Iron and steel industry ultra-low emission intelligent control system based on Internet of Things

Abstract

The invention relates to the technical field of emission control, and discloses a steel industry ultra-low emission intelligent control system based on the Internet of Things, and the system comprises a sequence construction module which is used for constructing a raw material sequence, a process sequence and an environment sequence; the prediction model training module is used for constructing a training sample of a prediction model; the objective function construction module is used for constructing a desulfurization efficiency model and a denitration efficiency model as objective functions; the code generation module randomly generates individual codes through a periodic mapping function; the adjustment list generation module outputs the codes of the individuals with the maximum fitness value in the initialized population as an adjustment list through an optimization algorithm; according to the method, the raw material parameters, the process parameters and the environmental parameters are comprehensively analyzed through the prediction model to predict the sulfur dioxide emission amount and the nitrogen oxide emission amount at multiple moments in the future, and the equipment parameters are automatically adjusted in combination with an optimization algorithm, so that the pollutant removal amount is maximized, the equipment energy consumption is minimized, and the equipment operation cost is minimized.

Description

Technical Field

[0001] The present invention relates to the technical field of emission control, and more specifically, it relates to an ultra-low emission intelligent control system for the iron and steel industry based on the Internet of Things. Background Art

[0002] The iron and steel industry is an important basic industry of the national economy and also the industrial field with the most concentrated energy consumption and air pollution emissions. In the iron and steel production process flow, it mainly includes key processes such as sintering, ironmaking, steelmaking, and rolling. Different types of air pollutants will be released during the processes of high-temperature smelting, fuel combustion, and material transportation in each link. Among them, the sintering process belongs to a high-temperature, open continuous combustion process, and a large amount of air is drawn through the sintering material layer by a suction fan, so it has become the most intensive link of pollutant emissions in iron and steel enterprises.

[0003] Currently, the pollution control of sintering tail gas mainly relies on treatment equipment such as desulfurization (FGD), denitrification (SCR / SNCR), and dust removal, and indirectly controls the pollutant emissions by adjusting parameters such as the suction fan frequency, ammonia injection amount, and desulfurization slurry injection amount. However, in the actual production process, the adjustment of the above parameters usually depends on the empirical judgment of operators, resulting in the following problems: 1. Parameter adjustment often based on the existing emission fluctuations, there is an obvious problem of adjustment response lag; 2. Workers adjust by experience, it is difficult to accurately match the current working conditions (such as factors such as raw material fluctuations and environmental changes), resulting in low pollutant treatment efficiency; 3. When pursuing emission compliance, it is easy to ignore the operating energy efficiency of equipment such as fans and spray systems, resulting in unnecessary energy consumption. Summary of the Invention

[0004] The present invention provides an ultra-low emission intelligent control system for the iron and steel industry based on the Internet of Things to solve the technical problems in the above background art.

[0005] The present invention provides an ultra-low emission intelligent control system for the iron and steel industry based on the Internet of Things, including: A sequence construction module, which is used to collect raw material parameters, process parameters, and environmental parameters simultaneously at preset time intervals within a preset time period, and respectively perform normalization to construct a raw material sequence, a process sequence, and an environmental sequence; A prediction model training module, which is used to use the raw material sequence, the process sequence, and the environmental sequence as sample data of the prediction model, and collect the sulfur dioxide emissions and nitrogen oxide emissions at the next N time moments as sample labels of the prediction model; A target function construction module, which is used to construct a desulfurization efficiency model and a denitrification efficiency model as target functions respectively according to the sulfur dioxide emissions, nitrogen oxide emissions, first adjustment parameter, and second adjustment parameter at the next N time moments; An encoding generation module, which is used to randomly generate the encodings of individuals in the initial population that meet the constraint conditions through a periodic mapping function; An adjustment list generation module, which is used to calculate the fitness value corresponding to an individual through an objective function according to the encoding of the individual until the iteration termination condition of the optimization algorithm is met, and output the encoding of the individual with the largest fitness value in the initial population as the adjustment list; The adjustment list is represented by the desulfurization slurry injection volume, ammonia injection volume, induced draft fan frequency, air duct baffle opening degree, and bag filter backwashing frequency at the next N moments.

[0006] Further, the iteration termination condition is that the number of iterations reaches the preset maximum number of iterations or the difference between the maximum fitness values of two consecutive iterations is less than the preset difference threshold, where the preset maximum number of iterations and the preset difference threshold are both user-defined parameters; the number of future moments N = future time period / future time interval, where the future time period, future time interval, preset time period, preset time interval, and the number of individuals in the initial population are all user-defined parameters; the lengths of the raw material sequence, process sequence, and environmental sequence are the same, and are all equal to the preset time period divided by the preset time interval.

[0007] Further, the prediction model consists of a first branch, a second branch, a third branch, a feature fusion layer, N first fully connected layers, and N second fully connected layers; The first branch is used to perform time series analysis on the raw material sequence to obtain a first vector; The first branch is constructed based on a bidirectional gated recurrent unit network model; The second branch is used to perform time series analysis on the process sequence to obtain a second vector; The second branch is constructed based on a temporal convolutional network model; The third branch is used to perform time series analysis on the environmental sequence to obtain a third vector; The third branch is constructed based on a gated recurrent unit network model; The feature fusion layer is used to perform feature fusion on the first vector, the second vector, and the third vector to obtain a fusion vector; The N first fully connected layers all input the fusion vector and output the sulfur dioxide emissions at the next N moments respectively; The N second fully connected layers all input the fusion vector and output the nitrogen oxide emissions at the next N moments respectively; Wherein the dimension numbers of the first vector, the second vector, the third vector, and the fusion vector are all user-defined parameters, and the activation functions of the first fully connected layer and the second fully connected layer are both GELU activation functions.

[0008] Further, the feature fusion layer uniformly transforms the first vector, the second vector, and the third vector into vectors with the same number of dimensions through matrix operations, then stacks the three into a matrix form, and finally converts them into a fusion vector through matrix operations.

[0009] Further, the first adjustment parameters include: the injection amount of desulfurization slurry, the frequency of the induced draft fan, the opening degree of the air duct baffle, and the backwashing frequency of the bag filter; the desulfurization efficiency model first constructs a weighted power sum function of the first adjustment parameters at a single moment, then fits the mapping relationship between it and the sulfur dioxide removal amount through the least squares method, and finally obtains the sulfur dioxide removal amounts at the next N moments through the fitted desulfurization efficiency model according to the first adjustment parameters at the next N moments, and then sums them up after subtracting the sulfur dioxide emission amounts at the next N moments respectively.

[0010] Further, the second adjustment parameters include: the ammonia injection amount, the frequency of the induced draft fan, the opening degree of the air duct baffle, and the backwashing frequency of the bag filter; the denitrification efficiency model first constructs a weighted power sum function of the second adjustment parameters at a single moment, then fits the mapping relationship between it and the nitrogen oxide removal amount through the least squares method, and finally obtains the nitrogen oxide removal amounts at the next N moments through the fitted denitrification efficiency model according to the second adjustment parameters at the next N moments, and then sums them up after subtracting the nitrogen oxide emission amounts at the next N moments respectively.

[0011] Further, the encoding of the individuals in the initial population is the same as the representation of the adjustment list, and the constraint condition is that the injection amount of desulfurization slurry, the ammonia injection amount, the frequency of the induced draft fan, the opening degree of the air duct baffle, and the backwashing frequency of the bag filter at the next N moments randomly generated cannot exceed the corresponding preset upper and lower limits.

[0012] Further, the periodic mapping function periodically generates a perturbation coefficient with a value range between 0 and 1 according to the serial number of the individual, multiplies the perturbation coefficient by the difference between the preset upper limit value and the lower limit value of each parameter in the adjustment list, and then sums it with the corresponding lower limit value to generate the encoding of the individual, where the serial number of the individual is represented by an increasing serial number starting from 1 and ending at G, and G represents the number of individuals in the initial population.

[0013] Further, the objective function is the weighted sum of the output value of the desulfurization efficiency model, the output value of the denitrification efficiency model, the reciprocal of the energy consumption value corresponding to each parameter in the adjustment list, and the reciprocal of the operating cost, where the energy consumption value and the operating cost corresponding to each parameter in the adjustment list are both user-defined parameters.

[0014] Further, the optimization algorithm is the sparrow search algorithm.

[0015] The beneficial effects of the present invention are as follows: By using a prediction model to comprehensively analyze the raw material parameters, process parameters, and environmental parameters, the present invention predicts the sulfur dioxide emissions and nitrogen oxide emissions at multiple future moments, and combines an optimization algorithm to achieve automatic adjustment of equipment parameters, maximizing the pollutant removal amount, minimizing the equipment energy consumption, and minimizing the equipment operation cost. Description of the Drawings

[0016] Figure 1 is a schematic diagram of an ultra-low emission intelligent control system for the steel industry based on the Internet of Things according to the present invention.

[0017] In the figure: sequence construction module 101, prediction model training module 102, objective function construction module 103, encoding generation module 104, adjustment list generation module 105. Detailed Embodiments

[0018] Now, the subject matter described herein will be discussed with reference to example embodiments. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. Without departing from the scope of protection of the content of this specification, the functions and arrangements of the elements discussed can be changed. Each example can omit, substitute, or add various processes or components as needed. Additionally, the features described relative to some examples can also be combined in other examples.

[0019] It should be noted that unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by those of ordinary skill in the art within the field to which the present invention pertains. The "first", "second", and similar terms used in one or more embodiments of the present invention do not denote any order, quantity, or importance, but are only used to distinguish different components. Words such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. Words such as "connected" or "coupled" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Upper", "lower", "left", "right", etc. are only used to represent relative position relationships, and when the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0020] As Figure 1 shown, an ultra-low emission intelligent control system for the steel industry based on the Internet of Things includes: A sequence construction module 101, which is used to collect raw material parameters, process parameters, and environmental parameters simultaneously at preset time intervals within a preset time period, and respectively perform normalization to construct a raw material sequence, a process sequence, and an environmental sequence; A prediction model training module 102, which is used to take the raw material sequence, the process sequence and the environment sequence as the sample data of the prediction model, and collect the sulfur dioxide emission and nitrogen oxide emission at the next N moments as the sample labels of the prediction model; An objective function construction module 103, which is used to construct a desulfurization efficiency model and a denitration efficiency model as objective functions respectively according to the sulfur dioxide emission, nitrogen oxide emission, the first adjustment parameter and the second adjustment parameter at the next N moments; An encoding generation module 104, which is used to randomly generate the encoding of the individuals in the initial population that meet the constraint conditions through a periodic mapping function; An adjustment list generation module 105, which is used to calculate the fitness value corresponding to the individual through the objective function according to the encoding of the individual until the iteration termination condition of the optimization algorithm is met, and output the encoding of the individual with the largest fitness value in the initial population as the adjustment list; The adjustment list is represented by the desulfurization slurry injection volume, ammonia injection volume, induced draft fan frequency, air duct baffle opening and bag filter backwashing frequency at the next N moments.

[0021] In an embodiment of the present invention, the iteration termination condition is that the number of iterations reaches the preset maximum number of iterations or the difference between the maximum fitness values of two consecutive iterations is less than the preset difference threshold, where the preset maximum number of iterations and the preset difference threshold are both user-defined parameters. Preferably, the preset maximum number of iterations is set to 200, and the preset difference threshold is set to 1.

[0022] In an embodiment of the present invention, the number of future moments N = future time period / future time interval, where the future time period, the future time interval, the preset time period, the preset time interval and the number of individuals in the initial population are all user-defined parameters. Preferably, the preset time period and the future time period are both set to 3 hours, the future time interval and the preset time interval are both set to 10 minutes, then the number of future moments N = 18, and the number of individuals in the initial population is set to 50.

[0023] In an embodiment of the present invention, the lengths of the raw material sequence, the process sequence and the environment sequence are the same, and are all equal to the preset time period divided by the preset time interval. The raw material parameters include: iron content in iron ore powder, sulfur content in iron ore powder, nitrogen content in iron ore powder, solid fuel ratio, raw material particle size, return ore ratio, limestone ratio, feed water content, etc.; the process parameters include: induced draft fan frequency, sintering temperature, sintering layer thickness, desulfurization slurry injection volume, ammonia injection volume, etc.; the environment parameters include: environmental temperature, rainfall, wind speed, etc.

[0024] It should be noted that the proportion of solid fuel represents the ratio of coke powder or pulverized coal in the raw materials, which affects the combustion heat intensity; the particle size of the raw materials represents the average particle diameter of the raw materials, which indirectly affects the flue gas flow field; the proportion of returned ore represents the proportion of sintered returned materials in the new batching, which affects the layer structure and composition stability; the proportion of limestone provides alkaline substances to participate in the desulfurization reaction; the water content of the incoming material affects the material adhesiveness; the injection amount of desulfurization slurry determines the supply rate of desulfurization reactants and directly affects the removal rate of sulfur dioxide; the ammonia injection amount determines the supply of raw materials for the reduction reaction and directly affects the removal rate of nitrogen oxides. In addition, the above parameters can be obtained by collecting through relevant instruments and sensors. For example, the iron content, sulfur content, and nitrogen content of iron ore powder can be obtained by portable X-ray fluorescence analysis, the sintering temperature can be obtained by an infrared thermometer, and the sintering layer thickness can be obtained by a laser displacement sensor, which will not be elaborated here.

[0025] In an embodiment of the present invention, the raw material parameters, process parameters, and environmental parameters are respectively normalized by the Min-Max normalization method, or can also be normalized by the Z-score method, so as to eliminate the dimension difference and improve the generalization ability of the prediction model, which will not be elaborated here.

[0026] In an embodiment of the present invention, the prediction model is composed of a first branch, a second branch, a third branch, a feature fusion layer, N first fully connected layers, and N second fully connected layers; The first branch is used to perform time series analysis on the raw material sequence to obtain a first vector; The second branch is used to perform time series analysis on the process sequence to obtain a second vector; The third branch is used to perform time series analysis on the environmental sequence to obtain a third vector; The feature fusion layer is used to perform feature fusion on the first vector, the second vector, and the third vector to obtain a fusion vector; The N first fully connected layers all input the fusion vector and respectively output the sulfur dioxide emissions at the next N time moments; The N second fully connected layers all input the fusion vector and respectively output the nitrogen oxide emissions at the next N time moments; Wherein the dimension numbers of the first vector, the second vector, the third vector, and the fusion vector are all custom parameters. Preferably, the dimension numbers of the first vector, the second vector, the third vector, and the fusion vector are respectively set to 64, 32, 16, and 128, and the activation functions of the first fully connected layer and the second fully connected layer are both GELU activation functions, or can also be ReLU activation functions, which will not be elaborated here.

[0027] In one embodiment of the present invention, the first branch is constructed based on the BiGRU (bidirectional gated recurrent unit network) model, the second branch is constructed based on the TCN (temporal convolutional network) model, and the third branch is constructed based on the GRU (gated recurrent unit network) model. They can also be uniformly constructed based on the Transformer (self-attention converter) model.

[0028] In one embodiment of the present invention, the feature fusion layer uniformly converts the first vector, the second vector, and the third vector into vectors of the same number of dimensions through matrix operations, then stacks the three into a matrix form, and finally converts them into a fused vector through matrix operations.

[0029] It should be noted that, for example, if the vectors are uniformly converted to 128 dimensions, then according to the above content, the first vector needs to be multiplied by a 64×128 matrix, the second vector needs to be multiplied by a 32×128 matrix, and the third vector needs to be multiplied by a 16×128 matrix. Then, after stacking the three, a 3×128 matrix is obtained, and finally, the fused vector is converted through matrix operations. The calculation formula is as follows: ,in represents the stacked matrix, and represent the first weight parameter and the second weight parameter respectively, and represent the first bias parameter and the second bias parameter respectively, represents a natural constant, express Activation function.

[0030] It should be noted that It needs to be designed as a 1×3 size vector, then Need to synchronize the design to a vector of size 1×3, It needs to be designed into a 128×128 size matrix, then The synchronization design needs to be a vector of size 1×128.

[0031] In one embodiment of the present invention, the first adjustment parameter includes: desulfurization slurry injection amount, exhaust fan frequency, duct baffle opening and bag filter backblowing frequency; the desulfurization efficiency model first constructs a weighted power sum function of the first adjustment parameter at a single moment, and then fits the mapping relationship between it and the sulfur dioxide removal amount through the least squares method. Finally, according to the first adjustment parameter at the next N moments, the sulfur dioxide removal amount at the next N moments is obtained through the fitted desulfurization efficiency model, and then the sulfur dioxide emissions at the next N moments are subtracted respectively and summed up.

[0032] Specifically, the calculation formula of the desulfurization efficiency model is as follows: , where represents the desulfurization efficiency coefficient, , , and respectively represent the desulfurized slurry injection volume, induced draft fan frequency, air duct baffle opening degree, and pulse jet bag filter backwashing frequency at the i-th future moment, , , and respectively represent the first weight coefficient, second weight coefficient, third weight coefficient, and fourth weight coefficient of the desulfurization efficiency model, , , and respectively represent the first power coefficient, second power coefficient, third power coefficient, and fourth power coefficient of the desulfurization efficiency model, represents the sulfur dioxide removal amount at the i-th future moment.

[0033] In an embodiment of the present invention, the second adjustment parameter includes: ammonia injection volume, induced draft fan frequency, air duct baffle opening degree, and pulse jet bag filter backwashing frequency; the denitration efficiency model first constructs a weighted power sum function of the second adjustment parameter at a single moment, then fits the mapping relationship between it and the nitrogen oxide removal amount through the least squares method, and finally obtains the nitrogen oxide removal amounts at the future N moments through the fitted denitration efficiency model according to the second adjustment parameters at the future N moments, and then sums them after subtracting the nitrogen oxide emission amounts at the future N moments respectively.

[0034] Specifically, the calculation formula of the denitration efficiency model is as follows: , where represents the denitration efficiency coefficient, , , and respectively represent the ammonia injection volume, induced draft fan frequency, air duct baffle opening degree, and pulse jet bag filter backwashing frequency at the i-th future moment, , , and respectively represent the first weight coefficient, second weight coefficient, third weight coefficient, and fourth weight coefficient of the denitration efficiency model, , , and respectively represent the first power coefficient, second power coefficient, third power coefficient, and fourth power coefficient of the denitration efficiency model, represents the nitrogen oxide removal amount at the i-th future moment.

[0035] In an embodiment of the present invention, the encoding of individuals in the initialized population is the same as the representation of the adjustment list, and the constraint conditions are that the desulfurization slurry injection volume, ammonia injection volume, induced draft fan frequency, air duct baffle opening, and bag filter backwashing frequency at the next N moments randomly generated cannot exceed the corresponding preset upper and lower limits.

[0036] For example, the upper and lower limits of the desulfurization slurry injection volume are set to 50 - 300 L / min, the upper and lower limits of the ammonia injection volume are set to 20 - 150 L / min, the induced draft fan frequency is set to 30 - 50 Hz, the air duct baffle opening is set to 20 - 100%, and the bag filter backwashing frequency is set to 0.5 - 6.0 times / min. Specific adjustments are made according to the equipment model and will not be elaborated here.

[0037] In an embodiment of the present invention, the periodic mapping function periodically generates a perturbation coefficient with a value range between 0 and 1 according to the serial number of the individual. After multiplying the perturbation coefficient by the difference between the preset upper limit and lower limit of each parameter in the adjustment list, the sum with the corresponding lower limit is used to generate the encoding of the individual, where the serial number of the individual is represented by an increasing serial number starting from 1 and ending at G, and G represents the number of individuals in the initialized population.

[0038] Specifically, the calculation formula of the periodic mapping function is as follows: , where , represents the perturbation coefficient, represents the serial number of the individual, represents the number of future moments, represents the modulo function.

[0039] For example, if the number of individuals in the initialized population G = 50 and the number of future moments N = 18, then the perturbation coefficients for the first and second moments of the individual with serial number 1 are 0.239 and 0.273 respectively. Assuming that the upper and lower limits of the desulfurization slurry injection volume are set to 50 - 300 L / min, then the corresponding individual encoding values are (300 - 50)×0.239 + 50 = 109.75 and (300 - 50)×0.273 + 50 = 118.25 respectively.

[0040] It should be noted that compared with the encoding of individuals initialized purely randomly, the present invention generates a perturbation coefficient through the periodic mapping function and then acts on the preset upper and lower limits of each parameter in the adjustment list, which can effectively avoid the problem that some regions in the solution space are too dense or sparse, thus facilitating global search.

[0041] In one embodiment of the present invention, the objective function is a weighted sum of the output value of the desulfurization efficiency model, the output value of the denitrification efficiency model, the reciprocal of the energy consumption value corresponding to each parameter in the adjustment list, and the reciprocal of the operating cost, where the energy consumption value and the operating cost corresponding to each parameter in the adjustment list are both user-defined parameters.

[0042] Specifically, the calculation formula of the objective function is as follows: , where represents the fitness value of an individual, represents the desulfurization efficiency coefficient output by the desulfurization efficiency model, represents the denitrification efficiency coefficient output by the denitrification efficiency model, and respectively represent the energy consumption value and the operating cost of the j-th parameter at the i-th future moment in the adjustment list, , , and respectively represent the first weight coefficient, the second weight coefficient, the third weight coefficient, and the fourth weight coefficient of the objective function, and are all user-defined parameters.

[0043] In one embodiment of the present invention, the optimization algorithm is the sparrow search algorithm, or it can also be the particle swarm algorithm, which will not be elaborated here.

[0044] It should be noted that the setting of the interval and the threshold size is for the convenience of comparison. Among them, the size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each group of sample data, as long as it does not affect the proportional relationship between the parameter and the quantified value. And the above formulas are all calculations of taking the numerical value after dimensionless, and the formulas are all obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0045] The above describes the embodiments of this embodiment, but this embodiment is not limited to the above specific implementation manners. The above specific implementation manners are only illustrative and not restrictive. Under the inspiration of this embodiment, those of ordinary skill in the art can also make many forms, all of which fall within the protection scope of this embodiment.

Claims

1. An intelligent control system for ultra-low emissions in the iron and steel industry based on the Internet of Things, characterized in that, Including: A sequence construction module, which is used to simultaneously collect raw material parameters, process parameters, and environmental parameters at preset time intervals within a preset time period, and respectively perform normalization to construct a raw material sequence, a process sequence, and an environmental sequence; A prediction model training module, which is used to use the raw material sequence, the process sequence, and the environmental sequence as sample data of the prediction model, and collect the sulfur dioxide emission and nitrogen oxide emission at the next N time moments as the sample labels of the prediction model; An objective function construction module, which is used to respectively construct a desulfurization efficiency model and a denitrification efficiency model as objective functions according to the sulfur dioxide emission, nitrogen oxide emission, first adjustment parameter, and second adjustment parameter at the next N time moments; An encoding generation module, which is used to randomly generate the encoding of the individuals in the initial population that meet the constraint conditions through a periodic mapping function; An adjustment list generation module, which is used to calculate the fitness value corresponding to the individual through the objective function according to the encoding of the individual until the iteration termination condition of the optimization algorithm is met, and output the encoding of the individual with the largest fitness value in the initial population as the adjustment list; The adjustment list is represented by the desulfurization slurry injection volume, ammonia injection volume, induced draft fan frequency, air duct baffle opening, and bag filter backwashing frequency at the next N time moments.

2. The intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The iteration termination condition is that the number of iterations reaches the preset maximum number of iterations or the difference between the maximum fitness values of two consecutive iterations is less than the preset difference threshold, where the preset maximum number of iterations and the preset difference threshold are both user-defined parameters; the number of future time moments N = future time period / future time interval, where the future time period, future time interval, preset time period, preset time interval, and the number of individuals in the initial population are all user-defined parameters; the lengths of the raw material sequence, the process sequence, and the environmental sequence are the same, and are all equal to the preset time period divided by the preset time interval.

3. An intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The prediction model consists of a first branch, a second branch, a third branch, a feature fusion layer, N first fully connected layers, and N second fully connected layers; The first branch is used to perform time series analysis on the raw material sequence to obtain a first vector; The first branch is constructed based on a bidirectional gated recurrent unit network model; The second branch is used to perform time series analysis on the process sequence to obtain a second vector; The second branch is constructed based on a time series convolutional network model; The third branch is used to perform time series analysis on the environmental sequence to obtain a third vector; The third branch is constructed based on a gated recurrent unit network model; The feature fusion layer is used to perform feature fusion on the first vector, the second vector, and the third vector to obtain a fusion vector; All N first fully connected layers input the fusion vector and respectively output the sulfur dioxide emissions at the next N time moments; All N second fully connected layers input the fusion vector and respectively output the nitrogen oxide emissions at the next N time moments; Where the dimension numbers of the first vector, the second vector, the third vector, and the fusion vector are all user-defined parameters, and the activation functions of the first fully connected layer and the second fully connected layer are both GELU activation functions.

4. The intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 3, characterized in that, The feature fusion layer uniformly transforms the first vector, the second vector, and the third vector into vectors with the same dimension number through matrix operations, then stacks the three into a matrix form, and finally converts them into a fusion vector through matrix operations.

5. The intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The first adjustment parameter includes: the desulfurization slurry injection volume, the frequency of the induced draft fan, the opening degree of the air duct baffle, and the backwashing frequency of the bag filter; the desulfurization efficiency model first constructs a weighted power sum function of the first adjustment parameter at a single moment, then fits the mapping relationship between it and the sulfur dioxide removal amount by the least squares method, and finally obtains the sulfur dioxide removal amount at the next N moments through the fitted desulfurization efficiency model according to the first adjustment parameter at the next N moments, and then sums them after subtracting the sulfur dioxide emissions at the next N moments respectively.

6. An intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The second adjustment parameter includes: the ammonia injection volume, the frequency of the induced draft fan, the opening degree of the air duct baffle, and the backwashing frequency of the bag filter; the denitrification efficiency model first constructs a weighted power sum function of the second adjustment parameter at a single moment, then fits the mapping relationship between it and the nitrogen oxide removal amount by the least squares method, and finally obtains the nitrogen oxide removal amount at the next N moments through the fitted denitrification efficiency model according to the second adjustment parameter at the next N moments, and then sums them after subtracting the nitrogen oxide emissions at the next N moments respectively.

7. An intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The encoding of the individuals in the initial population is the same as the representation of the adjustment list, and the constraint condition is that the desulfurization slurry injection volume, ammonia injection volume, frequency of the induced draft fan, opening degree of the air duct baffle, and backwashing frequency of the bag filter at the next N moments randomly generated cannot exceed the corresponding preset upper and lower limits.

8. An intelligent control system for ultra-low emissions in the iron and steel industry based on the Internet of Things according to claim 7, characterized in that, The periodic mapping function periodically generates a perturbation coefficient with a value range between 0 and 1 according to the serial number of the individual, multiplies the perturbation coefficient by the difference between the preset upper limit and the lower limit of each parameter in the adjustment list, and then sums it with the corresponding lower limit to generate the encoding of the individual, where the serial number of the individual is represented by an increasing serial number starting from 1 and ending at G, and G represents the number of individuals in the initial population.

9. The intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The objective function is the weighted sum of the output value of the desulfurization efficiency model, the output value of the denitrification efficiency model, the reciprocal of the energy consumption value corresponding to each parameter in the adjustment list, and the reciprocal of the operating cost, where the energy consumption value and the operating cost corresponding to each parameter in the adjustment list are both user-defined parameters.

10. An intelligent control system for ultra-low emissions in the steel industry based on the Internet of Things according to claim 1, characterized in that, The optimization algorithm is the sparrow search algorithm.

Images