Method for controlling a wet flue gas desulfurization apparatus, control device for a wet flue gas desulfurization apparatus, remote monitoring system equipped with the control device for a wet flue gas desulfurization apparatus, information processing device, and information processing system

Abstract

To simply implement control for circulation of absorbent in an absorption tower of wet flue gas desulfurization equipment.SOLUTION: Wet flue gas desulfurization equipment comprises an absorption tower, a circulation pump, and an absorber slurry supply unit. A control method of the equipment constructs a learning model by machine learning for the absorber concentration of an absorbent, a circulation flow volume of the absorbent, and a relationship between an explanatory variable including at least one parameter having a correlation with a power generator output and an objective variable being sulfur dioxide concentration in a future absorption tower exit. Then, the wet flue gas desulfurization equipment calculates a prediction value of the sulfur dioxide concentration by the learning model in each power generator output, and creates a table showing an input amount of the absorber and the circulation flow volume of the absorber so that the prediction value becomes a reference value or below. Then the equipment determines an input amount of the absorber corresponding to a state of the power generator and a control target value of the circulation flow volume of the absorber on the basis of the table.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a method for controlling a wet flue gas desulfurization apparatus, a control device for a wet flue gas desulfurization apparatus, a remote monitoring system including the control device for a wet flue gas desulfurization apparatus, an information processing device, and an information processing system. [Background technology]

[0002] In a wet flue gas desulfurization system, flue gas generated in a combustion device such as a boiler is introduced into the absorption tower of the desulfurization system and brought into gas-liquid contact with an absorbing solution circulating through the absorption tower. During this gas-liquid contact process, the absorbent (e.g., calcium carbonate) in the absorbing solution reacts with the sulfur dioxide (SO2) in the flue gas, absorbing the SO2 into the absorbing solution and removing the SO2 from the flue gas (de-sulfurizing the flue gas). Meanwhile, the absorbing solution that has absorbed the SO2 falls and accumulates in a storage tank below the absorption tower. The storage tank is supplied with absorbent, and the absorbent, whose absorption capacity has been restored by the supplied absorbent, is then supplied by a circulation pump to the top of the absorption tower for gas-liquid contact with the flue gas (absorbing SO2). Because the circulation pump that circulates the absorbing solution consumes a lot of power, in the past, in order to reduce power consumption, the required circulating flow rate of the absorbing solution was calculated based on the flow rate of the flue gas flowing into the absorption tower and the SO2 concentration in the flue gas, and the number of operating circulating pumps was controlled accordingly.

[0003] Patent Document 1 discloses a technique for appropriately adjusting the operating conditions of a circulation pump for circulating an absorbing solution in an absorption tower of such a wet flue gas desulfurization system. In this document, using operating data obtained from a combustion device such as a boiler and a wet flue gas desulfurization system, the correlation between the operating data and the SO2 concentration at the outlet of the absorber tower, and the correlation between the operating data and the absorbent concentration in the absorbing solution are modeled by machine learning, and the circulation flow rate of the absorbing solution and the absorbent concentration are controlled to be optimized based on tables obtained by these two learning models. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-11163 Summary of the Invention [Problem to be solved by the invention]

[0005] However, in Patent Document 1, when constructing a learning model, parameters highly correlated with the SO2 concentration at the outlet of the absorber or the absorbent concentration are not identified. As a result, there is a risk of prediction disturbances occurring, and the widening of the learning range of operating data may increase the computational load. Furthermore, the correlation between the operating data and the SO2 concentration at the outlet of the absorber, and the correlation between the operating data and the absorbent concentration in the absorption solution must be machine-learned to construct two learning models, which tends to complicate the computation.

[0006] At least one embodiment of the present disclosure has been made in consideration of the above-described circumstances, and aims to provide a control method for a wet flue gas desulfurization apparatus, a control device, a remote monitoring system, an information processing device, and an information processing system that can easily perform control for circulating an absorption liquid in an absorption tower of a wet flue gas desulfurization apparatus. [Means for solving the problem]

[0007] In order to solve the above problems, a control method for a wet flue gas desulfurization apparatus according to at least one embodiment of the present disclosure includes: A control method for a wet flue gas desulfurization apparatus that performs desulfurization by bringing exhaust gas generated in a combustion apparatus into gas-liquid contact with an absorption liquid in an absorption tower, comprising: constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; calculating a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creating a table that indicates the amount of absorbent introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; determining, based on the table, control target values ​​for the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator; Equipped with.

[0008] In order to solve the above problems, the control device of a wet flue gas desulfurization apparatus according to at least one embodiment of the present disclosure includes: A control device for a wet flue gas desulfurization device that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower, a learning model construction unit for constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creates a table that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; Equipped with.

[0009] In order to solve the above problem, a remote monitoring system according to at least one embodiment of the present disclosure includes: a control device for a wet flue gas desulfurization apparatus according to at least one embodiment of the present disclosure; a remote monitoring device electrically connected to the control device of the wet flue gas desulfurization device; Equipped with.

[0010] In order to solve the above problem, an information processing device according to at least one embodiment of the present disclosure includes: An information processing device that executes processing related to control of a wet flue gas desulfurization device that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower, a learning model construction unit for constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creates a table that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; Equipped with.

[0011] In order to solve the above problem, an information processing system according to at least one embodiment of the present disclosure includes: An information processing system including an information processing device that executes processing related to control of a wet flue gas desulfurization device that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower, and a terminal that can communicate with the information processing device, The information processing device includes: a learning model construction unit that, in response to a request from the terminal, constructs a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter that is correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable that is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creates a table that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; Equipped with. [Effects of the Invention]

[0012] According to at least one embodiment of the present disclosure, an object is to provide a control method for a wet flue gas desulfurization apparatus, a control device, a remote monitoring system, an information processing device, and an information processing system that can easily perform control for circulating an absorption liquid in an absorption tower of a wet flue gas desulfurization apparatus. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a configuration diagram of a wet flue gas desulfurization apparatus according to an embodiment. [Figure 2] 1 is a configuration diagram of a remote monitoring system according to an embodiment. [Figure 3] 3 is a flowchart showing a control method for a wet flue gas desulfurization apparatus according to one embodiment. [Figure 4] FIG. 4 is a flowchart showing a correction calculation for a response variable of the learning model in step S3 of FIG. 3. [Figure 5] FIG. 4 is a flowchart showing a correction calculation for explanatory variables of the learning model in step S3 of FIG. 3. [Figure 6] FIG. 6 is a flowchart showing a correction calculation for the explanatory variables of the learning model using the correction values ​​calculated in FIG. 5. [Figure 7] FIG. 7 is a diagram showing a flow of determining optimization conditions for performing correction calculation in FIG. 6. [Figure 8]This is an example of calculation showing the predicted value of SO2 concentration calculated by the learning model when the CaCO3 concentration and the number of fixed-displacement circulation pumps in operation are changed for each generator output. [Figure 9] 9 is an example of a table created based on the calculation results of FIG. 8. [Figure 10] This is an example of calculation showing the predicted value of SO2 concentration calculated by the learning model when the CaCO3 concentration and the capacity of the variable displacement circulation pump are changed for each generator output. [Figure 11] 11 is an example of a table created based on the calculation results of FIG. 10. [Figure 12] 1 is a configuration diagram of an information processing system according to an embodiment. [Figure 13] 13 is a diagram showing the internal configuration of the information processing device of FIG. 12 together with a control device. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, several embodiments of the present invention will be described with reference to the drawings. However, the scope of the present invention is not limited to the following embodiments. The dimensions, materials, shapes, relative positions, etc. of the components described in the following embodiments are merely illustrative examples and are not intended to limit the scope of the present invention thereto.

[0015] FIG. 1 is a configuration diagram of a wet flue gas desulfurization system 10 according to one embodiment. The wet flue gas desulfurization system 10 is an apparatus for desulfurizing exhaust gas generated in a combustion device 1. The combustion device 1 is, for example, a boiler for generating steam, and is configured as part of a power plant that can generate electricity by supplying the steam generated in the combustion device 1 to a generator 5. The wet flue gas desulfurization system 10 includes an absorption tower 11 that communicates with the combustion device 1 via piping 2, a plurality of circulation pumps 12a, 12b, 12c, etc. (In FIG. 1 , three circulation pumps are representatively illustrated, but the number is not limited. Furthermore, these will be referred to collectively as "circulation pumps 12" as appropriate) that are provided on a circulation piping 3 for an absorption liquid circulating within the absorption tower 11, an absorbent slurry supply unit 13 that supplies a slurry (absorbent slurry) of calcium carbonate (CaCO3) that is an absorbent contained in the absorption liquid into the absorption tower 11, and a gypsum recovery unit 14 that recovers gypsum from the absorption liquid. The absorption tower 11 is provided with an outlet pipe 16 through which the exhaust gas desulfurized by the operation described below flows out of the absorption tower 11 as an outflow gas, and the outlet pipe 16 is provided with a gas analyzer 17 for measuring the SO2 concentration in the outflow gas.

[0016] The absorbent slurry supply unit 13 includes an absorbent slurry production facility 21 for producing an absorbent slurry, an absorbent slurry supply pipe 22 that connects the absorbent slurry production facility 21 and the absorption tower 11, and an absorbent slurry supply amount control valve 23 that controls the flow rate of the absorbent slurry flowing through the absorbent slurry supply pipe 22. The gypsum recovery unit 14 includes a gypsum separator 25, a gypsum slurry withdrawal pipe 26 that connects the gypsum separator 25 and the absorption tower 11, and a gypsum slurry withdrawal pump 27 provided in the gypsum slurry withdrawal pipe 26.

[0017] The wet flue gas desulfurization system 10 is provided with a control device 15 for the wet flue gas desulfurization system 10. The control device 15 has an operation data receiving unit 30 electrically connected to an operation data acquiring unit 20 including various detectors for acquiring various operation data of the combustion device 1 and the wet flue gas desulfurization system 10 (for example, temperatures and pressures at various locations, flow rates of various fluids, etc.). The operation data acquiring unit 20 includes a gas analyzer 17.

[0018] The control device 15 includes a learning model construction unit 38 electrically connected to the operation data receiving unit 30, a table creation unit 31 electrically connected to the learning model construction unit 38, a control target value determination unit 32 electrically connected to the table creation unit 31, a circulation pump adjustment unit 33 electrically connected to the control target value determination unit 32, an absorbent slurry supply control unit 34 electrically connected to the control target value determination unit 32, and a learning model correction unit 35 electrically connected to the learning model construction unit 38. The circulation pump adjustment unit 33 is electrically connected to each of the circulation pumps 12. The absorbent slurry supply control unit 34 is electrically connected to the absorbent slurry supply amount control valve 23.

[0019] 2 shows the configuration of a remote monitoring system 40 for remotely monitoring the control status of the wet flue gas desulfurization system 10 (see FIG. 1). The remote monitoring system 40 includes a distributed control system (DCS) 41 for each piece of equipment that constitutes the combustion system 1 (see FIG. 1) and the wet flue gas desulfurization system 10 (see FIG. 1), an edge server 42 that is electrically connected to the DCS 41 and has a control device 15 mounted thereon, and a remote monitoring device 43 such as a desktop personal computer or tablet computer that is electrically connected to the edge server 42 via the cloud or a virtual private network (VPN) so as to be able to communicate with the edge server 42. The control status of the wet flue gas desulfurization system 10 can be remotely monitored by the remote monitoring device 43, which is usually located away from the edge server 42.

[0020] Next, the operation of the wet flue gas desulfurization system 10 to desulfurize the exhaust gas generated in the combustion system 1 will be described. As shown in Fig. 1, exhaust gas generated in combustion apparatus 1 flows through pipe 2 and enters absorption tower 11, and rises within absorption tower 11. By operating at least one circulation pump 12, the absorption liquid flows through circulation pipe 3 and enters absorption tower 11, and the absorption liquid flows downward within absorption tower 11. The absorption liquid that flows downward within absorption tower 11 accumulates within absorption tower 11, is caused to flow out of absorption tower 11 by circulation pump 12, and flows through circulation pipe 3. In this manner, the absorption liquid circulates within absorption tower 11.

[0021] In the absorption tower 11, the rising flue gas and the flowing down absorbing liquid come into gas-liquid contact. SO2 contained in the flue gas is reacted with the following reaction formula: SO2+CaCO3+2H2O+1 / 2O2→CaSO4·2H2O+CO2 As shown above, it reacts with the CaCO3 in the absorption solution, causing gypsum (CaSO4·2H2O) to precipitate in the absorption solution.

[0022] In this way, part of the SO2 in the flue gas is removed as gypsum in the absorption solution, i.e., the flue gas is desulfurized, so that the SO2 concentration in the effluent gas flowing out from the absorber 11 via the outlet pipe 16 is lower than the SO2 concentration in the flue gas flowing into the absorber 11 via the pipe 2. The effluent gas flowing out from the absorber 11 flows through the outlet pipe 16 and is released into the atmosphere, and along the way, the SO2 concentration is measured by the gas analyzer 17, and the measurement result is transmitted to the operation data receiving unit 30 of the control device 15.

[0023] Unless there is a large change in the CaCO concentration in the absorbing solution, the SO2 concentration in the effluent gas tends to decrease as the circulation flow rate of the absorbing solution circulating in the absorption tower 11 increases. By controlling the circulation flow rate by controlling the number of operating circulation pumps 12 using a control method described below, the control device 15 can control the SO2 concentration in the effluent gas, for example, so that the SO2 concentration in the effluent gas is equal to or less than a preset value.

[0024] The gypsum precipitated in the absorption solution in the absorption tower 11 is extracted as a gypsum slurry from the absorption tower 11 by a gypsum slurry extraction pump 27, and the gypsum slurry flows through a gypsum slurry extraction pipe 26 and flows into a gypsum separator 25. In the gypsum separator 25, the gypsum and water are separated, the gypsum is recovered, and the water is sent to a drainage facility (not shown).

[0025] Since CaCO3 in the absorption solution reacts with SO2 to form gypsum, the CaCO3 concentration in the absorption solution decreases as the flue gas is desulfurized. Using a control method described below, the control device 15 controls the aperture of the absorbent slurry supply amount control valve 23, and supplies the absorbent slurry produced in the absorbent slurry production facility 21 into the absorption tower 11 via the absorbent slurry supply pipe 22. This keeps the CaCO3 concentration in the absorption solution within a preset range, suppressing large fluctuations in the CaCO3 concentration during the flue gas desulfurization.

[0026] Next, a description will be given of a method for controlling the wet flue gas desulfurization system 10 by the control device 15. Fig. 3 is a flowchart showing a method for controlling the wet flue gas desulfurization system 10 according to one embodiment.

[0027] First, in step S1, various operational data for the combustion device 1 and the wet flue gas desulfurization system 10 are collected. Then, in step S2, a learning model is constructed by machine learning regarding the relationship between the various operational data and the future SO concentration in the effluent gas flowing from the absorber 11. Next, in step S3, the learning model constructed in step S2 is corrected. Then, in step S4, a table is created using the corrected learning model obtained in step S3. In the subsequent step S5, based on the table created in step S4, control target values ​​are determined for the circulation flow rate of the absorbent and the supply rate of the CaCO3 absorbent slurry so that the SO2 concentration in the effluent gas is below a predetermined set value. In step S6, the operating conditions of the circulation pump 12 are adjusted based on the control target value of the circulation flow rate determined in step S5. In step S7, the absorbent slurry supply rate control valve 23 is controlled based on the control target value of the supply rate of the CaCO3 absorbent slurry determined in step S5. This controls the SO2 concentration in the effluent gas to be below a predetermined set value.

[0028] Note that the correction of the learning model in step S3 may be omitted if the accuracy of the learning model before correction is sufficient. Also, while Fig. 3 illustrates the case where step S7 is performed after step S6, step S7 may be performed before step S6, or steps S6 and S7 may be performed simultaneously.

[0029] Next, each step of the method for controlling the wet flue gas desulfurization system 10 by the control device 15 will be described in detail. 1, in step S1, the operation data acquisition unit 20 acquires various operation data of the combustion device 1 and the wet flue gas desulfurization device 10, and the acquired various operation data is then transmitted to the control device 15 and received by the operation data receiving unit 30, whereby the control device 15 collects the various operation data. As described above, the operation data acquisition unit 20 includes the gas analyzer 17, and therefore the various operation data includes the SO2 concentration in the outflow gas.

[0030] In step S2, the learning model construction unit 38 constructs a learning model by machine learning regarding the relationship between the various operating data collected by the operating data acquisition unit 20 and the future SO2 concentration in the outflow gas. The learning model is constructed as a regression model using a regression method such as multiple regression, ridge regression, lasso regression, or elastic net. In this embodiment, as an example of the learning model, a regression model expressed as a linear polynomial as shown in the following equation is constructed. SO2 concentration at the outlet of the absorber = k1 × explanatory variable 1 + k2 × explanatory variable 2 + ··· + kn × explanatory variable n + b (1) By using a linear polynomial as a learning model in this way, it is possible to achieve higher explainability (interpretability) than a complex simulation model, and it is also possible to effectively reduce the computational load. Note that n is an arbitrary natural number, k1 to kn are coefficients, and b is an arbitrary intercept.

[0031] The learning model obtained by machine learning is constructed as a model showing the correlation between explanatory variables consisting of multiple parameters included in the operating data acquired by the operating data acquisition unit 20 and the future SO2 concentration in the outflow gas as a target variable. Here, the combination of multiple parameters included in the explanatory variables of the learning model can be arbitrarily selected from the following candidates. i) Generator output command value (external output command value for generator 5) ii) Generator output (output of generator 5) iii) Boiler air flow rate or boiler exhaust gas flow rate (air supply flow rate to combustion device 1 or exhaust gas flow rate from combustion device 1) iv) Desulfurization inlet SO2 concentration or boiler outlet SO2 concentration (SO2 concentration at the inlet of the absorption tower 11 or SO2 concentration at the outlet of the combustion device 1) v) Desulfurization outlet SO2 concentration or stack inlet SO2 (SO2 concentration at the outlet of the absorber 11) vi) CaCO3 concentration or pH of the absorption solution vii) Absorbent circulation flow rate (number of operating circulation pumps 12 or control value of discharge flow rate) These candidates are typically parameters that can be conventionally measured in many wet flue gas desulfurization systems 10 .

[0032] In this embodiment, the explanatory variables of the learning model are selected to include at least one of the following candidates: i) the generator output command value (an external output command value for the generator 5); iii) the boiler air flow rate or the boiler exhaust gas flow rate (the supply air flow rate to the combustion device 1 or the exhaust gas flow rate from the combustion device 1); or iv) the desulfurization inlet SO2 concentration or the boiler outlet SO2 concentration (the SO2 concentration at the inlet of the absorber 11 or the SO2 concentration at the outlet of the combustion device 1). More preferably, the explanatory variables of the learning model are selected to include, from the above candidates, iii) the boiler air flow rate or the boiler exhaust gas flow rate (the supply air flow rate to the combustion device 1 or the exhaust gas flow rate from the combustion device 1), and iv) the desulfurization inlet SO2 concentration or the boiler outlet SO2 concentration (the SO2 concentration at the inlet of the absorber 11 or the SO2 concentration at the outlet of the combustion device 1). This allows for the construction of a learning model that can predict with high accuracy the future SO2 concentration in the effluent gas, which is the objective function. Furthermore, by selecting some parameters as explanatory variables from the driving data acquired by the driving data acquisition unit 20 in this way, it is possible to efficiently narrow down the learning target data and reduce the computational burden of machine learning.

[0033] In step S3, the learning model constructed in step S2 is corrected by the learning model correction unit 35. The learning model is corrected based on the error between the predicted value of the SO2 concentration calculated by the learning model and the actual measured value, and is corrected for each of the explanatory variables and the objective function of the learning model.

[0034] FIG. 4 is a flowchart showing the correction calculation for the objective variable of the learning model in step S3 of FIG.

[0035] In the correction calculation for the objective variable, first, a predicted value Vp calculated using the learning model to be corrected is delayed by a correction time ΔVp corresponding to the measurement delay and output. The predicted value Vp calculated using the learning model does not include the time required to measure the actual measured value Vm by the gas analyzer 17. The correction time ΔVp corresponds to the time required to measure the actual measured value Vm by the gas analyzer 17, and by outputting the predicted value Vp with a delay, it becomes possible to compare the predicted value Vp with the actual measured value Vm measured by the gas analyzer 17. A corrected predicted value Vp' is obtained by outputting the predicted value Vp with a delay of the correction time ΔVp. Then, the ratio (Vm / Vp') of the corrected predicted value Vp' to the actual measured value Vm of the gas analyzer 17 is calculated, and a moving average of this ratio is calculated to obtain a correction value A for the objective function.

[0036] The correction value A calculated in this way is applied to the learning model. For example, in the case of the learning model expressed by the above formula (1), the learning model after correction is expressed by the following formula. SO2 concentration at the outlet of the absorber = (k1 × explanatory variable 1 + k2 × explanatory variable 2 + ··· + kn × explanatory variable n + b) × A (2)

[0037] Next, Fig. 5 is a flow diagram showing the correction calculation for the explanatory variables of the learning model in step S3 of Fig. 3. Here, as an example of the correction calculation, a case will be described in which the explanatory variable SG, which is a parameter correlated with the output of the generator 5 (generator output Y), is corrected.

[0038] A reference table Tr is prepared in advance, which defines the correlation between the generator output Y and the values ​​X1 to X4 of the explanatory variable SG. As shown in Fig. 5, the reference table Tr defines the values ​​X1 to X4 of the explanatory variable SG for each value of the generator output Y. In this embodiment, the reference table Tr includes characteristic functions which define the values ​​X1 to X4 of the explanatory variable SG for load points where the generator output Y is 25% MW, 50% MW, 75% MW, and 100% MW.

[0039] The learning model correction unit 35 inputs the generator output Y acquired from an external source into the reference table Tr, and outputs the corresponding explanatory variable SG. Meanwhile, the learning model correction unit 35 obtains the actual measured value SGm of the explanatory variable based on the results acquired by the operation data receiving unit 30, and obtains the ratio (SGm / SG) to the value SG of the explanatory variable output from the reference table Tr. Then, the moving average of this ratio is calculated to obtain the correction value B for the explanatory function.

[0040] The correction value B for the explanatory variable SG calculated in this way is applied to the learning model. Figure 6 is a flow diagram showing the correction calculation of the explanatory variable SG of the learning model using the correction value B calculated in Figure 5.

[0041] 6, when optimization condition C is met, the characteristic function defined in reference table Tr is corrected by correction value B. Optimization condition C includes at least one (e.g., all) conditions for determining whether the environment is ready for correcting reference table Tr using correction value B, and can determine, for example, from multiple perspectives whether the wet flue gas desulfurization system 10 and the control device 15 are operating in a stable state and no abnormalities have occurred.

[0042] Fig. 7 is a diagram showing a determination flow of optimization condition C for performing correction calculation in Fig. 6. In Fig. 7, as an example of optimization condition C, determination is made based on the following five conditions. (Condition 1) No abnormalities are observed in the explanatory variables. (Condition 2) The instrument used to obtain the actual measured value SGm of the explanatory variable is not currently being adjusted. (Condition 3) The generator output (or boiler load) is equal to or greater than a predetermined value (for example, 0% to 50%). (Condition 4) The plant is in a smoke-free state. (Condition 5) The initialization of the calculation software has been completed.

[0043] When such optimization condition C is satisfied, the learning model correction unit 35 corrects the explanatory variables for the characteristic functions defined in the reference table Tr by multiplying the value of the explanatory variable SG corresponding to the generator output by the correction value B. As a result of such correction, the characteristic functions defined in the reference table Tr are updated as shown in Fig. 6 (the dashed-dotted line indicates the characteristic function before correction, and the solid line indicates the characteristic function after correction).

[0044] In step S4, table T is created based on the learning model corrected in step S3. Table T is created by calculating, for each generator output, the amount of absorbent to be added and the circulation flow rate of the absorption solution so that the predicted value of the SO2 concentration calculated by the learning model satisfies a preset reference value (for example, is equal to or less than the reference value).

[0045] Here, a specific description will be given of a method for creating Table T in step S4 of Fig. 3. First, a case where the circulation pump 12 is a fixed displacement type and the circulation flow rate of the absorption solution is controlled by controlling the number of circulation pumps 12 will be described with reference to Figs. 8 and 9. Fig. 8 is a calculation example showing predicted values ​​of SO2 concentration calculated by a learning model when the CaCO3 concentration and the number of operating fixed displacement circulation pumps 12 are changed for each generator output, and Fig. 9 is an example of Table T created based on the calculation results of Fig. 8.

[0046] The CaCO3 concentration is a parameter corresponding to the amount of absorbent introduced, and its allowable range is between 2 and 5. The number of operating circulating pumps 12 is a parameter corresponding to the amount of circulating absorbent, and its allowable range is between 8 and 10. These allowable ranges may be preset, for example, based on plant design values.

[0047] In FIG. 8, first, let's look at the case where the generator output is "50%." When the CaCO3 concentration is "3" and the number of operating circulation pumps is "8," the predicted value of the SO2 concentration by the learning model is 105 ppm, exceeding the reference value (100 ppm). In this case, the table creation unit 31 increases the CaCO3 concentration by one level, which has a lower impact on operating costs than the number of operating circulation pumps, to "4." As a result, the predicted value of the SO2 concentration by the learning model becomes 100 ppm, which is below the reference value (100 ppm). Therefore, the table creation unit 31 determines that the optimal CaCO3 concentration when the generator output is "50%" is "4" and the number of operating circulation pumps is "8."

[0048] Next, let's look at the case where the generator output is "60%." When the CaCO3 concentration is "4" and the number of operating circulation pumps is "8," the predicted SO2 concentration by the learning model is 120 ppm, exceeding the reference value (100 ppm). In this case, the table creation unit 31 increases the CaCO3 concentration by one level to "5," which has a lower impact on operating costs than the number of operating circulation pumps. However, the predicted SO2 concentration by the learning model is 110 ppm, still exceeding the reference value (100 ppm). In this case, even if the CaCO3 concentration reaches the upper limit of the allowable range, "5," it cannot satisfy the reference value. Therefore, the number of operating circulation pumps is increased to "9" and the CaCO3 concentration is decreased to "3." As a result, the predicted SO2 concentration by the learning model becomes 100 ppm, below the reference value (100 ppm). Therefore, the table creation unit 31 determines that the optimal CaCO3 concentration when the generator output is "60%" is "3" and the number of operating pumps is "9".

[0049] In this way, the table creation unit 31 creates table T shown in Fig. 9 by identifying the optimal CaCO3 concentration and the number of operating circulation pumps for each generator output. In Fig. 9, table T is created so that as the generator output increases, the CaCO3 concentration decreases and increases at the timing when the number of operating circulation pumps 12 increases in a stepwise manner.

[0050] Next, a case where the circulation flow rate of the absorption solution is controlled by controlling the capacity of the circulation pump 12, when the circulation pump 12 is a variable displacement pump, will be described with reference to Figures 10 and 11. Figure 10 shows an example of calculations showing predicted values ​​of the SO2 concentration calculated by the learning model when the CaCO3 concentration and the capacity of the variable displacement circulation pump 12 are changed for each generator output, and Figure 11 shows an example of Table T created based on the calculation results of Figure 10.

[0051] The CaCO3 concentration is a parameter corresponding to the amount of absorbent introduced, and its allowable range is between 2 and 5. The circulation pump capacity is a parameter corresponding to the amount of circulating absorbent, and its allowable range is between 10 and 100. These allowable ranges may be preset, for example, based on plant design values.

[0052] In FIG. 10, first, let us consider the case where the generator output is "50%." When the CaCO3 concentration is "3" and the circulation pump capacity is "10," the predicted value of the SO2 concentration by the learning model is 105 ppm, exceeding the reference value (100 ppm). In this case, the table creation unit 31 increases the CaCO3 concentration by one level, which has a lower impact on operating costs than the circulation pump capacity, to "4." As a result, the predicted value of the SO2 concentration by the learning model becomes 100 ppm, which is below the reference value (100 ppm). Therefore, the table creation unit 31 determines that the optimal CaCO3 concentration when the generator output is "50%" is "4" and the circulation pump capacity is "10."

[0053] Next, when the generator output is 60%, if the CaCO3 concentration is 5 and the circulation pump capacity is 10, the predicted value of the SO2 concentration by the learning model is 120 ppm, exceeding the reference value (100 ppm). In this case, the table creation unit 31 changes the pump capacity to 15 because the CaCO3 concentration has reached the upper limit of the allowable range, 5. However, the predicted value of the SO2 concentration by the learning model is 110 ppm, still exceeding the reference value (100 ppm). In this case, the table creation unit 31 further changes the circulation pump capacity to 18. As a result, the predicted value of the SO2 concentration by the learning model becomes 100 ppm, below the reference value (100 ppm). Therefore, the table creation unit 31 determines that the optimal CaCO3 concentration at a generator output of 60% is 5 and the circulation pump capacity is 18.

[0054] In this way, the table creation unit 31 creates the table shown in Figure 11 by specifying the optimal CaCO3 concentration and circulation pump capacity for each generator output. In Figure 11, table T is created so that as the generator output increases, the pump capacity increases at the point when the increase in CaCO3 concentration is no longer sufficient.

[0055] 8 to 11 show a case where the search for the optimal CaCO3 concentration and the number of operating circulation pumps (or circulation pump capacity) for creating the table is performed so as to obtain a predicted value of SO2 concentration that satisfies the standard value while keeping the number of operating circulation pumps (or circulation pump capacity) as small as possible in order to minimize operating costs, but the search may also be performed from other perspectives.

[0056] In this embodiment, the table T is in the form of a graph as shown in FIGS. 9 and 11, but the table T does not necessarily have to be in this form, and may be in the form of a matrix, a mathematical formula, or the like.

[0057] In step S5, the control target value determination unit 32 acquires a command value for the generator output, and determines control target values ​​(respective control target values ​​for the input amount of absorbent and the circulation flow rate of the absorption liquid) corresponding to the generator output based on the table T created in step S4. Then, in steps S6 and S7, based on the control target values ​​determined in step S5, the circulation pumps 12a to 12c are controlled by the circulation pump adjustment unit 33, and the absorbent slurry supply control unit 34 controls the supply amount of absorbent.

[0058] In step S5, the circulation flow rate is adjusted by controlling the number of circulation pumps based on the table shown in FIG. 9 if the circulation pumps 12 are fixed displacement pumps, and by controlling the pump displacement based on table T shown in FIG. 11 if the circulation pumps 12 are variable displacement pumps.

[0059] As described above, according to the above embodiment, the amount of absorbent charged to the absorber 11 and the circulation flow rate of the absorbing solution circulating in the absorber 11 can be appropriately adjusted within a range in which the future SO2 concentration in the outflow gas will be equal to or less than a preset reference value. Such control calculations can determine the control target values ​​for the amount of absorbent charged to the absorber 11 and the circulation flow rate of the absorbing solution circulating in the absorber 11 based on a single learning model, and therefore the calculation load is low.

[0060] It is also possible to adopt a configuration in which each process executed by the control device 15 in the above embodiment is executed by an information processing device 44 (see FIG. 12) electrically connected to the edge server 42 via a cloud environment or VPN so as to be able to communicate with the edge server 42. In this case, the information processing device 44 includes an operation data receiving unit 30, a learning model construction unit 38, a learning model correction unit 35, a table creation unit 31, and a control target value determination unit 32, and may communicate the control target value determined by the control target value determination unit 32 to the circulation pump adjustment unit 33 and the absorbent slurry supply control unit 34 in the control device 15 to control the circulation pump and the supply amount of absorbent.

[0061] The operation data receiving unit 30 may also receive various operation data via the operation data relay unit 39 (FIG. 13) of the control device 15, or may receive various operation data from the operation data acquiring unit 20 as described above. In particular, when computing in a cloud environment, from a security perspective, the target control values ​​of the circulation pump and absorbent may not be directly controlled but may be displayed. For example, an operation index diagram generated in the cloud environment may be transmitted and displayed to a customer's device (terminal 45) via a dedicated app, and the customer may manually update the on-site operation index diagram. Meanwhile, the information processing device 44 may also include a circulation pump adjusting unit 33 and an absorbent slurry supply control unit 34 to remotely control the circulation pump and absorbent supply rate. Furthermore, the information processing device 44 may be configured to execute various processes in the information processing device 44 in response to a request from the terminal 45.

[0062] In addition, within the scope of the present disclosure, the components in the above-described embodiments may be replaced with well-known components as appropriate, and the above-described embodiments may be combined as appropriate.

[0063] The contents described in each of the above embodiments can be understood, for example, as follows.

[0064] (1) A method for controlling a wet flue gas desulfurization system according to one embodiment includes: A control method for a wet flue gas desulfurization apparatus (e.g., the wet flue gas desulfurization apparatus 10 of the above embodiment) that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower (e.g., the absorption tower 11 of the above embodiment), constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; calculating a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table (e.g., the reference table Tr in the above embodiment) that defines the relationship between the parameters and the output, and creating a table (e.g., the table T in the above embodiment) that indicates the amount of absorbent introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; determining, based on the table, control target values ​​for the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator; Equipped with.

[0065] According to the above aspect (1), the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorbing solution circulating in the absorption tower can be appropriately adjusted within a range in which the future SO2 concentration in the effluent gas will be equal to or less than a preset reference value. Such control calculations can determine the control target values ​​for the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorbing solution circulating in the absorption tower based on a single learning model, so the calculation load is low.

[0066] (2) In another embodiment, in the above embodiment (1), In the process of creating the table, the amount of absorbent introduced and the circulation flow rate of the absorption liquid are searched for so that the predicted value satisfies the reference value, within the preset allowable ranges of the amount of absorbent introduced and the absorption liquid, so as to reduce the operating cost of at least one circulation pump (e.g., circulation pump 12 in the above embodiment) for circulating the absorption liquid in the absorption tower.

[0067] According to the above aspect (2), the control target values ​​of the absorbent input amount and the absorbing solution circulation flow rate are searched for so as to reduce operating costs within a range in which the predicted value of sulfur dioxide discharged from the absorber is equal to or less than the reference value. By performing control based on the control target values ​​searched in this manner, it is possible to effectively reduce plant operating costs while suppressing sulfur dioxide discharged from the absorber within a necessary range.

[0068] (3) In another embodiment, in the above embodiment (1), The method further includes a step of correcting the learning model based on an error between a predicted value of the sulfur dioxide concentration of the exhaust gas at the outlet of the absorption tower by the learning model and an actual measured value.

[0069] According to the above aspect (3), the learning model is corrected based on the error between the predicted value obtained by the learning model and the actual measurement value, thereby improving the control accuracy.

[0070] (4) In another embodiment, in any one of the above (1) to (3), The at least one parameter includes at least one of an external command value of the output, the output, an air flow rate of the combustion device, and a sulfur dioxide concentration at an absorption tower inlet.

[0071] According to the above aspect (4), by including any one of these parameters in the explanatory variables of the learning model, it is possible to efficiently narrow down the data to be learned, thereby reducing the computational burden of machine learning and constructing a learning model that can make predictions with good accuracy.

[0072] (5) In another embodiment, in the above embodiment (4), The at least one parameter includes an air flow rate of the combustor and a sulfur dioxide concentration at the absorber inlet.

[0073] According to the above aspect (5), by including these parameters in the explanatory variables of the learning model, it is possible to efficiently narrow down the data to be learned, thereby further reducing the computational burden of machine learning and constructing a learning model that can make predictions with better accuracy.

[0074] (6) In another embodiment, in any one of the above (1) to (5), The learning model is expressed as a linear polynomial.

[0075] According to the above aspect (6), by using a linear polynomial as the learning model, the explainability (interpretability) is higher than that of a complex simulation model, and the calculation load can be effectively reduced.

[0076] (7) In another embodiment, in any one of the above (1) to (6), The reference table is defined as a function that indicates the relationship between the pre-specified parameters and the output.

[0077] According to the above aspect (7), the parameters having a correlation with the generator output are defined as a reference table indicating the correlation.

[0078] (8) In another embodiment, in any one of the above (1) to (7), The at least one circulation pump is of a fixed displacement type.

[0079] According to the above aspect (8), in a plant having a fixed displacement circulation pump, the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorption solution circulating within the absorption tower can be appropriately adjusted within a range in which the future sulfur dioxide concentration in the effluent gas will be equal to or less than a predetermined reference value.

[0080] (9) In another embodiment, in any one of the above (1) to (7), The at least one circulation pump is of a variable displacement type.

[0081] According to the above aspect (9), in a plant having a variable displacement circulation pump, the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorption solution circulating within the absorption tower can be appropriately adjusted within a range in which the future sulfur dioxide concentration in the effluent gas will be equal to or less than a predetermined reference value.

[0082] (10) A control device for a wet flue gas desulfurization apparatus according to one embodiment includes: A control device (e.g., the control device 15 of the above embodiment) for a wet flue gas desulfurization device (e.g., the wet flue gas desulfurization device 10 of the above embodiment) that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower (e.g., the absorption tower 11 of the above embodiment), a learning model construction unit (e.g., the learning model construction unit 38 in the above embodiment) for constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit (e.g., the table creation unit 31 in the above embodiment) that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table (e.g., the reference table Tr in the above embodiment) that defines the relationship between the parameters and the output, and creates a table (e.g., the table T in the above embodiment) that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit (for example, the control target value determination unit 32 in the above embodiment) for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; Equipped with.

[0083] According to the above aspect (10), the amount of absorbent introduced into the absorber tower and the circulation flow rate of the absorbing solution circulating in the absorber tower can be appropriately adjusted within a range in which the future SO2 concentration in the outflow gas will be equal to or less than a preset reference value. Such control calculations can determine the control target values ​​for the amount of absorbent introduced into the absorber tower 11 and the circulation flow rate of the absorbing solution circulating in the absorber tower based on a single learning model, so the calculation load is low.

[0084] (11) A remote monitoring system according to one aspect includes: A control device for a wet flue gas desulfurization apparatus according to the above aspect (10); a remote monitoring device electrically connected to the control device of the wet flue gas desulfurization device; Equipped with.

[0085] According to the above aspect (11), the control state of the wet flue gas desulfurization apparatus can be remotely monitored.

[0086] (12) An information processing device according to one aspect includes: An information processing device (for example, the information processing device 44 of the above embodiment) that executes processing related to control of a wet flue gas desulfurization device (for example, the wet flue gas desulfurization device 10 of the above embodiment) that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower (for example, the absorption tower 11 of the above embodiment), a learning model construction unit (e.g., the learning model construction unit 38 in the above embodiment) for constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit (e.g., the table creation unit 31 in the above embodiment) that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table (e.g., the reference table Tr in the above embodiment) that defines the relationship between the parameters and the output, and creates a table (e.g., the table T in the above embodiment) that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit (for example, the control target value determination unit 32 in the above embodiment) for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; Equipped with.

[0087] According to the above aspect (12), the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorbing solution circulating in the absorption tower can be appropriately adjusted within a range in which the future SO2 concentration in the effluent gas will be equal to or less than a preset reference value. Such control calculations can determine the control target values ​​for the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorbing solution circulating in the absorption tower based on a single learning model, and therefore the calculation load is low.

[0088] (13) An information processing system according to one aspect includes: An information processing system (e.g., the information processing system 46 of the above embodiment) including an information processing device (e.g., the information processing device 44 of the above embodiment) that executes processing related to control of a wet flue gas desulfurization device (e.g., the wet flue gas desulfurization device 10 of the above embodiment) that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower (e.g., the absorption tower 11 of the above embodiment), and a terminal (e.g., the terminal 45 of the above embodiment) that can communicate with the information processing device, The information processing device includes: a learning model construction unit (for example, the learning model construction unit 38 in the above embodiment) for constructing a learning model by machine learning, in response to a request from the terminal, regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit (e.g., the table creation unit 31 in the above embodiment) that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table (e.g., the reference table Tr in the above embodiment) that defines the relationship between the parameters and the output, and creates a table (e.g., the table T in the above embodiment) that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit (for example, the control target value determination unit 32 in the above embodiment) for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; Equipped with.

[0089] According to the above aspect (13), the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorbing solution circulating in the absorption tower can be appropriately adjusted within a range in which the future SO2 concentration in the effluent gas will be equal to or less than a preset reference value. Such control calculations can determine the control target values ​​for the amount of absorbent introduced into the absorption tower and the circulation flow rate of the absorbing solution circulating in the absorption tower based on a single learning model, and therefore the calculation load is low. [Explanation of symbols]

[0090] 1 Combustion equipment 2 Piping 3 Circulation piping 5. Generator 10 Wet flue gas desulfurization equipment 11 Absorption tower 12 Circulation pump 13 Absorbent slurry supply section 14 Gypsum Recovery Department 15 Control device 16 Outlet piping 17 Gas analyzer 20 Operation data acquisition unit 21 Absorbent slurry production facility 22 Absorbent slurry supply pipe 23 Absorbent slurry supply control valve 25 Gypsum separator 26 Gypsum slurry extraction pipe 27 Gypsum slurry extraction pump 30 Operation data receiver 31 Table Creation Section 32 Control target value determination unit 33 Circulation pump adjustment section 34 Absorbent slurry supply control unit 35 Learning model correction unit 38 Learning Model Construction Department 39 Operation data relay unit 40 Remote Monitoring System 42 Edge Servers 43 Remote monitoring equipment 44 Information processing equipment 45 terminals 46 Information Processing Systems

Claims

1. A control method for a wet flue gas desulfurization apparatus that performs desulfurization by bringing exhaust gas generated in a combustion apparatus into gas-liquid contact with an absorption liquid in an absorption tower, comprising: constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; calculating a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creating a table that indicates the amount of absorbent introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; determining, based on the table, control target values ​​for the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator; A method for controlling a wet flue gas desulfurization system.

2. 2. The control method for a wet flue gas desulfurization apparatus according to claim 1, wherein in the step of creating the table, a search is made for an amount of absorbent to be charged and a circulation flow rate of the absorption solution that will cause the predicted value to satisfy the reference value, within a preset allowable range of the amount of absorbent to be charged and the amount of the absorption solution, so as to reduce an operating cost of at least one circulation pump that circulates the absorption solution within the absorption tower.

3. 3. The control method for a wet flue gas desulfurization apparatus according to claim 1, further comprising a step of correcting the learning model based on an error between a predicted value of the sulfur dioxide concentration of the exhaust gas at the outlet of the absorption tower by the learning model and an actual measured value.

4. 4. The control method for a wet flue gas desulfurization apparatus according to claim 1, wherein the at least one parameter includes at least one of an external output command value for the generator, the output, an air flow rate of the combustion apparatus, and a sulfur dioxide concentration at an absorption tower inlet.

5. The method for controlling a wet flue gas desulfurization system according to claim 4 , wherein the at least one parameter includes an air flow rate of the combustion device and a sulfur dioxide concentration at the inlet of the absorber.

6. The method for controlling a wet flue gas desulfurization system according to claim 1 , wherein the learning model is expressed as a linear polynomial.

7. The method for controlling a wet flue gas desulfurization system according to claim 1 , wherein the reference table is defined as a function indicating a relationship between the parameter specified in advance and the output.

8. The wet flue gas desulfurization apparatus includes at least one circulation pump for circulating the absorption liquid in the absorption tower, The method for controlling a wet flue gas desulfurization system according to claim 1 , wherein the at least one circulation pump is a fixed displacement pump.

9. The wet flue gas desulfurization apparatus includes at least one circulation pump for circulating the absorption liquid in the absorption tower, The method for controlling a wet flue gas desulfurization system according to claim 1 , wherein the at least one circulation pump is a variable displacement pump.

10. A control device for a wet flue gas desulfurization device that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower, a learning model construction unit for constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creates a table that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; A control device for a wet flue gas desulfurization device comprising:

11. The control device for a wet flue gas desulfurization system according to claim 10; a remote monitoring device electrically connected to the control device of the wet flue gas desulfurization device; A remote monitoring system comprising:

12. An information processing device that executes processing related to control of a wet flue gas desulfurization device that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower, a learning model construction unit for constructing a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable which is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creates a table that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; An information processing device comprising:

13. An information processing system including an information processing device that executes processing related to control of a wet flue gas desulfurization device that performs desulfurization by bringing exhaust gas generated in a combustion device into gas-liquid contact with an absorption liquid in an absorption tower, and a terminal that can communicate with the information processing device, The information processing device includes: a learning model construction unit that, in response to a request from the terminal, constructs a learning model by machine learning regarding the relationship between explanatory variables including at least one parameter that is correlated with the absorbent concentration and circulation flow rate of the absorption liquid in the absorption tower and the output of a generator driven by the gas generated in the combustion device, and a target variable that is a future sulfur dioxide concentration at an outlet of the absorption tower; a table creation unit that calculates a predicted value of the sulfur dioxide concentration by the learning model for each output based on a reference table that defines the relationship between the parameters and the output, and creates a table that indicates the amount of absorbent to be introduced and the circulation flow rate of the absorption liquid so that the predicted value satisfies a reference value; a control target value determination unit for determining control target values ​​of the input amount of the absorbent and the circulation flow rate of the absorption liquid corresponding to the state of the generator based on the table; An information processing system comprising:

Images