Hydrogen-ammonia co-firing system, co-firing control device, and co-firing control method

Abstract

To provide a mixed combustion system, a mixed combustion control device, and a mixed combustion control method that can appropriately control a mixed combustion ratio.SOLUTION: A mixed combustion system comprises a combustion device for performing mixed combustion of hydrogen or ammonia and fossil fuel, an exhaust gas aftertreatment device for treating exhaust gas of the combustion device, and a mixed combustion control device for determining a mixed combustion ratio between the hydrogen or the ammonia and the fossil fuel to be burned by the combustion device so as to optimize one or a plurality of predetermined variables, and controlling combustion by the combustion device. The one or the plurality of variables include at least one of energy efficiency, energy cost, or carbon dioxide recovery efficiency.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a co-firing system, a co-firing control device, and a co-firing control method.

Background Art

[0002] Patent Document 1 describes the following system. That is, the system described in Patent Document 1 includes a gas turbine system. Further, this gas turbine system includes a gas turbine, a post-treatment system configured to receive exhaust gas from the gas turbine system, and a control device. Further, this control device receives an input, models the operating behavior of an industrial plant having a gas turbine and a post-treatment system based on the input, determines one or more operating parameter set values for the industrial plant, selects one or more operating parameter set values that reduce the output of a cost function, and is configured to apply one or more operating parameter set values to control the industrial plant. In the system described in Patent Document 1, by generating set values for the gas turbine and the post-treatment system in real time, for example, by improving efficiency or reducing emissions, it is possible to reduce fuel costs and lifetime operating costs while maintaining compliance over a long period and providing catalyst integrity monitoring.

[0003] By the way, as one of the technologies for reducing CO2 (carbon dioxide) emissions, there is a technology for reducing the amount of CO2 emissions by co-firing fossil fuel and hydrogen or ammonia in existing facilities that use fossil fuel. When co-firing fossil fuel and hydrogen or ammonia, appropriate control including the co-firing rate is required based on the differences in the combustion characteristics of fossil fuel and hydrogen or ammonia. The combustion characteristics referred to here are, for example, calorific value, combustion rate, and exhaust gas composition, and they affect the performance of combustion devices and exhaust gas treatment devices.

Prior Art Documents

Patent Documents

[0004] [Patent Document 1] U.S. Patent No. 10082060 [Overview of the project] [Problems that the invention aims to solve]

[0005] This disclosure has been made in view of the above circumstances and aims to provide a co-firing system, a co-firing control device, and a co-firing control method that can appropriately control the co-firing ratio. [Means for solving the problem]

[0006] To achieve the above objectives, the co-firing system according to the present disclosure comprises a combustion device for co-firing hydrogen or ammonia with a fossil fuel; an exhaust gas aftertreatment device for treating the exhaust gas of the combustion device; and a co-firing control device that controls combustion by the combustion device by determining the co-firing ratio of the hydrogen or ammonia and the fossil fuel burned by the combustion device to optimize one or more predetermined variables, wherein the one or more variables include at least one of energy efficiency, energy cost, or carbon dioxide capture efficiency.

[0007] The co-firing control device according to this disclosure is a co-firing system comprising a combustion device that co-fires hydrogen or ammonia with a fossil fuel, and an exhaust gas aftertreatment device that processes the exhaust gas of the combustion device, wherein the device controls combustion by determining the co-firing ratio of the hydrogen or ammonia and the fossil fuel burned by the combustion device in order to optimize at least one of energy efficiency, energy cost, or carbon dioxide recovery efficiency.

[0008] The co-firing control method according to this disclosure relates to a co-firing system comprising a combustion device that co-fires hydrogen or ammonia with a fossil fuel, and an exhaust gas aftertreatment device that processes the exhaust gas of the combustion device, wherein the combustion by the combustion device is controlled by determining the co-firing ratio of the hydrogen or ammonia and the fossil fuel burned by the combustion device in order to optimize at least one of energy efficiency, energy cost, or carbon dioxide recovery efficiency. [Effects of the Invention]

[0009] According to the co-firing system, co-firing control device, and co-firing control method of this disclosure, the co-firing ratio can be appropriately controlled. [Brief explanation of the drawing]

[0010] [Figure 1] This is a block diagram showing an example configuration of a co-firing system according to an embodiment of the present disclosure. [Figure 2] This block diagram shows an example configuration of a trained machine learning model according to the embodiment of this disclosure. [Figure 3] This block diagram shows another example configuration of a trained machine learning model according to the embodiments of this disclosure. [Figure 4] This block diagram shows yet another example configuration of a trained machine learning model according to the embodiments of this disclosure. [Figure 5] This flowchart shows an example of the operation of a co-firing control device according to the present disclosure. [Figure 6] This is a schematic block diagram showing the configuration of a computer according to the embodiments of this disclosure. [Modes for carrying out the invention]

[0011] Hereinafter, the co-firing system, co-firing control device, and co-firing control method according to the embodiments of this disclosure will be described with reference to Figures 1 to 6. Figure 1 is a block diagram showing an example configuration of the co-firing system according to the embodiment of this disclosure. Figures 2 to 4 are block diagrams showing an example configuration of a trained machine learning model according to the embodiment of this disclosure. Figure 5 is a flowchart showing an example operation of the co-firing control device according to the embodiment of this disclosure. Figure 6 is a schematic block diagram showing the configuration of a computer according to the embodiment of this disclosure. In each figure, the same or corresponding components are used with the same reference numerals, and explanations are omitted as appropriate.

[0012] (Mixed-fire system) As shown in Figure 1, the co-firing system 1 according to the embodiment of this disclosure includes, for example, a co-firing control device 101, a hydrogen / ammonia supply device 102, a fossil fuel supply device 103, valves 104 and 105, a combustion device 106, an exhaust gas aftertreatment device 107, and a monitoring device 110. In Figure 1, black arrows indicate the flow of fuel, white arrows indicate the flow of exhaust gas, dashed arrows indicate the flow of control signals, etc., and chained arrows indicate the flow of sensor signals, etc. The co-firing system 1 is installed, for example, in facilities such as plants, or in mobile bodies such as ships. In the following, if necessary, the case in which the co-firing system 1 is a power plant and the combustion device 106 is a gas turbine (GT) power generation device will be described as an example.

[0013] The hydrogen / ammonia supply device 102 supplies hydrogen or ammonia (fuel ammonia) 121 to the combustion device 106 via a valve 104. The hydrogen / ammonia supply device 102 includes, for example, a hydrogen production device, a hydrogen storage tank, an ammonia production device, an ammonia storage tank (storage device), etc. The hydrogen / ammonia supply device 102 can be a device that supplies only hydrogen, a device that supplies only ammonia, or a device that supplies both hydrogen and ammonia.

[0014] The fossil fuel supply device 103 supplies fossil fuels 122 such as LNG (liquefied natural gas) and coal to the combustion device 106 via a valve 105. The fossil fuel supply device 103 includes a fossil fuel storage device, etc.

[0015] The combustion device 106 includes, for example, one or more of a gas turbine (GT), a boiler, a power generation engine, or a marine engine, and co-fires hydrogen or ammonia 121 supplied from the hydrogen / ammonia supply device 102 and fossil fuel 122 supplied from the fossil fuel supply device 103. However, in this embodiment, when describing a specific example, as described above, the example given is a gas turbine power generation device in which the combustion device 106 includes a gas turbine and a generator driven by the gas turbine.

[0016] The exhaust gas 123, which includes the exhaust gas emitted from the combustion device 106 and the exhaust gas emitted from the hydrogen / ammonia supply device 102, is input to the exhaust gas aftertreatment device 107. However, if the hydrogen / ammonia supply device 102 does not emit exhaust gas (for example, if hydrogen or ammonia is produced in a hydrogen or ammonia production device that does not emit exhaust gas), or if there is no hydrogen or ammonia production device, the hydrogen / ammonia supply device 102 does not emit exhaust gas.

[0017] The exhaust gas aftertreatment device 107 includes a CO2 recovery device 1071, a denitrification device, a desulfurization device, etc., and recovers CO2 from the input exhaust gas 123, for example by the CO2 recovery device 1071, removes nitrogen oxides, and removes sulfur. The following description will be an example in which the exhaust gas aftertreatment device 107 includes a CO2 recovery device 1071, but the exhaust gas aftertreatment device 107 does not necessarily have to include a CO2 recovery device 1071. The CO2 (108) recovered by the exhaust gas aftertreatment device 107 is effectively utilized or stored and is not released into the atmosphere. On the other hand, the treated exhaust gas (109) treated by the exhaust gas aftertreatment device 107 is released into the atmosphere, for example, through a chimney. The treated exhaust gas (109) contains a low concentration of CO2 (109a).

[0018] The monitoring device 110 acquires sensor signals indicating predetermined physical quantities measured using various sensors in the hydrogen / ammonia supply device 102, combustion device 106, exhaust gas aftertreatment device 107, etc., as well as control signals indicating control states and control commands such as combustion characteristics and target output, and calculated values ​​such as output, efficiency, and CO2 recovery efficiency calculated based on the sensor signals (hereinafter, sensor signals, control signals, and calculated values ​​are collectively referred to as operating data), and outputs the operating data to the co-firing control device 101. Examples of sensor signals acquired by the co-firing control device 101 include the flow rates of fuel, air, exhaust gas, etc., temperature, pressure, output, rotational speed, and CO2 concentration of each part. The CO2 recovery efficiency is the ratio of CO2(108) to the sum of CO2(109a) and CO2(108). The CO2 recovery efficiency may be read as, for example, the CO2 recovery rate.

[0019] (Configuration of the co - firing control device) The co - firing control device 101 controls the combustion of the combustion device 106 by determining the co - firing ratio of hydrogen or ammonia and fossil fuel with which the combustion device 106 burns so as to optimize one or more predetermined variables. The one or more predetermined variables include, for example, at least one of power generation efficiency (energy efficiency), power generation cost (energy cost), and CO2 recovery efficiency. In the present embodiment, the co - firing control device 101 determines the co - firing ratio, for example, using the learned machine learning model M1. In this case, the co - firing control device 101 inputs the operation data output by the monitoring device 110 into the learned machine learning model M1, calculates the co - firing ratio as the output of the learned machine learning model M1 to determine the co - firing ratio, and controls the co - firing ratio by controlling valves 104, 105, etc.

[0020] In addition, the co - firing control device 101 notifies the determined co - firing ratio, etc. to the hydrogen / ammonia supply device 102, the combustion device 106, the exhaust gas post - treatment device 107, etc. The co - firing ratio is a value representing the mixing ratio of fossil fuel and hydrogen or ammonia, and can be, for example, the ratio of the calorific value of hydrogen or ammonia to the total calorific value of the input fuel to the combustion device 106. In the present embodiment, the co - firing control device 101 determines and controls the co - firing ratio according to the operating states of, for example, the hydrogen / ammonia supply device 102, the combustion device 106, the exhaust gas post - treatment device 107, etc. The combustion device 106 performs control considering power generation output, stable combustion, equipment life, etc. for the co-firing rate notified from the co-firing control device 101. For example, if the co-firing rate changes, the required air flow rate also changes. Also, desired combustion characteristics can be obtained with the fuel and air flow rate. In the combustion device 106, for example, based on the co-firing rate, fuel, and desired combustion characteristics, etc., the air flow rate is controlled with a separate operation terminal for each product. Also, when the hydrogen / ammonia supply device 102 includes a hydrogen / ammonia production device, control of the production amount according to the co-firing rate is performed. The exhaust gas post-treatment device 107 performs control of the treatment according to the exhaust gas amount that changes with the co-firing rate. Note that one or more of the parameters controlled according to the co-firing rate in the combustion device 106, the hydrogen / ammonia supply device 102, and the exhaust gas post-treatment device 107 may be determined in the co-firing control device 101 according to, for example, the parameters output by the monitoring device 110. In this case, for example, control that coordinates the entire combustion device 106, the hydrogen / ammonia supply device 102, and the exhaust gas post-treatment device 107 can be efficiently performed.

[0021] (Trained machine learning model) Figures 2 to 4 show three configuration examples of the trained machine learning model M1. Figures 2 to 4 show the trained machine learning model M1-1 for optimal power generation efficiency control, the trained machine learning model M2-1 for optimal power generation cost control, and the trained machine learning model M3-1 for optimal CO2 recovery efficiency control, respectively. The trained machine learning model M1 includes at least one of the trained machine learning model M1-1 for optimal power generation efficiency control, the trained machine learning model M2-1 for optimal power generation cost control, and the trained machine learning model M3-1 for optimal CO2 recovery efficiency control.

[0022] The trained machine learning model M1-1 for optimal power generation efficiency control shown in Figure 2 is a trained machine learning model that takes multiple operational data 1-1, operational data 1-2, ... acquired by the co-firing system 1 as input and the co-firing ratio as output, and outputs a co-firing ratio that optimizes power generation efficiency. The trained machine learning model M1-1 for optimal power generation efficiency control is a trained model that uses a neural network as an element, and the weighting coefficients between neurons in each layer of the neural network are optimized by machine learning so that the desired solution is output for a large amount of input data. The trained machine learning model M1-1 for optimal power generation efficiency control is composed of, for example, a program that performs calculations from input to output and a combination of weighting coefficients (parameters) used in those calculations. The trained machine learning model M1-1 for optimal power generation efficiency control is trained using training data that consists of multiple operational data 1-1, operational data 1-2, ... (input) and the co-firing ratio (output) as training data, for example, when the power generation efficiency is relatively high (for example, the power generation efficiency is greater than a predetermined threshold) obtained from past operational results or virtual operation using a simulator. In this case, power generation efficiency is the ratio of output energy (electrical power generated) to input energy (chemical energy contained in the fuel).

[0023] Furthermore, the pre-trained machine learning model M2-1 for optimal control of power generation costs and the pre-trained machine learning model M3-1 for optimal control of CO2 capture efficiency can also be constructed as models similar to the pre-trained machine learning model M1-1 for optimal control of power generation efficiency. In addition, the input data for the pre-trained machine learning model M1-1 for optimal control of power generation efficiency, namely operating data 1-1, operating data 1-2, ..., includes, for example, data such as generator output, fuel consumption used in efficiency calculations, and fuel lower heating value.

[0024] The trained machine learning model M2-1 for optimal power generation cost control shown in Figure 3 is a trained machine learning model that takes multiple operational data 2-1, operational data 2-2, ... obtained from the co-firing system 1 as input and the co-firing ratio as output, and outputs a co-firing ratio that optimizes power generation cost (= power generation cost per unit). Power generation cost also takes into account procurement costs such as hydrogen / ammonia production, transportation, and storage. Furthermore, since equipment maintenance costs are also taken into account, it is also effective to consider the impact of changes in the co-firing ratio on component lifespan. The trained machine learning model M2-1 for optimal power generation cost control is trained using training data that consists of multiple operational data 2-1, operational data 2-2, ... (input) and the co-firing ratio (output) as training data, for example, when the power generation cost obtained from past operational results or simulated operation using a simulator is relatively low (for example, when the power generation cost is less than a predetermined threshold). Furthermore, the input data for the pre-trained machine learning model M2-1 for optimal control of power generation costs, namely operating data 2-1, operating data 2-2, ..., includes, for example, data such as generator output and fuel costs used in calculating power generation costs.

[0025] The pre-trained machine learning model M3-1 for optimal CO2 recovery efficiency control shown in Figure 4 is a pre-trained machine learning model that takes multiple operational data 3-1, operational data 3-2, ... acquired by the co-firing system 1 as input and the co-firing ratio as output, and outputs a co-firing ratio that optimizes CO2 recovery efficiency. The pre-trained machine learning model M3-1 for optimal CO2 recovery efficiency control is trained using training data as training data, which consists of multiple operational data 3-1, operational data 3-2, ... (input) and the co-firing ratio (output) obtained from past operational results or simulated operation using a simulator, where the CO2 recovery efficiency is relatively high (for example, the CO2 recovery efficiency is higher than a predetermined threshold). The operational data 3-1, operational data 3-2, ... which are the input data for the pre-trained machine learning model M3-1 for optimal CO2 recovery efficiency control, include, for example, data such as generator output, CO2 emissions used in calculating CO2 recovery efficiency, and CO2 recovery amount. The data representing generator output is an example of data representing the output energy generated by the combustion device through co-firing, as per this disclosure. Furthermore, the data representing the amount of CO2 captured (or the amount of CO2 emitted and the amount of CO2 captured) is an example of data representing the acceptance status of the carbon dioxide capture equipment related to this disclosure.

[0026] Furthermore, the pre-trained machine learning model M1-1 for optimal power generation efficiency control is an example configuration of a pre-trained machine learning model for optimal energy efficiency control, which is machine-trained to output a co-firing ratio that optimizes energy efficiency, using multiple operational data acquired from a co-firing system as input and the co-firing ratio as output. Here, energy efficiency is the ratio of output energy (converted energy, etc.) to input energy. The pre-trained machine learning model M2-1 for optimal power generation cost control is an example configuration of a pre-trained machine learning model for optimal energy cost control, which is machine-trained to output a co-firing ratio that optimizes energy cost (cost per unit output energy), using multiple operational data acquired from a co-firing system as input and the co-firing ratio as output. The pre-trained machine learning model M3-1 for optimal CO2 recovery efficiency control is an example configuration of a pre-trained machine learning model for optimal carbon dioxide recovery efficiency control, which is machine-trained to output a co-firing ratio that optimizes carbon dioxide recovery efficiency of a carbon dioxide recovery device, using multiple operational data acquired from a co-firing system as input and the co-firing ratio as output.

[0027] (Example of operation of co-firing control device) Figure 5 shows an example of the operation of the co-firing control device 101. The co-firing control device 101 is capable of selectively using three trained machine learning models M1-1 for optimal control of power generation efficiency, M2-1 for optimal control of power generation cost, and M3-1 for optimal control of CO2 recovery efficiency as the trained machine learning model M1. In this example, the trained machine learning model M2-1 for optimal control of power generation cost is assumed to be the initially selected model. Furthermore, it is assumed that a predetermined lower limit is set for CO2 recovery efficiency in the co-firing system 1. In this example of operation, for example, during stable operation, the co-firing control device 101 determines the co-firing ratio to optimize cost and efficiency using the trained machine learning model M1-1 for optimal control of power generation efficiency or the trained machine learning model M2-1 for optimal control of power generation cost. Furthermore, when rapidly starting the gas turbine power generator to respond quickly to a power supply request, for example, the co-firing control device 101 determines the co-firing ratio to optimize CO2 recovery efficiency using the trained machine learning model M3-1 for optimal control of CO2 recovery efficiency. If the CO2 recovery device 1071 starts up slower than the gas turbine power generator, it is expected that the CO2 recovery efficiency will decrease for a certain period of time. The startup of such a CO2 recovery device is also an example of data that represents the acceptance state. Therefore, for example, when rapidly starting up the gas turbine power generator, the co-firing control device 101 maintains the CO2 recovery efficiency by using the trained machine learning model M3-1 for optimal control of CO2 recovery efficiency.

[0028] The process shown in Figure 5 is initiated, for example, in response to an operator's instruction. The co-firing control device 101 first selects a trained machine learning model M2-1 for optimal control of power generation costs (step S11). Subsequently, the co-firing control device 101 controls the co-firing rate using the trained machine learning model M1 selected in step S1 (step S2). The co-firing control device 101 also periodically determines whether or not a change in the trained model is necessary (step S3). In this example, a change in the trained model is necessary if, for example, a model other than the trained machine learning model M3-1 for optimal control of CO2 recovery efficiency is being used and the CO2 recovery efficiency falls below the lower limit. Alternatively, a change in the trained model is necessary if, for example, the trained machine learning model M3-1 for optimal control of CO2 recovery efficiency is being used and the CO2 recovery efficiency is well above the lower limit.

[0029] On the other hand, if a change in the trained model is necessary (step S3: YES), the co-firing control device 101 selects the trained machine learning model M3-1 for optimal CO2 recovery efficiency control if it is not using the trained machine learning model M3-1 for optimal CO2 recovery efficiency control (step S1), and selects a model other than the trained machine learning model M3-1 for optimal CO2 recovery efficiency control if it is using the trained machine learning model M3-1 (step S1). Thereafter, the co-firing control device 101 controls the co-firing rate using the trained machine learning model M1 selected in step S1 (step S2). The co-firing control device 101 also periodically determines whether a change in the trained model is necessary (step S3).

[0030] On the other hand, if no change is needed to the learned model (step S3: NO), the co-firing control device 101 determines whether or not to terminate the control of the co-firing ratio (step S4). Terminating the control of the co-firing ratio is, for example, when stopping the operation of the combustion device 106. If the control of the co-firing ratio is terminated (step S4: YES), the co-firing control device 101 terminates the process shown in Figure 5. If the control of the co-firing ratio is not terminated (step S4: NO), the co-firing control device 101 periodically determines whether or not a change is needed to the learned model (step S3).

[0031] Note that the operation example shown in Figure 5 is just one example of the operation of the co-firing control device 101, and the operation of the co-firing control device 101 is not limited to the example shown in Figure 5. For example, three pre-trained machine learning models M1-1 for optimal control of power generation efficiency, M2-1 for optimal control of power generation cost, and M3-1 for optimal control of CO2 capture efficiency may be used to calculate three co-firing ratios in parallel, and the maximum co-firing ratio, the average value based on two or three co-firing ratios, etc., may be determined as the co-firing ratio. Note that when using a calculated value based on two or three co-firing ratios, the co-firing ratio will be determined in a way that optimizes multiple variables.

[0032] (Effects, etc.) The co-firing system, co-firing control device, and co-firing control method configured above determine the co-firing ratio of hydrogen or ammonia burned by the combustion device 106 with fossil fuels to optimize at least one of energy efficiency (power generation efficiency), energy cost (power generation cost), or carbon dioxide capture efficiency, and control the combustion of the combustion device 106. Therefore, according to the co-firing system, co-firing control device, and co-firing control method of the embodiment, the co-firing ratio can be appropriately controlled.

[0033] As described above, according to this embodiment, the co-firing ratio of decarbonized fuel (hydrogen and ammonia) and conventional fuel can be changed according to the operating conditions. Furthermore, according to this embodiment, for example, the processes of hydrogen supply, power generation, and gas processing can be integrated and controlled to perform operations in accordance with decarbonization targets. In addition, CO2 emissions from the entire value chain, including fuel production, may be considered as a constraint condition for optimizing operation.

[0034] Furthermore, according to this embodiment, if the CO2 capture device requires waste heat from, for example, power generation equipment, it takes time from startup until CO2 can be captured. Power generation equipment is required to be operated to start up rapidly in response to electricity demand, and in that case, increasing the proportion of decarbonized fuels used becomes a means of achieving decarbonization targets. Also, once steady-state operation is achieved, it can be operated at a co-firing ratio determined by the balance between the cost ratio of decarbonized fuels and fossil fuels, and the cost of CO2 capture. In addition, as a decarbonization technology, the effectiveness can be maximized by understanding the predicted and actual effects of combinations of multiple means such as hydrogen / ammonia combustion and CO2 capture.

[0035] (Other embodiments) Although embodiments of this disclosure have been described in detail above with reference to the drawings, the specific configuration is not limited to these embodiments and may include design changes and the like that do not depart from the gist of this disclosure.

[0036] (Computer configuration) Figure 6 shows the configuration of a computer according to an embodiment of this disclosure. The computer 90 includes a processor 91, main memory 92, storage 93, and an interface 94. The above-described co-firing control device 101 is implemented in the computer 90. The operation of each of the above-described processing units is stored in storage 93 in the form of a program. The processor 91 reads the program from storage 93, loads it into main memory 92, and executes the above-described processing according to the program. The processor 91 also allocates storage areas in main memory 92 corresponding to each of the above-described storage units according to the program.

[0037] The program may be for implementing some of the functions that the computer 90 is to perform. For example, the program may perform functions in combination with other programs already stored in storage, or in combination with other programs implemented in other devices. In other embodiments, the computer may include a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) in addition to, or instead of, the above configuration. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), FPGA (Field Programmable Gate Array), etc. In this case, some or all of the functions implemented by the processor may be implemented by the integrated circuit.

[0038] Examples of storage 93 include HDDs (Hard Disk Drives), SSDs (Solid State Drives), magnetic disks, magneto-optical disks, CD-ROMs (Compact Disc Read Only Memory), DVD-ROMs (Digital Versatile Disc Read Only Memory), and semiconductor memory. Storage 93 may be an internal medium directly connected to the bus of the computer 90, or an external medium connected to the computer 90 via an interface 94 or a communication line. Furthermore, if this program is distributed to the computer 90 via a communication line, the computer 90 that receives the program may expand it into main memory 92 and execute the above processing. In at least one embodiment, storage 93 is a tangible storage medium that is not temporary.

[0039] <Note> The co-firing system 1 described in the above embodiment can be understood, for example, as follows.

[0040] (1) The co-firing system 1 according to the first embodiment comprises a combustion device 106 that co-fires hydrogen or ammonia with a fossil fuel, an exhaust gas aftertreatment device 107 that processes the exhaust gas of the combustion device 106, and a co-firing control device 101 that controls combustion by the combustion device by determining the co-firing ratio of the hydrogen or ammonia and the fossil fuel burned by the combustion device 106 to optimize one or more predetermined variables, wherein the one or more variables include at least one of energy efficiency (power generation efficiency, etc.), energy cost (power generation cost, etc.), or carbon dioxide capture efficiency. According to this embodiment and each of the following embodiments, the co-firing ratio can be appropriately controlled.

[0041] (2) The co-firing system 1 according to the second embodiment is the co-firing system 1 of (1), wherein the exhaust gas aftertreatment device includes a carbon dioxide recovery device (CO2 recovery device 1071).

[0042] (3) The co-firing system 1 according to the third embodiment is the co-firing system 1 of (2), wherein the co-firing control device 101 determines the co-firing ratio using at least one of the following: a trained machine learning model for optimal energy efficiency control (trained machine learning model for optimal power generation efficiency control) which is trained to output the co-firing ratio that optimizes the energy efficiency, with the input being a plurality of operating data acquired by the co-firing system and the output being the co-firing ratio; a trained machine learning model for optimal energy cost control (trained machine learning model for optimal power generation cost control) which is trained to output the co-firing ratio that optimizes the energy cost, with the input being a plurality of operating data acquired by the co-firing system and the output being the co-firing ratio; or a trained machine learning model for optimal carbon dioxide recovery efficiency control (trained machine learning model for optimal CO2 recovery efficiency control) which is trained to output the co-firing ratio that optimizes the carbon dioxide recovery efficiency of the carbon dioxide recovery device, with the input being a plurality of operating data acquired by the co-firing system and the output being the co-firing ratio.

[0043] (4) The co-firing system 1 according to the fourth embodiment is the co-firing system 1 of (1) to (3), wherein the plurality of operating data includes at least data representing the output energy (generator output, etc.) generated by the combustion device through co-firing.

[0044] (5) The co-firing system 1 according to the fifth embodiment is the co-firing system 1 of (4), wherein the plurality of operating data further include at least data representing the acceptance status of the carbon dioxide capture device (e.g., CO2 capture amount, CO2 emission amount, and CO2 capture amount). [Explanation of Symbols]

[0045] 1…Mixed-fire system 101... Co-firing control device 102…Hydrogen / Ammonia Supply System 103…Fossil fuel supply equipment 104, 105… valves 106... Combustion device 107... Exhaust gas aftertreatment device 110... Monitoring device 1071...CO2 capture device M1... Pre-trained machine learning model M1-1…Trained machine learning model for optimal control of power generation efficiency M1-2... Pre-trained machine learning model for optimal control of power generation costs M1-3…Trained machine learning model for optimal control of CO2 capture efficiency

Claims

1. A combustion device that co-fires hydrogen or ammonia with fossil fuels, An exhaust gas aftertreatment device for treating the exhaust gas of the combustion device, A co-firing control device that controls combustion by the combustion device by determining the co-firing ratio of the hydrogen or ammonia burned by the combustion device and the fossil fuel to optimize one or more predetermined variables, Equipped with, The one or more variables include at least one of energy efficiency, energy cost, or carbon dioxide capture efficiency. It is a mixed-fire system, The exhaust gas aftertreatment device includes a carbon dioxide recovery device, The aforementioned co-firing control device A trained machine learning model for energy efficiency optimization control, which takes multiple operating data acquired by the co-firing system as inputs and the co-firing ratio as output, outputs the co-firing ratio that optimizes the energy efficiency. A trained machine learning model for energy cost optimization control, which takes multiple operating data acquired by the co-firing system as input and the co-firing ratio as output, and outputs the co-firing ratio that optimizes the energy cost, or A trained machine learning model for optimizing carbon dioxide recovery efficiency control, which takes multiple operating data acquired from the co-firing system as inputs and the co-firing ratio as output, outputs the co-firing ratio that optimizes the carbon dioxide recovery efficiency of the carbon dioxide recovery device. The co-firing ratio is determined using at least one of the following: Co-firing system.

2. The aforementioned plurality of operating data include at least data representing the output energy generated by the combustion device through co-firing. The co-firing system according to claim 1.

3. The aforementioned plurality of operating data further include at least data representing the acceptance status of the carbon dioxide capture device. The co-firing system according to claim 2.

4. A combustion device that co-fires hydrogen or ammonia with fossil fuels, An exhaust gas aftertreatment device for treating the exhaust gas of the combustion device, Equipped with, The exhaust gas aftertreatment device includes a carbon dioxide recovery device. In a co-firing system, A co-firing control device that controls combustion by the combustion device by determining the co-firing ratio of the hydrogen or ammonia burned by the combustion device and the fossil fuel burned by the combustion device in order to optimize at least one of energy efficiency, energy cost, or carbon dioxide capture efficiency, The aforementioned co-firing control device is A trained machine learning model for energy efficiency optimization control, which takes multiple operating data acquired by the co-firing system as inputs and the co-firing ratio as output, outputs the co-firing ratio that optimizes the energy efficiency. A trained machine learning model for energy cost optimization control, which takes multiple operating data acquired by the co-firing system as input and the co-firing ratio as output, and outputs the co-firing ratio that optimizes the energy cost, or A trained machine learning model for optimizing carbon dioxide recovery efficiency control, which takes multiple operating data acquired from the co-firing system as inputs and the co-firing ratio as output, outputs the co-firing ratio that optimizes the carbon dioxide recovery efficiency of the carbon dioxide recovery device. The co-firing ratio is determined using at least one of the following: Co-firing control device.

5. A combustion device that co-fires hydrogen or ammonia with fossil fuels, An exhaust gas aftertreatment device for treating the exhaust gas of the combustion device, Equipped with, The exhaust gas aftertreatment device includes a carbon dioxide recovery device. In a co-firing system, A co-firing control method for controlling combustion by a combustion device, which determines the co-firing ratio of the hydrogen or ammonia burned by the combustion device and the fossil fuel burned by the combustion device in order to optimize at least one of energy efficiency, energy cost, or carbon dioxide capture efficiency, The aforementioned co-firing control method is: A trained machine learning model for energy efficiency optimization control, which takes multiple operating data acquired by the co-firing system as inputs and the co-firing ratio as output, outputs the co-firing ratio that optimizes the energy efficiency. A trained machine learning model for energy cost optimization control, which takes multiple operating data acquired by the co-firing system as input and the co-firing ratio as output, and outputs the co-firing ratio that optimizes the energy cost, or A trained machine learning model for optimizing carbon dioxide recovery efficiency control, which takes multiple operating data acquired from the co-firing system as inputs and the co-firing ratio as output, outputs the co-firing ratio that optimizes the carbon dioxide recovery efficiency of the carbon dioxide recovery device. The co-firing ratio is determined using at least one of the following: Co-firing control method.

Images