Multi-source fuel blending low-carbon combustion optimization method, equipment and product

Abstract

The invention discloses a multi-source fuel blending low-carbon combustion optimization method, device and product, and relates to the technical field of combustion optimization intelligent control, and the method comprises the steps: collecting and preprocessing the physicochemical property data of multi-source fuel, carrying out the parameter prediction of a multi-dimensional feature vector through a trained coal quality prediction model, and obtaining a multi-source fuel blending low-carbon combustion optimization model; the coal blending decision model is subjected to optimization calculation through a non-dominated sorting genetic algorithm, the air-coal ratio optimization model is subjected to optimization calculation through a particle swarm optimization algorithm, reliable data support can be provided for coal blending decision, the defects of the prior art in the coal quality prediction link are overcome, the optimal ratio of different coal qualities is calculated, and the coal quality prediction efficiency is improved. The scientific and intelligent level of coal blending decision making is improved, enterprise production benefits and environmental protection requirements are balanced, real-time dynamic adjustment can be achieved according to coal quality characteristics and combustion working conditions, the combustion efficiency is improved, and energy consumption and pollutant emission are reduced.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology for combustion optimization, and in particular to a low-carbon combustion optimization method, equipment and product for multi-source fuel blending. Background Technology

[0002] Coal-fired power generation, as the core of electricity supply, emits more than 10 billion tons of carbon annually, facing enormous pressure to reduce carbon emissions. In response to the "dual carbon" goals, coal-fired units need to transform towards low-carbon operation. Among these measures, blending coal with multiple fuel sources (such as biomass fuel, green ammonia, and solid waste-derived fuels) has become a key technological path, which can effectively reduce carbon emissions and pollutant emissions (such as NOx and SO2).

[0003] The existing technology has the following shortcomings: (1) Traditional coal quality prediction methods are difficult to provide stable and accurate coal quality data, resulting in a lack of reliable basis for subsequent coal blending decisions.

[0004] (2) The existing coal blending theory relies on extensive coal blending methods based on experience, which cannot adapt to changes in coal sources. There are data barriers between the coal blending optimization system and material information, making it difficult to comprehensively consider multiple objectives such as total electricity cost and pollutant emission concentration, resulting in the coal blending scheme failing to achieve a balance between economic and environmental benefits.

[0005] (3) Traditional air-coal ratio control uses fixed parameters and cannot be dynamically adjusted in real time according to coal quality characteristics and combustion conditions, resulting in low coal combustion efficiency and insufficient energy utilization. At the same time, it is difficult to effectively control pollutant emissions, which increases the environmental protection costs of enterprises. Summary of the Invention

[0006] The purpose of this application is to provide a low-carbon combustion optimization method, equipment, and product for multi-source fuel blending, which can achieve rapid and accurate coal quality prediction, provide reliable data support for coal blending decisions, make up for the shortcomings of existing technologies in the coal quality prediction stage, realize the calculation of the optimal blending ratio of different coal qualities, break down information silos, improve the scientific and intelligent level of coal blending decisions, balance enterprise production efficiency and environmental protection requirements, and can dynamically adjust in real time according to coal quality characteristics and combustion conditions, improve combustion efficiency, reduce energy consumption and pollutant emissions, and reduce enterprise environmental governance costs.

[0007] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for optimizing low-carbon combustion by blending multiple fuel sources, including: Collect and preprocess the physicochemical properties data of multi-source fuels to obtain multi-dimensional feature vectors, which include numerical coal quality parameters, spectral data and image data; The parameters of the multidimensional feature vector are predicted by a trained coal quality prediction model. The coal quality prediction model is constructed based on a long short-term memory network combined with a convolutional neural network in deep learning. The prediction results of the coal quality prediction model include the predicted values ​​of calorific value, volatile matter, ash content, sulfur content and moisture content. The coal blending decision model is optimized by using a non-dominated sorting genetic algorithm to obtain the optimal fuel ratio. The coal blending decision model is constructed with the dual objectives of minimizing the total cost of electricity and minimizing the pollutant emission concentration, combined with the prediction results and constraints. The optimal air-coal ratio is obtained by performing optimization calculations on the air-coal ratio optimization model using the particle swarm optimization algorithm. The input of the air-coal ratio optimization model includes the fuel ratio output by the coal blending decision model.

[0008] Optionally, the multi-source fuels include, but are not limited to: coal, biomass fuels, gaseous synthetic fuels, and solid waste-derived fuels; The physicochemical property data includes numerical coal quality parameters, spectral data, and image data. The numerical coal quality parameters include, but are not limited to, calorific value, volatile matter, ash content, sulfur content, and moisture content.

[0009] Optionally, the network structure of the coal quality prediction model includes: Input layer: Receives multidimensional feature vectors containing numerical coal quality parameters, spectral data, and image data; Long Short-Term Memory Network Layer: 2 layers, 128 neurons per layer, used to extract temporal features of the spectral data; Convolutional Neural Network Layers: 2 convolutional layers and 1 max pooling layer, used to extract spatial features of the image data. The convolutional kernels of the convolutional layers are 3×3 with a stride of 1 and an activation function of ReLU. The pooling window of the max pooling layer is 2×2. Feature fusion layer: concatenates the outputs of the long short-term memory network layer and the neural network layer; Fully connected layers: 3 layers, with 256, 128 and 64 neurons respectively, and ReLU activation function; Output layer: 5 neurons, outputting predicted values ​​for calorific value, volatile matter, ash content, sulfur content, and moisture content, respectively.

[0010] Optionally, the functional expression of the coal blending decision model is: ; In the formula, This represents the total cost per kilowatt-hour. The weighting coefficient represents the total cost per kilowatt-hour. This represents the emission concentration of the i-th pollutant. The weighting coefficients representing pollutant emission concentrations, and , Indicates the types and quantities of pollutants; Among them, the total cost per kilowatt-hour The calculation model is as follows: ; In the formula, Indicates fuel procurement costs, Indicates transportation costs, Indicates storage cost, Indicates processing cost, This indicates the cost of equipment depreciation; The pollutant emission concentration The calculation model is as follows: ; In the formula, This represents the predicted emissions of the i-th pollutant. Indicates heat output. Indicates volatile matter, Indicates ash content, Indicates sulfur content. Indicates moisture content. Indicates furnace temperature. This indicates the wind-to-coal ratio.

[0011] Optionally, the constraints include coal quality characteristics constraints, resource quantity constraints, process constraints, and environmental constraints; The coal quality characteristics constraint is as follows: ; ; ; In the formula, This indicates the lower limit of the calorific value of the mixed fuels allowed for stable combustion in the boiler. This indicates the received lower heating value of the blended fuel. This indicates the upper limit of the calorific value of the mixed fuels allowed for stable combustion in the boiler; This indicates the lower limit of volatile matter content in the mixed fuels that is permissible for stable combustion in a boiler. This indicates the dry, ash-free volatile matter content of the blended fuel. This indicates the upper limit of volatile matter content in the mixed fuels allowed for stable combustion in the boiler; This indicates the received ash content of the blended fuel. This indicates the maximum permissible ash content limit in boiler design; The resource constraint is as follows: , ; In the formula, Indicating the first step in the fuel blending scheme The availability of this type of fuel Indicates the first in inventory The availability of this type of fuel Indicates the quantity of each type of fuel; The process constraints are as follows: ; ; In the formula, Indicates the first The lower limit of the blending ratio of various fuels in the process. Indicates the first The mass blending ratio of each fuel in the blended fuel Indicates the first The upper limit of the blending ratio of various fuels in the process; The environmental constraints are as follows: ; In the formula, Indicates the first The emission concentration of various pollutants, Indicates the first Emission concentration limits for various pollutants.

[0012] Optionally, the specific process of optimizing the coal blending decision model using a non-dominated sorting genetic algorithm includes: During the algorithm initialization phase, a set of conditions is randomly generated for each non-dominated individual in the population. Weighting coefficients ; Weighting coefficients As part of the genes of each non-dominant individual, it is encoded, crossovered, and mutated together with the decision variables of fuel ratio to achieve the co-evolution of weights and fuel ratio; During the elite retention phase of each generation, based on the non-dominant ranking and crowding of non-dominant individuals, combined with the target spatial region guided by the weight coefficient, priority is given to retaining the weight combination of non-dominant individuals that guide the population toward the current Pareto front sparse region. The weight mutation strategy is dynamically adjusted based on the algorithm's running status. When population diversity decreases, the mutation intensity of the weight coefficients is increased. The optimization process is repeated until the algorithm reaches the preset number of iterations or the Pareto optimal solution remains unchanged for multiple consecutive generations. At least one Pareto optimal solution is then output, which is the optimized fuel ratio.

[0013] Optionally, the specific process of obtaining the optimized wind-coal ratio by performing optimization calculations on the wind-coal ratio optimization model using the particle swarm optimization algorithm includes: The comprehensive combustion characteristic parameters of the blended fuel are calculated by weighted average method. The blended fuel is the multi-source fuel in the fuel ratio output by the coal blending decision model. The comprehensive combustion characteristic parameters are input into the air-coal ratio optimization model. With total electricity cost and pollutant emission concentration as constraints and combustion efficiency as the optimization objective, the optimal air-coal ratio parameter combination is searched in the air-coal ratio optimization model using the particle swarm optimization algorithm. The air-coal ratio parameter combination includes the total primary air volume, the total secondary air volume, the air volume distribution coefficient of each burner, and the coal feed rate. The particle swarm optimization algorithm steps include: The total primary air volume, the total secondary air volume, and the air volume distribution coefficient of each burner are encoded as particle position vectors. Within the upper and lower limits of air volume that meet the safety of boiler operation, a group of particles are randomly initialized, and the velocity of each particle is randomly initialized. The position vector of each particle is decoded into the air-coal ratio control parameter, which is then input into the air-coal ratio optimization model to obtain the predicted values ​​of combustion efficiency, total electricity cost, and pollutant emission concentration under the air-coal ratio parameter. Using the predicted combustion efficiency as the base fitness, a penalty function is applied to particles that violate the total cost of electricity or pollutant emission concentration constraints, thereby reducing their effective fitness. Compare the current fitness of each particle with its historical best fitness to update the individual's optimal position; compare the fitness of all particles to update the global optimal position. Based on the individual optimal position and the global optimal position of the particle, the velocity and position of each particle are updated according to the preset inertia weight, individual learning factor and social learning factor. Boundary correction is performed on positions that exceed the upper and lower limits of airflow after the update, and speeds that exceed the speed range are truncated. Repeat the optimization process until the algorithm reaches the preset maximum number of iterations or the global optimal solution is no longer improved after multiple consecutive generations. Then, output the combination of air-coal ratio parameters corresponding to the global optimal position at this point as the optimal solution.

[0014] Optionally, the low-carbon combustion optimization method for multi-source fuel blending also includes real-time monitoring of boiler combustion status and adjusting the air-coal ratio or fuel blending ratio based on the monitoring results. The specific process includes: Collect furnace temperature, flue gas composition concentration, pressure, and flow parameters to form a real-time monitoring dataset; The adjustment parameters in the real-time monitoring dataset are compared with their corresponding preset target values ​​to calculate the real-time deviation. When the real-time deviation of any adjustment parameter continues to exceed its corresponding allowable deviation threshold for a preset duration, it is determined to be an operational abnormality and a first-level feedback control signal is generated. The first-level feedback control signal is sent to the wind-coal ratio optimization model first, and the wind-coal ratio optimization model is re-optimized by the particle swarm optimization algorithm to generate and execute the adjusted wind-coal ratio parameters. After implementing the adjusted air-coal ratio parameters, the adjustment parameters are continuously monitored; if the real-time deviation of the adjustment parameters still does not return to the allowable deviation threshold within the second preset time period, a secondary feedback control signal is generated. The secondary feedback control signal is sent to the coal blending decision model, and the coal blending decision model is re-optimized using a non-dominated sorting genetic algorithm to generate a new fuel ratio scheme and execute it. The process continues in a loop until the runtime exception is corrected.

[0015] In a second aspect, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the low-carbon combustion optimization method for multi-source fuel blending as described above.

[0016] Thirdly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the low-carbon combustion optimization method for multi-source fuel blending described above.

[0017] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a low-carbon combustion optimization method, equipment, and product for multi-source fuel blending. By collecting and preprocessing the physicochemical property data of multi-source fuels, a multi-dimensional feature vector is obtained. A trained coal quality prediction model is then used to predict the parameters of this feature vector, achieving rapid and accurate coal quality prediction. This provides reliable data support for coal blending decisions and overcomes the shortcomings of existing technologies in the coal quality prediction stage. A non-dominated sorting genetic algorithm is used to optimize the coal blending decision model, resulting in an optimized fuel ratio. The coal blending decision model has two objectives: minimizing total electricity cost and minimizing pollutant emission concentration. It is constructed by combining the prediction parameters and constraints output by the coal quality prediction model, realizing the calculation of the optimal ratio for different coal qualities. This breaks down information silos, improves the scientific and intelligent level of coal blending decisions, and balances enterprise production efficiency with environmental protection requirements. Finally, a particle swarm optimization algorithm is used to optimize the air-coal ratio model, resulting in an optimized air-coal ratio. This ratio can be dynamically adjusted in real time according to coal quality characteristics and combustion conditions, improving combustion efficiency, reducing energy consumption and pollutant emissions, and lowering enterprise environmental governance costs. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A schematic flowchart illustrating a low-carbon combustion optimization method for blending multiple fuels, provided in an embodiment of this application; Figure 2 for Figure 1 A schematic diagram of the network structure of the medium-coal quality prediction model; Figure 3 for Figure 1 A detailed flowchart of the optimization calculation of the coal blending decision model using a non-dominated sorting genetic algorithm; Figure 4 for Figure 1 A detailed flowchart of the particle swarm optimization algorithm; Figure 5 for Figure 1 A detailed flowchart for adjusting the air-coal ratio or fuel blending ratio based on monitoring results feedback; Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] In one exemplary embodiment, such as Figure 1 As shown, a low-carbon combustion optimization method for blending multiple fuels is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is described using a computer device as an example, and includes the following steps 101 to 104. Wherein: Step 101: Collect and preprocess the physicochemical properties data of multi-source fuels to obtain a multi-dimensional feature vector, which includes numerical coal quality parameters, spectral data and image data. Step 102: Predict the parameters of the multidimensional feature vector using the trained coal quality prediction model. The coal quality prediction model is constructed based on a long short-term memory network combined with a convolutional neural network in deep learning. The prediction results of the coal quality prediction model include the predicted values ​​of calorific value, volatile matter, ash content, sulfur content, and moisture content. Step 103: The coal blending decision model is optimized by using a non-dominated sorting genetic algorithm to obtain the optimal fuel ratio. The coal blending decision model is constructed with the dual objectives of minimizing the total cost of electricity and minimizing the pollutant emission concentration, combined with the prediction results and constraints. Step 104: The wind-coal ratio optimization model is optimized by using the particle swarm optimization algorithm to obtain the optimized wind-coal ratio. The input of the wind-coal ratio optimization model includes the fuel ratio output by the coal blending decision model.

[0023] Implementing steps 101 to 104 above enables rapid and accurate coal quality prediction, providing reliable data support for coal blending decisions, compensating for the shortcomings of existing technologies in the coal quality prediction stage, realizing the calculation of the optimal blending ratio of different coal qualities, breaking down information silos, improving the scientific and intelligent level of coal blending decisions, balancing enterprise production efficiency and environmental protection requirements, and enabling real-time dynamic adjustments based on coal quality characteristics and combustion conditions to improve combustion efficiency, reduce energy consumption and pollutant emissions, and reduce enterprise environmental governance costs.

[0024] As an optional implementation, the multi-source fuels include, but are not limited to: coal, biomass fuels, gaseous synthetic fuels, and solid waste-derived fuels; The physicochemical property data includes numerical coal quality parameters, spectral data, and image data. The numerical coal quality parameters include, but are not limited to, calorific value, volatile matter, ash content, sulfur content, and moisture content.

[0025] Specifically, multi-source fuels refer to combustible materials from multiple sources, of multiple types, and with different physicochemical properties. The categories of multi-source fuels and their corresponding key physicochemical properties are shown in Table 1.

[0026] Table 1. Categories of Multi-Source Fuels and Corresponding Key Physicochemical Properties

[0027] For solid fuels (such as coal, biomass fuels, and solid waste-derived fuels), during fuel transportation, online automatic sampling devices automatically collect the physicochemical properties of fuel samples at preset intervals (such as once every 30 minutes). For newly arrived batches of solid fuel, multi-point sampling is conducted during unloading, and spectral and image data are collected in real time using online sensors (such as near-infrared spectral sensors and laser-induced breakdown spectral sensors).

[0028] The numerical coal quality parameters and corresponding sensors are shown in Table 2.

[0029] Table 2 Numerical coal quality parameters and corresponding sensors

[0030] For gaseous fuels (such as ammonia and hydrogen), high-precision online sensors are installed at key nodes of the pipeline to automatically collect the physicochemical properties of fuel samples according to a preset cycle (such as once every 30 seconds).

[0031] The parameters of the gaseous fuel and the corresponding sensors are shown in Table 3.

[0032] Table 3 Monitoring parameters and corresponding sensors for gaseous fuels

[0033] The collected data underwent preprocessing, including missing value imputation, outlier removal, and data normalization. Additionally, background information such as fuel origin, mining batch, transportation and storage conditions, as well as sampling time and location, were collected and recorded. Then, a time-series dataset was constructed, with each data point including: timestamp, fuel batch, sampling location, numerical coal quality parameters, spectral data, and image data.

[0034] As an optional implementation method, such as Figure 2 As shown, the network structure of the coal quality prediction model includes: Input layer: Receives multidimensional feature vectors containing numerical coal quality parameters, spectral data, and image data; Long Short-Term Memory Network Layer: 2 layers, 128 neurons per layer, used to extract temporal features of the spectral data; Convolutional Neural Network Layers: 2 convolutional layers and 1 max pooling layer, used to extract spatial features of the image data. The convolutional kernels of the convolutional layers are 3×3 with a stride of 1 and an activation function of ReLU. The pooling window of the max pooling layer is 2×2. Feature fusion layer: concatenates the outputs of the long short-term memory network layer and the neural network layer; Fully connected layers: 3 layers, with 256, 128 and 64 neurons respectively, and ReLU activation function; Output layer: 5 neurons, outputting predicted values ​​for calorific value, volatile matter, ash content, sulfur content, and moisture content, respectively.

[0035] Specifically, Long Short-Term Memory (LSTM) networks can effectively process time-series data and capture the changing patterns of coal quality characteristics over time and batches, while Convolutional Neural Networks (CNNs) are used to analyze feature information in spectral and image data. After the coal quality prediction model is established, it needs to be trained using a training set. Mean Squared Error (MSE) is used as the loss function, and the Stochastic Gradient Descent (SGD) algorithm combined with an adaptive learning rate adjustment strategy is employed for training. During training, the model is evaluated in real-time using a validation set to prevent overfitting. When the prediction error on the test set reaches the preset accuracy requirement, the final model parameters are determined. The training, validation, and test sets are all derived from historical data collected from multiple fuel sources and are divided in a 7:2:1 ratio.

[0036] As an optional implementation, the functional expression of the coal blending decision model is: ; In the formula, This represents the total cost per kilowatt-hour. The weighting coefficient represents the total cost per kilowatt-hour. This represents the emission concentration of the i-th pollutant. The weighting coefficients representing pollutant emission concentrations, and , Indicates the types and quantities of pollutants; Among them, the total cost per kilowatt-hour The calculation model is as follows: ; In the formula, Indicates fuel procurement costs, Indicates transportation costs, Indicates storage cost, Indicates processing cost, This indicates the cost of equipment depreciation; The pollutant emission concentration The calculation model is as follows: ; In the formula, This represents the predicted emissions of the i-th pollutant. Indicates heat output. Indicates volatile matter, Indicates ash content, Indicates sulfur content. Indicates moisture content. Indicates furnace temperature. This indicates the wind-to-coal ratio.

[0037] As an optional implementation, the constraints include coal quality characteristics constraints, resource quantity constraints, process constraints, and environmental constraints. The coal quality characteristics constraint is as follows: ; ; ; In the formula, This indicates the lower limit of the calorific value of the mixed fuels allowed for stable combustion in the boiler. This indicates the received lower heating value of the blended fuel. This indicates the upper limit of the calorific value of the mixed fuels allowed for stable combustion in the boiler; This indicates the lower limit of volatile matter content in the mixed fuels that is permissible for stable combustion in a boiler. This indicates the dry, ash-free volatile matter content of the blended fuel. This indicates the upper limit of volatile matter content in the mixed fuels allowed for stable combustion in the boiler; This indicates the received ash content of the blended fuel. This indicates the maximum permissible ash content limit in boiler design; The resource constraint is as follows: , ; In the formula, Indicating the first step in the fuel blending scheme The availability of this type of fuel Indicates the first in inventory The availability of this type of fuel Indicates the quantity of each type of fuel; The process constraints are as follows: ; ; In the formula, Indicates the first The lower limit of the blending ratio of various fuels in the process. Indicates the first The mass blending ratio of each fuel in the blended fuel Indicates the first The upper limit of the blending ratio of various fuels in the process; The environmental constraints are as follows: ; In the formula, Indicates the first The emission concentration of various pollutants, Indicates the first Emission concentration limits for various pollutants.

[0038] Specifically, coal quality constraints are based on boiler combustion requirements, setting ranges for the calorific value, volatile matter, and ash content of the mixed fuels to ensure stable boiler combustion. Specifically, the weighted average calorific value of the mixed fuels must meet the requirements for stable boiler combustion; the weighted average volatile matter must be within the safe and efficient range of the boiler design to prevent ignition difficulties or excessively intense combustion; and the weighted average ash content of the mixed fuels must be limited to the upper limit of the boiler design and soot blowing system capacity to prevent severe slagging, ash accumulation, and wear. Resource constraints are based on existing fuel reserves to ensure that the usage in the fuel blending scheme does not exceed the available inventory. Process constraints are based on the enterprise's coal blending equipment, conveying equipment, and other process conditions, setting upper and lower limits for coal blending ratios to avoid equipment malfunctions due to unreasonable blending. Environmental constraints are based on national environmental regulations and local emission standards, combined with the enterprise's environmental protection facilities' treatment capacity, to determine the emission concentration limits for major pollutants (such as sulfur dioxide, nitrogen oxides, and particulate matter). Sulfur content is the direct determining variable for calculating sulfur dioxide (SO2) emission concentration.

[0039] As an optional implementation method, such as Figure 3 As shown, the specific process of optimizing the coal blending decision model using a non-dominated sorting genetic algorithm includes steps 201 to 205. Wherein: Step 201: In the algorithm initialization phase, randomly generate a set of conditions for each non-dominated individual in the population. Weighting coefficients ; Step 202, weighting coefficients As part of the genes of each non-dominant individual, it is encoded, crossovered, and mutated together with the decision variables of fuel ratio to achieve the co-evolution of weights and fuel ratio; Step 203: In the elite retention phase of each generation, based on the non-dominant ranking and crowding of non-dominant individuals, and combined with the target spatial region guided by the weight coefficient, priority is given to retaining the weight combination of non-dominant individuals that guide the population to develop towards the current Pareto front sparse region. Step 204: Dynamically adjust the weight mutation strategy according to the algorithm's running status. When the population diversity decreases, increase the mutation intensity of the weight coefficients. Step 205: Repeat steps 203 to 204 until the algorithm reaches the preset number of iterations or the Pareto optimal solution has not improved for multiple consecutive generations, and output at least one Pareto optimal solution, that is, the optimized fuel ratio.

[0040] Specifically, the decision variable for fuel blending is the blending ratio of multiple fuel sources. This ratio is a continuous real-valued variable, and its optimal value is determined using a non-dominated sorting genetic algorithm to achieve the dual objectives of minimizing total electricity cost and pollutant emission concentration. The decision variable must satisfy the sum-of-proportions constraint, i.e., the process constraint. In the non-dominated sorting genetic algorithm, each non-dominated individual is encoded as a set of fuel blending ratios and weight coefficients. For example, assuming a fuel blending ratio of 65% anthracite, 25% biomass, and 10% green ammonia, with total electricity cost accounting for 70% of the weight and pollutant emission concentration accounting for 30%, the corresponding non-dominated individual can be encoded as: [ =0.65, =0.25, =0.10, =0.7, =0.3].

[0041] As an optional implementation, the specific process of obtaining the optimized wind-coal ratio by performing optimization calculations on the wind-coal ratio optimization model using the particle swarm optimization algorithm includes: The comprehensive combustion characteristic parameters of the blended fuel are calculated by weighted average method. The blended fuel is the multi-source fuel in the fuel ratio output by the coal blending decision model. The comprehensive combustion characteristic parameters are input into the air-coal ratio optimization model. With total electricity cost and pollutant emission concentration as constraints and combustion efficiency as the optimization objective, the optimal air-coal ratio parameter combination is searched in the air-coal ratio optimization model using the particle swarm optimization algorithm. The air-coal ratio parameter combination includes the total primary air volume, the total secondary air volume, the air volume distribution coefficient of each burner, and the coal feed rate. like Figure 4 As shown, the particle swarm optimization algorithm includes steps 301 to 307. Wherein: Step 301: Encode the total primary air volume, the total secondary air volume, and the air volume distribution coefficient of each burner into the position vector of the particles. Randomly initialize a group of particles within the upper and lower limits of the air volume that meet the safety of boiler operation, and randomly initialize the velocity of each particle. Step 302: Decode the position vector of each particle into the air-coal ratio control parameter, input it into the air-coal ratio optimization model, and obtain the predicted values ​​of combustion efficiency, total electricity cost and pollutant emission concentration under the air-coal ratio parameter. Step 303: Using the predicted combustion efficiency as the base fitness, a penalty function is applied to particles that violate the total cost of electricity constraint or pollutant emission concentration constraint to reduce their effective fitness. Step 304: Compare the current fitness of each particle with its historical best fitness and update the individual optimal position; compare the fitness of all particles and update the global optimal position. Step 305: Based on the individual optimal position and the global optimal position of the particle, update the velocity and position of each particle according to the preset inertia weight, individual learning factor and social learning factor. Step 306: Perform boundary correction on positions that exceed the upper and lower limits of airflow after the update, and truncate speeds that exceed the speed range; Step 307: Repeat steps 302 to 306 until the algorithm reaches the preset maximum number of iterations or the global optimal solution no longer improves after multiple consecutive generations. Output the combination of wind-coal ratio parameters corresponding to the global optimal position at this time as the optimal solution.

[0042] Specifically, the combustion characteristic data (such as ignition temperature, burnout temperature, combustion rate constant, etc.) of each individual fuel in the multi-source fuel ratio output by the coal blending decision model are weighted and averaged to obtain comprehensive combustion characteristic parameters. A support vector regression (SVR) algorithm is used to establish a model relating the air-coal ratio to the comprehensive combustion characteristic parameters and boiler efficiency. Through training with an extended dataset, the model parameters are determined, enabling the model to accurately predict combustion efficiency, pollutant emission concentrations, and total cost of electricity (CTO) under different air-coal ratios. Regarding the acquisition of the extended dataset, in actual operating megawatt-scale units, arbitrarily changing the air-coal ratio and fuel type for data collection can easily lead to serious safety accidents such as coking, flameout, and overheating, causing huge economic losses. Therefore, computational fluid dynamics software is used to perform three-dimensional numerical simulations of the combustion process of mixed fuels in the boiler, simulating the distribution of temperature, velocity, and concentration fields in the furnace under different air-coal ratios. The relationship between combustion efficiency, pollutant formation mechanisms, and the air-coal ratio is analyzed, providing an extended dataset for training in air-coal ratio optimization.

[0043] In another exemplary embodiment, to achieve dynamic closed-loop control of the combustion process and ensure that the combustion process is always in an optimal state of high efficiency and low carbon emissions, after step 104 above, the method further includes real-time monitoring of the boiler combustion status and adjusting the air-coal ratio or fuel ratio based on the monitoring results. Figure 5 As shown, the specific process includes steps 401 to 406. Wherein: Step 401: Collect furnace temperature, flue gas component concentration, pressure and flow parameters to form a real-time monitoring dataset; Step 402: Compare the adjustment parameters in the real-time monitoring dataset with their corresponding preset target values, calculate the real-time deviation, and determine the operation as abnormal when the real-time deviation of any adjustment parameter continuously exceeds its corresponding allowable deviation threshold for a preset duration, and generate a first-level feedback control signal. Step 403: The first-level feedback control signal is sent to the wind-coal ratio optimization model first, and the wind-coal ratio optimization model is re-optimized by the particle swarm optimization algorithm to generate and execute the adjusted wind-coal ratio parameters. Step 404: After executing the adjusted air-coal ratio parameters, continue to monitor the adjustment parameters; if the real-time deviation of the adjustment parameters still does not return to the allowable deviation threshold within the second preset time period, then generate a secondary feedback control signal. Step 405: The secondary feedback control signal is sent to the coal blending decision model, and the coal blending decision model is re-optimized using a non-dominated sorting genetic algorithm to generate a new fuel ratio scheme and execute it. Step 406: Repeat steps 402 to 405 until the operational anomaly is corrected.

[0044] Specifically, sensors, including temperature sensors, pressure sensors, and flue gas composition analyzers, are deployed at key locations in the boiler combustion system. Temperature sensors are placed in the furnace, superheater, economizer, etc., to monitor the temperature distribution within the furnace in real time. Pressure sensors are installed in the air ducts, furnace, etc., to monitor changes in air pressure. The flue gas composition analyzer is used to detect the concentrations of pollutants such as sulfur dioxide, nitrogen oxides, and particulate matter in real time. The adjustment parameters in the real-time monitoring data set include at least the oxygen concentration, nitrogen oxide concentration, and furnace outlet flue gas temperature in the flue gas. The allowable deviation threshold is dynamically set based on the boiler's design parameters, environmental emission standards, and current load.

[0045] This application also provides an application scenario in which the above-mentioned low-carbon combustion optimization method for multi-source fuel blending is applied. Specifically, the low-carbon combustion optimization method for multi-source fuel blending provided in this embodiment can be applied to the intelligent blending scenario of multi-source fuels in coal-fired power plants.

[0046] For example, a large coal-fired power plant (1000MW installed capacity) has gradually introduced multi-source fuels such as biomass pellets, green ammonia, and solid waste-derived fuels in response to the "dual carbon" target. However, due to the complexity of fuel sources and large fluctuations in coal quality, traditional manual coal blending methods are difficult to balance between economy and environmental protection, and the control of the air-coal ratio relies on experience, resulting in large fluctuations in combustion efficiency and frequent NOx emission exceedances.

[0047] Applying the low-carbon combustion optimization method for multi-source fuel blending provided in this embodiment, comprehensive physicochemical property data of the incoming multi-source fuels are first collected. At key nodes such as coal conveyor belts, biomass feeding lines, green ammonia pipelines, and RDF conveying lines, various sensing devices, including near-infrared spectrometers, laser-induced breakdown spectrometers, online elemental analyzers, and high-precision gas sensors, are deployed to collect multi-dimensional information in real time, including numerical coal quality parameters (calorific value Q, volatile matter V, ash content A, sulfur content S, moisture content M), spectral data, and image data. After preprocessing such as missing value imputation, outlier removal, and normalization, the collected data forms a unified multi-dimensional feature vector.

[0048] The multidimensional feature vector is input in real time into a pre-trained coal quality prediction model. This model is based on deep learning technology, integrating a Long Short-Term Memory (LSTM) network and a Convolutional Neural Network (CNN). The LSTM layer is responsible for capturing the evolution of fuel characteristics over time, while the CNN layer is specifically used to extract spatial features from spectral and image data. The features output by both are concatenated in a feature fusion layer and then processed through multiple fully connected networks to accurately predict the key coal quality parameters of each fuel in the future. The mean absolute error (MAE) of the coal quality prediction is controlled within ±0.32 MJ / kg for calorific value and ±1.2% for volatile matter, providing high-precision data support for subsequent optimization.

[0049] After obtaining accurate prediction parameters, the coal blending decision optimization stage begins, constructing a dual-objective coal blending decision model centered on minimizing total electricity cost and pollutant emission concentration. Total electricity cost The entire process cost, including fuel procurement, transportation, storage, processing, and equipment depreciation, was meticulously calculated. Constraints were set for the coal blending decision model: coal quality characteristics (e.g., the calorific value of the blended fuel must be between 19.5-23.5 MJ / kg to ensure stable combustion), resource constraints (the consumption of each fuel must not exceed real-time inventory), process constraints (e.g., the biomass blending limit for a single coal mill is 20%), and strict environmental constraints (NOx and SO2 emission concentrations must be below national standard limits). A non-dominated sorting genetic algorithm (NSGA-II) was used for optimization calculations, which determined the blending ratio of each fuel and the target weight coefficients. The common coding is represented by individual genes, and co-evolution is achieved through operations such as selection, crossover, and mutation. In each generation, based on the non-dominated ranking and crowding of individuals, combined with the target space region guided by their weight coefficients, the elite preservation strategy prioritizes retaining solutions that contribute to expanding the diversity of the Pareto frontier. After multiple generations of evolution, the algorithm finally outputs a set of Pareto optimal solutions as the optimized fuel ratio. For example, under a certain 750MW load condition, a balanced solution is given as: 70% coal, 18% biomass, 7% green ammonia, and 5% RDF.

[0050] After obtaining the optimized fuel blending scheme, the air-coal ratio is further refined. First, based on the fuel blending scheme, the comprehensive combustion characteristic parameters of the mixed fuel (such as ignition temperature and burnout temperature) are calculated using a weighted average method. These parameters, along with the current boiler load, are input into the air-coal ratio optimization model. The air-coal ratio optimization model aims to maximize combustion efficiency, while taking total electricity cost and pollutant emissions as constraints. Optimization variables include total primary air volume, total secondary air volume, and air volume distribution coefficients for each burner. A particle swarm optimization (PSO) algorithm is used for efficient optimization. The air volume control parameters are encoded as particle positions. The particle swarm is initialized within the safe operating range of the boiler. The position of each particle is decoded into specific air-coal ratio control parameters and input into the air-coal ratio optimization model to obtain predicted combustion efficiency, cost, and emission data. The fitness function (based on combustion efficiency, with penalties for violating constraints) is used to evaluate particle performance. The particle velocity and position are dynamically updated by tracking individual historical optima and the global optimum of the swarm, ultimately finding the optimal air-coal ratio.

[0051] To achieve long-term stability and optimization of the combustion process, sensors are deployed at key locations such as the furnace and flue to monitor regulating parameters such as furnace temperature distribution and the concentrations of O2, CO, and NOx in the flue gas. Real-time monitoring data is compared with preset target values ​​and allowable deviation thresholds. If a regulating parameter (such as NOx concentration) is detected to be continuously exceeding limits, the system immediately initiates feedback control. First-level feedback: The air-coal ratio optimization model is triggered to quickly re-optimize (usually within 30 seconds), and the adjusted airflow parameters are executed, attempting to correct deviations by adjusting the air distribution. Second-level feedback: If the operating conditions do not return to normal within a specified time after the first-level adjustment, it is determined that the fuel base needs adjustment, which in turn triggers the coal blending decision model to re-optimize, generating a new fuel blending scheme and executing it. This two-level progressive feedback adjustment strategy of "adjusting air first, then coal" constitutes a dynamic closed-loop control system, ensuring that the boiler maintains a highly efficient, clean, and stable operating state when facing fluctuations in coal quality or load changes.

[0052] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 6As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data on blending and combustion of multi-source fuels. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When executed by the processor, the computer program implements a low-carbon combustion optimization method for blending multi-source fuels.

[0053] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0054] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0055] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0056] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0057] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0058] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0059] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0060] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for optimizing low-carbon combustion by blending multiple fuel sources, characterized in that, The low-carbon combustion optimization method for blending multiple fuels includes: Collect and preprocess the physicochemical properties data of multi-source fuels to obtain multi-dimensional feature vectors, which include numerical coal quality parameters, spectral data and image data; The parameters of the multidimensional feature vector are predicted by a trained coal quality prediction model. The coal quality prediction model is constructed based on a long short-term memory network combined with a convolutional neural network in deep learning. The prediction results of the coal quality prediction model include the predicted values ​​of calorific value, volatile matter, ash content, sulfur content and moisture content. The coal blending decision model is optimized by using a non-dominated sorting genetic algorithm to obtain the optimal fuel ratio. The coal blending decision model is constructed with the dual objectives of minimizing the total cost of electricity and minimizing the pollutant emission concentration, combined with the prediction results and constraints. The optimal air-coal ratio is obtained by performing optimization calculations on the air-coal ratio optimization model using the particle swarm optimization algorithm. The input of the air-coal ratio optimization model includes the fuel ratio output by the coal blending decision model.

2. The low-carbon combustion optimization method for blending multiple fuels according to claim 1, characterized in that, The multi-source fuels include, but are not limited to: coal, biomass fuels, gaseous synthetic fuels, and solid waste-derived fuels; The physicochemical property data includes numerical coal quality parameters, spectral data, and image data. The numerical coal quality parameters include, but are not limited to, calorific value, volatile matter, ash content, sulfur content, and moisture content.

3. The low-carbon combustion optimization method for multi-source fuel blending according to claim 1, characterized in that, The network structure of the coal quality prediction model includes: Input layer: Receives multidimensional feature vectors containing numerical coal quality parameters, spectral data, and image data; Long Short-Term Memory Network Layer: 2 layers, 128 neurons per layer, used to extract temporal features of the spectral data; Convolutional Neural Network Layers: 2 convolutional layers and 1 max pooling layer, used to extract spatial features of the image data. The convolutional kernels of the convolutional layers are 3×3 with a stride of 1 and an activation function of ReLU. The pooling window of the max pooling layer is 2×2. Feature fusion layer: concatenates the outputs of the long short-term memory network layer and the neural network layer; Fully connected layers: 3 layers, with 256, 128 and 64 neurons respectively, and ReLU activation function; Output layer: 5 neurons, outputting predicted values ​​for calorific value, volatile matter, ash content, sulfur content, and moisture content, respectively.

4. The low-carbon combustion optimization method for blending multiple fuels according to claim 1, characterized in that, The functional expression of the coal blending decision model is: ； In the formula, This represents the total cost per kilowatt-hour. The weighting coefficient represents the total cost per kilowatt-hour. This represents the emission concentration of the i-th pollutant. The weighting coefficients representing pollutant emission concentrations, and , Indicates the types and quantities of pollutants; Among them, the total cost per kilowatt-hour The calculation model is as follows: ； In the formula, Indicates fuel procurement costs, Indicates transportation costs, Indicates storage cost, Indicates processing cost, This indicates the cost of equipment depreciation; The pollutant emission concentration The calculation model is as follows: ； In the formula, This represents the predicted emissions of the i-th pollutant. Indicates heat output. Indicates volatile matter, Indicates ash content, Indicates sulfur content. Indicates moisture content. Indicates furnace temperature. This indicates the wind-to-coal ratio.

5. The low-carbon combustion optimization method for multi-source fuel blending according to claim 1, characterized in that, The constraints include coal quality characteristics constraints, resource quantity constraints, process constraints, and environmental constraints. The coal quality characteristics constraint is as follows: ； ； ； In the formula, This indicates the lower limit of the calorific value of the mixed fuels allowed for stable combustion in the boiler. This indicates the received lower heating value of the blended fuel. This indicates the upper limit of the calorific value of the mixed fuels allowed for stable combustion in the boiler; This indicates the lower limit of volatile matter content in the mixed fuels that is permissible for stable combustion in a boiler. This indicates the dry, ash-free volatile matter content of the blended fuel. This indicates the upper limit of volatile matter content in the mixed fuels allowed for stable combustion in the boiler; This indicates the received ash content of the blended fuel. This indicates the maximum permissible ash content limit in boiler design; The resource constraint is as follows: ， ； In the formula, Indicating the first step in the fuel blending scheme The availability of this type of fuel Indicates the first in inventory The availability of this type of fuel Indicates the quantity of each type of fuel; The process constraints are as follows: ； ； In the formula, Indicates the first The lower limit of the blending ratio of various fuels in the process. Indicates the first The mass blending ratio of each fuel in the blended fuel Indicates the first The upper limit of the blending ratio of various fuels in the process; The environmental constraints are as follows: ； In the formula, Indicates the first The emission concentration of various pollutants, Indicates the first Emission concentration limits for various pollutants.

6. The low-carbon combustion optimization method for multi-source fuel blending according to claim 4, characterized in that, The specific process of optimizing the coal blending decision model using a non-dominated sorting genetic algorithm includes: During the algorithm initialization phase, a set of conditions is randomly generated for each non-dominated individual in the population. Weighting coefficients ; Weighting coefficients As part of the genes of each non-dominant individual, it is encoded, crossovered, and mutated together with the decision variables of fuel ratio to achieve the co-evolution of weights and fuel ratio; During the elite retention phase of each generation, based on the non-dominant ranking and crowding of non-dominant individuals, combined with the target spatial region guided by the weight coefficient, priority is given to retaining the weight combination of non-dominant individuals that guide the population toward the current Pareto front sparse region. The weight mutation strategy is dynamically adjusted based on the algorithm's running status. When population diversity decreases, the mutation intensity of the weight coefficients is increased. The optimization process is repeated until the algorithm reaches the preset number of iterations or the Pareto optimal solution remains unchanged for multiple consecutive generations. At least one Pareto optimal solution is then output, which is the optimized fuel ratio.

7. The low-carbon combustion optimization method for blending multiple fuels according to claim 1, characterized in that, The specific process of obtaining the optimized air-coal ratio by performing optimization calculations on the air-coal ratio optimization model using the particle swarm optimization algorithm includes: The comprehensive combustion characteristic parameters of the blended fuel are calculated by weighted average method. The blended fuel is the multi-source fuel in the fuel ratio output by the coal blending decision model. The comprehensive combustion characteristic parameters are input into the air-coal ratio optimization model. With total electricity cost and pollutant emission concentration as constraints and combustion efficiency as the optimization objective, the optimal air-coal ratio parameter combination is searched in the air-coal ratio optimization model using the particle swarm optimization algorithm. The air-coal ratio parameter combination includes the total primary air volume, the total secondary air volume, the air volume distribution coefficient of each burner, and the coal feed rate. The particle swarm optimization algorithm steps include: The total primary air volume, the total secondary air volume, and the air volume distribution coefficient of each burner are encoded as particle position vectors. Within the upper and lower limits of air volume that meet the safety of boiler operation, a group of particles are randomly initialized, and the velocity of each particle is randomly initialized. The position vector of each particle is decoded into the air-coal ratio control parameter, which is then input into the air-coal ratio optimization model to obtain the predicted values ​​of combustion efficiency, total electricity cost, and pollutant emission concentration under the air-coal ratio parameter. Using the predicted combustion efficiency as the base fitness, a penalty function is applied to particles that violate the total cost of electricity or pollutant emission concentration constraints, thereby reducing their effective fitness. Compare the current fitness of each particle with its historical best fitness to update the individual's optimal position; compare the fitness of all particles to update the global optimal position. Based on the individual optimal position and the global optimal position of the particle, the velocity and position of each particle are updated according to the preset inertia weight, individual learning factor and social learning factor. Boundary correction is performed on positions that exceed the upper and lower limits of airflow after the update, and speeds that exceed the speed range are truncated. Repeat the optimization process until the algorithm reaches the preset maximum number of iterations or the global optimal solution is no longer improved after multiple consecutive generations. Then, output the combination of air-coal ratio parameters corresponding to the global optimal position at this point as the optimal solution.

8. The low-carbon combustion optimization method for blending multiple fuels according to claim 1, characterized in that, It also includes real-time monitoring of boiler combustion status, and adjusting the air-coal ratio or fuel blending based on the monitoring results. The specific process includes: Collect furnace temperature, flue gas composition concentration, pressure, and flow parameters to form a real-time monitoring dataset; The adjustment parameters in the real-time monitoring dataset are compared with their corresponding preset target values ​​to calculate the real-time deviation. When the real-time deviation of any adjustment parameter continues to exceed its corresponding allowable deviation threshold for a preset duration, it is determined to be an operational abnormality and a first-level feedback control signal is generated. The first-level feedback control signal is sent to the wind-coal ratio optimization model first, and the wind-coal ratio optimization model is re-optimized by the particle swarm optimization algorithm to generate and execute the adjusted wind-coal ratio parameters. After implementing the adjusted air-coal ratio parameters, the adjustment parameters are continuously monitored; if the real-time deviation of the adjustment parameters still does not return to the allowable deviation threshold within the second preset time period, a secondary feedback control signal is generated. The secondary feedback control signal is sent to the coal blending decision model, and the coal blending decision model is re-optimized using a non-dominated sorting genetic algorithm to generate a new fuel ratio scheme and execute it. The process continues in a loop until the runtime exception is corrected.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the low-carbon combustion optimization method for multi-source fuel blending according to any one of claims 1-8.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the low-carbon combustion optimization method for multi-source fuel blending as described in any one of claims 1-8.

Images