introduced. describe: ,in It is related to the local conversion rate or the thickness of the product layer. When the thickness of the local product layer reaches the pore throat radius, =1; The surface area is not considered when clogging occurs.

[0056] When considering blockage, the change in effective porosity can be described by the effective surface area: ; in, is the pore shrinkage rate constant, which controls the rate at which porosity decreases; f is the molar volume effect function, which reflects the effect of the ratio of the molar volume of solid product to reactant on pore shrinkage. When the molar volume of product is greater than that of reactant, f is positive, leading to a decrease in porosity. The reaction rate R is directly proportional to the effective surface area: ,in, is the reaction rate constant, C is the reactant concentration, and n is the order.

[0057] By incorporating porosity to introduce diffusion resistance for sulfur dioxide, the most significant effect of pore blockage is the increased resistance to the diffusion of gaseous reactants into the particle interior. The effective diffusion coefficient of sulfur dioxide... Represented as: ; in The intrinsic diffusion coefficient is the diffusion coefficient of a gas in a completely open channel or free space. The initial porosity, It is an empirical, structurally sensitive parameter. This reflects the degree to which a decrease in porosity hinders the diffusion path. Used to describe the increased tortuosity caused by congestion.

[0058] The differential equation for the reaction rate of gas-phase sulfur dioxide is obtained by using the law of mass conservation, the effective diffusion coefficient, and the reaction rate: .

[0059] The above model describes the reaction during the sulfidation process. In the initial stage, the reaction occurs at the pore surface, pores grow, surface area increases, and the reaction accelerates. In the intermediate stage, products begin to accumulate, and some micropores are blocked. According to the characteristics of the random pore model, the reaction rate reaches its peak and then begins to decline. At this point, the structural parameters in the model are actually dynamically increasing (because blockage increases the irregularity of the structure). In the final stage, due to severe pore blockage, the effective diffusion coefficient of sulfur dioxide drops to almost zero. A dense product layer may form on the outer surface of the particles, making it impossible for reactants to reach the internal active sites (due to incomplete chemical transformation), even though these sites may still exist, thus macroscopically terminating the reaction.

[0060] Using a stochastic pore model that considers pore blockage, structural parameters and diffusion coefficients are treated as variables that change with the reaction process (conversion rate or time). The essence of reaction termination is the formation of diffusion dead zones. The model correlates the sharp drop in macroscopic reaction rate with the physical blockage of micropores by quantifying the diffusion barrier caused by porosity decay.

[0061] In a fluidized bed furnace, the residence time distribution of particles, the gas-solid contact mode, and the temperature and concentration distribution within the furnace have a decisive impact on the overall desulfurization efficiency. Therefore, it is necessary to establish a macroscopic gas-solid two-phase fluid dynamics model for the fluidized bed furnace. This model includes: Two-phase sub-model: The two-phase sub-model divides the bed into a bubble phase and an emulsion phase and describes the mass exchange between the two phases; the bubble phase moves upward in the form of bubbles, contains a small amount of solids, and the upward flow of gas is plug flow; the emulsion phase contains dense solids, the gas flow velocity is close to the minimum fluidization velocity, and the gas flow is regarded as fully mixed or plug flow. The mass exchange between the two phases is determined by the mass transfer coefficient between the bubble phase and the emulsion phase.

[0062] A sub-model of particle residence time distribution is needed, as the residence time distribution of particles within the fluidized bed furnace is crucial for determining the utilization rate of the desulfurizing agent. Particles undergo multiple cycles within the furnace, resulting in a total residence time significantly longer than a single pass. Typically, a multi-stage series fully mixed-flow reactor model or a plug flow model with backflow is used for approximation. It is assumed that the combustion chamber, cyclone separator, and return feeder constitute a closed-loop circulation system. The particle age distribution function is also considered. This can be obtained by solving the mass balance of the circulating circuit. Let the average time for a particle to pass through the combustion chamber once be... The cyclone separator efficiency is The average number of cycles for the particles is Total average stay For fine particles, Lower concentrations result in shorter residence times for coarse particles; higher concentrations result in longer residence times for coarse particles. In practical modeling, the combustion chamber is considered as a reactor with a certain degree of backmixing, and an axial diffusion model is used to describe the axial concentration distribution of particles, thereby obtaining the residence time distribution of particles of different sizes.

[0063] Gas-solid phase proton transfer model: Sulfur dioxide must overcome gas film resistance to transfer from the bulk gas phase to the surface of CaO particles. In a fluidized bed, due to the vigorous movement of the particles, the gas film resistance is usually small and can be approximated as negligible, meaning that the sulfur dioxide concentration on the particle surface is considered equal to the concentration in the bulk gas phase.

[0064] The fluidized bed furnace hydrodynamic sub-model is simplified to a one-dimensional axial model. The combustion chamber is divided into several micro-elements along its height. Within each micro-element, uniform gas-solid mixing is assumed. A system of ordinary differential equations is established based on mass and energy balance. With the combustion chamber height z as the independent variable, the axial distribution of sulfur dioxide concentration and bed temperature is established. Specifically, based on the gas-phase sulfur dioxide mass balance, the sulfur dioxide concentration varies with the combustion chamber height z as follows: ; in To represent apparent air velocity, The overall mass transfer coefficient is... This refers to the specific surface area of ​​the particles. This represents the particle surface concentration (usually assumed to be 0 because the reaction is very fast). This is a gas-phase homogeneous reaction (neglected).

[0065] Based on the axial distribution of sulfur dioxide concentration and bed temperature, the axial variation of solid phase (CaO) conversion rate is obtained: ; in The average particle velocity is taken into account, considering the internal circulation of particles within the circulating fluidized bed.

[0066] These equations need to be solved in conjunction with the heat balance equation.

[0067] Fluidized bed furnace heat balance model: Heat balance is key to linking combustion, heat transfer, and desulfurization. The temperature distribution inside the furnace directly affects the reaction rate, while the endothermic / exothermic effects of the desulfurization reaction in turn affect the temperature. Therefore, it is necessary to establish coupled heat balance equations.

[0068] For the entire combustion chamber, the heat input includes: heat released during coal combustion. Air physical heat Physical heat of returned ash Limestone physical heat .

[0069] Heat output includes: heat carried away by flue gas Physical heat of fly ash and bottom ash Combustion chamber heat dissipation The desulfurization reaction is net endothermic. (The difference between the heat absorbed by calcination and the heat released by sulfidation); the evaporation heating surface absorbs heat. (Effectively utilize a portion of the heat).

[0070] Therefore, the heat balance equation can be written as: ; In actual operation, bed temperature is the main controlled variable, maintained stable by adjusting the coal feed rate and air volume. Therefore, the heat balance model is transformed into a differential equation regarding bed temperature for dynamic simulation. Both radiation and convection heat transfer occur simultaneously within the fluidized bed furnace combustion chamber. Particle convection dominates in the dense phase region, while radiation increases in the dilute phase region. Considering the zoning of bed emissivity parameters, a further simplification is achieved using the lumped parameter method, employing empirical formulas to calculate the overall heat transfer coefficient, thus linking the heat absorption of the evaporation heating surface to the bed temperature.

[0071] The limestone calcination model, sulfidation reaction model, fluidized bed furnace gas-solid two-phase fluid dynamics model, and fluidized bed furnace heat balance model are coupled to form a holistic mathematical model describing the fluidized bed furnace desulfurization process. The following solution strategy is adopted: the combustion chamber is divided into several control volumes along the axial direction; within each control volume, it is assumed that the gas phase concentration, solid phase conversion rate, and temperature are uniform; a set of ordinary differential equations for mass balance and energy balance are established for each control volume; given boundary conditions (inlet coal quantity, limestone quantity, air volume, return material quantity, etc.), an iterative method is used to solve the steady-state solution; for dynamic simulations, numerical integration methods (such as the Euler method and the Runge-Kutta method) are used to solve the differential equations. The desulfurization efficiency, bed temperature, and flue gas temperature calculated by the fluidized bed furnace desulfurization process model can be used as sample data for subsequent energy consumption models.

[0072] S300 uses a desulfurization model to generate virtual training data covering a wide range of operating conditions. Combined with actual fluidized bed furnace DCS operation data, training data for an energy consumption model based on a deep neural network is obtained. The energy consumption model predicts the comprehensive desulfurization energy intensity index and desulfurization efficiency based on decision variables.

[0073] While fluidized bed desulfurization process models can reveal the inherent laws governing the desulfurization process, their computational complexity makes them insufficient for real-time optimization control. Therefore, a fast energy consumption model based on deep neural networks is needed. This application employs a hybrid approach of using a fluidized bed desulfurization process model and data-driven methods. The model generates a large number of virtual samples, which are then combined with actual operational data to train a deep neural network-based energy consumption model for rapid prediction of desulfurization efficiency and overall desulfurization energy intensity.

[0074] The actual operating data uses historical data from the actual DCS: long-term operating data is extracted from the distributed control system (DCS) of the target fluidized bed furnace, including: coal feed rate, limestone feed rate, primary and secondary air volume, bed temperature, differential pressure inside and outside the combustion chamber, flue gas temperature, flue gas carbon dioxide concentration, flue gas oxygen content, load, etc. The fluidized bed furnace desulfurization process model performs extensive simulation calculations (such as Latin hypercube sampling) within the set parameter range to generate input-output datasets covering various operating conditions, compensating for boundary conditions that are difficult to cover in the actual data.

[0075] Real-world DCS data often contains noise, outliers, and dynamic transitions, which can affect accuracy if used directly for modeling. Therefore, preprocessing is necessary, including: using the 3σ criterion or box plot method to remove samples deviating from the normal range; applying moving averages or low-pass filtering to time-series data to reduce random noise; using a sliding window variance test to screen for steady-state segments where parameter changes are less than a set threshold to reflect characteristics under steady-state conditions; scaling variables to the [0,1] or [-1,1] interval to eliminate dimensionality effects; and calculating the correlation coefficients between each variable and the target variable (such as sulfur dioxide emission concentration and energy intensity), removing weakly correlated variables, and reducing the input dimensionality.

[0076] Given the highly nonlinear input-output relationship and large data volume in the fluidized bed desulfurization process, a multi-layer feedforward neural network was chosen as the energy consumption model. The energy consumption model includes: Input layer: The number of nodes is equal to the number of selected feature variables, including: bed temperature, calcium-to-sulfur ratio, primary air rate, average limestone particle size, fluidized bed load, and coal sulfur content. Hidden layer: The number of nodes in the hidden layer is set to 2-3 times the number of input nodes, and the optimal value is determined through trial and error. ReLU or tanh is selected as the activation function. Output layer: The output layer has two nodes, corresponding to desulfurization efficiency and overall desulfurization energy intensity, respectively.

[0077] The datasets obtained from simulation and actual operation were divided into a training set (70%), a validation set (15%), and a test set (15%). Mean squared error was used as the loss function, and Adam was selected as the optimizer. During training, the validation set error was monitored, and early stopping was employed to prevent overfitting.

[0078] Actual operating data often reflects the normal operation of the fluidized bed furnace, with limited boundary conditions. When training an energy consumption model purely based on data, the resulting model has limited generalization ability, and its predictions may become distorted when operating conditions exceed the training data range. While the fluidized bed furnace desulfurization process model is computationally slow, it ensures physical plausibility and can effectively supplement boundary condition data to train the energy consumption model, thereby enhancing its generalization ability.

[0079] Constraint embedding: Physical constraints (such as desulfurization efficiency cannot exceed 1 and energy consumption intensity must be positive) are added to the loss function of the neural network to make the prediction conform to physical common sense.

[0080] S400 utilizes an energy consumption model to construct a multivariate collaborative optimization model, with the objective function being to minimize the comprehensive energy consumption intensity of desulfurization, and the constraints being sulfur dioxide emission concentration, bed temperature safety range, and equipment constraints.

[0081] Based on the energy consumption model, a multivariate collaborative optimization algorithm is designed to minimize energy intensity while satisfying emission and safety constraints, and to find the optimal combination of operating parameters.

[0082] The decision variables for multivariate collaborative optimization problems include: Bed temperature, range [850, 950], calcium-sulfur ratio, range [1.5, 3.0], primary air rate (proportion of primary air volume to total air volume), range [0.4, 0.7], average limestone particle size, range [0.1, 1.0].

[0083] In addition, the fluidized bed furnace load is determined by external demand and is used as a fixed parameter during optimization, with optimization performed separately for different loads.

[0084] The objective function of the multivariate collaborative optimization problem is: The primary objective is to minimize the overall energy intensity index of desulfurization. ; The components of the comprehensive energy consumption intensity index for desulfurization are calculated using a fluidized bed desulfurization process model.

[0085] Secondary objectives may include maximizing desulfurization efficiency. .

[0086] The constraints of multivariable collaborative optimization problems include: Emission constraints: The concentration of sulfur dioxide at the outlet is lower than the emission standard. (Ultra-low emission standards), or according to specific environmental protection requirements.

[0087] Safety constraints: Lower limit of bed temperature upper limit The lower limit of the primary air rate ensures fluidization, while the upper limit prevents blow-through.

[0088] Equipment constraints: Maximum output limit of the blower, corresponding air volume range. Maximum feeding rate of the limestone feeder.

[0089] Process constraints: Desulfurization efficiency not exceeding 1, calcium-sulfur ratio not less than 1.

[0090] The solution to a multi-objective optimization problem is not a single optimal solution, but a set of Pareto optimal solutions. Each point on the Pareto front corresponds to a trade-off. The NSGA-II algorithm solves multi-objective optimization problems and has the following characteristics: A parent population of a predetermined size is randomly generated. Each individual is represented as a vector of decision variables.

[0091] The fitness is calculated using the comprehensive energy consumption intensity index of desulfurization and the desulfurization efficiency, and the optimization direction is unified by operations such as taking the negative or reciprocal of the desulfurization efficiency.

[0092] The parent population is sorted non-dominated, and each individual is assigned rank and crowding distance.

[0093] Parents are selected through tournament selection based on rank and crowding degree. The rules of tournament selection are: when the ranks are the same, individuals with larger crowding degree distances are selected; when the ranks are different, individuals with smaller ranks are selected. Simulated binary crossover and polynomial mutation are performed to generate offspring populations of equal size.

[0094] By merging the parent and offspring generations, the overall population size is doubled. The merged population is then subjected to a non-dominated sort, and individuals with the largest distance in that layer are selected based on crowding distance to fill the set population size, forming a new population.

[0095] Iterate until the maximum number of generations is reached. Finally, return to the Pareto front.

[0096] Although NSGA-II can directly search for the Pareto front, post-analysis is still required to reveal synergistic relationships between variables. Specific examples include: Local sensitivity analysis determines the impact of individual variable changes on energy consumption intensity and identifies key variables under different loads, using either of the following methods: 1) Calculating the partial derivatives of the objective function with respect to each input variable using an energy consumption model, and then calculating the gradient near the optimal solution; or 2) Performing global sensitivity analysis to assess the main effects and interaction effects of each variable.

[0097] Optimization is performed for different boiler load conditions, and a mapping table between load and the optimal set of operating parameters is established. The Pareto front obtained by offline operation of the optimization algorithm is embedded into the real-time control system. When the operating conditions change, the corresponding optimal parameter setpoint is selected by interpolation from the Pareto front based on the current load and coal quality information, and used as the setpoint of the underlying controller.

[0098] Example 2 like Figure 4 As shown, this embodiment of the invention provides a multi-variable collaborative energy intensity optimization device for a fluidized bed desulfurization process, comprising: at least one processing unit, the processing unit being connected to a storage unit via a bus unit, the storage unit serving as a computer-readable storage medium, and capable of storing software programs, computer-executable programs, and modules, such as the software program, computer-executable program, and modules corresponding to the multi-variable collaborative energy intensity optimization method for a fluidized bed desulfurization process in this embodiment of the invention. The processing unit implements the aforementioned multi-variable collaborative energy intensity optimization method for a fluidized bed desulfurization process by running the software program, computer-executable program, and modules stored in the storage unit, including: A fluidized bed desulfurization process model was established, which includes a limestone calcination model, a sulfidation reaction model, a fluidized bed gas-solid two-phase fluid dynamics model, and a fluidized bed heat balance model, in order to model the mechanism of the decision variables of the fluidized bed process. Virtual training data covering a wide range of operating conditions is generated using a desulfurization model. Combined with actual fluidized bed furnace DCS operation data, training data for an energy consumption model based on a deep neural network is obtained. The energy consumption model predicts energy consumption intensity and desulfurization efficiency based on decision variables. A multivariate collaborative optimization model was constructed with the objective function of minimizing the comprehensive energy consumption intensity of desulfurization, and the constraints of sulfur dioxide emission concentration, bed temperature safety range and equipment constraints. The multi-objective optimization algorithm is used to solve the multivariate collaborative optimization model to obtain the optimal combination of operating parameters, and the optimal combination of operating parameters is applied to the fluidized bed furnace control system.

[0099] Example 3 This invention provides a computer-readable storage medium storing a computer program, characterized in that, when executed, the computer program implements a multi-variable coordinated energy intensity optimization method for the fluidized bed desulfurization process, comprising: A fluidized bed desulfurization process model was established, which includes a limestone calcination model, a sulfidation reaction model, a fluidized bed gas-solid two-phase fluid dynamics model, and a fluidized bed heat balance model, in order to model the mechanism of the decision variables of the fluidized bed process. Virtual training data covering a wide range of operating conditions is generated using a desulfurization model. Combined with actual fluidized bed furnace DCS operation data, training data for an energy consumption model based on a deep neural network is obtained. The energy consumption model predicts energy consumption intensity and desulfurization efficiency based on decision variables. A multivariate collaborative optimization model was constructed with the objective function of minimizing the comprehensive energy consumption intensity of desulfurization, and the constraints of sulfur dioxide emission concentration, bed temperature safety range and equipment constraints. The multi-objective optimization algorithm is used to solve the multivariate collaborative optimization model to obtain the optimal combination of operating parameters, and the optimal combination of operating parameters is applied to the fluidized bed furnace control system.

[0100] In the embodiments provided by this invention, it should be understood that the disclosed structures and methods can be implemented in other ways. For example, the structural embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another assembly system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, structures, or units, and may be electrical, mechanical, or other forms.

[0101] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0102] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0103] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for optimizing energy intensity in a fluidized bed desulfurization process using multi-variable synergy, characterized in that, include: A fluidized bed desulfurization process model was established, which includes a limestone calcination model, a sulfidation reaction model, a fluidized bed gas-solid two-phase fluid dynamics model, and a fluidized bed heat balance model, in order to model the mechanism of the decision variables of the fluidized bed process. Decision variables include: bed temperature, calcium-to-sulfur ratio, primary air rate, and average limestone particle size; Virtual training data covering a wide range of operating conditions is generated using a desulfurization model. Combined with actual fluidized bed furnace DCS operation data, training data for an energy consumption model based on a deep neural network is obtained. The energy consumption model predicts the comprehensive energy consumption intensity index and desulfurization efficiency of desulfurization based on decision variables. A multivariate collaborative optimization model was constructed using an energy consumption model, with the objective function being to minimize the comprehensive energy consumption intensity of desulfurization and maximize the desulfurization efficiency, and the constraints being sulfur dioxide emission concentration, bed temperature safety range, and equipment constraints. The multi-objective optimization algorithm is used to solve the multivariate collaborative optimization model to obtain the optimal combination of operating parameters, and the optimal combination of operating parameters is applied to the fluidized bed furnace control system.

2. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 1, characterized in that, The description of the limestone calcination process includes the following steps: heat is transferred from the gas flow to the particle surface; chemical decomposition occurs at the reaction interface; carbon dioxide diffuses through the product layer to the particle surface; carbon dioxide diffuses from the particle surface to the bulk gas flow; assuming the particles are spheres with an initial radius... initial density The limestone calcination reaction begins on the outer surface and gradually progresses inward, forming unreacted nuclei. Based on the random pore model, the limestone conversion rate... The change over time can be described by the following equation: ； in: The surface reaction rate constant during the calcination process depends on the bed temperature; This represents the initial specific surface area of ​​the limestone. The pore structure parameters are determined by the initial porosity and pore length, and are obtained by fitting the pore structure data of limestone after calcination through experiments such as mercury intrusion porosimetry. Z1 is a parameter related to the diffusion resistance of the product layer during calcination; Z2 is the molar volume ratio of the product to the reactants during the calcination reaction.

3. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 1, characterized in that, In the vulcanization reaction model, a stochastic pore model considering pore blockage introduces the variation of structural parameters over time, relating the reaction rate to pore surface area and diffusion resistance; wherein, the variation of structural parameters over time includes: When considering pore blockage, due to product layer accumulation, the effective pore radius r decreases with reaction time, expressed as: ,in, It refers to the thickness of the product layer when the pore radius shrinks to a critical value. At that time, the throat of the orifice is blocked, and the orifice no longer participates in the reaction; Considering congestion, the congestion probability function is used. Describe the effective surface area : ,in It is related to the local conversion rate or the thickness of the product layer. When the thickness of the local product layer reaches the pore throat radius, =1; The surface area is not considered when clogging occurs; When considering blockage, effective porosity is described by effective surface area: ； in, is the pore shrinkage rate constant, which controls the rate at which porosity decreases; f is the molar volume effect function, which reflects the effect of the ratio of the molar volume of solid product to reactant on pore shrinkage. When the molar volume of product is greater than that of reactant, f is positive, leading to a decrease in porosity. Correlating reaction rate with pore surface area and diffusion resistance includes: The reaction rate R is directly proportional to the effective surface area: ,in, is the reaction rate constant, C is the reactant concentration, and n is the order; The effective diffusion coefficient of sulfur dioxide is determined by introducing diffusion resistance to sulfur dioxide through porosity. Represented as: ； in The intrinsic diffusion coefficient is the diffusion coefficient of a gas in a completely open channel or free space. The initial porosity, It is an empirical, structurally sensitive parameter. This reflects the degree to which a decrease in porosity hinders the diffusion path. Used to describe the increased tortuosity caused by congestion; The differential equation for the reaction rate of gas-phase sulfur dioxide is obtained by using the law of mass conservation, the effective diffusion coefficient, and the reaction rate: 。 4. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 1, characterized in that, The fluid dynamics model of the fluidized bed furnace gas-solid two-phase fluid includes: a two-phase sub-model that divides the bed into a bubble phase and an emulsion phase and describes the mass exchange between the two phases; a particle residence time distribution sub-model that describes the residence time distribution of particles in the fluidized bed furnace; and a fluid dynamics simulation sub-model of the fluidized bed furnace.

5. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 4, characterized in that, The combustion chamber, cyclone separator, and return feeder constitute a closed-loop circulation system; the particle age distribution function is obtained by solving the mass balance of the closed-loop circulation system, including: assuming the average time for a particle to pass through the combustion chamber once is... The cyclone separator efficiency is The average number of cycles for the particles is Total average stay The axial concentration distribution of particles is described by an axial diffusion model, thereby obtaining the residence time distribution of particles of different sizes.

6. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 4, characterized in that, The fluid dynamics sub-model of the fluidized bed furnace is simplified into a one-dimensional axial model. The combustion chamber is divided into several micro-elements along the height. It is assumed that the gas and solid are uniformly mixed in each micro-element. A set of ordinary differential equations is established through mass and energy balance. The axial distribution of sulfur dioxide concentration and bed temperature is established with the combustion chamber height z as the independent variable.

7. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 1, characterized in that, The solution strategy for the fluidized bed desulfurization process model includes: dividing the combustion chamber into several control volumes along the axial direction; assuming uniform gas phase concentration, solid phase conversion rate, and temperature within each control volume; establishing a set of ordinary differential equations for mass balance and energy balance for each control volume; given boundary conditions, including inlet coal quantity, limestone quantity, air volume, and return material quantity, and solving the steady-state solution using an iterative method; for dynamic simulation, solving the differential equations using a numerical integration method, and using the desulfurization efficiency, bed temperature, and flue gas temperature calculated by the fluidized bed desulfurization process model as sample data for subsequent energy consumption models.

8. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 1, characterized in that, The process of training the energy consumption surrogate model includes: using a multi-layer feedforward neural network as the surrogate model architecture; input layer nodes include bed temperature, calcium-sulfur ratio, primary air rate, average limestone particle size, boiler load, and coal sulfur content; output layer nodes include desulfurization efficiency and comprehensive desulfurization energy intensity index; using the fluidized bed desulfurization process model, Latin hypercube sampling is performed within the feasible region to generate virtual samples, which are then mixed with actual DCS historical data to form a training set; physical constraints are added to the loss function of the neural network to ensure that the prediction results conform to physical common sense.

9. The energy intensity optimization method for multi-variable synergy in fluidized bed desulfurization process according to claim 1, characterized in that, The formula for calculating the comprehensive energy consumption intensity index of desulfurization is as follows: ； in, This refers to the amount of sulfur dioxide removed. To reduce heat loss due to combustion efficiency; Power consumption for auxiliary equipment, including primary air fans, secondary air fans, induced draft fans, limestone grinding and conveying equipment; The standard coal equivalent coefficient for electricity; This refers to the amount of limestone consumed. This represents the unit implicit energy of limestone.

10. A multi-variable synergistic energy intensity optimization device for fluidized bed desulfurization process, comprising: At least one processing unit, the processing unit being connected to a storage unit via a bus unit, the storage unit storing a computer program that can run on a processor, characterized in that the processing unit implements the energy intensity optimization method for multivariate synergy in the fluidized bed desulfurization process as described in any one of claims 1-9 by running the computer program stored in the storage unit.

Images