Electric power carbon emission reduction prediction method, system and equipment based on carbon beneficiation mechanism

Abstract

The invention discloses an electric power carbon emission reduction prediction method, system and device based on a carbon beneficiation mechanism. The method comprises the following steps: firstly, obtaining a multi-dimensional dynamic data set of a target area in a preset certain time period; based on the multi-dimensional dynamic data set, obtaining a load structure data sequence, a dynamic reference carbon emission and a dynamic carbon popularity behavior influence factor sequence of the target area in a preset time period; and finally, inputting the load structure data sequence, the dynamic reference carbon emission and the dynamic carbon popularity behavior influence factor sequence of the target area in a preset time period into a carbon emission reduction prediction model as carbon reduction characteristics to obtain a carbon emission reduction prediction result in a next time period. The electric power carbon emission reduction prediction method realizes electric power carbon emission reduction prediction with comprehensive and accurate data, scientific and reasonable reference, quantitative dynamic behavior influence and accurate and reliable prediction model, has relatively high application value, can provide powerful support for implementation of a carbon benefit mechanism and scientific management of a power grid, and promotes low-carbon development of the electric power industry.

Description

Technical Field

[0001] This invention belongs to the field of data processing and analysis technology, specifically relating to a method, system, and equipment for predicting carbon emission reductions in the power sector based on a carbon inclusive mechanism. Background Technology

[0002] Carbon credit mechanisms, as a market-based means of incentivizing public participation in emissions reduction, have demonstrated enormous potential in the power industry. By quantifying, recording, and incentivizing users' emissions-reducing behaviors such as energy conservation and the use of green electricity, carbon credit mechanisms effectively stimulate widespread emissions reduction motivation. In the power industry, accurately predicting the carbon emissions reductions resulting from carbon credit activities is crucial for scientifically assessing emissions reduction effectiveness, optimizing carbon credit reward strategies, and supporting low-carbon grid dispatch.

[0003] The effective operation of carbon incentive mechanisms relies on accurate prediction of carbon emission reductions in the power sector. However, existing methods for predicting power sector carbon emission reductions suffer from two core flaws: First, the treatment of carbon incentive behaviors is too static and aggregated. Existing technologies typically treat user behavior as isolated, instantaneous events, leading to either an overestimation or underestimation of the behavior's contribution to emission reductions. Second, the application of grid carbon emission factors is too homogeneous and coarse. Existing technologies generally use the regional grid average carbon emission factor for calculations, failing to reflect the significant differences in carbon emission intensity under different operating conditions of the grid in real time. In reality, during peak load periods, the grid often utilizes marginal units with high carbon emissions, where the emission reduction benefits per unit of electricity saved are far greater than during off-peak load periods. Ignoring this difference makes it impossible to accurately assess the true emission reduction contribution of different behaviors under different operating scenarios, thus affecting prediction accuracy and weakening the scientific validity of incentive strategies developed based on this.

[0004] Therefore, there is an urgent need for a new method for predicting carbon emission reductions in the power sector, which can effectively overcome the limitations of existing technologies and more accurately predict the carbon emission reduction effects under the carbon inclusive mechanism. Summary of the Invention

[0005] The purpose of this invention is to address the aforementioned problems in the existing technology by providing a method, system, and device for predicting carbon emission reductions in the power sector based on a carbon inclusive mechanism.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] In a first aspect, this invention proposes a method for predicting carbon emission reductions in the power sector based on a carbon inclusive mechanism, including:

[0008] S1. Obtain a multidimensional dynamic dataset of the target area within a preset time period. The multidimensional dynamic dataset includes time-series data of electricity consumption, time-series data of carbon inclusive behavior, and dynamic carbon emission factor data.

[0009] S2. Based on the multidimensional dynamic dataset, obtain the load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence for the target area over a preset time period.

[0010] S3. Input the load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence of the target area for a preset time period as carbon reduction features into the carbon emission reduction prediction model to obtain the carbon emission reduction prediction results for the next time period.

[0011] S2 includes:

[0012] The total electricity consumption at each collection time point in the power consumption time series data is decomposed into base load electricity, peak load electricity and valley load electricity. The base load ratio, peak load ratio and valley load ratio are calculated. The base load ratio, peak load ratio and valley load ratio at each collection time point are arranged and combined according to time series to form a load structure data sequence.

[0013] Using time-series electricity consumption data and dynamic carbon emission factor data as inputs, a dynamic baseline carbon emission is determined through a load grading calculation model.

[0014] Based on time-series data of carbon inclusive behavior, a dynamic sequence of influencing factors of carbon inclusive behavior is obtained through a time-varying impact model of carbon inclusive behavior.

[0015] The load classification calculation model is as follows:

[0016] ;

[0017] In the above formula, This represents the dynamic baseline carbon emissions for the target region within a preset time period. The length of the preset time period, , , At the time points of collection Base load power, peak load power, and valley load power. , , At the time points of collection The base load carbon emission factor, peak load carbon emission factor, and valley load carbon emission factor.

[0018] The carbon inclusive time-varying impact model is as follows:

[0019] ;

[0020] ;

[0021] In the above formula, To collect data at the time point The dynamic carbon inclusive behavior impact factors are calculated, and the dynamic carbon inclusive behavior impact factors at each data collection time point are arranged in time series to form a dynamic carbon inclusive behavior impact factor sequence. , The length of the preset time period, To collect data at the time point The total number of carbon-inclusive events within the country. To collect data at the time point A set of timestamps of all carbon-inclusive events. For behavioral events timestamp instantaneous intensity, It is an exponentially decaying function. For behavioral events Preset attenuation coefficient, For behavioral events The preset baseline emission reduction coefficient, For behavioral events At the time of collection The load type coupling coefficient, Load type, including base load, peak load, and valley load. For behavioral events In load type The average occurrence intensity on, For load type The average carbon emission factor, This represents the average carbon emission factor.

[0022] Secondly, this invention proposes a power carbon emission reduction prediction system based on a carbon inclusive mechanism, including a multi-dimensional dynamic dataset acquisition module, a dataset processing module, and a carbon emission reduction prediction module.

[0023] The multidimensional dynamic dataset acquisition module is used to acquire a multidimensional dynamic dataset of the target area within a preset time period. The multidimensional dynamic dataset includes electricity consumption time series data, carbon inclusive behavior time series data, and dynamic carbon emission factor data.

[0024] The dataset processing module is used to obtain a load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence for a target area over a preset time period based on a multidimensional dynamic dataset.

[0025] The carbon emission reduction prediction module is used to input the load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence of the target area for a preset time period as carbon reduction features into the carbon emission reduction prediction model to obtain the carbon emission reduction prediction result for the next time period.

[0026] The dataset processing module includes a load structure data sequence formation unit, a dynamic baseline carbon emission determination unit, and a dynamic carbon inclusive behavior influencing factor sequence acquisition unit.

[0027] The load structure data sequence forming unit is used to decompose the total electricity consumption at each collection time point in the power consumption time series data into base load electricity, peak load electricity and valley load electricity, calculate the base load ratio, peak load ratio and valley load ratio, and arrange and combine the base load ratio, peak load ratio and valley load ratio at each collection time point according to the time sequence to form a load structure data sequence.

[0028] The dynamic baseline carbon emission determination unit is used to take electricity consumption time-series data and dynamic carbon emission factor data as inputs and determine the dynamic baseline carbon emission through a load grading calculation model.

[0029] The unit for obtaining the dynamic carbon inclusive behavior impact factor sequence is used to obtain the dynamic carbon inclusive behavior impact factor sequence based on the time series data of carbon inclusive behavior and through the carbon inclusive time-varying impact model.

[0030] The load classification calculation model is as follows:

[0031] ;

[0032] In the above formula, This represents the dynamic baseline carbon emissions for the target region within a preset time period. The length of the preset time period, , , At the time points of collection Base load power, peak load power, and valley load power. , , At the time points of collection The base load carbon emission factor, peak load carbon emission factor, and valley load carbon emission factor.

[0033] The carbon inclusive time-varying impact model is as follows:

[0034] ;

[0035] ;

[0036] In the above formula, To collect data at the time point The dynamic carbon inclusive behavior impact factors are calculated, and the dynamic carbon inclusive behavior impact factors at each data collection time point are arranged in time series to form a dynamic carbon inclusive behavior impact factor sequence. , The length of the preset time period, To collect data at the time point The total number of carbon-inclusive events within the country. To collect data at the time point A set of timestamps of all carbon-inclusive events. For behavioral events timestamp instantaneous intensity, It is an exponentially decaying function. For behavioral events Preset attenuation coefficient, For behavioral events The preset baseline emission reduction coefficient, For behavioral events At the time of collection The load type coupling coefficient, Load type, including base load, peak load, and valley load. For behavioral events In load type The average occurrence intensity on, For load type The average carbon emission factor, This represents the average carbon emission factor.

[0037] Thirdly, the present invention proposes a power carbon emission reduction prediction device based on a carbon inclusive mechanism, including a processor and a memory;

[0038] The memory is used to store computer program code and to transmit the computer program code to the processor;

[0039] The processor is used to execute the aforementioned method for predicting electricity carbon emission reductions based on the carbon inclusive mechanism, according to the instructions in the computer program code.

[0040] Fourthly, the present invention provides a computer storage medium on which a computer program is stored;

[0041] When the computer program is executed by the processor, it implements the steps of the aforementioned method for predicting carbon emission reductions in electricity based on the carbon inclusive mechanism.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] This invention proposes a method, system, and device for predicting carbon emission reductions in the power sector based on a carbon inclusive mechanism. The method first acquires a multidimensional dynamic dataset of a target region over a predetermined time period. This dataset includes time-series data on electricity consumption, time-series data on carbon inclusive behavior, and dynamic carbon emission factor data. Then, based on the multidimensional dynamic dataset, it obtains a load structure data sequence, a dynamic baseline carbon emission, and a sequence of dynamic carbon inclusive behavior influencing factors for the target region over the predetermined time period. Finally, it uses these data as carbon reduction features and inputs them into a carbon emission reduction prediction model to obtain the predicted carbon emission reduction for the next time period. On the one hand, this method establishes a high-fidelity load classification calculation model closely linked to the real-time operation of the power grid, based on time-series data of electricity consumption and dynamic carbon emission factor data, to determine a dynamic baseline carbon emission. Compared with using a single average factor, this baseline carbon emission can more accurately capture the greater emission reduction benefits brought about by power saving during peak load periods with high carbon emission intensity, providing a scientific reference for accurately quantifying the true value of carbon inclusive behavior. On the other hand, based on time-series data of carbon inclusive behavior, this method introduces behavior effect decay and load type coupling coefficient to construct a time-varying impact model of carbon inclusive behavior, obtaining a dynamic carbon inclusive behavior impact factor sequence. Behavior effect decay simulates the time-dependent characteristics of the emission reduction effect of carbon inclusive behavior through an exponential decay function. The load type coupling coefficient is used to quantify the correlation between the occurrence time of carbon inclusive behavior and the carbon emission intensity of different load states of the power grid, amplifying the emission reduction value of behavior occurring during high carbon intensity periods and reducing the emission reduction value of behavior occurring during low carbon intensity periods, making the quantification of carbon inclusive behavior more accurate and reasonable. Attached Figure Description

[0044] Figure 1 This is an overall flowchart of the method described in this invention.

[0045] Figure 2 This is a structural diagram of the system described in this invention.

[0046] Figure 3 This is a structural diagram of the device described in this invention. Detailed Implementation

[0047] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0048] This invention proposes a method, system, and equipment for predicting electricity carbon emission reductions based on a carbon inclusive mechanism. It acquires a multi-dimensional dynamic dataset of a target area over a preset time period, including time-series data of electricity consumption, time-series data of carbon inclusive behavior, and dynamic carbon emission factor data. Based on the time-series data of electricity consumption, it decomposes it into base load, peak load, and valley load, outputting a load structure data sequence. Based on the time-series data of electricity consumption and the dynamic carbon emission factor data, it determines the dynamic baseline carbon emission through a load grading calculation model. Based on the time-series data of carbon inclusive behavior, it obtains the dynamic carbon inclusive behavior influence factor sequence through a time-varying carbon inclusive model. The load structure data sequence, dynamic baseline carbon emission, and dynamic carbon inclusive behavior influence factor sequence of the preset time period are used as carbon reduction features and input into the carbon emission reduction prediction model to obtain the carbon emission reduction prediction result for the next time period. Therefore, it achieves comprehensive and accurate data, a scientifically sound baseline, dynamically quantified behavioral influences, and an accurate and reliable prediction model for predicting electricity carbon emission reductions. It has high application value and can provide strong support for the implementation of the carbon inclusive mechanism and the scientific management of the power grid, promoting the low-carbon development of the power industry.

[0049] Example 1:

[0050] like Figure 1 As shown, the method for predicting carbon emission reductions in the power sector based on the carbon inclusive mechanism proceeds in the following steps:

[0051] 1. Obtain a multidimensional dynamic dataset of the target region within a preset time period;

[0052] The target area can be a city or an industrial park, dynamically defined according to specific needs. The sampling frequency (time unit) for the preset time period can be set to hours or minutes, or customized based on actual conditions and specific requirements. The preset time period is divided into a series of continuous, fixed-length time units based on the sampling time points. Each time unit corresponds to the time interval between the current sampling time point and the previous sampling time point. The length of the time unit can be set according to data collection and prediction needs, such as 15 minutes or 1 hour. For each time unit within the preset time period, a multidimensional dynamic dataset is collected at each time interval, meaning each sampling time point corresponds to one multidimensional dynamic dataset.

[0053] Data acquisition for the multidimensional dynamic dataset is achieved through smart meter systems, user self-reporting platforms, third-party carbon credit platforms, and public databases. Data cleaning and standardization processes are then performed to ensure data integrity and consistency.

[0054] The multidimensional dynamic dataset includes time-series data on electricity consumption, time-series data on carbon credit incentive behavior, and dynamic carbon emission factor data, comprehensively reflecting the electricity use and carbon emissions in the target area within a preset time period. Specifically, the time-series data on electricity consumption is defined as the total electricity consumption of the target area at each collection point within the preset time period; the time-series data on carbon credit incentive behavior is defined as the behavior type, timestamp, duration, and instantaneous intensity of each carbon credit incentive event within the preset time period; and the dynamic carbon emission factor data is defined as the baseload carbon emission factor corresponding to each collection point within the preset time period. Peak load carbon emission factor Valley Carbon Emission Factor You can choose the carbon emission factor data corresponding to the end of the time unit, or you can choose the average value of the carbon emission factor within the time unit.

[0055] For example, the carbon inclusion behavior time series data includes records of carbon inclusion behaviors initiated by users through APP, smart home devices, etc. Each record includes: the type of carbon inclusion behavior, the timestamp of the behavior, the duration of the behavior, and the instantaneous intensity of the behavior, such as the real-time power when energy-saving equipment is turned on, and the rated power of the air conditioner when it is turned off is 2.5kW. The types of carbon inclusion behaviors include, but are not limited to, energy-saving equipment usage behavior, green electricity purchase behavior, and demand response participation behavior, which refer to the behavior of users reducing carbon emissions through energy saving, using green electricity, etc.

[0056] For example, dynamic carbon emission factor data includes real-time carbon emission factors corresponding to base load, peak load, and valley load, obtained from grid operators and synchronized with electricity consumption time-series data. , , The unit is: These factors reflect the marginal carbon emission intensity of grid-connected power supply under different load conditions. Generally speaking, .

[0057] 2. Based on the multidimensional dynamic dataset, the load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence of the target area for a preset time period are obtained;

[0058] The total electricity consumption at each collection time point in the electricity consumption time series data Decomposed into base charge (Minimum daily load, i.e., the minimum load required to maintain basic grid operation), peak load power (Peak daily load, i.e., electricity consumption during the period of maximum daily load), Off-peak load electricity consumption (Remaining load, i.e., electricity consumption during the remaining time period), calculate the base load ratio at each data collection time point. Peak load ratio Valley-Herb Ratio Furthermore, the base load ratio, peak load ratio, and valley load ratio at each collection time point are arranged and combined according to time sequence to form a load structure data sequence, which reflects the power grid operation structure, making the analysis of electricity consumption data more accurate and providing a solid foundation for subsequent calculations.

[0059] The process of decomposing the total electricity consumption at each collection time point in the electricity consumption time series data into base load electricity, peak load electricity, and valley load electricity can be obtained based on the threshold method of statistical percentiles of the load curve. The specific steps include:

[0060] First, prepare the data: collect daily load curve data of the target area over the past year or a typical cycle (such as four seasons), with the time resolution consistent with the collection time point, such as 96 points / day;

[0061] Then determine the load threshold:

[0062] Base load threshold Calculate the 5th percentile value of all historical load data points. This means that 95% of the load data points are above this value. Therefore, this load level can be approximately considered as the "base load" that the power grid needs to maintain.

[0063] Peak load threshold Calculate the 95th percentile value of all historical load data points. This means that only 5% of the load points are above this value, so this load level can be approximately considered to represent the "peak load" of the power grid.

[0064] Finally, decompose by time interval: Let the total load value for each time unit be... ,like Then all the electricity generated within that time unit is classified as off-peak electricity; if Then all the electricity within that time unit is classified as base load; if Then, all the electricity generated within that time unit is classified as peak load electricity.

[0065] The total electricity consumption at each collection time point in the time series data can be decomposed into base load electricity, peak load electricity, and valley load electricity. This can also be obtained based on pattern recognition methods using cluster analysis. The specific steps include:

[0066] First, data preparation is carried out: historical daily load curves are collected, and unsupervised machine learning clustering algorithms (such as K-Means clustering) are used to classify all daily load curves according to their shape characteristics, such as typical patterns like "peak day", "flat day", and "valley day".

[0067] Then define the load type:

[0068] The typical load level in the "off-peak day" pattern is defined as off-peak load;

[0069] The typical load level in the "flat day" pattern is defined as the base load;

[0070] The typical load level in the "peak day" pattern is defined as peak load.

[0071] Finally, the electricity is decomposed: for each time unit, according to the load mode day to which it belongs, its load value is assigned to the corresponding load type, and the electricity is divided into the corresponding base load electricity, peak load electricity, and valley load electricity.

[0072] It should be noted that the total electricity consumption can be decomposed into base load, peak load, and valley load. One method can be selected based on the actual situation, or other load decomposition methods can be selected by referring to relevant existing technologies.

[0073] Using time-series electricity consumption data and dynamic carbon emission factor data as inputs, a dynamic baseline carbon emission is determined through a load grading calculation model.

[0074] The load classification calculation model is as follows:

[0075] ;

[0076] In the above formula, This represents the dynamic baseline carbon emissions for the target area within a preset time period, indicating the carbon emissions the power grid would have generated under actual operating conditions without carbon credit initiatives. The length of the preset time period, , , At the time points of collection The base load, peak load, and valley load (kWh) are used to accurately describe the structure of electricity consumption. , , At the time points of collection The base load carbon emission factor, peak load carbon emission factor, and valley load carbon emission factor are used to reflect the marginal carbon emission intensity of the power grid under different load conditions, making the benchmark calculation more consistent with the physical reality of the power grid.

[0077] Thus, a high-fidelity dynamic benchmark carbon emission was established, which is closely linked to the real-time operation of the power grid. Compared with the use of a single average factor, this benchmark can keenly capture the greater emission reduction benefits brought about by energy saving during peak load periods with high carbon emission intensity. Through a load classification calculation model, combined with the electricity consumption and carbon emission factors under different load conditions, the impact of electricity consumption structure on carbon emissions is accurately described, providing a scientific reference system for accurately quantifying the true value of carbon inclusive behavior.

[0078] Based on time-series data of carbon inclusive behavior, dynamic carbon inclusive behavior impact factors are obtained through a time-varying impact model. The dynamic carbon inclusive behavior impact factors at each data collection time point are arranged in time series to form a dynamic carbon inclusive behavior impact factor sequence. ;

[0079] The carbon inclusive time-varying impact model is as follows:

[0080] ;

[0081] In the above formula, To collect data at the time point The dynamic carbon inclusive behavior influencing factor characterizes all historical behaviors at a given time point. The remaining emission reduction impact (after considering decay and scenario coupling, at time point) (net carbon emission contribution) The length of the preset time period, To collect data at the time point The total number of carbon-inclusive events within the country. To collect data at the time point A set of timestamps of all carbon-inclusive events. For behavioral events timestamp The instantaneous intensity is used to record the original force of a behavioral event. The function is an exponential decay function, simulating the carbon inclusive behavior effect from the time of occurrence. At the appointed time The decay process transforms static behavioral data into dynamic time series. For behavioral events The preset attenuation coefficient, obtained by fitting historical data (or determined based on expert experience), quantifies the rate at which the emission reduction effect of the behavior decays over time (e.g., the energy-saving effect diminishes as the room temperature rises after the lights are turned off). It is key to capturing the dynamics of the behavior. For behavioral events The preset baseline emission reduction factor, which can be determined through experiments or historical statistical data, defines the unit emission reduction capacity of this behavior under ideal conditions. For behavioral events At the time of collection Load type coupling coefficient;

[0082] in, This represents a behavior effect decay model, for behavioral events. From the time of occurrence At the appointed time Integrate the instantaneous intensity and multiply by an exponential decay term. ; Based on the relationship between the historical occurrence period of the behavior and the carbon emission factor, the following calculations were made: Analyzing which load periods this type of behavior mainly occurred in the past, the emission reduction value of behavior occurring during high carbon intensity periods was amplified, while the emission reduction value of behavior occurring during low carbon intensity periods was reduced.

[0083] The carbon emission factor is obtained by comparing the weighted average carbon emission factor for different load types with the system average carbon emission factor. The calculation formula is as follows:

[0084] ;

[0085] In the above formula, Load type, including base load, peak load, and valley load; For behavioral events In load type The average occurrence intensity, i.e., the weighted average carbon emission factor for different load types, For load type The average carbon emission factor, The system's average carbon emission factor;

[0086] The acquisition method is as follows: within a preset history window (the preset history window is much longer than the time period), the behavior events are... In all those belonging to " "The intensity of occurrence of the load type within the time unit is averaged;"

[0087] The acquisition method is as follows: within the preset history window, all data marked as " "The carbon emission factors of the time units of the load type are averaged;

[0088] The method for obtaining this information is as follows: within a preset window, the ratio of total carbon emissions to total electricity consumption for all time units.

[0089] molecular The calculation is for: behavioral events The average intensity over various historical load periods is the weighted sum of the average intensity and the typical carbon intensity for the corresponding periods. This reflects the "actual" carbon emission context faced by this type of behavior; the denominator is... The calculation is: if the behavioral event The occurrence of these events is uniformly distributed, independent of load type, and is subject to the expected average carbon emission background. Therefore, Its function is to measure behavioral events Whether the activity primarily occurs during periods of high or low carbon intensity loads, i.e., whether the value of a unit of emission reduction effort made by a certain type of behavior, due to its inherent timing preference, is higher than (>1) or lower than (<1) the social average (=1), depends on whether the behavior occurs more frequently during peak load periods. A value greater than 1 amplifies its emission reduction contribution, while a value less than 1 diminishes it.

[0090] For example, It is a dimensionless amplification or reduction factor. When it is greater than 1, it indicates that such behavioral events have historically occurred more often during periods of high carbon intensity, and their emission reduction value is amplified; when it is less than 1, it indicates that such behavioral events have occurred more often during periods of low carbon intensity, and their emission reduction value is reduced. This describes the frequency and intensity of such behavioral events historically occurring under a certain load type (such as peak load periods). It is a historical statistic, not real-time data. Over a sufficiently long historical period (set as a historical window), this type of behavioral event is considered among all events belonging to the "..." category. "The intensity of load types occurring within a time interval is averaged. For example, for the behavior of 'adjusting office air conditioning,' the value may be high because this behavior often occurs during the peak electricity consumption period on weekday afternoons." Typical carbon emission intensity representing a certain load type (such as peak load) is a characteristic value based on historical data, rather than real-time values. In the historical data, for all values ​​marked as " "Calculate the average carbon emission factor for time units of load type." As a benchmark, representing the overall average carbon emissions level of the entire power grid during the same historical period, it is obtained by dividing the total carbon emissions of all time intervals by the total electricity consumption in historical data, or by directly dividing the total carbon emissions of all time intervals. Find the average.

[0091] Therefore, by introducing a behavior effect decay model and a load type coupling coefficient, the impact of carbon inclusive behavior can be dynamically quantified. The behavior effect decay model simulates the time-dependent characteristics of the emission reduction effect of carbon inclusive behavior through an exponential decay function, taking into account the decay of the behavior effect over time, and accurately quantifies the dynamic change process of the impact of carbon inclusive behavior, fundamentally avoiding the prediction bias caused by static processing. The load type coupling coefficient is used to quantify the correlation between the time period of carbon inclusive behavior and the carbon emission intensity of different load states of the power grid, amplifying the emission reduction value of behavior occurring during high carbon intensity periods and reducing the emission reduction value of behavior occurring during low carbon intensity periods, making the quantification of carbon inclusive behavior more accurate and reasonable. Through load segmentation and coupling mechanism, a precise correlation between carbon inclusive behavior and the real-time carbon emission intensity of the power grid is established, realizing differentiated assessment of emission reduction value under different operating scenarios.

[0092] By deeply integrating dynamic behavior modeling with load segmentation mechanisms, a technological leap from "trend judgment" to "accurate mechanism simulation" has been achieved. Its core value lies in establishing a precise correlation between carbon inclusive behavior and grid carbon emission characteristics, solving the long-standing problems of accuracy and practicality in carbon emission reduction prediction, and providing solid technical support for the effective implementation of the carbon inclusive mechanism.

[0093] 3. Input the load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence of the target area for a preset time period as carbon reduction features into the carbon emission reduction prediction model to obtain the carbon emission reduction prediction results for the next time period.

[0094] The carbon emission reduction prediction model is obtained as follows:

[0095] Collect a large amount of multidimensional dynamic datasets of historical time periods in the target area, obtain the carbon reduction features of each historical time period, and label each historical carbon reduction feature as the target ground value for model training.

[0096] The label for each historical carbon reduction feature is set as follows: the actual carbon emission reduction record value for the next time period corresponding to the historical time period. If the actual carbon emission reduction record value for the next time period does not exist, it can be calculated using the following formula:

[0097] ;

[0098] In the above formula, Time period The actual carbon emission reductions, i.e., the historical period. Labeling of carbon reduction characteristics Historical period Dynamic benchmark carbon emissions, Historical period Actual observed carbon emissions;

[0099] The historical carbon reduction features after labeling are used as the training sample set. The carbon emission reduction prediction model is trained using a pre-selected neural network structure. The model parameters are continuously optimized to obtain the final carbon emission reduction prediction model.

[0100] For example, the pre-selected neural network structure can be a convolutional neural network or a Transformer time series prediction model as the base network structure, used to learn the relationship between carbon reduction features and predicted carbon emission reductions.

[0101] By collecting a large amount of historical multidimensional dynamic datasets over a wide range of time periods, historical carbon reduction features are constructed and labeled. A neural network structure is then used for training to obtain a carbon emission reduction prediction model. This training method based on a large amount of historical data enables the model to learn the relationship between different factors and carbon emission reduction, improving the accuracy and reliability of the prediction results. The obtained carbon emission reduction prediction results can be synchronized to the management center to generate carbon inclusive policy reports, optimize grid dispatch suggestions and plans, etc., providing strong support for the implementation of the carbon inclusive mechanism and the scientific management of the power grid, and contributing to the promotion of low-carbon development in the power industry.

[0102] Example 2:

[0103] like Figure 2 As shown, the power carbon emission reduction prediction system based on the carbon inclusive mechanism includes a multi-dimensional dynamic dataset acquisition module, a dataset processing module, and a carbon emission reduction prediction module.

[0104] The multidimensional dynamic dataset acquisition module is used to acquire a multidimensional dynamic dataset of the target area within a preset time period. The multidimensional dynamic dataset includes electricity consumption time series data, carbon inclusive behavior time series data, and dynamic carbon emission factor data.

[0105] The dataset processing module is used to obtain a load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence for a target area over a preset time period based on a multidimensional dynamic dataset.

[0106] The carbon emission reduction prediction module is used to input the load structure data sequence, dynamic baseline carbon emissions, and dynamic carbon inclusive behavior influencing factor sequence of the target area for a preset time period as carbon reduction features into the carbon emission reduction prediction model to obtain the carbon emission reduction prediction result for the next time period.

[0107] The dataset processing module includes a load structure data sequence formation unit, a dynamic baseline carbon emission determination unit, and a dynamic carbon inclusive behavior influencing factor sequence acquisition unit.

[0108] The load structure data sequence forming unit is used to decompose the total electricity consumption at each collection time point in the power consumption time series data into base load electricity, peak load electricity and valley load electricity, calculate the base load ratio, peak load ratio and valley load ratio, and arrange and combine the base load ratio, peak load ratio and valley load ratio at each collection time point according to the time sequence to form a load structure data sequence.

[0109] The dynamic baseline carbon emission determination unit is used to take electricity consumption time-series data and dynamic carbon emission factor data as inputs and determine the dynamic baseline carbon emission through a load grading calculation model.

[0110] The unit for obtaining the dynamic carbon inclusive behavior impact factor sequence is used to obtain the dynamic carbon inclusive behavior impact factor sequence based on the time series data of carbon inclusive behavior and through the carbon inclusive time-varying impact model.

[0111] The load classification calculation model is as follows:

[0112] ;

[0113] In the above formula, This represents the dynamic baseline carbon emissions for the target region within a preset time period. The length of the preset time period, , , At the time points of collection Base load power, peak load power, and valley load power. , , At the time points of collection The base load carbon emission factor, peak load carbon emission factor, and valley load carbon emission factor.

[0114] The carbon inclusive time-varying impact model is as follows:

[0115] ;

[0116] ;

[0117] In the above formula, To collect data at the time point The dynamic carbon inclusive behavior impact factors are calculated, and the dynamic carbon inclusive behavior impact factors at each data collection time point are arranged in time series to form a dynamic carbon inclusive behavior impact factor sequence. , The length of the preset time period, To collect data at the time point The total number of carbon-inclusive events within the country. To collect data at the time point A set of timestamps of all carbon-inclusive events. For behavioral events timestamp instantaneous intensity, It is an exponentially decaying function. For behavioral events Preset attenuation coefficient, For behavioral events The preset baseline emission reduction coefficient, For behavioral events At the time of collection The load type coupling coefficient, Load type, including base load, peak load, and valley load. For behavioral events In load type The average occurrence intensity on, For load type The average carbon emission factor, This represents the average carbon emission factor.

[0118] Example 3:

[0119] like Figure 3 As shown, the power carbon emission reduction prediction device based on the carbon inclusive mechanism includes a processor and a memory;

[0120] The memory is used to store computer program code and to transmit the computer program code to the processor;

[0121] The processor is used to execute the electricity carbon emission reduction prediction method based on the carbon inclusive mechanism described in Embodiment 1 according to the instructions in the computer program code.

[0122] Example 4:

[0123] A computer storage medium on which computer programs are stored;

[0124] When the computer program is executed by the processor, it implements the steps for predicting electricity carbon emission reductions based on the carbon inclusive mechanism described in this scheme.

Claims

1. A method for predicting carbon emission reduction of power based on carbon favor mechanism, characterized in that, the method comprises: S1, obtaining a multi-dimensional dynamic data set of a target region in a preset time period, the multi-dimensional dynamic data set comprising power consumption time series data, carbon favor behavior time series data, and dynamic carbon emission factor data; S2, obtaining a load structure data sequence, a dynamic benchmark carbon emission amount, and a dynamic carbon favor behavior influence factor sequence of the target region in the preset time period based on the multi-dimensional dynamic data set; S3, inputting the load structure data sequence, the dynamic benchmark carbon emission amount, and the dynamic carbon favor behavior influence factor sequence of the target region in the preset time period into a carbon emission reduction prediction model as carbon reduction characteristics to obtain a carbon emission reduction prediction result of a next time period. 2.The method for predicting carbon emission reduction of power based on carbon favor mechanism according to claim 1, characterized in that, the S2 comprises: decomposing total power consumption at each collection time point in the power consumption time series data into base load power consumption, peak load power consumption, and valley load power consumption, calculating base load proportion, peak load proportion, and valley load proportion, and arranging and combining the base load proportion, the peak load proportion, and the valley load proportion at each collection time point in time sequence to form the load structure data sequence; inputting the power consumption time series data and the dynamic carbon emission factor data as input to determine the dynamic benchmark carbon emission amount through a load grading calculation model; obtaining the dynamic carbon favor behavior influence factor sequence through a carbon favor time-varying influence model based on the carbon favor behavior time series data. 3.The method for predicting carbon emission reduction of power based on carbon favor mechanism according to claim 2, characterized in that, the load grading calculation model is: ； In the above formula, is the dynamic benchmark carbon emission amount of the target area in the preset time period, is the length of the preset time period, , , respectively are the base load power, peak load power and valley load power at the collection time point , , , respectively are the base load carbon emission factor, peak load carbon emission factor and valley load carbon emission factor at the collection time point . 4.The method for predicting carbon emission reduction of power based on carbon favor mechanism according to claim 2, characterized in that, the carbon favor time-varying influence model is: ； ； In the above formula, is the dynamic carbon favor behavior impact factor at the collection time point , the dynamic carbon favor behavior impact factor at each collection time point constitutes a dynamic carbon favor behavior impact factor sequence in time sequence , is the length of the preset time period, is the total number of carbon favor behavior events within the collection time point , is the timestamp set of all carbon favor behavior events occurring within the collection time point , is the instantaneous intensity of the behavior event at the timestamp , is the exponential decay function, is the preset decay coefficient of the behavior event , is the preset baseline emission reduction coefficient of the behavior event , is the load type coupling coefficient of the behavior event at the collection time point , is the load type, including base load, peak load, and valley load, is the average occurrence intensity of the behavior event on the load type , is the average carbon emission factor of the load type , is the average carbon emission factor. 5.A system for predicting carbon emission reduction of power based on carbon favor mechanism, characterized in that, the system comprises a multi-dimensional dynamic data set acquisition module, a data set processing module, and a carbon emission reduction prediction module; the multi-dimensional dynamic data set acquisition module is configured to acquire a multi-dimensional dynamic data set of a target region in a preset time period, the multi-dimensional dynamic data set comprising power consumption time series data, carbon favor behavior time series data, and dynamic carbon emission factor data; the data set processing module is configured to obtain a load structure data sequence, a dynamic benchmark carbon emission amount, and a dynamic carbon favor behavior influence factor sequence of the target region in the preset time period based on the multi-dimensional dynamic data set; the carbon emission reduction prediction module is configured to input the load structure data sequence, the dynamic benchmark carbon emission amount, and the dynamic carbon favor behavior influence factor sequence of the target region in the preset time period into a carbon emission reduction prediction model as carbon reduction characteristics to obtain a carbon emission reduction prediction result of a next time period. 6.The system for predicting carbon emission reduction of power based on carbon favor mechanism according to claim 5, characterized in that, the data set processing module comprises a load structure data sequence forming unit, a dynamic benchmark carbon emission amount determining unit, and a dynamic carbon favor behavior influence factor sequence obtaining unit. The load structure data sequence forming unit is configured to decompose total power consumption at each collection time point in the power consumption time series data into base load, peak load and valley load, calculate base load proportion, peak load proportion and valley load proportion, and arrange and combine the base load proportion, peak load proportion and valley load proportion at each collection time point in time sequence to form the load structure data sequence; The dynamic baseline carbon emission amount determining unit is configured to take the power consumption time series data and dynamic carbon emission factor data as input, and determine the dynamic baseline carbon emission amount through the load grading calculation model; The dynamic carbon preferential treatment behavior influence factor sequence obtaining unit is configured to obtain the dynamic carbon preferential treatment behavior influence factor sequence through the carbon preferential treatment time-varying influence model based on the carbon preferential treatment time series data.

7. The power carbon emission reduction amount prediction system based on the carbon preferential treatment mechanism according to claim 6, characterized in that, the load grading calculation model is: ； In the above formula, is the dynamic benchmark carbon emission amount of the target area in the preset time period, is the length of the preset time period, , , respectively are the base load power, peak load power and valley load power at the collection time point , , , respectively are the base load carbon emission factor, peak load carbon emission factor and valley load carbon emission factor at the collection time point .

8. The power carbon emission reduction amount prediction system based on the carbon preferential treatment mechanism according to claim 6, characterized in that, the carbon preferential treatment time-varying influence model is: ； ； In the above formula, is the dynamic carbon favor behavior impact factor at the collection time point , and the dynamic carbon favor behavior impact factors at each collection time point constitute a dynamic carbon favor behavior impact factor sequence in time sequence , is the length of the preset time period, is the total number of carbon favor behavior events within the collection time point , is the timestamp set of all carbon favor behavior events occurring within the collection time point , is the instantaneous intensity of the behavior event at the timestamp , is an exponential decay function, is the preset decay coefficient of the behavior event , is the preset reference emission reduction coefficient of the behavior event , is the load type coupling coefficient of the behavior event at the collection time point , is the load type, including base load, peak load, and valley load, is the average occurrence intensity of the behavior event on the load type , is the average carbon emission factor of the load type , is the average carbon emission factor.

9. A power carbon emission reduction amount prediction device based on the carbon preferential treatment mechanism, characterized in that, comprising a processor and a memory; the memory is configured to store computer program code and transmit the computer program code to the processor; the processor is configured to execute the power carbon emission reduction amount prediction method based on the carbon preferential treatment mechanism according to the instructions in the computer program code.

10. A computer storage medium having a computer program stored thereon, characterized in that: the computer program is configured to implement the steps of the power carbon emission reduction amount prediction method based on the carbon preferential treatment mechanism according to any one of claims 1-4 when executed by a processor.

Images