Control method for multi-strand flue gas denitration process of smelting reduction furnace and denitration process

Abstract

The invention provides a control method for a denitration process of multi-strand flue gas of a smelting reduction furnace, the multi-strand flue gas comprises hot blast stove combustion flue gas, rotary drying kiln flue gas and pulverized coal drying heating flue gas, and an NOX concentration prediction model, a flue gas amount prediction model and a flue gas temperature prediction model are established for the three strands of flue gas; in the NOX concentration prediction model, NOX concentration prediction of each strand of flue gas is carried out, and the NOX concentration is calculated and summarized; calculating the flue gas amount of each strand of flue gas in the flue gas amount prediction model, and calculating the summarized flue gas amount; in the flue gas temperature prediction model, the SCR denitration inlet flue gas temperature is calculated; calculating the consumption of a reducing agent for NOX removal according to the summarized NOX concentration and the summarized flue gas amount so as to adjust the opening degree of a reducing agent adjusting valve; frequency modulation is conducted on an induced draft fan according to the collected flue gas amount; and calculating the required gas consumption according to the collected flue gas amount and the SCR denitration inlet flue gas temperature so as to adjust the opening degree of a gas adjusting valve. The invention further provides a denitration process which comprises the control method.

Description

Technical Field

[0001] The present invention relates to a denitration process, and in particular to a control method for a multi-flue-gas denitration process in a smelting reduction furnace and a denitration process. Background Art

[0002] Nitrogen oxides (NO X ) are poisonous to the ecological environment and human health and are one of the main pollutants in the industrial flue gas of a smelting reduction furnace. The current mainstream denitration process is selective catalytic reduction (SCR). In the SCR denitration device, after the flue gas is heat-exchanged by the GGH, ammonia enters the flue through the ammonia injection grid device to be mixed with the flue gas, and then the flue gas heating system is used to raise the temperature of the flue gas to ~280°C. Finally, the NH3 / NO X mixed flue gas enters the catalyst layer, and a catalytic reduction reaction occurs under the action of a medium and low temperature catalyst, and denitration is finally achieved.

[0003] At present, the determination of the consumption of the reducing agent and the gas consumption of the coal gas is mainly based on measuring the inlet flue gas volume, NO X concentration, and flue gas temperature of the device, and then calculating the corresponding material consumption through software and adjusting the opening degrees of the reducing agent (urea / ammonia water) and coal gas pipelines.

[0004] However, the flue gas of the smelting reduction furnace consists of three flue gases, namely, the flue gas from the rotary drying kiln outlet, the flue gas from the hot blast stove combustion, and the flue gas from the pulverized coal drying and heating. The flue gas volume, NO X concentration, and flue gas temperature fluctuate with the change of the main machine load. Therefore, the instrument measurement always has limitations and cannot represent the global NO X concentration. To achieve up-to-standard emissions, the actual operation can only adopt the method of excessive ammonia injection, which will cause the ammonia concentration at the outlet to exceed the standard seriously and form new pollution. The calculation and flow regulation of the coal gas consumption also need to rely on the instrument to measure the flue gas volume and flue gas temperature. At the same time, the catalyst reaction temperature will seriously affect the catalytic reduction reaction. The heating coal gas consumption needs to be determined according to the flue gas volume and temperature rise. Under different load conditions, the flue gas volume is in a fluctuating state. If the coal gas regulating valve does not respond in time, it will cause waste of coal gas resources. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a control method for a multi-flue-gas denitration process in a smelting reduction furnace and a denitration process, which can solve the limitations of instrument measurement of flue gas properties, realize intelligent control of the reducing agent dosage, and avoid ammonia escape; at the same time, reduce the coal gas consumption for flue gas heating and the fan frequency modulation, and achieve the effect of energy conservation and consumption reduction.

[0006] To solve the above technical problem, the present invention provides a control method for a multi-flue-gas denitration process in a smelting reduction furnace. The multi-flue-gas includes the flue gas from the hot blast stove combustion, the flue gas from the rotary drying kiln, and the flue gas from the pulverized coal drying and heating. For the three flue gases, NOX Concentration prediction model, flue gas volume prediction model, flue gas temperature prediction model;

[0007] In the NO X concentration prediction model, NO generation models are established for three flue gases respectively to predict the NO concentration of each flue gas, and the total NO concentration is calculated; X X X concentration;

[0008] In the flue gas volume prediction model, the component parameters of hydrogen-rich gas and the air volume blown by the blower are collected to calculate the flue gas volume of each flue gas respectively, and the total flue gas volume is calculated;

[0009] In the flue gas temperature prediction model, the calorific value of the hydrogen-rich gas components is collected to calculate the generated flue gas temperature for each flue gas, and the heat-exchanged flue gas temperature of the rotary drying kiln flue gas and the pulverized coal drying and heating flue gas is calculated according to the generated flue gas temperature; then, the SCR denitration inlet flue gas temperature is calculated based on the calculated combustion flue gas temperature of the hot blast stove and the heat-exchanged flue gas temperature of the rotary drying kiln flue gas and the pulverized coal drying and heating flue gas;

[0010] According to the total NO X concentration and the total flue gas volume, calculate the consumption of the reducing agent for NO removal to adjust the opening of the reducing agent regulating valve; frequency-modulate the induced draft fan according to the total flue gas volume; calculate the required gas consumption according to the total flue gas volume and the SCR denitration inlet flue gas temperature to adjust the opening of the gas regulating valve. X

[0011] In some embodiments, the specific prediction steps of the NO X concentration prediction model are as follows:

[0012] S1: Collect the concentration-related parameters of three flue gases respectively;

[0013] S2: Establish a prediction model. For each flue gas, N groups of parameters collected in step S1 are selected as the total samples, and a part of the samples are randomly selected as the training samples to train the prediction model, and the remaining samples are selected as the test samples to test the prediction model;

[0014] S3: Improve the prediction accuracy and generalization ability by adjusting the constant parameters of the prediction model;

[0015] S4: Output the NO X concentration y11 of the combustion flue gas of the hot blast stove, the NO X concentration y21 of the rotary drying kiln flue gas, and the NO X concentration y31 of the pulverized coal drying and heating flue gas.

[0016] In some embodiments, the specific prediction steps of the flue gas volume prediction model are as follows:​​​

[0017] S1: Collect the parameters related to the flue gas volume of the three flue gases respectively;

[0018] S2: Establish a prediction model. For each flue gas, N groups of parameters collected in step S1 are selected as the total sample. Among them, some samples are randomly selected as the training samples to train the prediction model, and the remaining samples are selected as the test samples to test the prediction model;

[0019] S3: Improve the prediction accuracy and generalization ability by adjusting the constant parameters of the prediction model;

[0020] S4: Output the flue gas volume y12 of the hot blast stove combustion, the flue gas volume y22 of the rotary drying kiln, and the flue gas volume y32 of the pulverized coal drying and heating.

[0021] In some of the embodiments, the specific prediction steps of the flue gas temperature prediction model are as follows:

[0022] S1: Collect the parameters related to the flue gas temperature of the three flue gases respectively;

[0023] S2: Establish a prediction model. For each flue gas, N groups of parameters collected in step S1 are selected as the total sample. Among them, some samples are randomly selected as the training samples to train the prediction model, and the remaining samples are selected as the test samples to test the prediction model;

[0024] S3: Improve the prediction accuracy and generalization ability by adjusting the constant parameters of the prediction model;

[0025] S4: Output the flue gas temperature y13 of the hot blast stove combustion, the flue gas temperature y23 of the rotary drying kiln, and the flue gas temperature y33 of the pulverized coal drying and heating.

[0026] In some of the embodiments, the concentration-related parameters in S1 include the inclination angle of the combustion nozzle, the inclination angle of the air nozzle, the air-gas preheating temperature, the air coefficient, the air-gas nozzle spacing, the opening parameter of the gas damper, and the opening parameter of the air damper.

[0027] In some of the embodiments, for the flue gas of the hot blast stove combustion, the parameters related to the flue gas volume in S1 include the fuel composition, the fuel consumption, the fuel calorific value, the air coefficient, the oxygen content in the flue gas, the moisture content in the flue gas, the opening parameter of the gas damper, and the opening parameter of the air damper;

[0028] For the flue gas of the rotary drying kiln, the parameters related to the flue gas volume in S1 include the fuel composition, the fuel consumption, the fuel calorific value, the air coefficient, the amount of ore powder entering the kiln, the moisture content of the ore powder, the opening parameter of the gas damper, and the opening parameter of the air damper;

[0029] For the flue gas used for drying and heating pulverized coal, the parameters related to the flue gas volume in S1 include fuel composition, fuel consumption, fuel calorific value, air coefficient, amount of ore powder fed into the kiln, moisture content of ore powder, opening parameter of the gas damper, and opening parameter of the air damper.

[0030] In some embodiments, the parameters related to the flue gas temperature in S1 include fuel composition, fuel consumption, fuel calorific value, feed temperature, discharge temperature, air coefficient, oxygen content in the flue gas, moisture content in the flue gas, opening parameter of the gas damper, opening parameter of the air damper, and desulfurization water spray amount;

[0031] In some embodiments, the prediction model in S2 adopts a radial basis neural network model.

[0032] In some embodiments, the constant parameter in S3 includes the spread constant.

[0033] In some embodiments, a feedback mechanism is further included, and the reducing agent regulating valve is finely adjusted through the CEMS concentration at the flue gas outlet.

[0034] The present invention also provides a multi-strand flue gas denitrification process for a smelting reduction furnace, including the control method of the multi-strand flue gas denitrification process for the smelting reduction furnace described above.

[0035] Compared with the prior art, the present invention adopts the method of establishing a prediction model, establishes a front-end NO X concentration prediction model, flue gas temperature prediction and flue gas volume prediction model and combines them with the reducing agent regulating valve and gas regulating valve at the back end, and the frequency modulation of the induced draft fan. By predicting the NO X concentration, flue gas volume, and flue gas temperature, parameters such as the reducing agent dosage are calculated to adjust the reducing agent regulating valve, gas regulating valve, and the frequency modulation of the induced draft fan, which not only solves the limitations of instrument measurement but also solves the system hysteresis, intelligently controls the reducing agent, gas consumption, etc., reduces the ammonia escape phenomenon, reduces the gas consumption for flue gas heating and the fan frequency modulation, so as to achieve the effect of energy conservation and consumption reduction. Description of the Drawings

[0036] Figure 1 Shows the existing flue gas flow process of the reduction furnace;

[0037] Figure 2 Shows the existing flue gas treatment process of the reduction furnace;

[0038] Figure 3 Shows the model architecture flow chart of the control method of the present invention. Detailed Embodiments

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0040] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", "inner", "outer", "top / bottom end", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0041] In the description of the present invention, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "provided with", "sheathed / connected", "connected", etc. should be understood in a broad sense. For example, "connected" can be a wall-mounted connection, a detachable connection, or an integral connection. It can be a mechanical connection, an electrical connection, a direct connection, or an indirect connection through an intermediate medium. It can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0042] As Figure 1 and 2 shown, the existing reduction furnace flue gas consists of three parts: rotary drying kiln flue gas, hot blast stove combustion flue gas, and pulverized coal drying and heating flue gas. Both the rotary drying kiln and pulverized coal drying and heating use hydrogen-rich gas as the combustion medium. After combustion, the heat source is provided for pulverized coal drying and ore powder pre-reduction. The flue gas after releasing heat converges with the hot blast stove combustion flue gas and enters the flue gas desulfurization and denitrification device for flue gas treatment.

[0043] In the SCR denitrification device, after the flue gas is heat-exchanged by the GGH, ammonia enters the flue through the ammonia injection grid device and mixes with the flue gas. Then, the flue gas heat supply system is used to raise the temperature of the flue gas to ~280 °C. Finally, the NH3 / NO X mixed flue gas enters the catalyst layer, and a catalytic reduction reaction occurs under the action of medium and low temperature catalysts, and finally denitrification is achieved.

[0044] In order to be able to intelligently control the dosage of the reducing agent ammonia, reduce ammonia slip, reduce the gas consumption of flue gas heating gas and the frequency modulation of the fan, and save energy consumption, this embodiment provides a multi-stream flue gas denitrification process for a smelting reduction furnace. The process adopts a control method for the multi-stream flue gas denitrification process of the smelting reduction furnace. The control method establishes NO X concentration prediction models, flue gas volume prediction models, and flue gas temperature prediction models for the three streams of flue gas, namely, the combustion flue gas of the hot blast stove, the flue gas of the rotary drying kiln, and the pulverized coal drying and heating flue gas;

[0045] Since the generation of NO X mainly has three ways: fuel-type NO X , thermal-type NO X , and prompt NO X . In the rotary drying kiln and pulverized coal drying and heating, the high-temperature flue gas after combustion (~1050 °C) is used for raw material drying. Therefore, the NO X in these two streams of flue gas is mainly thermal-type NO X .

[0046] In the SRV smelting reduction furnace, the high-temperature flue gas (~1200 °C) generated by the combustion of hydrogen-rich gas enables the coke or pulverized coal to burn fully, reduces the ore, and refines the molten iron. Therefore, in the flue gas of the smelting reduction furnace, the NO X is mainly thermal-type NO X and fuel-type NO X .

[0047] Therefore, in the NO X concentration prediction model, it is necessary to establish corresponding NO X generation models for the three streams of flue gas respectively to predict the NO X concentration of each stream of flue gas, and then calculate and summarize the NO X concentration.

[0048] In the flue gas volume prediction model, collect the component parameters of hydrogen-rich gas and the air volume blown in by the blower to calculate the flue gas volume of each stream of flue gas respectively, and calculate and summarize the flue gas volume;

[0049] In the flue gas temperature prediction model, collect the calorific value of the hydrogen-rich gas component to calculate the generated flue gas temperature for each stream of flue gas respectively, and calculate the heat-exchanged flue gas temperature of the rotary drying kiln flue gas and the pulverized coal drying and heating flue gas according to the calculated generated flue gas temperature; then calculate the SCR denitrification inlet flue gas temperature according to the calculated generated hot blast stove combustion flue gas temperature and the heat-exchanged flue gas temperature of the rotary drying kiln flue gas and the pulverized coal drying and heating flue gas based on the law of conservation of energy;

[0050] After the prediction and calculation of the above three models, the summarized NO X concentration and the summarized flue gas volume can be used to calculate the NO XThe consumption of the removed reducing agent is used to adjust the opening degree of the reducing agent regulating valve; the induced draft fan is frequency modulated according to the total flue gas volume; the required gas consumption is calculated according to the total flue gas volume and the SCR denitration inlet flue gas temperature to adjust the opening degree of the gas regulating valve.

[0051] Currently, the prediction models include the Radial Basis Function Neural Network Model (RBF), the Elman Neural Network Model, the Support Vector Machine Model (SVM), etc. In this embodiment, the Radial Basis Function Neural Network Model (RBF) is used as the prediction model.

[0052] Specifically, the NO X Specific prediction steps of the concentration prediction model are as follows:

[0053] S1: Collect the concentration-related parameters of three flue gases respectively:

[0054] For the hot blast stove combustion flue gas, the concentration-related parameters include the combustion nozzle inclination angle, the air nozzle inclination angle, the air-gas preheating temperature, the air coefficient, the air-gas nozzle spacing, the gas damper opening parameter, the air damper opening parameter, the feeding amount, etc.

[0055] For the rotary drying kiln flue gas, the concentration-related parameters include the combustion nozzle inclination angle, the air nozzle inclination angle, the air-gas preheating temperature, the air coefficient, the air-gas nozzle spacing, the gas damper opening parameter, the air damper opening parameter, etc.

[0056] For the pulverized coal drying and heating flue gas, the concentration-related parameters include the combustion nozzle inclination angle, the air nozzle inclination angle, the air-gas preheating temperature, the air coefficient, the air-gas nozzle spacing, the gas damper opening parameter, the air damper opening parameter, etc.

[0057] S2: Establish a Radial Basis Function Neural Network Model (RBF). For each flue gas, 100 groups of plant operation parameters collected in step S1 are respectively selected as the total samples. Among them, 90 groups of samples are randomly selected as the training samples to train the prediction model, and the remaining 10 groups of samples are selected as the test samples to test the prediction model.

[0058] S3: Improve the prediction accuracy and generalization ability by adjusting the constant parameters of multiple different prediction models, such as the spread constant spread.

[0059] S4: Output the NO X concentration y11 of the hot blast stove combustion flue gas, the NO X concentration y21 of the rotary drying kiln flue gas, and the NO X concentration y31 of the pulverized coal drying and heating flue gas.

[0060] Specific prediction steps of the flue gas volume prediction model are as follows:

[0061] S1: Collect the parameters related to the flue gas volume of three flue gases respectively:

[0062] For the combustion flue gas of the hot blast stove, the parameters related to the flue gas volume include fuel composition (H2, H2O, CH4, O2, CO, C2H4, C2H6, C3H6), fuel consumption, fuel calorific value, air coefficient, oxygen content in the flue gas, moisture content in the flue gas, gas damper opening parameter, and air damper opening parameter.

[0063] For the flue gas of the rotary drying kiln, the parameters related to the flue gas volume include fuel composition (H2, H2O, CH4, O2, CO, C2H4, C2H6, C3H6), fuel consumption, fuel calorific value, air coefficient, ore powder input amount into the kiln, ore powder moisture content, gas damper opening parameter, and air damper opening parameter.

[0064] For the coal powder drying and heating flue gas, the parameters related to the flue gas volume include fuel composition (H2, H2O, CH4, O2, CO, C2H4, C2H6, C3H6), fuel consumption, fuel calorific value, air coefficient, ore powder input amount into the kiln, ore powder moisture content, gas damper opening parameter, and air damper opening parameter.

[0065] S2: Establish a radial basis function neural network model (RBF). For each flue gas, 100 groups of plant operation parameters collected in step S1 are selected as the total samples. Among them, 90 groups of samples are randomly selected as training samples to train the prediction model, and the remaining 10 groups of samples are selected as test samples to test the prediction model.

[0066] S3: Improve the prediction accuracy and generalization ability by adjusting the constant parameters of multiple different prediction models, such as the spread constant spread.

[0067] S4: Output the combustion flue gas volume y12 of the hot blast stove, the flue gas volume y22 of the rotary drying kiln, and the coal powder drying and heating flue gas volume y32.

[0068] The specific prediction steps of the flue gas temperature prediction model are as follows:

[0069] S1: Collect the parameters related to the flue gas temperature of three flue gases respectively:

[0070] For the combustion flue gas of the hot blast stove, the parameters related to the flue gas temperature include fuel composition, fuel consumption, fuel calorific value, feed temperature, discharge temperature, air coefficient, oxygen content in the flue gas, moisture content in the flue gas, gas damper opening parameter, air damper opening parameter, and desulfurization water spray amount.

[0071] For the flue gas of the rotary drying kiln, the parameters related to the flue gas temperature include fuel composition, fuel consumption, fuel calorific value, feed temperature, discharge temperature, air coefficient, oxygen content in the flue gas, moisture content in the flue gas, gas damper opening parameter, air damper opening parameter, and desulfurization water spray amount.

[0072] For the pulverized coal drying and heating flue gas, the flue gas temperature-related parameters include fuel composition, fuel consumption, fuel calorific value, feed temperature, discharge temperature, air coefficient, flue gas oxygen content, flue gas moisture content, gas damper opening parameter, air damper opening parameter, and desulfurization water injection volume.

[0073] S2: Establish a radial basis function neural network model (RBF). For each flue gas, 100 groups of plant operation parameters collected in step S1 are respectively selected as the total samples. Among them, 90 groups of samples are randomly selected as training samples to train the prediction model, and the remaining 10 groups of samples are selected as test samples to test the prediction model.

[0074] S3: Improve the prediction accuracy and generalization ability by adjusting the constant parameters of multiple different prediction models, such as the spread constant spread.

[0075] S4: Output the combustion flue gas temperature y13 of the hot blast stove, the flue gas temperature y23 of the rotary drying kiln, and the pulverized coal drying and heating flue gas temperature y33.

[0076] According to the values of y11, y21, y31, y12, y22, and y32, calculate the total flue gas volume and NOx concentration, and thus calculate the amount of NO to be removed X amount, and then calculate the consumption of the reducing agent, and actuate the ammonia regulating valve in advance. At the same time, establish a feedback mechanism, and finely adjust the ammonia regulating valve according to the outlet CEMS concentration.

[0077] Calculate the total flue gas volume according to the values of y12, y22, and y32; according to the values of the flue gas temperature y13, y23, y33, y12, y22, and y32 of each flue gas and the law of conservation of heat, calculate the flue gas temperature at the inlet of SCR denitration. Then, according to the difference between the flue gas temperature at the outlet of the GGH raw flue gas side and the flue gas temperature at the inlet of SCR denitration, calculate the gas consumption of the gas for flue gas heating, and actuate the gas regulating valve in advance. At the same time, it is also necessary to monitor the flue gas temperature at the SCR inlet and further fine-tune the gas regulating valve.

[0078] Predict the flue gas volume at the inlet of the induced draft fan according to the values of y12, y22, and y32, and perform frequency modulation on the induced draft fan in advance to achieve energy conservation and consumption reduction.

[0079] As mentioned above, it is only a preferred specific embodiment of the present invention, but the design concept of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention who makes non-substantive modifications to the present invention using this concept shall fall within the scope of infringement of the present invention.

Claims

1. A control method for a multi-stream flue gas denitrification process of a smelting reduction furnace, wherein the multi-stream flue gas includes hot blast furnace combustion flue gas, rotary drying kiln flue gas, and coal powder drying and heating flue gas, characterized in that: Establish NO for three flue gases X Concentration prediction model, flue gas volume prediction model, flue gas temperature prediction model; In the NO X In the concentration prediction model, NO X Generate a model for NO per flue gas X Concentration prediction and calculation of summary NO X concentration; In the flue gas volume prediction model, the hydrogen-rich coal gas component parameters and the air volume blown by the blower are collected to calculate the flue gas volume of each flue gas respectively, and the total flue gas volume is calculated; In the flue gas temperature prediction model, the calorific value of the hydrogen-rich coal gas component is collected to calculate the flue gas temperature for each flue gas, and the flue gas temperature after heat exchange of the rotary drying kiln flue gas and the pulverized coal drying heating flue gas is calculated according to the calculated flue gas temperature; Then, the SCR denitrification inlet flue gas temperature is calculated based on the calculated hot blast furnace combustion flue gas temperature and the flue gas temperature after heat exchange between the rotary drying kiln flue gas and the pulverized coal drying heating flue gas; According to the summary NO X The concentration and the summed flue gas volume are calculated as NO X The reduced agent consumption is removed to adjust the opening of the reduced agent regulating valve; the induced draft fan is frequency-modulated according to the aggregated flue gas volume; the required gas consumption is calculated according to the aggregated flue gas volume and the SCR denitrification inlet flue gas temperature to adjust the opening of the gas regulating valve.

2. The control method for the multi-stream flue gas denitrification process of a smelting reduction furnace according to claim 1, characterized in that: The NO X The specific prediction steps of the concentration prediction model are: S1: Collect concentration-related parameters of three flue gases respectively; S2: Establish a prediction model. For each flue gas, select N groups of parameters collected in step S1 as total samples, randomly select some samples as training samples to train the prediction model, and select the remaining samples as test samples to test the prediction model; S3: Improve prediction accuracy and generalization ability by adjusting the constant parameters of the prediction model; S4: Output hot air furnace combustion flue gas NO X Concentration y11, rotary drying kiln flue gas NO X Concentration y21, coal powder drying and heating flue gas NO X Concentration y31.

3. The control method for the multi-stream flue gas denitrification process of a smelting reduction furnace according to claim 1, characterized in that: The specific prediction steps of the flue gas volume prediction model are: S1: Collect relevant parameters of the flue gas volume of three flue gases respectively; S2: Establish a prediction model. For each flue gas, select N groups of parameters collected in step S1 as total samples, randomly select some samples as training samples to train the prediction model, and select the remaining samples as test samples to test the prediction model; S3: Improve prediction accuracy and generalization ability by adjusting the constant parameters of the prediction model; S4: Output hot blast furnace combustion flue gas volume y12, rotary drying kiln flue gas volume y22, coal powder drying heating flue gas volume y32.

4. The control method for the multi-stream flue gas denitrification process of a smelting reduction furnace according to claim 1, characterized in that: The specific prediction steps of the flue gas temperature prediction model are: S1: Collect the flue gas temperature related parameters of three flue gases respectively; S2: Establish a prediction model. For each flue gas, select N groups of parameters collected in step S1 as total samples, randomly select some samples as training samples to train the prediction model, and select the remaining samples as test samples to test the prediction model; S3: Improve prediction accuracy and generalization ability by adjusting the constant parameters of the prediction model; S4: Output hot blast furnace combustion flue gas temperature y13, rotary drying kiln flue gas temperature y23, coal powder drying heating flue gas temperature y33.

5. The control method for the multi-stream flue gas denitrification process of a smelting reduction furnace according to claim 2, characterized in that: The concentration-related parameters in S1 include the combustion nozzle inclination angle, the air nozzle inclination angle, the air-gas preheating temperature, the air coefficient, the air-gas nozzle spacing, the gas damper opening parameter, the air damper opening parameter, and the feed rate.

6. The control method for the multi-stream flue gas denitrification process of a smelting reduction furnace according to claim 3, characterized in that: The flue gas volume related parameters in S1 for the hot blast furnace combustion flue gas include fuel composition, fuel consumption, fuel calorific value, air coefficient, flue gas oxygen content, flue gas moisture content, gas damper opening parameter, and air damper opening parameter; The flue gas volume related parameters in S1 for the rotary drying kiln flue gas include fuel composition, fuel consumption, fuel calorific value, air coefficient, mineral powder feeding amount, mineral powder moisture content, gas damper opening parameter, and air damper opening parameter; The flue gas volume related parameters in S1 for coal powder drying and heating flue gas include fuel composition, fuel consumption, fuel calorific value, air coefficient, mineral powder feeding amount, mineral powder moisture content, gas damper opening parameters, and air damper opening parameters.

7. The control method for the multi-stream flue gas denitrification process of a smelting reduction furnace according to claim 4, characterized in that: The flue gas temperature related parameters in S1 include fuel composition, fuel consumption, fuel calorific value, feed temperature, discharge temperature, air coefficient, flue gas oxygen content, flue gas moisture content, gas damper opening parameters, air damper opening parameters, and desulfurization water injection volume.

8. The control method for a multi-stream flue gas denitration process of a smelting reduction furnace according to any one of claims 2 to 7, characterized in that: The prediction model in S2 adopts a radial basis neural network model.

9. According to the control method of the multi-stream flue gas denitrification process of a smelting reduction furnace according to any one of claims 2 to 7, the constant parameters in S3 include a dispersion constant.

10. The control method for a multi-stream flue gas denitration process of a smelting reduction furnace according to any one of claims 2 to 7, characterized in that: It also includes a feedback mechanism for finely adjusting the reducing agent regulating valve according to the CEMS concentration at the flue gas outlet.

11. A multi-stream flue gas denitrification process for a smelting reduction furnace, characterized in that: The invention comprises a control method for a multi-stream flue gas denitrification process of a smelting reduction furnace as described in any one of claims 1 to 10.

Images