CN117422167A - Electric power carbon emission predictive analysis method based on tree model - Google Patents
Electric power carbon emission predictive analysis method based on tree model Download PDFInfo
- Publication number
- CN117422167A CN117422167A CN202311217093.9A CN202311217093A CN117422167A CN 117422167 A CN117422167 A CN 117422167A CN 202311217093 A CN202311217093 A CN 202311217093A CN 117422167 A CN117422167 A CN 117422167A
- Authority
- CN
- China
- Prior art keywords
- model
- carbon emission
- data
- prediction
- carbon
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 title claims abstract description 121
- 229910052799 carbon Inorganic materials 0.000 title claims abstract description 121
- 238000004458 analytical method Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 64
- 238000011156 evaluation Methods 0.000 claims abstract description 31
- 238000005457 optimization Methods 0.000 claims abstract description 19
- 238000012216 screening Methods 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 11
- 238000004140 cleaning Methods 0.000 claims abstract description 6
- 239000003245 coal Substances 0.000 claims description 26
- 238000012549 training Methods 0.000 claims description 16
- 238000012795 verification Methods 0.000 claims description 12
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 claims description 9
- 229910052717 sulfur Inorganic materials 0.000 claims description 9
- 239000011593 sulfur Substances 0.000 claims description 9
- 238000003066 decision tree Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000007637 random forest analysis Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 claims description 6
- 229910052739 hydrogen Inorganic materials 0.000 claims description 6
- 239000001257 hydrogen Substances 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000000513 principal component analysis Methods 0.000 claims description 4
- 238000012800 visualization Methods 0.000 claims description 4
- 230000001364 causal effect Effects 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000001816 cooling Methods 0.000 claims description 3
- 238000010219 correlation analysis Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 claims description 3
- 230000006872 improvement Effects 0.000 abstract description 17
- 230000008859 change Effects 0.000 abstract description 4
- 230000010485 coping Effects 0.000 abstract description 2
- 238000009472 formulation Methods 0.000 abstract description 2
- 239000000203 mixture Substances 0.000 abstract description 2
- 238000007726 management method Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000010438 heat treatment Methods 0.000 description 5
- 230000010354 integration Effects 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000007405 data analysis Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 239000005431 greenhouse gas Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Abstract
The invention provides a tree model-based electric power carbon emission predictive analysis method, which comprises the following steps of: accessing, cleaning and preprocessing the original data to obtain preprocessed data; screening out characteristic variables with the carbon emission prediction correlation higher than a correlation threshold based on the pretreatment data and the carbon emission accounting methodology, and eliminating redundant or irrelevant characteristics; constructing a carbon emission prediction and optimization model by combining a carbon emission accounting methodology; adopting a tree model structure, analyzing the model to identify key influencing factors; comparing the actual carbon emission data with the predicted result to obtain an evaluation result; and adjusting model parameters and characteristic engineering according to the evaluation result. The method has the advantages of realizing remarkable improvement in the aspects of carbon emission prediction accuracy, key influence factor identification capability, targeted carbon emission efficiency improvement strategy formulation, model interpretability and the like. This will help provide a more accurate, efficient and intelligent solution for carbon emissions management, contributing to coping with global climate change and reducing carbon emissions.
Description
Technical Field
The invention relates to the field of electric power carbon emission, in particular to a tree model-based electric power carbon emission prediction analysis method.
Background
With the increasing global climate change problem, the reduction of greenhouse gas emissions and the improvement of carbon emission efficiency have become the common concern of various countries. In order to better manage carbon emissions and achieve low carbon development, carbon emission quota systems have been implemented in many countries, encouraging businesses to reduce carbon emissions and to implement carbon asset management. In this context, building a carbon asset intelligent analysis model based on data is significant in assisting enterprises in formulating carbon emission improvement strategies, predicting carbon emissions and carbon quotas, and identifying key influencing factors.
In the prior art, carbon emission and carbon quota management are mainly performed through data statistics and manual analysis, and although some effects are achieved by the methods, the following problems and disadvantages still exist: the data processing efficiency is low: the traditional carbon emission data analysis method mainly relies on manual data statistics and processing, which has low efficiency, and is easy to cause errors when processing a large amount of data, thereby influencing the accuracy of results. Model accuracy is limited: most of the existing carbon emission prediction models adopt simple statistical methods such as linear regression, the accuracy of the methods is low when the nonlinear problems are processed, and the actual requirements of carbon emission prediction and management are difficult to meet. Lack of targeted improvement strategies: the existing carbon emission management method is often remained on a macroscopic level, lacks a targeted analysis and improvement strategy for specific emission units, and limits the improvement of carbon emission efficiency. It is difficult to identify key influencing factors: the existing carbon emission data analysis method only pays attention to a few influencing factors, ignores other possible key influencing factors, and causes poor prediction and management effects.
Disclosure of Invention
In view of the above, the present invention has been made to provide a tree model-based predictive analysis method of electric carbon emissions that overcomes or at least partially solves the above-mentioned problems.
According to an aspect of the present invention, there is provided a tree model-based predictive analysis method of electric carbon emission, the predictive analysis method including:
accessing, cleaning and preprocessing the original data to obtain preprocessed data;
screening out characteristic variables with the carbon emission prediction correlation higher than a correlation threshold based on the pretreatment data and the carbon emission accounting methodology, and eliminating redundant or irrelevant characteristics;
constructing a carbon emission prediction and optimization model by combining a carbon emission accounting methodology;
adopting a tree model structure, analyzing the model to identify key influencing factors;
comparing the actual carbon emission data with the predicted result to obtain an evaluation result;
and adjusting model parameters and characteristic engineering according to the evaluation result.
Optionally, the preprocessing of the raw data specifically includes:
screening: according to actual requirements, screening out key data related to carbon emission prediction, and eliminating irrelevant data information;
missing value processing: filling missing values in the data by adopting an interpolation method, a Lagrange interpolation method or other proper methods so as to reduce the influence of the missing values on the model;
outlier processing: abnormal values in the data are detected and processed through a box diagram and 3 sigma principle method, so that the data are more stable and reliable;
data type conversion: the different types of data are uniformly converted into a format suitable for model training and prediction.
Optionally, the screening out feature variables with the carbon emission prediction correlation higher than the correlation threshold based on the pretreatment data and the carbon emission accounting methodology, and removing redundant or irrelevant features specifically includes:
based on the preprocessing data and the carbon emission accounting methodology, adopting a correlation analysis, principal component analysis and maximum information coefficient method to screen out characteristic variables with the correlation higher than a correlation threshold value with the carbon emission prediction, and eliminating redundant or irrelevant characteristics.
Optionally, the characteristic variables specifically include:
the control variables comprise unit numbers, months, unit types, installed capacity, pressure parameters/unit types and steam turbine exhaust cooling modes;
the factor variables comprise low heat value of the coal to be charged, moisture of the air-dried base of the coal to be charged, moisture of the base received by the coal to be charged, sulfur content of the base received by the coal to be charged, carbon content of the element of the integrated sample air-dried base, hydrogen content of the element of the integrated sample air-dried base, moisture of the integrated sample air-dried base, high heat value of the integrated sample air-dried base, sulfur content of the integrated sample air-dried base, low heat value of the base received by the integrated sample and carbon content of the element of the base received by the coal to be charged.
Optionally, the constructing the carbon emission prediction and optimization model specifically includes:
model selection: according to the data characteristics and the problem demands, comparing the advantages and disadvantages of different models, and selecting a proper model to predict the carbon emission; optional models include decision trees, random forests, XGBoost;
model training: training the selected model using the preprocessed data set. Adopting a cross verification or other verification method to adjust model parameters to obtain optimal model performance;
model evaluation: selecting a proper evaluation index according to the prediction target;
model optimization: optimizing the model according to the model evaluation result;
model output: and outputting a prediction result, and taking the related evaluation index as a basis for evaluating the performance of the model.
Optionally, the optimizing the model specifically includes: model parameters, feature selection and feature construction are adjusted.
Optionally, the tree model structure is adopted, and the analysis of the model to identify key influencing factors specifically includes:
visualization tree model structure: the selected tree model is visually displayed, so that splitting conditions and weights of various variables in the model and the roles of the variables in the decision process can be intuitively observed;
evaluation of variable importance: comparing their contribution to model predictions by calculating the importance scores of the variables in the model;
identifying key influencing factors: selecting key influencing factors according to the variable importance evaluation result;
interpretation of results: and analyzing the key influence factors, acquiring the relation between the key influence factors and the predicted targets, and setting forth causal logic and potential mechanisms.
The invention provides a tree model-based electric power carbon emission predictive analysis method, which comprises the following steps of: accessing, cleaning and preprocessing the original data to obtain preprocessed data; screening out characteristic variables with the carbon emission prediction correlation higher than a correlation threshold based on the pretreatment data and the carbon emission accounting methodology, and eliminating redundant or irrelevant characteristics; constructing a carbon emission prediction and optimization model by combining a carbon emission accounting methodology; adopting a tree model structure, analyzing the model to identify key influencing factors; comparing the actual carbon emission data with the predicted result to obtain an evaluation result; and adjusting model parameters and characteristic engineering according to the evaluation result. The method has the advantages of realizing remarkable improvement in the aspects of carbon emission prediction accuracy, key influence factor identification capability, targeted carbon emission efficiency improvement strategy formulation, model interpretability and the like. This will help provide a more accurate, efficient and intelligent solution for carbon emissions management, contributing to coping with global climate change and reducing carbon emissions.
The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a tree model-based electric power carbon emission prediction analysis method according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The terms "comprising" and "having" and any variations thereof in the description embodiments of the invention and in the claims and drawings are intended to cover a non-exclusive inclusion, such as a series of steps or elements.
The technical scheme of the invention is further described in detail below with reference to the accompanying drawings and the examples.
Example 1
As shown in fig. 1, the invention provides a method for predicting and optimizing carbon emission of a power plant based on tree model integration, which mainly solves the following technical problems: improving the accuracy of carbon emission prediction, optimizing the identification capability of key influencing factors, formulating a targeted carbon emission efficiency improvement strategy and improving the interpretability of a model. The specific technical scheme is as follows:
1. data preprocessing: first, the raw data is accessed, cleaned and preprocessed to ensure data quality and integrity. The data preprocessing includes the following operations: (1) screening: and screening out key data related to carbon emission prediction according to actual requirements, and eliminating irrelevant data information. (2) missing value processing: and filling missing values in the data by adopting an interpolation method, a Lagrange interpolation method or other proper methods so as to reduce the influence of the missing values on the model. (3) outlier handling: abnormal values in the data are detected and processed through the box diagram, the 3 sigma principle and other methods, so that the data are more stable and reliable. (4) data type conversion: the different types of data are uniformly converted into a format suitable for model training and prediction.
2. Characteristic engineering: based on the existing data and carbon emission accounting methodology, feature variables with higher correlation with carbon emission prediction are screened out through methods such as correlation analysis, principal Component Analysis (PCA), maximum information coefficient and the like, and redundant or irrelevant features are removed. The control variables include unit number, month, unit type, installed capacity (MW), pressure parameter/unit type
The factor variables comprise low heat value of the coal entering the furnace, air-dry basis moisture of the coal entering the furnace, moisture of the base received by the coal entering the furnace, carbon content of elements of the air-dry basis of the comprehensive sample, hydrogen content of elements of the air-dry basis of the comprehensive sample, moisture of the air-dry basis of the comprehensive sample, high heat value of the air-dry basis of the comprehensive sample, moisture of the sulfur of the air-dry basis of the comprehensive sample, low heat value of the base received by the comprehensive sample, carbon content of the elements of the base received by the coal entering the furnace and the like.
3. Model construction: and combining a carbon emission accounting methodology, and adopting a tree model to integrate algorithms such as decision trees, random forests, XGBoost and the like to construct a carbon emission prediction and optimization model. The detailed steps of model construction are as follows: (1) model selection: and according to the data characteristics and the problem demands, comparing the advantages and disadvantages of different models, and selecting a proper model to predict the carbon emission. Alternative models include decision trees, random forests, XGBoost, and the like. (2) model training: training the selected model using the preprocessed data set. The model parameters are adjusted to obtain optimal model performance using cross-validation or other validation methods. (3) model evaluation: depending on the predicted target, suitable evaluation indices such as R2, MAE (mean absolute error), RMSE (root mean square error), etc. are selected. The prediction targets can be variables such as carbon emission, carbon quota quantity, carbon asset gap and the like, and can also be built by self. In the evaluation process, the model needs to meet the standard of fitting accuracy R2>0.8 so as to ensure the model prediction accuracy. (4) model optimization: and optimizing the model according to the model evaluation result. The optimization method comprises the steps of adjusting model parameters, feature selection, feature construction and the like so as to improve the model performance. (5) model output: and outputting a prediction result, and taking relevant evaluation indexes (R2, MAE, RMSE and the like) as the basis of model performance evaluation.
4. Key influencing factor identification and interpretation: by employing a tree model structure, the model is analyzed to identify key influencing factors. The tree model structure has the characteristics of stronger visualization, and can clearly show the effect and importance of each variable in the model. By analyzing the significance and interpretation of each factor, key influencing factors are effectively identified, and enterprises and government departments are facilitated to understand model prediction results and logic behind the model prediction results more deeply. The detailed steps are as follows: (1) visualizing a tree model structure: the selected tree model is visually displayed, so that the splitting conditions and weights of all variables in the model and the roles of the variables in the decision process can be visually observed. (2) assessing variable importance: by calculating the importance scores of the individual variables in the model, their contributions to the model predictions are compared. The importance score may be calculated by an attribute of the model itself or by a feature importance algorithm. (3) identifying key influencing factors: and selecting variables with higher scores as key influencing factors according to the variable importance evaluation results. These key factors play a central role in predicting carbon emissions and should be focused and studied. (4) interpretation of results: the key influencing factors are deeply analyzed, the relation between the key influencing factors and the predicted targets is explored, and causal logic and potential mechanisms are explained. In addition, an additional analysis of the impact of non-critical factors may be performed to obtain a more comprehensive understanding.
5. Model verification and optimization: and (3) evaluating the prediction accuracy and reliability of the model by comparing the actual carbon emission data with the prediction result. And according to the evaluation result, adjusting model parameters and characteristic engineering to continuously optimize the model performance. The optimized model is deployed into a carbon emission management system of the power plant so as to realize real-time monitoring and prediction of carbon emission, and a more accurate, efficient and intelligent carbon emission management solution is provided for the power plant. And continuously improving the model according to the actual application condition, wherein the continuous improvement comprises the steps of adjusting a prediction strategy, optimizing a key influence factor identification method, updating a carbon emission efficiency improvement strategy and the like so as to improve the actual application value of the model.
Example 2
Aiming at the case that a certain power enterprise group has 240 coal-fired generator sets, the invention provides a concrete implementation mode of a power plant carbon emission prediction and optimization method based on tree model integration. The following is a detailed implementation:
data cleaning: the raw data is first cleaned and preprocessed. The total of 2980 original data is obtained, and 45 covered field variables are obtained, wherein the included field variables comprise unit type, fuel coal consumption, coal-fired test data and the like. In the data cleaning process, the missing value is processed, the abnormal value is removed, and the data type is converted. Thus, the quality and the integrity of the data can be ensured, and more accurate data can be provided for subsequent feature engineering and model construction.
Feature variable selection: the potential influencing factors of the variables are screened based on the existing data and the carbon emission accounting methodology. The selected characteristic variables comprise unit number, month, unit type, installed capacity (MW), pressure parameter/unit type, steam turbine exhaust cooling mode, low-level calorific value of coal-to-be-fired, air-to-be-fired dry base moisture, base moisture received by coal-to-be-fired, base sulfur received by coal-to-be-fired, and the comprehensive sample air-dry basis element carbon content, the comprehensive sample air-dry basis element hydrogen content, the comprehensive sample air-dry basis moisture, the comprehensive sample air-dry basis high-order heating value, the comprehensive sample air-dry basis sulfur content, the comprehensive sample received basis low-order heating value, the furnace coal received basis element carbon content and the like. These characteristic variables will be the control and factor variables of the model input.
Model construction: and combining a carbon emission accounting methodology, and adopting a tree model to integrate algorithms such as decision trees, random forests, XGBoost and the like to construct a carbon emission prediction and optimization model. In the embodiment, the input data comprises low-level calorific value of the coal entering the furnace, air-dry basis moisture of the coal entering the furnace, moisture of the base received by the coal entering the furnace, carbon content of elements of the air-dry basis of the comprehensive sample, hydrogen content of elements of the air-dry basis of the comprehensive sample, moisture of the air-dry basis of the comprehensive sample, high-level calorific value of the air-dry basis of the comprehensive sample, moisture of the air-dry basis of the comprehensive sample, low-level calorific value of the base received by the comprehensive sample and the like; the predicted result is the base element carbon content received by the coal being charged. The prediction precision requirement reaches more than 0.8.
Model training: (1) data set partitioning: the data set is divided into training and testing sets in time order, typically 70% of the data is selected for model training and 30% of the data is used for model testing. And (2) model training and parameter adjustment: the method of tree model integration is adopted, the algorithms of decision trees, random forests, XGBoost and the like are combined for model training, training set data are used for model fitting, and parameter tuning is carried out, so that the prediction accuracy and stability of the model are improved. (3) model verification and evaluation: and (3) performing model verification by using the test set data, and evaluating the prediction accuracy and reliability of the model. And (3) adopting indexes such as fitting precision R2, MAE (mean absolute error), RMSE (root mean square error) and the like to evaluate the performance of the model, and improving and optimizing the model. According to the model prediction result, the evaluation indexes are as follows: r2 (R-square) =0.851, which indicates that the model can interpret the 85.1% of the total variance of the data variation, indicating that the model has higher fitting accuracy for the prediction of the carbon content of the base element received by the coal being charged. MAE (mean absolute error) =0.015, which means that the mean difference between the model predicted value and the true value is 0.015, and the error is small. RMSE (root mean square error) =0.003, which means that the average deviation of the model predicted value from the true value is 0.003, and the error is smaller. In a comprehensive view, the evaluation index shows that the model has strong prediction capability on the carbon content of the base element received by the coal entering the furnace, has high accuracy and stability, and can be used as an important reference basis for accounting and optimizing carbon emission by enterprises and government departments.
The model may explain the analysis: and analyzing the feature importance of the prediction result by adopting a Shapley feature importance analysis method. Shapley is a method of measuring the importance of features to a predicted outcome, and may represent the contribution of each feature to the predicted outcome. By calculating the shape value of each feature value, the relative importance ranking of different features in the prediction result can be obtained. The analysis results are shown in the following table. It can be seen that the comprehensive sample receives the basic low-level heating value, the coal-fired receives the basic moisture and the coal-fired low-level heating value are three characteristic variables with the greatest influence on the prediction result, and the Shapley values of the three characteristic variables are 0.059, 0.052 and 0.042 respectively, which are consistent with the selection of the characteristic variables. These results can help enterprises and government departments to better understand the model prediction results and logic behind the model prediction results, and help to purposefully formulate carbon emission efficiency improvement strategies, and improve the accuracy and efficiency of carbon emission management of power plants.
Model verification and optimization: the model is verified and optimized during the implementation process to ensure its accuracy and reliability. The verification phase may be performed by the following steps: verifying by using data of other months to verify the generalization capability of the model in different time periods; internal verification is carried out on the model by adopting methods such as cross verification and the like so as to evaluate the stability and generalization capability of the model; and comparing the error of the model prediction result with the actual value, and evaluating the prediction accuracy of the model. In the verification stage, adjustment and optimization are required according to specific conditions. The optimization method may include the following aspects: feature selection: and eliminating characteristic variables with small influence on the prediction result, and reducing the complexity and noise interference of the model. Parameter adjustment: and adjusting parameters in the model to obtain a better prediction effect. Such as maximum depth in the decision tree, number of trees in the random forest, etc. Model integration: and combining prediction results of a plurality of models to perform model fusion so as to improve the accuracy and stability of prediction. Data enhancement: the data enhancement technology is utilized to increase the training data quantity and improve the generalization capability and the prediction accuracy of the model. In the optimization process, the above methods need to be selected and combined according to specific situations to obtain the best prediction result. Meanwhile, attention is required to be paid to the change of the evaluation index, and the optimization method is adjusted in time so as to achieve the expected prediction effect. Table 1 predicts the results.
| Characteristic variable | Shapley value | Importance ranking |
| The comprehensive sample receives the basic low-order heating value | 0.059 | 1 |
| The coal fed into the furnace receives the basic water | 0.052 | 2 |
| Low heat value of coal being fed into furnace | 0.042 | 3 |
| High-order calorific value of comprehensive sample air-dry basis | 0.028 | 4 |
| The carbon content of the base element is received by the coal fed into the furnace | 0.019 | 5 |
| The comprehensive sample receives basic sulfur | 0.016 | 6 |
| Comprehensive sample air-dry basis elemental carbon content | 0.012 | 7 |
| Comprehensive sample air-dry basis element hydrogen content | 0.009 | 8 |
| Comprehensive sample air dry basis moisture | 0.008 | 9 |
| The coal fed into the furnace receives the sulfur content | 0.006 | 10 |
The beneficial effects are that: prediction accuracy improves: by adopting an integrated tree model algorithm, such as decision trees, random forests, XGBoost and the like, the method has higher accuracy in the aspect of carbon emission prediction. Compared with the traditional linear regression or neural network model, the tree model has stronger nonlinear fitting capability and robustness, and can better capture complex relations and potential rules in data. Experimental results show that the fitting accuracy R2 of the invention reaches more than 0.8, which is obviously superior to the prior art.
Key influencing factor recognition capability enhancement: the invention adopts the tree model structure to analyze the significance and interpretation of each factor, and more effectively identifies key influencing factors. This helps businesses and government agencies better understand the model predictions and the logic behind them to develop more targeted carbon emission management strategies.
Model interpretability improves: the tree model structure has the characteristics of stronger visualization, and can clearly show the effect and importance of each variable in the model. This allows enterprises and government authorities to more easily understand the predictive outcome of the model and the relationships of various factors in the decision process, improving the interpretability of the model.
Energy consumption and resources are saved: through the real-time monitoring and prediction of the carbon emission, enterprises can more accurately know the carbon emission condition and formulate corresponding carbon emission reduction strategies. This will help to reduce energy consumption, reduce raw material waste, improve production efficiency to realize more environmental protection and sustainable development.
Model continuous improvement: the present invention provides a mechanism for continuous improvement, including adjusting prediction strategies, optimizing key impact factor identification methods, updating carbon emission efficiency improvement strategies, and the like. The model can be optimized according to the actual application situation, and the actual application value of the model is improved.
The foregoing detailed description of the invention has been presented for purposes of illustration and description, and it should be understood that the invention is not limited to the particular embodiments disclosed, but is intended to cover all modifications, equivalents, alternatives, and improvements within the spirit and principles of the invention.
Claims (7)
1. A tree model-based predictive analysis method for electric carbon emissions, the predictive analysis method comprising:
accessing, cleaning and preprocessing the original data to obtain preprocessed data;
screening out characteristic variables with the carbon emission prediction correlation higher than a correlation threshold based on the pretreatment data and the carbon emission accounting methodology, and eliminating redundant or irrelevant characteristics;
constructing a carbon emission prediction and optimization model by combining a carbon emission accounting methodology;
adopting a tree model structure, analyzing the model to identify key influencing factors;
comparing the actual carbon emission data with the predicted result to obtain an evaluation result;
and adjusting model parameters and characteristic engineering according to the evaluation result.
2. The method for predicting and analyzing the carbon emission of the electric power based on the tree model as claimed in claim 1, wherein the preprocessing of the raw data specifically comprises:
screening: according to actual requirements, screening out key data related to carbon emission prediction, and eliminating irrelevant data information;
missing value processing: filling missing values in the data by adopting an interpolation method, a Lagrange interpolation method or other proper methods so as to reduce the influence of the missing values on the model;
outlier processing: abnormal values in the data are detected and processed through a box diagram and 3 sigma principle method, so that the data are more stable and reliable;
data type conversion: the different types of data are uniformly converted into a format suitable for model training and prediction.
3. The method for predicting and analyzing the carbon emission of the electric power based on the tree model according to claim 1, wherein the step of screening out the feature variables with the carbon emission prediction correlation higher than the correlation threshold based on the preprocessing data and the carbon emission accounting methodology, and the step of eliminating the redundant or irrelevant features specifically comprises the following steps:
based on the preprocessing data and the carbon emission accounting methodology, adopting a correlation analysis, principal component analysis and maximum information coefficient method to screen out characteristic variables with the correlation higher than a correlation threshold value with the carbon emission prediction, and eliminating redundant or irrelevant characteristics.
4. The tree model-based electric carbon emission predictive analysis method according to claim 1, wherein the characteristic variables specifically include:
the control variables comprise unit numbers, months, unit types, installed capacity, pressure parameters/unit types and steam turbine exhaust cooling modes;
the factor variables comprise low heat value of the coal to be charged, moisture of the air-dried base of the coal to be charged, moisture of the base received by the coal to be charged, sulfur content of the base received by the coal to be charged, carbon content of the element of the integrated sample air-dried base, hydrogen content of the element of the integrated sample air-dried base, moisture of the integrated sample air-dried base, high heat value of the integrated sample air-dried base, sulfur content of the integrated sample air-dried base, low heat value of the base received by the integrated sample and carbon content of the element of the base received by the coal to be charged.
5. The method for predicting and analyzing carbon emissions in electric power based on tree model according to claim 1, wherein the method for constructing a carbon emission prediction and optimization model specifically comprises:
model selection: according to the data characteristics and the problem demands, comparing the advantages and disadvantages of different models, and selecting a proper model to predict the carbon emission; optional models include decision trees, random forests, XGBoost;
model training: training the selected model using the preprocessed data set. Adopting a cross verification or other verification method to adjust model parameters to obtain optimal model performance;
model evaluation: selecting a proper evaluation index according to the prediction target;
model optimization: optimizing the model according to the model evaluation result;
model output: and outputting a prediction result, and taking the related evaluation index as a basis for evaluating the performance of the model.
6. The method for predicting and analyzing the carbon emission of electric power based on the tree model according to claim 5, wherein the optimizing the model specifically comprises: model parameters, feature selection and feature construction are adjusted.
7. The method for predicting and analyzing carbon emissions in electric power based on tree model as claimed in claim 1, wherein said analyzing the model to identify key influencing factors comprises:
visualization tree model structure: the selected tree model is visually displayed, so that splitting conditions and weights of various variables in the model and the roles of the variables in the decision process can be intuitively observed;
evaluation of variable importance: comparing their contribution to model predictions by calculating the importance scores of the variables in the model;
identifying key influencing factors: selecting key influencing factors according to the variable importance evaluation result;
interpretation of results: and analyzing the key influence factors, acquiring the relation between the key influence factors and the predicted targets, and setting forth causal logic and potential mechanisms.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311217093.9A CN117422167A (en) | 2023-09-20 | 2023-09-20 | Electric power carbon emission predictive analysis method based on tree model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311217093.9A CN117422167A (en) | 2023-09-20 | 2023-09-20 | Electric power carbon emission predictive analysis method based on tree model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117422167A true CN117422167A (en) | 2024-01-19 |
Family
ID=89527377
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311217093.9A Pending CN117422167A (en) | 2023-09-20 | 2023-09-20 | Electric power carbon emission predictive analysis method based on tree model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117422167A (en) |
-
2023
- 2023-09-20 CN CN202311217093.9A patent/CN117422167A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111144609A (en) | Boiler exhaust emission prediction model establishing method, prediction method and device | |
| CN113205207A (en) | XGboost algorithm-based short-term power consumption load fluctuation prediction method and system | |
| CN109116444B (en) | PCA-kNN-based air quality model PM2.5Forecasting method | |
| Bangalore et al. | An approach for self evolving neural network based algorithm for fault prognosis in wind turbine | |
| CN113051744A (en) | Evaluation system and evaluation method for regional atmospheric pollution regulation and control scheme | |
| CN113298288A (en) | Power supply station operation and maintenance cost prediction method integrating time sequence and neural network | |
| CN114021483A (en) | Ultra-short-term wind power prediction method based on time domain characteristics and XGboost | |
| CN111695666A (en) | Wind power ultra-short term conditional probability prediction method based on deep learning | |
| Tan et al. | A methodology for augmenting sparse pairwise comparison matrices in AHP: applications to energy systems | |
| CN113408659A (en) | Building energy consumption integrated analysis method based on data mining | |
| CN115905319B (en) | Automatic identification method and system for abnormal electricity fees of massive users | |
| CN117422167A (en) | Electric power carbon emission predictive analysis method based on tree model | |
| CN110852496A (en) | Natural gas load prediction method based on LSTM recurrent neural network | |
| CN112801388B (en) | Power load prediction method and system based on nonlinear time series algorithm | |
| CN114970353A (en) | MSWI process dioxin emission soft measurement method based on missing data filling | |
| Schreiber et al. | Quantifying the influences on probabilistic wind power forecasts | |
| Gamst et al. | Time aggregation techniques applied to a capacity expansion model for real-life sector coupled energy systems | |
| CN110852544B (en) | Reliability evaluation method and device for wind generating set | |
| Raymand et al. | Machine learning-based estimation of buildings' characteristics employing electrical and chilled water consumption data: Pipeline optimization | |
| Zhao et al. | An interpretable ultra-short-term wind power prediction model based on the feature matrix reconstruction through regression trees | |
| Friederich et al. | A Framework for Validating Data-Driven Discrete-Event Simulation Models of Cyber-Physical Production Systems | |
| Gujjarlapudi et al. | Data driven machine learning models for short‐term load forecasting considering electrical vehicle load | |
| CN116720662B (en) | Distributed energy system applicability evaluation method based on set pair analysis | |
| Folorunso et al. | A predictive model for estimating petroleum consumption using machine learning approach | |
| Ardiansyah et al. | Potential classification prediction of solar and wind energy in Indonesia using machine learning with random forest algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |