Coal-fired unit carbon emission minute-level online monitoring method based on generating capacity
By constructing a basic model of carbon emissions based on power generation and the KNN method, and combining it with real-time data to correct carbon emissions, the problem of real-time monitoring of carbon emissions from coal-fired power units was solved, achieving minute-level online monitoring, reducing costs and improving monitoring accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- STATE GRID ZHEJIANG ELECTRIC POWER CO MARKETING SERVICE CENT
- Filing Date
- 2025-08-20
- Publication Date
- 2026-04-24
AI Technical Summary
Existing technologies cannot achieve real-time monitoring of carbon emissions from coal-fired power units and require additional hardware equipment, leading to increased costs and inconvenient installation. Furthermore, the accuracy is affected by factors such as flue gas turbulence and humidity.
By constructing a basic model of carbon emissions based on power generation, using the KNN method to obtain the deviation relationship of carbon emissions, and combining real-time coal-fired unit data to correct carbon emissions, minute-level online monitoring is achieved, avoiding additional hardware investment.
It enables real-time, high-precision monitoring of carbon emissions from coal-fired power units, adapts to unit load fluctuations and coal quality changes, reduces costs, and has good implementability and economic efficiency.
Smart Images
Figure CN121920642A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of carbon emission technology for coal-fired power generating units, and in particular to a minute-level online monitoring method for carbon emissions from coal-fired power generating units based on power generation. Background Technology
[0002] Carbon emissions are a general term or abbreviation for greenhouse gas emissions. As a key sector for carbon emissions, the power industry will bear greater responsibility for emission reduction, allowing the power sector to reach its peak emissions later in the future, thus supporting the entire society in reaching its peak emissions as early as possible.
[0003] Currently, power plant carbon emission calculation methods are divided into statistical values and real-time values. Statistical values mainly calculate the total carbon emissions over a period of time, and their calculation methods include emission factor method, material balance method, life cycle method, and input-output model method. Real-time values are obtained by collecting real-time data to acquire the unit's carbon emissions. Related technologies involve installing sensors in the boiler tail flue and using CEMS for carbon emission monitoring and measurement. However, statistical values can only calculate carbon emissions over a period of time and lack the characteristic of real-time acquisition, while real-time values require separate equipment, which increases costs and is inconvenient to install. Furthermore, their accuracy is affected by flue gas turbulence and humidity. Summary of the Invention
[0004] This application addresses the technical problems in existing technologies, such as the inability to obtain real-time carbon emissions from generating units and the need for additional hardware investment. It provides a minute-level online monitoring method for carbon emissions from coal-fired generating units based on power generation. This method utilizes the minute-level on-grid power generation data already available in the power system to achieve minute-level online monitoring of carbon emissions from fossil carbon source units. It does not require additional hardware investment and has good implementability and economic advantages.
[0005] To achieve the above technical objectives, this application provides a minute-level online monitoring method for carbon emissions from coal-fired power units based on power generation, comprising the following steps: constructing a basic carbon emission model based on power generation and carbon emissions from historical coal-fired power unit data; obtaining the deviation relationship of carbon emissions based on historical power generation and carbon emissions using the KNN method; correcting the basic carbon emission model based on the deviation relationship of carbon emissions to obtain a corrected carbon emission model; and outputting real-time corrected carbon emission values based on real-time coal-fired power unit data and the corrected carbon emission model.
[0006] Furthermore, the construction of a basic carbon emission model based on historical coal-fired power generation and carbon emission data includes: constructing a first coordinate value in the historical sample coordinates based on the power generation data of historical coal-fired power units; constructing a second coordinate value in the historical sample coordinates based on the carbon emission data of historical coal-fired power units; constructing historical sample coordinates by matching the first and second coordinate values according to time series; and constructing a basic carbon emission model based on the historical sample coordinates using a regression algorithm.
[0007] Furthermore, the basic carbon emission model constructed based on the regression algorithm and historical sample coordinates includes: constructing a basic carbon emission model based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units.
[0008] Furthermore, a basic model for carbon emissions is constructed based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired units. This includes: constructing the relationship between coal consumption per kilowatt-hour and load rate based on historical coal consumption per kilowatt-hour and load rate curves of coal-fired units with the same rated capacity using the least squares method; obtaining a load rate correction coefficient based on the relationship between coal consumption per kilowatt-hour and load rate and coal consumption per kilowatt-hour; and obtaining CO2 emissions per kilowatt-hour based on the load rate correction coefficient, coal consumption per kilowatt-hour, and carbon content.
[0009] Furthermore, the basic model for carbon emissions based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired units also includes obtaining carbon content based on independent variable coefficients, volatile matter content on the received basis, lower heating value, fixed carbon content on the received basis, and ash content on the received basis.
[0010] Furthermore, the basic model for carbon emissions based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired units also includes: obtaining the carbon oxidation rate of coal-fired units based on the annual slag volume, average carbon content of slag, annual fly ash production, average carbon content of fly ash, average dust removal efficiency of dust removal system, coal consumption, average lower heating value of coal, and carbon content per unit calorific value of coal.
[0011] Furthermore, obtaining the deviation relationship of carbon emissions based on historical power generation and historical carbon emissions using the KNN method includes: obtaining a similar set of power generation based on the Euclidean distance of historical power generation using the KNN method; obtaining a carbon emission correction value based on the carbon emissions in the similar set of power generation; and constructing the deviation relationship of carbon emissions based on the carbon emission correction value and historical power generation.
[0012] Furthermore, obtaining a similar set of power generation based on the Euclidean distance of historical power generation using the KNN method includes: calculating the similarity between each historical power generation based on the Euclidean distance; and for each historical power generation, selecting a preset number of similar sets of power generation based on the similarity ranking.
[0013] Furthermore, the carbon emission correction model is obtained by modifying the basic carbon emission model based on the deviation relationship of carbon emissions. This includes: using a fitting method to fit the coordinates in the deviation relationship of carbon emissions into a curve to obtain the carbon emission correction model.
[0014] Furthermore, the real-time carbon emission correction value output based on real-time coal-fired unit data and the carbon emission correction model includes: constructing the on-grid power generation of coal-fired units based on time-series matching of real-time coal-fired unit data; and constructing the real-time carbon emission correction value based on the carbon emission correction model and the on-grid power generation of coal-fired units.
[0015] This application relates to a minute-level online monitoring method for carbon emissions from coal-fired power plants based on power generation. It utilizes minute-level on-grid electricity data from power plants already available in the power system to achieve minute-level online monitoring of carbon emissions from fossil fuel-based power plants. By combining historical carbon emission data, a dynamic regression model is constructed to achieve real-time, high-precision monitoring of carbon emissions. No additional hardware investment is required, offering significant advantages in implementability and economic efficiency. It is particularly suitable for the refined management and emission reduction control of carbon emissions in large-scale coal-fired power plants. Attached Figure Description
[0016] Figure 1 This is a schematic diagram of historical values of a coal-fired power unit under one embodiment of this application.
[0017] Figure 2 For the purposes of this application Figure 1 The diagram shows a power-carbon emission curve obtained from the application-based model in one embodiment.
[0018] Figure 3 For the purposes of this application Figure 1 The diagram shows a power-carbon emission curve obtained from the application optimization model in one embodiment.
[0019] Figure 4 For the purposes of this application Figure 1 The diagram shows a minute-level power generation, power data, and calculated and measured carbon emission values for a coal-fired unit A in one embodiment.
[0020] Figure 5 For the purposes of this application Figure 1 The diagram shows a comparison between the measured and calculated values of real-time carbon emission monitoring for a coal-fired unit A in one embodiment.
[0021] Figure 6 For the purposes of this application Figure 1 The diagram shows a carbon emission intensity curve of a coal-fired unit A in one embodiment. Detailed Implementation
[0022] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0023] like Figure 1-6 As shown, a method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation includes: A basic model of carbon emissions is constructed based on historical data of power generation and carbon emissions from coal-fired power units; Based on the KNN method, the deviation relationship of carbon emissions is obtained from historical power generation and carbon emissions; A carbon emission correction model is obtained by correcting the basic carbon emission model based on the deviation relationship of carbon emissions. The real-time carbon emission correction value is output based on real-time coal-fired unit data and carbon emission correction model.
[0024] In this embodiment, a basic carbon emission model is constructed by combining historical coal-fired unit data on power generation and carbon emissions. A deviation relationship in carbon emissions is obtained based on historical power generation and carbon emissions using the KNN method, leading to a carbon emission correction model. Real-time carbon emission correction values are then output based on real-time coal-fired unit data and the carbon emission correction model. This improves the accuracy of online real-time carbon emission monitoring to the minute level, significantly outperforming traditional hourly monitoring methods based on thermal calculations. The real-time carbon emission correction values output from the real-time coal-fired unit data and the carbon emission correction model automatically capture the impact of dynamic factors such as unit load fluctuations and coal quality changes on carbon emissions, avoiding the systematic biases caused by fixed emission factors in traditional methods. Simultaneously, the model can be gradually optimized to adapt to long-term trends such as unit aging and changes in combustion efficiency.
[0025] The basic carbon emission model constructed based on historical coal-fired power unit data includes: constructing the first coordinate value in the historical sample coordinate system based on the power generation data; constructing the second coordinate value in the historical sample coordinate system based on the carbon emissions data; and constructing the historical sample coordinate system (X) by matching the first and second coordinate values according to time sequence. i Y i ), where X i Y represents the power generation from historical coal-fired power unit data. i The carbon emissions are based on historical coal-fired power unit data; a basic model of carbon emissions is constructed based on the coordinates of historical samples using a regression algorithm.
[0026] In this embodiment, the actual amount of carbon involved in combustion is calculated by the carbon content of coal and the carbon oxidation rate, reflecting the transformation process of carbon elements from fuel to emissions. A load rate correction coefficient is introduced, and the load rate correction coefficient is fitted by historical coal-fired unit data to obtain the coal consumption deviation under different load conditions, ensuring the reliability of carbon emission calculations across the entire load range.
[0027] Specifically, a correlation function between load factor and coal consumption per kilowatt-hour is constructed, and the output of the correlation function is used as the input for calculating CO2 emissions per kilowatt-hour. The function parameters are solved by the least squares method to minimize the sum of squared errors between the fitted curve and historical coal-fired unit data, thus obtaining a basic carbon emission model that includes the quantitative relationship between coal consumption per kilowatt-hour and load factor.
[0028] Specifically, a basic carbon emission model is constructed based on historical sample coordinates using a regression algorithm. This basic carbon emission model is then used as the current regression curve, given a specific electricity generation X. i The corresponding carbon emissions y are calculated based on the real-time carbon emission model of coal-fired power units. i Specifically, the current regression curve construction process involves selecting all historical coal-fired power unit data, obtaining the power generation and carbon emissions, and filling them into historical sample coordinates to form a complete set of historical sample coordinates. Then, based on a regression algorithm, this set is used to calculate a regression curve that reflects the relationship between power generation and carbon emissions. Thus, given a power generation X... i The corresponding carbon emissions y can then be calculated. i Furthermore, it requires no additional hardware, reducing costs and offering strong adaptability. Adaptability can be achieved simply by adjusting the historical sample library, avoiding the complexity of traditional methods that require remodeling for each type of generator set.
[0029] The carbon emission baseline model, constructed based on historical sample coordinates using a regression algorithm, includes the following: a baseline model is built using historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and the carbon oxidation rate of coal-fired power units. This includes creating a real-time carbon emission calculation formula for coal-fired power units, as shown in Formula 1. (Formula) (1); where E C P represents real-time total carbon emissions; e represents real-time electricity generation; c1 CO2 emissions per kilowatt-hour.
[0030] like Figure 2 As shown, the basic model for carbon emissions based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units also includes creating a formula for calculating CO2 emissions per kilowatt-hour, as shown in Formula 2: (Formula) (2); where, M is the molar mass of carbon dioxide; Cη is the molar mass of carbon; C is the carbon content; ms is the coal consumption per kilowatt-hour, generally selected from the latest statistical data of the power plant operating at high load rate (load rate above 86%) in the current month; η l As the load factor correction coefficient, based on the historical coal consumption per kilowatt-hour and load factor curves of coal-fired power units with the same rated capacity, the least squares method is used to fit the data to obtain the relationship between coal consumption per kilowatt-hour and load factor.
[0031] like Figure 2 As shown, the basic model for carbon emissions based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units also includes creating a formula for calculating carbon content, as shown in Formula 3: (Formula) (3); where, It is a constant. , , and For the independent variable coefficients, please refer to Table 1 for details; Var is the volatile matter content on the received basis; Qnet,ar is the lower heating value; FCar is the fixed carbon content on the received basis; Aar is the ash content on the received basis.
[0032] Table 1. Different coal types , , , and Value
[0033] The basic carbon emission model, built based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units, also includes creating a formula for calculating the carbon oxidation rate of coal-fired power units, as shown in Formula 4: (Formula) (4); where Gz is the annual slag quantity; Cz is the average carbon content of the slag; Ga is the annual fly ash production; Ca is the average carbon content of the fly ash; The average dust removal efficiency of the dust removal system; FC coal This refers to the consumption of coal; NCV coal The lower heating value of coal; CC coal This refers to the carbon content per unit calorific value of coal.
[0034] like Figure 3 As shown, obtaining the deviation relationship of carbon emissions based on historical power generation and carbon emissions using the KNN method includes: obtaining a similar set of power generation based on the Euclidean distance of historical power generation using the KNN method; obtaining a carbon emission correction value based on the carbon emissions in the similar set of power generation; and constructing the deviation relationship of carbon emissions based on the carbon emission correction value and historical power generation.
[0035] like Figure 3As shown, obtaining the deviation relationship of carbon emissions based on historical power generation and historical carbon emissions using the KNN method also includes selecting a test point X on the current regression curve. i ;Analyze the test point X i The test point X is obtained. i Corresponding sample point Y i ; Calculate the test point X respectively i to sample point Y i The Euclidean distance; based on each sample point Y i Distance from test point X i The Euclidean distance is used to select the K nearest target sample points, which are then sorted in descending order. K is a positive integer, and its value needs to be determined through experiments or cross-validation based on the quantity and distribution characteristics of historical sample data to ensure the representativeness of similar samples. In a specific embodiment, K≤20; the Y-coordinate of the selected K nearest target sample points is calculated. i average Return to select a test point X on the current regression curve. i Until each test point X i Each was selected once; repeating the operation yielded a series of averages. This leads to a deviation in carbon emission levels.
[0036] like Figure 3 As shown, constructing a similarity set of power generation based on Euclidean distance and power generation quantity includes calculating the similarity between various historical power generation quantities based on Euclidean distance; for each historical power generation quantity, selecting a preset number of similarity sets of power generation quantities based on similarity ranking. Specifically, a formula for calculating Euclidean distance is created, as shown in Formula 5: (Formula) (5); where D(X) i Y i ) represents the Euclidean distance between the sample point and the test point; K represents the number of nearest neighbor samples selected by the KNN method.
[0037] like Figure 3 As shown, obtaining a carbon emission correction model based on the deviation relationship of carbon emissions from the basic carbon emission model includes: fitting a curve to the coordinates in the deviation relationship of carbon emissions using a fitting method to obtain the carbon emission correction model. The fitting method includes, but is not limited to, least squares method, weighted least squares method, etc. Alternatively, in other embodiments, the carbon emission correction model is constructed based on the current regression curve in the basic carbon emission model based on the deviation relationship of carbon emissions from the basic carbon emission model, using a judgment-based approach; constructing the carbon emission correction model based on the current regression curve in the basic carbon emission model based on the deviation relationship of carbon emissions from the basic carbon emission model using a judgment-based logic includes judging the coordinate Y in the deviation relationship of carbon emissions from the basic carbon emission model. i average If the carbon emissions deviate from the coordinate Y in the relationship i average With the sample points Y in the current regression curve i If they coincide, then select the sample point Y from the regression curve. i If the coordinate Y in the deviation relationship of carbon emissions i average With the sample points Y in the current regression curve i If they do not overlap, then select the coordinate Y in the deviation relationship of carbon emissions. i average .
[0038] The real-time carbon emission correction value is output based on real-time coal-fired unit data and a carbon emission correction model, including: constructing the on-grid power generation of coal-fired units based on time-series matching of real-time coal-fired unit data; and constructing the real-time carbon emission correction value based on the carbon emission correction model and the on-grid power generation of coal-fired units. For example, if the power generation from coal-fired unit data with an update frequency of N minutes is selected, where N is a positive number and 1≤N≤15; the carbon emissions from the coal-fired unit data with an update frequency of N minutes are calculated; and minute-level real-time coal-fired unit data is obtained.
[0039] like Figure 4 As shown, in a specific embodiment, a coal-fired unit A equipped with a real-time carbon emission monitoring device was selected for carbon emission model verification. The unit has a rated capacity of 1000 MWe, and its coal consumption per kilowatt-hour at high load rate was calculated to have an average CO2 emission of 0.822 per kilowatt-hour. It uses wet desulfurization and has no CCUS device in the plant.
[0040] like Figure 5-6 As shown, coal samples were taken from the plant and subjected to industrial analysis and testing. The carbon content of the coal was measured to be 80.5%, and the oxidation rate was calculated as 99%. The real-time carbon emissions of Unit A were calculated using the calculation formula of the basic carbon emission model, and the results were compared with the measured values after the installation of real-time carbon emission monitoring.
[0041] The mean absolute error (MAE) is used to characterize the closeness between the calculated result and the measurement result. The smaller the value, the better the fit. Its calculation is shown in Formula 6: (Formula) (6); The carbon emission difference at each time point was calculated using the evaluation method described above. The calculated MAE value of the formula is 3.87, while the calculated MAE value of the formula (1-5) using the basic carbon emission model is 3.00. In comparison, the calculated MAE value in this application is reduced by 0.87, showing a significant optimization effect. It enables real-time and high-precision monitoring of carbon emissions and requires no additional investment in hardware equipment, thus possessing good implementability and economic advantages.
[0042] The technical features of the above embodiments can be combined arbitrarily, and the execution order of the method steps is not restricted. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0043] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, characterized in that, Includes the following steps: A basic model of carbon emissions is constructed based on historical data of power generation and carbon emissions from coal-fired power units; Based on the KNN method, the deviation relationship of carbon emissions is obtained from historical power generation and carbon emissions; A carbon emission correction model is obtained by correcting the basic carbon emission model based on the deviation relationship of carbon emissions. The real-time carbon emission correction value is output based on real-time coal-fired unit data and carbon emission correction model.
2. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 1, is characterized in that... The basic carbon emission model constructed based on historical coal-fired power generation and carbon emission data includes: The first coordinate value in the historical sample coordinates is constructed based on the power generation data of historical coal-fired units. The second coordinate value in the historical sample coordinate system is constructed based on the carbon emission data of historical coal-fired power units. Based on the temporal matching of the first and second coordinate values, construct historical sample coordinates; A basic model of carbon emissions is constructed based on historical sample coordinates using a regression algorithm.
3. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 2, is characterized in that... The carbon emission baseline model constructed based on historical sample coordinates using the regression algorithm includes: A basic model of carbon emissions is constructed based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units.
4. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 3, is characterized in that... The basic carbon emission model constructed based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units includes: Based on the least squares method, the relationship between the change of coal consumption per kilowatt-hour and load rate is constructed according to the historical coal consumption per kilowatt-hour and load rate curves of coal-fired power units with the same rated capacity. Based on the relationship between coal consumption per kilowatt-hour and load factor, and using coal consumption per kilowatt-hour to obtain load factor correction coefficient; The CO2 emissions per kilowatt-hour are obtained based on the load factor correction factor, coal consumption per kilowatt-hour, and carbon content.
5. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 4, is characterized in that... The basic carbon emission model constructed based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units also includes: Carbon content is obtained based on independent variable coefficients, volatile matter content of the received basis, lower heating value, fixed carbon content of the received basis, and ash content of the received basis.
6. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 5, is characterized in that... The basic carbon emission model, constructed based on historical sample coordinates, CO2 emissions per kilowatt-hour, carbon content, and carbon oxidation rate of coal-fired power units, also includes: The carbon oxidation rate of coal-fired power units is obtained based on the annual slag volume, average carbon content of slag, annual fly ash production, average carbon content of fly ash, average dust removal efficiency of dust removal system, coal consumption, average lower heating value of coal, and carbon content per unit calorific value of coal.
7. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 6, is characterized in that... The deviation relationship of carbon emissions obtained based on historical power generation and historical carbon emissions using the KNN method includes: Based on the KNN method, a set of similar power generation is obtained according to the Euclidean distance of historical power generation. Carbon emission correction values are obtained based on carbon emissions from a similar set of electricity generation. The deviation relationship of carbon emissions is constructed based on the revised carbon emission values and historical power generation.
8. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 7, is characterized in that... The method of obtaining a similar set of power generation based on the Euclidean distance of historical power generation using the KNN method includes: The similarity between historical power generation is calculated based on Euclidean distance. For each historical power generation, a set of power generation similarities with a preset number of power generation values is selected based on similarity ranking.
9. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 7, is characterized in that... The carbon emission correction model obtained by correcting the basic carbon emission model based on the deviation relationship of carbon emissions includes: The carbon emission correction model is obtained by fitting the coordinates in the deviation relationship of carbon emissions to a curve using the fitting method.
10. The method for minute-level online monitoring of carbon emissions from coal-fired power units based on power generation, as described in claim 9, is characterized in that... The step of outputting real-time carbon emission correction values based on real-time coal-fired unit data and a carbon emission correction model includes: Based on time-series matching of real-time coal-fired power unit data, the on-grid power generation of coal-fired power units is constructed. Real-time carbon emission correction values are constructed based on the carbon emission correction model and the on-grid power generation of coal-fired units.