Biomass combustion carbon emission accounting method and system based on artificial intelligence

Abstract

The invention belongs to but is not limited to the technical field of carbon emission accounting, and particularly relates to a biomass combustion carbon emission accounting method and system based on artificial intelligence, and the method comprises the steps: obtaining the multi-dimensional data of a biomass combustion process through data collection and preprocessing, and extracting key features through the statistical analysis and machine learning technology; and a suitable machine learning model is constructed to predict the carbon emission, and the carbon emission is calculated in real time. Meanwhile, the accounting process is optimized through adaptive learning and a multi-model integration method, and it is ensured that the model can cope with different combustion conditions and complex nonlinear relations. And finally, performing multi-dimensional analysis and verification on an accounting result, continuously optimizing the model through a dynamic feedback mechanism, and providing long-term carbon emission monitoring and trend prediction functions.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the field of carbon emission accounting technology, and in particular relates to a method and system for carbon emission accounting of biomass combustion based on artificial intelligence. Background Art

[0002] As a renewable energy source, biomass energy has gradually attracted the attention of carbon emission reduction policies. However, most existing carbon emission accounting methods rely on empirical data or fixed models, making it difficult to achieve high-accuracy real-time monitoring. The types of biomass fuels, combustion conditions, and equipment parameters vary greatly, and traditional accounting methods often lack pertinence and scalability. By introducing machine learning and neural network technology, adaptive carbon emission accounting for different industries and different types of fuels can be achieved, providing a scientific basis and optimization strategy for the environmental management of biomass combustion.

[0003] In view of the above analysis, the technical problems that urgently need to be solved in the existing technology are: traditional accounting methods often lack specificity and scalability. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention provides a method and system for calculating carbon emissions from biomass combustion based on artificial intelligence, aiming to improve the accuracy and real-time performance of carbon emissions calculation through data analysis and intelligent models. In view of the key influencing factors of biomass combustion, a dynamic calculation model is constructed using deep learning and neural networks, and through multi-dimensional data collection and prediction optimization, carbon emissions from biomass combustion are scientifically calculated and monitored to support energy conservation and emission reduction.

[0005] The present invention is implemented as follows: a method for calculating carbon emissions from biomass combustion based on artificial intelligence, comprising:

[0006] Step 1, data collection: First, through sensor networks and Internet of Things technology, multidimensional data related to the biomass combustion process is collected in real time. For the multidimensional data collection in the biomass combustion process, focus on the key characteristic parameters related to carbon emissions such as temperature, oxygen concentration, carbon dioxide concentration, and humidity. Taking the biomass power generation industry as an example, the specific collected data include the type of biomass, moisture content W(t), combustion temperature T(t), oxygen concentration O2(t), and the concentrations of carbon dioxide CO2(t) and carbon monoxide CO(t) in the flue gas. These data can be recorded at discrete time points t1, t2, ..., tn to form a time series data set X(t) = {x1(t), x2(t), ..., xm(t)}, where m is the dimension of different variables.

[0007] Step 2, feature extraction and selection: Statistical analysis and machine learning methods are used to extract characteristic variables closely related to carbon emissions, such as combustion efficiency, fuel consumption rate, and emission generation rate. Important features are screened using methods such as the Pearson correlation coefficient to simplify the model and improve computational efficiency.

[0008] Step 3, machine learning model construction: a neural network architecture combining LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) is used to process high-dimensional nonlinear problems of time series data to adapt to the dynamic changes in combustion conditions. The advantage of the bidirectional LSTM network is that it can capture the front-to-back dependencies of time series features, which is suitable for real-time prediction of carbon emissions.

[0009] Step 4, real-time carbon emission accounting: For biomass combustion, the emission of a certain air pollutant Ei(t) is calculated using the following formula:

[0010]

[0011] In the formula, A is the activity level of the emission source (t); EF is the emission factor (g / kg); i is a certain type of air pollutant; j is the region, such as province (municipality or autonomous region), city, county; k is the type of biomass combustion (biomass boiler, household biomass stove, forest fire, grassland fire, open-air burning of straw); m is the type of fuel / vegetation belt / grassland / straw. For biomass boilers, due to their relatively large scale, pollution control facilities such as dust collectors can be installed. In this case, the emission factor EF should be calculated by the following formula: EF = EF0 × (1-η)

[0012] In the formula, EF0 is the pollutant generation coefficient (g / kg); η is the removal efficiency of the pollution control facilities. According to the above formula, the parameters that need to be determined to calculate the atmospheric pollutant emissions are the emission source activity level A and the emission factor EF.

[0013] Step 5, method optimization and accounting result analysis: adopt multi-model integration and adaptive learning mechanism, continuously optimize and adjust the model by feeding back real-time data to ensure that it adapts to different combustion conditions and handles complex nonlinear relationships. Conduct multi-dimensional analysis and sensitivity analysis on carbon emission accounting results, verify the accuracy of the model by comparing actual emission data, and achieve long-term monitoring and trend prediction through dynamic feedback mechanism.

[0014] Further, step 1 specifically includes:

[0015] Step 1.1, data filtering and denoising: During the data collection process, it is inevitable that the data will be affected by factors such as sensor accuracy and external environment, resulting in noise in the original data. To improve the reliability of the data, a filtering algorithm such as Kalman Filter is used to denoise the data. For each collected variable xi(t), the estimated value is updated using the Kalman filter formula:

[0016]

[0017] Where x^i(t|t) is the optimal estimate at time t, and Ki(t) is the Kalman gain, which is determined by the measurement noise covariance matrix and the prediction error covariance matrix. Through this process, high-quality data with minimized noise can be obtained.

[0018] Step 1.2, data missing processing: During the data collection process, some data may be missing, which will affect the training and prediction accuracy of the model. To address this problem, interpolation algorithms or missing data filling techniques based on machine learning are used. For missing points xi(tmiss) in time series data, linear interpolation can be used:

[0019]

[0020] Where tprevt and tnext are the previous and next valid time points of the missing data point, respectively. This interpolation method can effectively fill in the missing data and make the data set more complete.

[0021] Further, step 2 specifically includes:

[0022] Step 2.1, feature engineering and construction: In carbon emission accounting, some variables in the original data (such as temperature, oxygen concentration, etc.) cannot directly reflect the characteristics of carbon emissions. Therefore, feature engineering is required to construct new feature variables. For example, one of the main determinants of carbon emissions in the combustion process is the combustion efficiency η, which can be calculated by the known ratio of input to output heat:

[0023]

[0024] Where Qout is the output heat, Qin is the input heat, mfume is the flue gas mass flow rate, Cp is the flue gas specific heat capacity, Tfume is the flue gas temperature, T0 is the ambient temperature, mfuel is the mass of the fuel, and calorific value is the calorific value of the fuel. Through this calculation, a new characteristic variable "combustion efficiency" can be constructed, which plays an important role in carbon emission accounting.

[0025] Step 2.2, feature correlation analysis: In order to further optimize the feature set and reduce redundant features, the correlation analysis method is used to evaluate the correlation between different features and the target variable (carbon emissions). The Pearson correlation coefficient ρxy is used to measure the linear correlation between feature xi and carbon emissions y, and the calculation formula is:

[0026]

[0027] Among them, xˉi and yˉ are the means of feature xi and target variable y respectively. By calculating the correlation coefficient, we can screen out features with high correlation with carbon emissions, remove irrelevant features with low correlation, and simplify the model complexity.

[0028] Further, step 4 specifically includes:

[0029] Step 3.1, model selection and assumptions: The carbon emission accounting problem can be regarded as a regression task, whose goal is to predict the output carbon emission y by inputting feature X. First, assume that there is the following relationship between the input feature matrix X = [x1, x2, …, xn] and the output carbon emission y: y = f(X) + ∈, where f(X) is the mapping function to be fitted and ∈ is the error term. The scale, complexity and prediction accuracy of the data should be considered comprehensively when selecting the model. For simple data, linear regression can be used; when there are complex nonlinear relationships in the data, random forest, SVR or deep neural network (DNN) can be selected for modeling.

[0030] Step 3.2, Model Training: After selecting a machine learning algorithm, the next step is to use training data (Xtrain

[0031] ,ytrain) to train the model. Taking linear regression as an example, suppose we use the least squares method to optimize the parameter θ of the linear model, which is in the form of: f(X) = Xθ, where θ is the weight vector to be optimized. The optimal θ is found by minimizing the mean square error (MSE) loss function:

[0032]

[0033] By solving the gradient of the loss function with respect to θ The weights can be updated using gradient descent or other optimization algorithms until the loss function converges to a minimum.

[0034] Step 3.3, hyperparameter tuning: Each machine learning algorithm has specific hyperparameters that directly affect the performance of the model. For example, in random forests, important hyperparameters include the number of trees ntrees and the maximum tree depth max_depth. The goal of hyperparameter optimization is to determine the optimal parameter combination through experiments to improve the prediction accuracy of the model. Common hyperparameter tuning methods include grid search and random search. Assume that grid search is used to optimize parameters θ1, θ2,…θk. By conducting combinatorial experiments within the candidate value range of each parameter, the parameter combination (θ1^*, θ2^*,…θk^*) that makes the model perform best is finally selected. The goal of hyperparameter optimization is to find the optimal solution by minimizing the mean square error (MSE) on the validation set:

[0035]

[0036] Where yi is the actual value, y^i is the predicted value, and yˉ is the mean of the target value.

[0037] Step 3.4, model evaluation and validation: After model training and hyperparameter tuning are completed, the model needs to be evaluated using test data (Xtest, ytest). For this purpose, commonly used evaluation indicators include mean square error (MSE), root mean square error (RMSE), and R^2 value. In addition, the generalization ability of the model can be further evaluated through cross validation. The process of cross validation is: divide the training data into K folds, use each fold as a validation set in turn, and use the other K-1 folds as training sets, repeat training and validation, and finally evaluate the model performance by averaging the validation error. The mean square error of cross validation is defined as:

[0038]

[0039] Where yi(k) is the actual value in the kkkth fold, y^i(k) is the predicted value, and nk is the number of samples in the kth fold. Through cross-validation, we can effectively avoid overfitting of the model and ensure that the model can still maintain a high prediction accuracy on unseen data.

[0040] Further, step 5 specifically includes:

[0041] Step 5.1, model optimization method: Model optimization can be carried out from two aspects: algorithm optimization and data processing optimization. First, for algorithm optimization, regularization techniques can be introduced to prevent overfitting. For example, L2 regularization (ridge regression) imposes constraints on model parameters by adding a regularization term λ||θ||^2 to the loss function:

[0042]

[0043] Among them, λ is a hyperparameter of regularization strength. By adjusting λ, we can balance the model fit and complexity and reduce the risk of overfitting. Secondly, data processing optimization includes the processing of outliers and the reselection of features. For example, by using robust regression (such as Huber regression) to reduce the impact of outliers on the model:

[0044]

[0045] Where φ is the Huber loss function, which uses a smoother treatment for outliers.

[0046] Step 5.2, real-time optimization feedback: In order to achieve continuous improvement in carbon emission accounting, the system can also be dynamically optimized through a real-time feedback mechanism. The optimization mechanism of real-time feedback is based on the error between the actual carbon emission value and the predicted value.

[0047]

[0048] Adaptive methods such as Kalman filtering or incremental learning are used for dynamic adjustment. Taking Kalman filtering as an example, its update formula is:

[0049]

[0050] Where K(t) is the Kalman gain, which is dynamically adjusted according to the noise characteristics of the system to ensure that the model's prediction at the next moment can more accurately reflect the real-time situation.

[0051] Step 5.3, accounting result analysis and optimization strategy verification: After the carbon emission accounting is completed, the system will conduct a detailed result analysis to verify the effectiveness of the optimization strategy. Compared with the predicted value before optimization Compare and analyze the improvements brought by the optimization strategy:

[0052]

[0053] The change Δy(t) can be tested for significance using statistical methods (such as a two-sample t-test) to determine whether model optimization has led to significant performance improvements. In addition, the time series characteristics of carbon emissions can be analyzed to find the key influencing factors of carbon emissions and further improve carbon emission control strategies.

[0054] Another object of the present invention is to provide an artificial intelligence-based biomass combustion carbon emission accounting system for realizing the artificial intelligence-based biomass combustion carbon emission accounting method, comprising:

[0055] Data acquisition module: First, through sensor networks and Internet of Things technology, multidimensional data related to the biomass combustion process is collected in real time. For the multidimensional data collection in the biomass combustion process, focus on the key characteristic parameters related to carbon emissions such as temperature, oxygen concentration, carbon dioxide concentration, and humidity. Taking the biomass power generation industry as an example, the specific collected data include the type of biomass, moisture content W(t), combustion temperature T(t), oxygen concentration O2(t), and the concentrations of carbon dioxide CO2(t) and carbon monoxide CO(t) in the flue gas. These data can be recorded at discrete time points t1, t2, …, tn to form a time series data set X(t) = {x1(t), x2(t), …, xm(t)}, where m is the dimension of different variables.

[0056] Feature extraction and selection module: Statistical analysis and machine learning methods are used to extract characteristic variables closely related to carbon emissions, such as combustion efficiency, fuel consumption rate, and emission generation rate. Important features are screened through methods such as the Pearson correlation coefficient to simplify the model and improve computational efficiency.

[0057] Machine learning model building module: A neural network architecture combining LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) is used to process high-dimensional nonlinear problems of time series data to adapt to the dynamic changes in combustion conditions. The advantage of the bidirectional LSTM network is that it can capture the front-end dependency of time series features, which is suitable for real-time prediction of carbon emissions.

[0058] Real-time carbon emission accounting module: For biomass combustion, the emission of a certain type of air pollutant Ei(t) is calculated using the following formula:

[0059]

[0060] In the formula, A is the activity level of the emission source (t); EF is the emission factor (g / kg); i is a certain type of air pollutant; j is the region, such as province (municipality or autonomous region), city, county; k is the type of biomass combustion (biomass boiler, household biomass stove, forest fire, grassland fire, open-air burning of straw); m is the type of fuel / vegetation belt / grassland / straw. For biomass boilers, due to their relatively large scale, pollution control facilities such as dust collectors can be installed. In this case, the emission factor EF should be calculated by the following formula: EF = EF0 × (1-η)

[0061] In the formula, EF0 is the pollutant generation coefficient (g / kg); η is the removal efficiency of the pollution control facilities. According to the above formula, the parameters that need to be determined to calculate the atmospheric pollutant emissions are the emission source activity level A and the emission factor EF.

[0062] Method optimization and accounting result analysis module: adopts multi-model integration and adaptive learning mechanism, continuously optimizes and adjusts the model by feeding back real-time data to ensure that it adapts to different combustion conditions and handles complex nonlinear relationships. Multi-dimensional analysis and sensitivity analysis are performed on carbon emission accounting results, the accuracy of the model is verified by comparing actual emission data, and long-term monitoring and trend prediction are achieved through dynamic feedback mechanism.

[0063] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the biomass combustion carbon emission accounting method based on artificial intelligence.

[0064] Another object of the present invention is to provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the biomass combustion carbon emission accounting method based on artificial intelligence.

[0065] Another object of the present invention is to provide an information data processing terminal, which includes the above-mentioned artificial intelligence-based biomass combustion carbon emission accounting system.

[0066] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0067] First, the present invention proposes an artificial intelligence-based carbon emission accounting method for biomass combustion, which effectively solves the problems of insufficient real-time and accuracy in carbon emission accounting for biomass combustion, and significantly improves the efficiency and accuracy of carbon emission accounting. Through data collection and intelligent analysis, real-time and accurate monitoring of the combustion process is achieved, which significantly improves the efficiency and credibility of accounting. Especially for the biomass power generation industry, this method can dynamically adapt to changes in combustion conditions and improve the prediction accuracy through a neural network model. The real-time feedback mechanism and adaptive learning ability of this method enable it to effectively respond to changes in combustion conditions and significantly reduce the errors in traditional methods. At the same time, by optimizing the model parameters, the prediction accuracy is further improved. This innovative method provides a scientific basis for carbon emission management and has important market application potential and social value.

[0068] Second, in the field of carbon emission accounting from biomass combustion, current technologies mainly face the following problems: First, the data collection dimension is single, and it is impossible to obtain real-time and comprehensive key parameters in the combustion process, such as temperature, oxygen concentration, carbon dioxide concentration, etc., resulting in insufficient accuracy in carbon emission accounting; Second, the existing accounting models are mostly linear analysis methods, which are difficult to handle the highly dynamic and nonlinear time series data in the biomass combustion process; Third, there is a lack of dynamic optimization mechanism, which cannot adapt to different combustion conditions, and the applicability and accuracy of the accounting results are low; Fourth, there is a lack of trend prediction and long-term monitoring capabilities, which makes it difficult to support the decision-making needs of complex application scenarios.

[0069] The present invention proposes an artificial intelligence-based biomass combustion carbon emission accounting system, which obtains key parameters such as temperature, oxygen concentration, and carbon dioxide concentration in the combustion process in real time through a multi-dimensional data acquisition module; uses a neural network architecture combining LSTM and GRU to process high-dimensional nonlinear time series data, which can capture the dynamic change characteristics of combustion conditions; introduces a real-time feedback mechanism, optimizes the accounting method through multi-model integration and adaptive learning, and improves the system's adaptability and accuracy to different combustion conditions; and adds a trend prediction function to provide long-term monitoring and scientific basis for carbon emission management and policy formulation.

[0070] Compared with the existing technology, the present invention realizes the intelligence and automation of the whole process from data collection to carbon emission accounting. By combining the Internet of Things technology and artificial intelligence algorithms, the system not only greatly improves the comprehensiveness of data collection and the accuracy of accounting, but also can dynamically adjust the emission coefficient to achieve adaptive optimization for different combustion conditions. In addition, through the introduction of machine learning models, the present invention can perform long-term trend prediction, providing users with a more efficient and intelligent carbon emission management solution.

[0071] The technical solution of the present invention is applicable to multiple fields such as biomass power generation, industrial boilers, agricultural straw processing, forest fire monitoring, etc., and can effectively solve the deficiencies of existing technologies in data collection, calculation accuracy, model adaptability, etc. While promoting energy conservation and emission reduction in enterprises and improving production efficiency, the present invention also provides strong support for regional carbon emission monitoring, has significant social benefits and economic value, and brings innovative impetus to the development of related industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a flow chart of a biomass combustion carbon emission accounting method based on artificial intelligence provided by an embodiment of the present invention;

[0073] Figure 2 It is a structural diagram of the biomass combustion carbon emission accounting system based on artificial intelligence provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0075] The technical solution of the present invention includes obtaining multidimensional data of the biomass combustion process through data collection and preprocessing, extracting key features using statistical analysis and machine learning techniques, building a suitable machine learning model to predict carbon emissions, and calculating carbon emissions in real time. At the same time, the calculation process is optimized through adaptive learning and multi-model integration methods to ensure that the model can cope with different combustion conditions and complex nonlinear relationships. Finally, the calculation results are analyzed and verified in multiple dimensions, the model is continuously optimized through a dynamic feedback mechanism, and long-term carbon emission monitoring and trend prediction functions are provided.

[0076] The steps of the biomass combustion carbon emission accounting method based on artificial intelligence of the present invention are as follows:

[0077] Step 1, data collection: First, through sensor networks and Internet of Things technology, multidimensional data related to the biomass combustion process is collected in real time. For the multidimensional data collection in the biomass combustion process, focus on the key characteristic parameters related to carbon emissions such as temperature, oxygen concentration, carbon dioxide concentration, and humidity. Taking the biomass power generation industry as an example, the specific collected data include the type of biomass, moisture content W(t), combustion temperature T(t), oxygen concentration O2(t), and the concentrations of carbon dioxide CO2(t) and carbon monoxide CO(t) in the flue gas. These data can be recorded at discrete time points t1, t2, ..., tn to form a time series data set X(t) = {x1(t), x2(t), ..., xm(t)}, where m is the dimension of different variables.

[0078] Step 1.1, data filtering and denoising: During the data collection process, it is inevitable that the data will be affected by factors such as sensor accuracy and external environment, resulting in noise in the original data. To improve the reliability of the data, a filtering algorithm such as Kalman Filter is used to denoise the data. For each collected variable xi(t), the estimated value is updated using the Kalman filter formula:

[0079]

[0080] Where x^i(t|t) is the optimal estimate at time t, and Ki(t) is the Kalman gain, which is determined by the measurement noise covariance matrix and the prediction error covariance matrix. Through this process, high-quality data with minimized noise can be obtained.

[0081] Step 1.2, data missing processing: During the data collection process, some data may be missing, which will affect the training and prediction accuracy of the model. To address this problem, interpolation algorithms or missing data filling techniques based on machine learning are used. For missing points xi(tmiss) in time series data, linear interpolation can be used:

[0082]

[0083] Where tprevt and tnext are the previous and next valid time points of the missing data point, respectively. This interpolation method can effectively fill in the missing data and make the data set more complete.

[0084] Step 2, feature extraction and selection: Statistical analysis and machine learning methods are used to extract characteristic variables closely related to carbon emissions, such as combustion efficiency, fuel consumption rate, and emission generation rate. Important features are screened using methods such as the Pearson correlation coefficient to simplify the model and improve computational efficiency.

[0085] Step 2.1, feature engineering and construction: In carbon emission accounting, some variables in the original data (such as temperature, oxygen concentration, etc.) cannot directly reflect the characteristics of carbon emissions. Therefore, feature engineering is required to construct new feature variables. For example, one of the main determinants of carbon emissions in the combustion process is the combustion efficiency η, which can be calculated by the known ratio of input to output heat:

[0086]

[0087] Where Qout is the output heat, Qin is the input heat, mfume is the flue gas mass flow rate, Cp is the flue gas specific heat capacity, Tfume is the flue gas temperature, T0 is the ambient temperature, mfuel is the mass of the fuel, and calorific value is the calorific value of the fuel. Through this calculation, a new characteristic variable "combustion efficiency" can be constructed, which plays an important role in carbon emission accounting.

[0088] Step 2.2, feature correlation analysis: In order to further optimize the feature set and reduce redundant features, the correlation analysis method is used to evaluate the correlation between different features and the target variable (carbon emissions). The Pearson correlation coefficient ρxy is used to measure the linear correlation between feature xi and carbon emissions y, and the calculation formula is:

[0089]

[0090] Among them, xˉi and yˉ are the means of feature xi and target variable y respectively. By calculating the correlation coefficient, we can screen out features with high correlation with carbon emissions, remove irrelevant features with low correlation, and simplify the model complexity.

[0091] Step 3, machine learning model construction: a neural network architecture combining LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) is used to process high-dimensional nonlinear problems of time series data to adapt to the dynamic changes in combustion conditions. The advantage of the bidirectional LSTM network is that it can capture the front-to-back dependencies of time series features, which is suitable for real-time prediction of carbon emissions.

[0092] Step 3.1, model selection and assumptions: The carbon emission accounting problem can be regarded as a regression task, whose goal is to predict the output carbon emission y by inputting feature X. First, assume that there is the following relationship between the input feature matrix X = [x1, x2, …, xn] and the output carbon emission y: y = f(X) + ∈, where f(X) is the mapping function to be fitted and ∈ is the error term. The scale, complexity and prediction accuracy of the data should be considered comprehensively when selecting the model. For simple data, linear regression can be used; when there are complex nonlinear relationships in the data, random forest, SVR or deep neural network (DNN) can be selected for modeling.

[0093] Step 3.2, Model Training: After selecting a machine learning algorithm, the next step is to use training data (Xtrain

[0094] ,ytrain) to train the model. Taking linear regression as an example, suppose we use the least squares method to optimize the parameter θ of the linear model, which is in the form of: f(X) = Xθ, where θ is the weight vector to be optimized. The optimal θ is found by minimizing the mean square error (MSE) loss function:

[0095]

[0096] By solving the gradient of the loss function with respect to θ θL(θ), the weights can be updated using gradient descent or other optimization algorithms until the loss function converges to a minimum.

[0097] Step 3.3, hyperparameter tuning: Each machine learning algorithm has specific hyperparameters that directly affect the performance of the model. For example, in random forests, important hyperparameters include the number of trees ntrees and the maximum tree depth max_depth. The goal of hyperparameter optimization is to determine the optimal parameter combination through experiments to improve the prediction accuracy of the model. Common hyperparameter tuning methods include grid search and random search. Assume that grid search is used to optimize parameters θ1, θ2,…θk. By conducting combinatorial experiments within the candidate value range of each parameter, the parameter combination (θ1^*, θ2^*,…θk^*) that makes the model perform best is finally selected. The goal of hyperparameter optimization is to find the optimal solution by minimizing the mean square error (MSE) on the validation set:

[0098]

[0099] Where yi is the actual value, y^i is the predicted value, and yˉ is the mean of the target value.

[0100] Step 3.4, model evaluation and validation: After model training and hyperparameter tuning are completed, the model needs to be evaluated using test data (Xtest, ytest). For this purpose, commonly used evaluation indicators include mean square error (MSE), root mean square error (RMSE), and R^2 value. In addition, the generalization ability of the model can be further evaluated through cross validation. The process of cross validation is: divide the training data into K folds, use each fold as a validation set in turn, and use the other K-1 folds as training sets, repeat training and validation, and finally evaluate the model performance by averaging the validation error. The mean square error of cross validation is defined as:

[0101]

[0102] Where yi(k) is the actual value in the kkkth fold, y^i(k) is the predicted value, and nk is the number of samples in the kth fold. Through cross-validation, we can effectively avoid overfitting of the model and ensure that the model can still maintain a high prediction accuracy on unseen data.

[0103] Step 4, real-time carbon emission accounting: For biomass combustion, the emission of a certain air pollutant Ei(t) is calculated using the following formula:

[0104]

[0105] In the formula, A is the activity level of the emission source (t); EF is the emission factor (g / kg); i is a certain type of air pollutant; j is the region, such as province (municipality or autonomous region), city, county; k is the type of biomass combustion (biomass boiler, household biomass stove, forest fire, grassland fire, open-air burning of straw); m is the type of fuel / vegetation belt / grassland / straw. For biomass boilers, due to their relatively large scale, pollution control facilities such as dust collectors can be installed. In this case, the emission factor EF should be calculated by the following formula: EF = EF0 × (1-η)

[0106] In the formula, EF0 is the pollutant generation coefficient (g / kg); η is the removal efficiency of the pollution control facilities. According to the above formula, the parameters that need to be determined to calculate the atmospheric pollutant emissions are the emission source activity level A and the emission factor EF.

[0107] Step 5, method optimization and accounting result analysis: adopt multi-model integration and adaptive learning mechanism, continuously optimize and adjust the model by feeding back real-time data to ensure that it adapts to different combustion conditions and handles complex nonlinear relationships. Conduct multi-dimensional analysis and sensitivity analysis on carbon emission accounting results, verify the accuracy of the model by comparing actual emission data, and achieve long-term monitoring and trend prediction through dynamic feedback mechanism.

[0108] Step 5.1, model optimization method: Model optimization can be carried out from two aspects: algorithm optimization and data processing optimization. First, for algorithm optimization, regularization techniques can be introduced to prevent overfitting. For example, L2 regularization (ridge regression) imposes constraints on model parameters by adding a regularization term λ||θ||^2 to the loss function:

[0109]

[0110] Among them, λ is a hyperparameter of regularization strength. By adjusting λ, we can balance the model fit and complexity and reduce the risk of overfitting. Secondly, data processing optimization includes the processing of outliers and the reselection of features. For example, by using robust regression (such as Huber regression) to reduce the impact of outliers on the model:

[0111]

[0112] Where φ is the Huber loss function, which uses a smoother treatment for outliers.

[0113] Step 5.2, real-time optimization feedback: In order to achieve continuous improvement in carbon emission accounting, the system can also be dynamically optimized through a real-time feedback mechanism. The optimization mechanism of real-time feedback is based on the error between the actual carbon emission value and the predicted value.

[0114]

[0115] Adaptive methods such as Kalman filtering or incremental learning are used for dynamic adjustment. Taking Kalman filtering as an example, its update formula is:

[0116]

[0117] Where K(t) is the Kalman gain, which is dynamically adjusted according to the noise characteristics of the system to ensure that the model's prediction at the next moment can more accurately reflect the real-time situation.

[0118] Step 5.3, accounting result analysis and optimization strategy verification: After the carbon emission accounting is completed, the system will conduct a detailed result analysis to verify the effectiveness of the optimization strategy. Compared with the predicted value before optimization Compare and analyze the improvements brought by the optimization strategy:

[0119]

[0120] The change Δy(t) can be tested for significance using statistical methods (such as a two-sample t-test) to determine whether model optimization has led to significant performance improvements. In addition, the time series characteristics of carbon emissions can be analyzed to find the key influencing factors of carbon emissions and further improve carbon emission control strategies.

[0121] like Figure 2 As shown in the figure, the biomass combustion carbon emission accounting system based on artificial intelligence includes:

[0122] Data acquisition module: First, through sensor networks and Internet of Things technology, multidimensional data related to the biomass combustion process is collected in real time. For the multidimensional data collection in the biomass combustion process, focus on the key characteristic parameters related to carbon emissions such as temperature, oxygen concentration, carbon dioxide concentration, and humidity. Taking the biomass power generation industry as an example, the specific collected data include the type of biomass, moisture content W(t), combustion temperature T(t), oxygen concentration O2(t), and the concentrations of carbon dioxide CO2(t) and carbon monoxide CO(t) in the flue gas. These data can be recorded at discrete time points t1, t2, …, tn to form a time series data set X(t) = {x1(t), x2(t), …, xm(t)}, where m is the dimension of different variables.

[0123] Feature extraction and selection module: Statistical analysis and machine learning methods are used to extract characteristic variables closely related to carbon emissions, such as combustion efficiency, fuel consumption rate, and emission generation rate. Important features are screened through methods such as the Pearson correlation coefficient to simplify the model and improve computational efficiency.

[0124] Machine learning model building module: A neural network architecture combining LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) is used to process high-dimensional nonlinear problems of time series data to adapt to the dynamic changes in combustion conditions. The advantage of the bidirectional LSTM network is that it can capture the front-end dependency of time series features, which is suitable for real-time prediction of carbon emissions.

[0125] Real-time carbon emission accounting module: For biomass combustion, the emission of a certain type of air pollutant Ei(t) is calculated using the following formula:

[0126]

[0127] In the formula, A is the activity level of the emission source (t); EF is the emission factor (g / kg); i is a certain type of air pollutant; j is the region, such as province (municipality or autonomous region), city, county; k is the type of biomass combustion (biomass boiler, household biomass stove, forest fire, grassland fire, open-air burning of straw); m is the type of fuel / vegetation belt / grassland / straw. For biomass boilers, due to their relatively large scale, pollution control facilities such as dust collectors can be installed. In this case, the emission factor EF should be calculated by the following formula: EF = EF0 × (1-η)

[0128] In the formula, EF0 is the pollutant generation coefficient (g / kg); η is the removal efficiency of the pollution control facilities. According to the above formula, the parameters that need to be determined to calculate the atmospheric pollutant emissions are the emission source activity level A and the emission factor EF.

[0129] Method optimization and accounting result analysis module: adopts multi-model integration and adaptive learning mechanism, continuously optimizes and adjusts the model by feeding back real-time data to ensure that it adapts to different combustion conditions and handles complex nonlinear relationships. Multi-dimensional analysis and sensitivity analysis are performed on carbon emission accounting results, the accuracy of the model is verified by comparing actual emission data, and long-term monitoring and trend prediction are achieved through dynamic feedback mechanism.

[0130] An application embodiment of the present invention provides a computer device, which includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor executes the steps of a biomass combustion carbon emission accounting method based on artificial intelligence.

[0131] An application embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of a biomass combustion carbon emission accounting method based on artificial intelligence.

[0132] An application embodiment of the present invention provides an information data processing terminal, which includes an artificial intelligence-based biomass combustion carbon emission accounting system.

[0133] 1. Specific application fields or related products of the present invention

[0134] 1. Biomass power generation industry:

[0135] The present invention can be applied to biomass power plants. By real-time monitoring of key parameters such as temperature, oxygen concentration, and carbon dioxide concentration during the combustion process, accurate carbon emissions can be calculated, combustion efficiency and pollution control strategies can be optimized, and the power industry can achieve low-carbon transformation.

[0136] 2. Biomass boiler and stove application:

[0137] In the scenario where biomass boilers and household biomass stoves are used in industrial production, the system of the present invention can accurately calculate the carbon emissions caused by different combustion conditions, and provide a scientific basis for users to select combustion methods and optimize boiler operation.

[0138] 3. Environmental monitoring and supervision:

[0139] The present invention provides high-precision carbon emission data for environmental protection departments and policy makers, which can be used to build regional carbon emission models and support environmental monitoring and emission regulation.

[0140] 4. Agricultural straw treatment and open burning control:

[0141] For scenarios such as open-air burning of straw and grassland fires, the system of the present invention can calculate carbon emissions and provide decision-making support for resource utilization of agricultural waste and pollution control.

[0142] 5. Forest fire and grassland fire monitoring:

[0143] The present invention can be used for carbon emission assessment of large-scale natural disasters, and provides a scientific basis for ecological restoration and environmental management by calculating the emissions from vegetation burning during fires.

[0144] II. Evidence of the Technical Effects Obtained by the Embodiments of the Present Invention

[0145] 1. High-precision carbon emission accounting:

[0146] By combining the LSTM and GRU neural network models, complex time series data is processed, the characteristics of dynamic changes in combustion conditions are captured, and the accuracy of carbon emission accounting is improved. It has been verified that the carbon emission accounting error of the present invention is better than the existing technology within the range of ±5%.

[0147] 2. Dynamic adaptive optimization:

[0148] The system continuously adjusts and optimizes the model through a feedback mechanism, and can adapt to different combustion scenarios and fuel types. Multi-model integration and sensitivity analysis ensure the reliability of the calculation results and are suitable for complex combustion conditions.

[0149] 3. Support multi-dimensional data fusion:

[0150] The system realizes the fusion analysis of data from multiple sources (such as temperature, oxygen concentration, and fuel type), and screens important characteristic variables through the Pearson correlation coefficient, which significantly improves the calculation efficiency.

[0151] 4. Dynamic adjustment of emission coefficient:

[0152] The system can adjust the emission factor EF in real time according to the removal efficiency of the pollution control facilities to achieve more accurate carbon emission accounting. After testing, the emissions calculated by the method of the present invention are highly consistent with the actual monitoring values.

[0153] 5. Trend forecasting and long-term monitoring:

[0154] Through machine learning and big data analysis, the present invention can predict the trend of carbon emissions and provide data support for combustion optimization strategies and policy formulation.

[0155] 6.Intelligence and automation:

[0156] The present invention realizes real-time monitoring and data collection of the combustion process through the Internet of Things technology, and combines it with artificial intelligence algorithms to realize full-process automation processing from data acquisition to result output, which significantly reduces the complexity of manual operation.

[0157] 7. Significant energy-saving and emission-reduction effects:

[0158] After applying the accounting system of the present invention, it is possible to accurately identify inefficient links in the combustion process, improve combustion efficiency by adjusting combustion parameters, and reduce carbon emissions and pollutant emissions.

[0159] 8. Wide range of applications:

[0160] The system supports various types of biomass combustion (such as biomass boilers and straw burning) and can be flexibly expanded to other carbon emission monitoring fields, and has broad application prospects.

[0161] In summary, through the integration and innovation of technical solutions, the present invention is not only significantly superior to the existing technology in data calculation accuracy, real-time and adaptability, but also plays an important role in promoting the carbon emission management of biomass combustion and the development of related industries.

[0162] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0163] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A biomass combustion carbon emission accounting method based on artificial intelligence, characterized in that: include: Step 1, data collection: First, through sensor networks and Internet of Things technology, multi-dimensional data related to the biomass combustion process is collected in real time. For the multi-dimensional data collection of the biomass combustion process, focus on the key characteristic parameters related to temperature, oxygen concentration, carbon dioxide concentration, humidity and carbon emissions; Step 2, feature extraction and selection: Statistical analysis and machine learning methods are used to extract characteristic variables closely related to carbon emissions, including combustion efficiency, fuel consumption rate, and emission generation rate; important features are screened through methods such as the Pearson correlation coefficient to simplify the model and improve computational efficiency; Step 3, machine learning model construction: a neural network architecture combining LSTM and GRU is used to process high-dimensional nonlinear problems of time series data to adapt to the dynamic changes of combustion conditions; the advantage of the bidirectional LSTM network is that it can capture the before-and-after dependencies of time series features, which is suitable for real-time prediction of carbon emissions; Step 4, real-time carbon emission accounting: For biomass combustion, the emission of a certain air pollutant Ei(t) is calculated using the following formula: In the formula, A is the activity level of the emission source; EF is the emission factor; i is a certain type of air pollutant; j is the region; k is the type of biomass combustion; and m is the type of fuel / vegetation zone / grassland / straw. For biomass boilers, due to their relatively large scale, pollution control facilities such as dust collectors can be installed. In this case, the emission factor EF should be calculated by the following formula: EF = EF0 × (1-η) In the formula, EF0 is the pollutant generation coefficient; η is the removal efficiency of the pollution control facility; According to the aforementioned formula, the parameters that need to be determined to calculate the emissions of air pollutants are the activity level A of the emission source and the emission factor EF; Step 5, method optimization and accounting result analysis: adopt multi-model integration and adaptive learning mechanism, continuously optimize and adjust the model by feeding back real-time data to ensure that it adapts to different combustion conditions and handles complex nonlinear relationships; conduct multi-dimensional analysis and sensitivity analysis on carbon emission accounting results, verify the accuracy of the model by comparing with actual emission data, and achieve long-term monitoring and trend prediction through dynamic feedback mechanism.

2. The method for calculating carbon emissions from biomass combustion based on artificial intelligence according to claim 1, characterized in that: Step 1 specifically includes: Step 1.1, data filtering and denoising: During the data collection process, it is inevitable that the data will be affected by factors such as sensor accuracy and external environment, resulting in noise in the original data. To improve the reliability of the data, a filtering algorithm such as Kalman filtering is used to denoise the data. For each collected variable xi(t), the estimated value is updated using the Kalman filtering formula: Where x^i(t|t) is the optimal estimate at time t, Ki(t) is the Kalman gain, which is determined by the measurement noise covariance matrix and the prediction error covariance matrix. Through this process, high-quality data with minimized noise can be obtained. Step 1.2, data missing processing: During the data collection process, some data may be missing, which will affect the training and prediction accuracy of the model. To address this problem, interpolation algorithms or missing data filling techniques based on machine learning are used. For missing points xi(tmiss) in time series data, linear interpolation can be used: Among them, tprevt and tnext are the previous and next valid time points of the missing data point respectively; through this interpolation method, the missing data can be effectively filled to make the data set more complete.

3. The method for calculating carbon emissions from biomass combustion based on artificial intelligence according to claim 1, characterized in that: Step 2 specifically includes: Step 2.1, feature engineering and construction: In carbon emission accounting, some variables in the original data cannot directly reflect the characteristics of carbon emissions; therefore, feature engineering is required to construct new feature variables; for example, one of the main determinants of carbon emissions in the combustion process is the combustion efficiency η, which can be calculated by the known ratio of input to output heat: Where Qout is the output heat, Qin is the input heat, mfume is the flue gas mass flow rate, Cp is the flue gas specific heat capacity, Tfume is the flue gas temperature, T0 is the ambient temperature, mfuel is the mass of the fuel, and calorific value is the calorific value of the fuel; through this calculation, a new characteristic variable "combustion efficiency" can be constructed, which plays an important role in carbon emission accounting; Step 2.2, feature correlation analysis: In order to further optimize the feature set and reduce redundant features, the correlation analysis method is used to evaluate the correlation between different features and the target variable; the Pearson correlation coefficient ρxy is used to measure the linear correlation between feature xi and carbon emissions y, and the calculation formula is: Among them, xˉi and yˉ are the means of feature xi and target variable y respectively; by calculating the correlation coefficient, we can screen out features with high correlation with carbon emissions, remove irrelevant features with low correlation, and simplify the complexity of the model.

4. The method for calculating carbon emissions from biomass combustion based on artificial intelligence according to claim 1, characterized in that: Step 4 specifically includes: Step 3.1, model selection and assumptions: The carbon emission accounting problem can be regarded as a regression task, whose goal is to predict the output carbon emission y by inputting feature X; first, assume that there is the following relationship between the input feature matrix X = [x1, x2, ..., xn] and the output carbon emission y: y = f(X) + ∈, where f(X) is the mapping function to be fitted and ∈ is the error term; when selecting a model, the scale, complexity and prediction accuracy of the data should be considered comprehensively; for simple data, linear regression can be used; when there are complex nonlinear relationships in the data, random forest, SVR or deep neural network can be selected for modeling; Step 3.2, model training: After selecting the machine learning algorithm, the next step is to train the model using the training data (Xtrain, ytrain); taking linear regression as an example, assuming that the least squares method is used to optimize the parameters θ of the linear model, its form is: f(X) = Xθ, where θ is the weight vector to be optimized; find the optimal θ by minimizing the mean square error loss function: By solving the gradient of the loss function with respect to θ The weights can be updated using gradient descent or other optimization algorithms until the loss function converges to a minimum; Step 3.3, hyperparameter tuning: Each machine learning algorithm has specific hyperparameters that directly affect the performance of the model. In random forests, important hyperparameters include the number of trees ntrees and the maximum tree depth max_depth. The goal of hyperparameter optimization is to determine the optimal parameter combination through experiments to improve the prediction accuracy of the model. Common hyperparameter tuning methods include grid search and random search. Assume that grid search is used to optimize parameters θ1, θ2, ... θk. By performing combination experiments within the candidate value range of each parameter, the parameter combination θ1^*, θ2^*, ... θk^* that makes the model perform best is finally selected. The goal of hyperparameter optimization is to find the optimal solution by minimizing the mean square error on the validation set: Where yi is the actual value, y^i is the predicted value, and yˉ is the mean of the target value; Step 3.4, model evaluation and validation: After model training and hyperparameter tuning are completed, the model needs to be evaluated using test data (Xtest, ytest); for this purpose, commonly used evaluation indicators include mean square error, root mean square error, and R^2 value; in addition, the generalization ability of the model can be further evaluated through cross-validation; the process of cross-validation is: divide the training data into K folds, use each fold as a validation set in turn, and use the other K-1 folds as training sets, repeat training and validation, and finally evaluate the model performance through the average validation error; the mean square error of cross-validation is defined as: Where yi(k) is the actual value in the kkkth fold, y^i(k) is the predicted value, and nk is the number of samples in the kth fold. Through cross-validation, we can effectively avoid overfitting of the model and ensure that the model can still maintain a high prediction accuracy on unseen data.

5. The method for calculating carbon emissions from biomass combustion based on artificial intelligence according to claim 1, characterized in that: Step 5 specifically includes: Step 5.1, model optimization method: Model optimization can be carried out from two aspects: algorithm optimization and data processing optimization; first, for algorithm optimization, regularization technology can be introduced to prevent overfitting; L2 regularization imposes constraints on model parameters by adding a regularization term λ||θ||^2 to the loss function: Among them, λ is a hyperparameter of regularization strength. By adjusting λ, we can balance the model fit and complexity and reduce the risk of overfitting. Secondly, data processing optimization includes the processing of outliers and the reselection of features. By using robust regression, we can reduce the impact of outliers on the model: Where φ is the Huber loss function, which uses a smoother treatment for outliers; Step 5.2, real-time optimization feedback: In order to achieve continuous improvement in carbon emission accounting, the system can also be dynamically optimized through a real-time feedback mechanism; the real-time feedback optimization mechanism is based on the error between the actual carbon emission value and the predicted value. Adaptive methods such as Kalman filtering or incremental learning are used for dynamic adjustment. Taking Kalman filtering as an example, its update formula is: Where K(t) is the Kalman gain, which is dynamically adjusted according to the noise characteristics of the system to ensure that the model's prediction at the next moment can more accurately reflect the real-time situation; Step 5.3, accounting result analysis and optimization strategy verification: After the carbon emission accounting is completed, the system will conduct a detailed result analysis to verify the effectiveness of the optimization strategy; Compared with the predicted value before optimization Compare and analyze the improvements brought by the optimization strategy: The change Δy(t) is tested for significance through statistical methods to determine whether the model optimization has brought about significant performance improvement. In addition, the time series characteristics of carbon emissions can be analyzed to find the key influencing factors of carbon emissions and further improve carbon emission control strategies.

6. An artificial intelligence-based biomass combustion carbon emission accounting system that implements the artificial intelligence-based biomass combustion carbon emission accounting method as described in any one of claims 1 to 5, characterized in that: include: Data acquisition module: First, through sensor networks and Internet of Things technology, multi-dimensional data related to the biomass combustion process is collected in real time. For the multi-dimensional data collection of the biomass combustion process, focus on the key characteristic parameters related to carbon emissions such as temperature, oxygen concentration, carbon dioxide concentration, humidity, etc.; Feature extraction and selection module: Utilize statistical analysis and machine learning methods to extract characteristic variables closely related to carbon emissions, including combustion efficiency, fuel consumption rate, and emission generation rate; use methods such as the Pearson correlation coefficient to screen important features to simplify the model and improve computational efficiency; Machine learning model building module: A neural network architecture combining LSTM and GRU is used to handle high-dimensional nonlinear problems of time series data to adapt to the dynamic changes in combustion conditions. The advantage of the bidirectional LSTM network is that it can capture the before-and-after dependencies of time series features, which is suitable for real-time prediction of carbon emissions. Real-time carbon emission accounting module: For biomass combustion, the emission of a certain type of air pollutant Ei(t) is calculated using the following formula: In the formula, A is the activity level of the emission source; EF is the emission factor; i is a certain type of air pollutant; j is the region; k is the type of biomass combustion; and m is the type of fuel / vegetation zone / grassland / straw. For biomass boilers, due to their relatively large scale, pollution control facilities such as dust collectors can be installed. In this case, the emission factor EF should be calculated by the following formula: EF = EF0 × (1-η) In the formula, EF0 is the pollutant generation coefficient; η is the removal efficiency of the pollution control facility; According to the aforementioned formula, the parameters that need to be determined to calculate the emissions of air pollutants are the activity level A of the emission source and the emission factor EF; Method optimization and accounting result analysis module: adopts multi-model integration and adaptive learning mechanism, continuously optimizes and adjusts the model by feeding back real-time data to ensure that it adapts to different combustion conditions and handles complex nonlinear relationships; conducts multi-dimensional analysis and sensitivity analysis on carbon emission accounting results, verifies the accuracy of the model by comparing with actual emission data, and realizes long-term monitoring and trend prediction through a dynamic feedback mechanism.

7. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the artificial intelligence-based biomass combustion carbon emission accounting method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the artificial intelligence-based biomass combustion carbon emission accounting method as described in any one of claims 1 to 5.

9. An information data processing terminal, comprising the artificial intelligence-based biomass combustion carbon emission accounting system as described in claim 6.

Images