Real-time monitoring method, system and device for urban-level power consumption carbon emission based on power system streaming data, readable storage medium and program product

Abstract

The application relates to a city-level power consumption carbon emission real-time monitoring method, system and device based on power system streaming data, a readable storage medium and a program product. By identifying target abnormal data in streaming data, repairing the target abnormal data based on constraint state estimation to generate repaired data; based on the repaired data, a time-varying power flow matrix is constructed, the effective net power flow of the time-varying power flow matrix is used as a distribution coefficient, the carbon emission injection amount of a power generation system is distributed to a plurality of consumer ends on the power consumption side, and the average carbon intensity of the consumer ends is obtained; by applying a load change amount in the repaired data, the emission increment corresponding to the load change amount is calculated, and the marginal emission factor is calculated according to the emission increment and the load change amount; based on the average carbon intensity and the marginal emission factor, carbon emission monitoring information is output. In this way, high-frequency time resolution carbon emission monitoring is realized, the effectiveness of the carbon emission monitoring information is guaranteed, and the authenticity of the emission reduction effect is improved.

Description

Technical Field

[0001] This application relates to the field of power system data processing technology, and in particular to a method, system, device, readable storage medium, and program product for real-time monitoring of urban-level electricity consumption carbon emissions based on power system streaming data. Background Technology

[0002] Carbon emission accounting for urban electricity consumption is a crucial foundation for urban carbon emission monitoring and energy carbon management. With the advancement of new power generation systems, the high proportion of renewable energy connected to the grid, and the normalization of inter-regional trading, the operating status of power generation systems exhibits significant high-frequency and time-varying characteristics, leading to dramatic fluctuations in the carbon emission intensity of urban electricity consumption.

[0003] Currently, urban-level electricity emissions accounting primarily employs the annual or quarterly average emission factor method, directly converting electricity consumption into static carbon emissions based on the regional power grid's average emission factor. In scenarios oriented towards real-time operation of power generation systems, refined carbon management, and low-carbon decision optimization, static carbon emissions, characterized by long-term static values, mask critical high-emission periods resulting from fluctuations in the proportion of renewable energy and dynamic adjustments in cross-regional transactions. This leads to an inability to quantify the true emission reduction effect, resulting in a distorted assessment of emission reduction outcomes. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, system, equipment, readable storage medium, and program product for real-time monitoring of urban-level electricity consumption carbon emissions based on power system streaming data to address the aforementioned technical issues. This would ensure the effectiveness of carbon emission monitoring information and improve the authenticity of emission reduction results.

[0005] Firstly, this application provides a method for real-time monitoring of carbon emissions from urban-level electricity consumption based on streaming data from the power system, including:

[0006] Acquire streaming data of the power generation system operation;

[0007] Identify target anomalous data in streaming data, repair the target anomalous data based on constraint state estimation, and generate repaired data;

[0008] Based on the repaired data, a time-varying power flow matrix is ​​constructed. Using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, the carbon emission injection of the power generation system is allocated to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity of the consumption end.

[0009] By applying load change to the repaired data, the emission increment corresponding to the load change is calculated, and the marginal emission factor is calculated based on the emission increment and load change.

[0010] Based on average carbon intensity and marginal emission factor, carbon emission monitoring information is output.

[0011] In one embodiment, the steps of identifying target anomalous data in streaming data and repairing the target anomalous data based on constraint state estimation to generate repaired data include:

[0012] Identify time-series anomalies in streaming data based on time-series consistency, and identify residual anomalies in streaming data based on physical consistency;

[0013] Based on time-series anomalous data and residual anomalous data, as well as data constraints for streaming data, identify target anomalous data in streaming data;

[0014] With the goal of minimizing the repair cost, the abnormal target data is repaired to obtain repaired data.

[0015] In one embodiment, the streaming data includes at least one of a first type of data, a second type of data, and a third type of data. The step of constructing a time-varying power flow matrix based on the repaired data includes:

[0016] When the streaming data is the first type of data that includes cross-regional data, a time-varying power flow matrix is ​​constructed based on the exchanged power between cross-regional data.

[0017] When the streaming data is a second type of data that includes data from the same region, the line power flow in the data from the same region is aggregated and combined with the network loss term to construct a time-varying power flow matrix.

[0018] When the streaming data is a third type of data including topologyless data, the equivalent transmission coefficient is obtained by constraining the topologyless data, and a time-varying power flow matrix is ​​constructed based on the equivalent transmission coefficient.

[0019] In one embodiment, the step of allocating the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, and obtaining the average carbon intensity at the consumption ends, includes:

[0020] Calculate the carbon emission injection of the power generation system;

[0021] The carbon emission injection is decomposed using the effective net power flow of the time-varying power flow matrix as the allocation coefficient to obtain the electricity consumption parameters corresponding to multiple consumer ends.

[0022] A carbon tracking linear equation is constructed based on carbon emission injection and electricity consumption parameters from multiple consumer ends; the carbon tracking linear equation satisfies the carbon conservation law.

[0023] Solve the carbon tracking linear equation to obtain the average carbon intensity at the consumer end.

[0024] In one embodiment, the streaming data includes streaming data collected from multiple generator sets in a power generation system, and the streaming data includes generator set processing data corresponding to the multiple generator sets.

[0025] The steps for calculating the marginal emission factor by applying load changes to the repaired data, calculating the emission increment corresponding to the load change, and calculating the marginal emission factor based on the emission increment and load change include:

[0026] Based on the generator set output data, the operating type of the power generation system is determined; the operating type includes the first operating type, the second operating type, and the third operating type.

[0027] When the operation type is the first operation type in which multiple generator sets are allocated according to a preset ratio, the marginal participation factors corresponding to the load change of multiple generator sets are integrated to obtain the marginal emission factor.

[0028] When the second operating type is multiple generator sets based on actual scheduling allocation, the marginal emission factor is obtained by solving the scheduling constraints according to the load change.

[0029] When the operating type is the third operating type, which is assigned by multiple generator sets based on historical data, the marginal emission factor is determined based on the linear relationship between load and marginal factor in the historical data.

[0030] In one embodiment, after the step of outputting carbon emission monitoring information based on average carbon intensity and marginal emission factor, the method further includes:

[0031] Calculate the confidence intervals for average carbon intensity and marginal emission factor;

[0032] Monitor the changes in average carbon intensity and marginal emission factor, and output an abnormal alarm event when the changes exceed the confidence interval;

[0033] Based on abnormal alarm events, locate the abnormal locations that affect average carbon intensity and marginal emission factor.

[0034] Secondly, this application also provides a carbon emission monitoring system for electricity consumption, comprising:

[0035] The acquisition module is used to acquire streaming data of the power generation system operation;

[0036] The evaluation and repair module is used to identify target anomalous data in streaming data, repair the target anomalous data based on constraint state estimation, and generate repaired data.

[0037] The first calculation module is used to construct a time-varying power flow matrix based on the repaired data, and use the effective net power flow of the time-varying power flow matrix as the allocation coefficient to allocate the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity of the consumption end.

[0038] The second calculation module is used to calculate the emission increment corresponding to the load change by applying the load change to the repair data, and to calculate the marginal emission factor based on the emission increment and the load change.

[0039] The information output module is used to output carbon emission monitoring information based on average carbon intensity and marginal emission factor.

[0040] Thirdly, this application also provides an apparatus including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the city-level electricity consumption carbon emission real-time monitoring method based on power system streaming data provided in the first aspect.

[0041] Fourthly, this application also provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the city-level electricity consumption carbon emission real-time monitoring method based on power system streaming data provided in the first aspect.

[0042] Fifthly, this application also provides a program product, including a computer program, which, when executed by a processor, performs the steps of the city-level electricity consumption carbon emission real-time monitoring method based on power system streaming data provided in the first aspect.

[0043] The aforementioned method, system, equipment, readable storage medium, and program products for real-time monitoring of urban-level electricity consumption carbon emissions based on power system streaming data acquire streaming data of the power generation system operation; identify target anomalies in the streaming data; repair the target anomalies based on constraint state estimation to generate repaired data; construct a time-varying power flow matrix based on the repaired data; use the effective net power flow of the time-varying power flow matrix as the allocation coefficient to distribute the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side, obtaining the average carbon intensity at the consumption end; calculate the emission increment corresponding to the load change by applying load change to the repaired data; calculate the marginal emission factor based on the emission increment and load change; and output carbon emission monitoring information based on the average carbon intensity and marginal emission factor. In this way, by acquiring streaming data of the power generation system operation and performing anomaly repair, constructing a time-varying power flow matrix to allocate carbon emission injection, and simultaneously calculating the average carbon intensity and marginal emission factor, high-frequency time-resolution carbon emission monitoring is achieved, accurately reflecting rapid changes in the power supply structure, ensuring the effectiveness of carbon emission monitoring information, and improving the authenticity of emission reduction effects. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A diagram illustrating the application environment of a city-level real-time monitoring method for carbon emissions from electricity consumption based on streaming data from the power system in one embodiment.

[0046] Figure 2 This is a schematic diagram of the first process in one embodiment of a method for real-time monitoring of carbon emissions from urban-level electricity consumption based on streaming data from the power system.

[0047] Figure 3 This is a schematic diagram of the second process in one embodiment of a method for real-time monitoring of carbon emissions from urban-level electricity consumption based on streaming data from the power system.

[0048] Figure 4 This is a schematic diagram of the third process in one embodiment of a method for real-time monitoring of carbon emissions from urban-level electricity consumption based on streaming data of the power system.

[0049] Figure 5 This is a schematic diagram of a carbon emission monitoring system for electricity consumption in one embodiment.

[0050] Figure 6 This is a schematic diagram of the internal structure of a device in one embodiment. Detailed Implementation

[0051] 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.

[0052] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0053] The city-level real-time monitoring method for carbon emissions from electricity consumption based on power system streaming data provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on another network server.

[0054] Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, aircraft, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc.

[0055] Server 104 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides cloud computing services.

[0056] In high-frequency operation data processing of power systems, existing methods for calculating carbon emissions from urban-level electricity consumption suffer from drawbacks such as coarse time granularity and inability to characterize intraday dynamic changes in power supply structure. Specifically, traditional methods rely on annual or quarterly average emission factors, making it difficult to respond to high-frequency fluctuations in consumption-side carbon intensity caused by renewable energy output fluctuations, unit start-up and shutdown ramp-up, and dynamic adjustments in inter-regional power exchange, thus masking high emission risks during critical periods. Furthermore, the calculation results only provide average carbon intensity, lacking the ability to calculate marginal emission factors, resulting in the inability to quantify new emissions caused by load changes, affecting the accuracy of emission reduction assessments for demand response and time-of-use pricing strategies. Furthermore, the physical consistency between the cross-regional / cross-city power flow attribution mechanism and the network loss allocation rules is weak. The attribution of the source of input power is unclear, and the boundary between external power transmission and local consumption is blurred, resulting in deviations in the calculation of carbon intensity on the consumption side. High-frequency data sources are heterogeneous, with differences in sampling frequency, time alignment errors, missing measurements, and noise in unit output, load, and exchanged power. There is a lack of data repair mechanisms based on power balance constraints and rate of change constraints, which leads to power imbalance and severe fluctuations in results. The full recalculation strategy has too high computational overhead in high-frequency update scenarios in multiple cities, making it difficult to meet the near real-time service requirements. Moreover, there is a lack of closed-loop verification with measured emission data, resulting in insufficient traceability of results.

[0057] For example, in the operation of a power grid in a city in East China, when wind power output drops rapidly due to sudden weather changes, the power structure shifts towards thermal power units. Traditional methods cannot capture this structural change during this period, and the carbon intensity on the consumption side is systematically underestimated. There is a time alignment error between the unit output data provided by the dispatch system and the load data of the metering system, and load measurement records are frequently missing. Directly using these records for carbon intensity calculation leads to excessive residual power imbalance. Cross-regional power input is transmitted through multiple levels of the power grid, and the attribution mechanism is based solely on transaction volume statistics, without matching the actual power flow and grid topology. The network loss allocation rules do not consider the differences between the power supply side and the power consumption side, resulting in incorrect attribution of the source of input power and further amplifying the error in carbon intensity on the consumption side. In scenarios where real-time output of carbon emission monitoring information is required, the calculation process of constructing the full allocation matrix is ​​too time-consuming and cannot support minute-level update requirements.

[0058] If the above problems are not addressed, carbon emission monitoring results will not be able to accurately characterize the operating status of the power system, leading to distorted assessments of the emission reduction effects of load shifting strategies; insufficient stability of the results will make it difficult to meet the data reliability requirements of regulatory agencies; accumulated calculation delays will hinder the implementation of real-time energy optimization decisions; and the deviation will continue to widen in long-term operation, affecting the authority and auditability of carbon accounting results, thereby weakening the effectiveness of energy and carbon management measures.

[0059] Based on this, in an exemplary embodiment, such as Figure 2 As shown, a method for real-time monitoring of carbon emissions from urban-level electricity consumption based on streaming data of the power system is provided, and this method is applied to... Figure 1 Taking server 104 as an example, this method for real-time monitoring of city-level electricity consumption carbon emissions based on power system streaming data can include steps 21 to 25, wherein:

[0060] Step 21: Obtain streaming data of the power generation system operation.

[0061] Here, the streaming data of the power generation system operation can be various types of operational data collected from the power system in real time or near real time, such as the output data of generator sets, the load data of each node in the power grid, the exchange of electricity between different regions or cities, and the emission factors or emission model parameters of generator sets related to carbon emissions.

[0062] Streaming data is generated and transmitted in the form of a continuous data stream, which can be used to reflect the operating status of a power generation system within a specific time slice.

[0063] As an example, the output data of each generator set, the load data of each region, and the power exchange data across regions can be manually recorded periodically and then entered into the system.

[0064] As another example, data acquisition systems, such as Supervisory Control and Data Acquisition (SCADA) systems, can be used to periodically query and acquire data on generator sets, loads, and exchanged electricity at fixed time intervals (e.g., every hour).

[0065] In one embodiment, streaming data may include unit output data, unit emission factors (or emission model parameters), node loads, and exchanged power. For example, streaming data may be an operational data stream of a power generation system, and may include at least the unit output data P corresponding to multiple generating units included in the power generation system. g,t Unit emission factors or emission model parameters ef g Urban or power grid node load L i,t Cross-regional / cross-city electricity exchange or equivalent exchange power X ij,t .

[0066] In one embodiment, after the step of acquiring streaming data of power system operation, the streaming data can be preprocessed to obtain the acquisition time slices of different dimensions of data (e.g., unit output data, unit emission factor, node load and exchange power) in the streaming data, and perform time alignment processing on the different dimensions of data according to the acquisition time slices to map the different dimensions of data to a unified time slice to obtain streaming data.

[0067] For example, the streaming data can exist in the form of observation vectors, such as y t ={P g,t L i,t Xi j,t ef g}

[0068] In other implementations, when the power generation system in the streaming data corresponds to a node-level or substation-level system, a mapping table can be established to aggregate the node set to the city, and the power flow corresponding to each node can be summarized according to the node-city mapping.

[0069] Step 22: Identify target anomalous data in the streaming data, repair the target anomalous data based on constraint state estimation, and generate repaired data.

[0070] Among them, the target abnormal data can be data points or data segments in the above streaming data that are inconsistent with the data pattern, do not meet physical laws or logical consistency due to reasons such as sensor failure, communication interruption, data entry error or system operation abnormality.

[0071] Here, the target outlier data can specifically manifest as missing values, outliers, mutation values, or values ​​that are inconsistent with other related data.

[0072] Furthermore, after identifying the target anomaly data in the streaming data, the anomaly data can be repaired to obtain the corresponding repaired data. It should be understood that constraint state estimation can optimize the data by combining the streaming data, which serves as the raw observation data, with the physical constraints of the power system (e.g., power balance, equipment capacity limitations, operating rate of change limitations, etc.) to repair the target anomaly data and obtain the repaired data.

[0073] Here, repaired data can refer to streaming data that has been identified and corrected for anomalies, missing parts, or inconsistencies after constraint state estimation processing. Repaired data is physically consistent and can more accurately reflect the actual operating state of the power system.

[0074] As one implementation method, a numerical threshold can be set as a constraint state. Any data point that exceeds the set upper and lower limits is marked as abnormal, and the abnormal data value is repaired within the constraint state.

[0075] As another approach, statistical methods can be used, such as calculating the mean and standard deviation of the data, identifying data points that deviate from the mean by more than a certain standard deviation as anomalies, and then using linear interpolation to correct them.

[0076] In addition, a series of logical rules can be preset. For example, the output of the generator set cannot be negative. When the data violates these rules, it will be forcibly corrected to the range allowed by the rules.

[0077] Step 23: Based on the repaired data, construct a time-varying power flow matrix. Using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, allocate the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity of the consumption ends.

[0078] The time-varying power flow matrix can be a matrix used to describe the power transmission relationship between cities or nodes in a power network. The elements of the time-varying power flow matrix represent the effective net power flow or equivalent power flow from one city or node to another city or node within a specific time slice.

[0079] It should be understood that since the operating state of the power system changes over time, the time-varying power flow matrix also changes over time slices, which can be used to reflect the real-time characteristics of power flow.

[0080] Here, effective net power flow refers to the net amount of electrical energy actually transferred between two cities or nodes in a power network. Effective net power flow takes into account the direction and magnitude of electrical energy transmission and is a core element in constructing a time-varying power flow matrix, used to characterize the flow path and distribution ratio of electrical energy in the network.

[0081] In addition, carbon emission injection can refer to the total amount of carbon emissions directly or indirectly generated by generator sets on the power generation side. Carbon emission injection is usually calculated based on the output of the generator set and its corresponding carbon emission factor, representing the carbon emission source in the power production process.

[0082] It should be noted that the multiple consumption ends on the electricity consumption side can refer to various end users or regions that receive and consume electricity in the power system, such as cities, industrial parks, large enterprises or residential communities. The consumption end is the object of final attribution and accounting for carbon emissions.

[0083] It should be noted that average carbon intensity can refer to the carbon emissions corresponding to each unit of electricity consumed by a certain consumer end within a specific time period. Average carbon intensity can be used to reflect the overall carbon footprint of the electricity consumed by that consumer end.

[0084] As one approach, it can be assumed that carbon emissions are allocated according to the proportion of total electricity consumption at each consumer end. That is, without considering the actual flow path of electricity in the network, the total carbon emissions from the power generation side are directly distributed to each consumer end proportionally.

[0085] As another approach, a static power flow matrix can be constructed based on historical average grid topology and power flow data, and used as an allocation coefficient to distribute carbon emission injections to various consumer ends.

[0086] In addition, carbon emissions can be allocated based on the power flow between nodes by directly measuring the power flow between them and using it as matrix elements.

[0087] Step 24: By applying the load change amount to the repair data, calculate the emission increment corresponding to the load change amount, and calculate the marginal emission factor based on the emission increment and the load change amount.

[0088] Among them, load change can be the increase or decrease of electricity load at a certain consumer end within a certain period of time on the electricity consumption side. Load change can be an actual occurrence or a disturbance applied for analysis purposes.

[0089] Here, the emission increment can be the corresponding change in the total carbon emissions of the power generation system due to changes in load on the electricity consumption side. The emission increment can be used to reflect the impact of new or reduced electricity demand on overall carbon emissions.

[0090] It should be understood that the marginal emission factor can refer to the increase in carbon emissions of the power generation system caused by each additional unit of electricity consumption at a certain consumption end on the electricity consumption side within a specific time slice. The marginal emission factor can be used to reflect the carbon emission characteristics of the marginal power sources called upon by the power system in response to additional load under the current operating state, and has reference value for evaluating the emission reduction effect of load adjustment strategies.

[0091] One exemplary approach is to use the average carbon emission factor of the power generation system as the marginal emission factor, that is, to assume that any load change will result in the same emission change as the average level.

[0092] Another exemplary approach is to assume that when the load changes, it is always a specific coal-fired unit in the system that responds, and the emission increment is calculated based on the emission factor of that unit to determine the corresponding marginal emission factor.

[0093] In addition, a simplified model can be established by performing linear regression analysis on historical load and emission data to predict the emission increment caused by load changes, thereby determining the corresponding marginal emission factor.

[0094] Step 25: Output carbon emission monitoring information based on average carbon intensity and marginal emission factor.

[0095] Carbon emission monitoring information refers to various data, reports, charts, or alarms regarding carbon emissions from electricity consumption, provided to users or management departments based on calculations and analyses of the aforementioned average carbon intensity and marginal emission factors. Carbon emission monitoring information helps users understand their electricity consumption carbon footprint and supports energy use decisions.

[0096] One approach is to display the calculated average carbon intensity and marginal emission factor values ​​directly on a text interface, or to generate periodic reports containing these values, such as daily or weekly reports, and send them to relevant users via email. Alternatively, these data can be stored in a database for users to manually query and download as needed.

[0097] Suppose that within a city region, it is necessary to monitor the carbon emissions from electricity consumption at multiple consumption points (e.g., industrial park A, commercial area B, and residential area C). The city region's electricity supply consists of both local power generation systems and inter-regional power exchange with external power grids.

[0098] First, it is possible to continuously acquire streaming data on the operation of the power generation system, including output data from local generator units, load data from various consumer terminals, and data on the electricity exchanged with the external power grid. For example, this data may be transmitted to a data center at minute-level frequencies via various sensors and metering devices.

[0099] Subsequently, the streaming data undergoes quality assessment, identifying any anomalous data to be processed. For example, if the output data of a generator suddenly jumps from its normal value to zero, or if the load data of a consumer remains unchanged for an extended period, it is identified as an anomaly. These anomalous data are then repaired based on constraint state estimation. Specifically, the physical constraints of the power system can be considered, such as the balance between total generation, total load, and network losses, as well as the upper and lower limits of generator output and ramp rate limitations. Through optimization algorithms, the anomalous data is adjusted to satisfy these physical constraints while approximating the original observations as closely as possible, thus generating physically consistent repaired data. For example, for generators whose output jumps to zero, the output value can be estimated and filled based on their historical operating patterns and current system load.

[0100] Next, based on this restored data, a time-varying power flow matrix is ​​constructed to describe how electricity flows between local generators, consumer terminals, and the external power grid. For example, if the external power grid supplies electricity to the urban area, this matrix reflects the inflow of that electricity. Simultaneously, the carbon emission injection of the power generation system is calculated, typically based on the output of the local generators and their corresponding carbon emission factors (e.g., the carbon emission factor of coal-fired units). Then, using the effective net power flow reflected in the time-varying power flow matrix as the allocation coefficient, the carbon emission injection of the power generation system is allocated to the three consumer terminals: industrial park A, commercial area B, and residential area C. By solving the carbon tracking linear equation, the average carbon intensity of each consumer terminal, i.e., the carbon emissions contained in each kilowatt-hour of electricity, is obtained.

[0101] Furthermore, to assess the impact of load changes on carbon emissions, a load change factor can be applied to the repaired data. For example, assuming the load of industrial park A increases by 1 MW, the emission increment corresponding to the load change is calculated to represent how the simulated power system responds to the increased load. This includes, for example, which generators increase output to meet demand, and the emission characteristics of these generators. Based on the calculated emission increment and the applied load change factor, the marginal emission factor of industrial park A is calculated.

[0102] Finally, based on the calculated average carbon intensity and marginal emission factor, the output includes carbon emission monitoring information, which at least includes the average carbon intensity and marginal emission factor of each consumption end, as well as a trend graph of the change over time. This information is used to guide industrial park A in optimizing its energy use, for example, by increasing production load during periods of lower marginal emission factor, thereby reducing energy consumption emissions.

[0103] Based on the above implementation method, firstly, by acquiring streaming data of the power generation system operation and identifying and repairing the abnormal data to be processed, the problems of missing data, noise, and inconsistencies in caliber that exist in the existing technology are solved. For example, when the output data of the generator set shows a jump, it can be repaired based on constraint state estimation, ensuring the physical consistency and reliability of the data, avoiding power imbalance and carbon conservation failure caused by data quality problems in traditional methods, and improving data quality and calculation accuracy.

[0104] Secondly, a time-varying power flow matrix is ​​constructed based on the repaired data, and the effective net power flow is used as the allocation coefficient to distribute the carbon emission injection of the power generation system to multiple consumption terminals on the electricity consumption side, thereby obtaining the average carbon intensity at the consumption side. Here, the time-varying power flow matrix is ​​used to characterize the actual flow path of electricity between various consumption terminals within the urban area, and a fine attribution of carbon emissions is performed accordingly. This solves the problems of existing technologies that mostly use annual / quarterly average emission factors or allocation based on transaction volume, which are difficult to correspond to actual power flows, and the attribution of cross-regional / cross-city power flows and the characterization of grid losses are insufficient, resulting in weak physical consistency. It can provide a more physically consistent and interpretable carbon intensity at the consumption side.

[0105] Furthermore, by incorporating load changes into the repaired data, the corresponding emission increment is calculated, and thus the marginal emission factor is determined. Based on this, when the load of industrial park A increases, the resulting carbon emissions are calculated, and the marginal emission factor is obtained. This solves the problem that existing technologies often only output average carbon intensity and cannot provide the marginal emission factor. The marginal emission factor is valuable for energy optimization scenarios such as demand response and load shifting, and can directly assess the emission reduction effect of load changes.

[0106] Finally, carbon emission monitoring information is output based on average carbon intensity and marginal emission factor. In the example above, industrial park A can adjust its production plan based on this monitoring information, such as increasing electricity consumption during periods of lower marginal emission factor, thereby achieving economic operation. This provides real-time and accurate decision-making basis for the refined management of carbon emissions from city-level electricity consumption, overcoming the shortcomings of existing technologies in terms of low temporal resolution and difficulty in reflecting structural changes.

[0107] In one embodiment, after the step of outputting carbon emission monitoring information based on average carbon intensity and marginal emission factor, the method further includes:

[0108] (1) Calculate the confidence intervals for average carbon intensity and marginal emission factor.

[0109] Here, statistical or simulation methods are used to determine the possible range of values ​​for these key indicators at a certain confidence level, in order to quantify the reliability or uncertainty of the estimates of average carbon intensity and marginal emission factor.

[0110] For example, based on historical monitoring data, parametric statistical methods can be used, such as assuming that the data follows a normal distribution, and using the mean and standard deviation to calculate the Z-score or t-score to construct confidence intervals.

[0111] In addition, Monte Carlo simulation or Bayesian update propagation can be used to randomly sample input variables (such as power generation, fuel type, load change, etc.) that affect average carbon intensity and marginal emission factor, run the carbon emission calculation model multiple times, and obtain the empirical distribution of average carbon intensity and marginal emission factor, thereby determining their confidence intervals.

[0112] (2) Monitor the changes in average carbon intensity and marginal emission factor, and output an abnormal alarm event when the changes exceed the confidence interval.

[0113] Real-time tracking of these indicators is crucial for timely detection of potential anomalies. Specifically, the absolute difference or percentage change rate between the current average carbon intensity or marginal emission factor and its corresponding value from the previous moment or period (e.g., the previous day or week) can be calculated. Furthermore, time series analysis methods such as moving averages or exponentially weighted moving averages can be used to calculate the trend or fluctuation range of the indicators within a specific time window, thereby obtaining the magnitude of their changes.

[0114] When the change exceeds the confidence interval, an abnormal alarm event is output. Based on the preset judgment logic, the fluctuation of the average carbon intensity or marginal emission factor is automatically identified as exceeding its normal or acceptable range, and the corresponding alarm notification is triggered.

[0115] For example, the currently monitored average carbon intensity or marginal emission factor can be compared with the upper and lower limits of a pre-calculated confidence interval. When the average carbon intensity or marginal emission factor exceeds this range, it is judged as abnormal and an alarm is generated.

[0116] Alternatively, a dynamic threshold can be set that adjusts according to real-time changes in the confidence interval. When the change exceeds the dynamic threshold, an abnormal alarm event is output.

[0117] For example, when the output jump exceeds the threshold or the residual is abnormal, the contribution can be decomposed by unit, by exchange edge, or by city connection edge to generate alarm events and cause results.

[0118] (3) Based on the abnormal alarm events, locate the abnormal locations that affect the average carbon intensity and marginal emission factor.

[0119] Once an anomaly is detected, the root cause can be further analyzed to provide more specific diagnostic information so that targeted interventions can be implemented.

[0120] For example, the original data source that caused the abnormal alarm event can be traced, and the operating data, fuel consumption data, load change data, or changes in the power grid topology of a specific generator set can be examined to determine the specific link or equipment where the abnormality occurred.

[0121] In addition, a pre-set rule base or expert system can be used to match possible causes of anomalies and corresponding location information based on the type and characteristics of the anomaly alarm. For example, if the marginal emission factor rises abnormally, it may be related to the start-up or increased output of a specific high-carbon emission unit, thereby locating the generator unit corresponding to the anomaly.

[0122] In the above implementation, based on the calculation of average carbon intensity and marginal emission factor on the electricity consumption side, a reliability assessment and anomaly diagnosis mechanism for key indicators is further introduced. Statistical boundaries are set for the normal fluctuation range of these indicators. Furthermore, the real-time changes in these indicators are continuously monitored and compared with preset confidence intervals. When the change exceeds the normal range defined by the confidence interval, an anomaly alarm event is output, thereby achieving timely and automatic identification of potential anomalies. Upon receiving an anomaly alarm event, the specific anomaly affecting average carbon intensity and marginal emission factor can be traced and located based on the event's characteristics, enhancing the practicality and effectiveness of electricity consumption carbon emission monitoring. In this way, not only can the current carbon emission level be understood, but its reliability can also be assessed, and the source of the problem can be quickly located when abnormal fluctuations occur, achieving precise carbon emission management and control.

[0123] In some of the above implementation methods, a method is proposed to identify target anomalous data in streaming data and repair the target anomalous data based on constraint state estimation to generate repaired data. However, in actual power system operation data, the types of anomalous data are diverse, which may include temporal inconsistencies or physical irrationalities. If these anomalous data are not effectively distinguished and identified and repaired in conjunction with data constraints, the accuracy of the repair results may be insufficient, thereby affecting the accuracy of subsequent carbon emission monitoring.

[0124] Based on this, in an exemplary embodiment, such as Figure 3 As shown, step 22, identifying target anomalous data in the streaming data, repairing the target anomalous data based on constraint state estimation, and generating repaired data includes steps 221 to 223, wherein:

[0125] Step 221: Identify time-series anomalous data in streaming data based on time-series consistency, and identify residual anomalous data in streaming data based on physical consistency.

[0126] Temporal consistency can be used to identify anomalies in streaming data by analyzing the continuity, stationarity, or conformity to specific trends of data over time.

[0127] One approach is to use a sliding window method, comparing the statistical characteristics (such as mean and variance) of the current data point with those of the previous N data points. If the values ​​exceed a preset threshold, the data point is marked as an anomaly. Another approach is to use a time series forecasting model to predict the data value at the next time step, identifying data points where the deviation between the actual and predicted values ​​exceeds a certain range as time series anomalies.

[0128] It should be understood that performing time-series consistency checks on streaming data can identify time-series anomalous data in the streaming data; wherein, time-series anomalous data includes at least one of data jumps, missing data, and drift.

[0129] Here, physical consistency can be achieved by verifying whether the data conforms to the inherent physical laws and operational constraints of the power system to detect anomalies in streaming data.

[0130] For example, a power system state estimation model can be used to substitute the measurement data and calculate the residual between the measured value and the estimated value. If the residual exceeds a preset threshold, it is marked as a residual abnormality.

[0131] As another example, the power flow calculation results of the power system can be used to check whether the power balance of each node, the direction and magnitude of the power flow of the line conform to the physical laws. Those that do not conform are identified as residual abnormalities.

[0132] It should be noted that performing physical consistency checks on streaming data can calculate residual anomalies. These residual anomalies include at least one of the following: power balance residuals, boundary exchange constraint residuals, and load aggregation consistency residuals.

[0133] Step 222: Identify target anomalous data in the streaming data based on time-series anomalous data, residual anomalous data, and data constraints for streaming data.

[0134] Among them, target abnormal data can refer to the abnormal data that needs to be repaired by accurately locating it, taking into account different types of abnormalities and known system constraints.

[0135] For example, after identifying time-series outliers and residual outliers, a data quality score can be generated based on these data. If the data quality score is lower than a preset threshold, constraint state estimation can be performed on the streaming data to identify the target outliers in the streaming data.

[0136] Step 223: With the goal of minimizing the repair cost, repair the target abnormal data to obtain repaired data.

[0137] When repairing outlier data, it is essential not only to ensure the accuracy of the repaired data but also to minimize alterations to the original data and preserve its original characteristics. For example, the repair cost can be defined as the sum of the squares of the Euclidean distances between data points before and after repair. An optimization algorithm (such as the least squares method) can be used to find the repair value that minimizes this cost while satisfying data constraints. Alternatively, the repair cost can also be a weighted sum of the relative errors between data points before and after repair, with the weights set according to the importance or degree of anomaly of the data points.

[0138] Specifically, after identifying the target abnormal data, decision variables can be constructed based on the target abnormal data. With the objective of minimizing the repair cost function, constraint states are solved for multiple repair terms in the constructed decision variables, and the repair data is output. The repair data includes at least the unit output repair value, the city load repair value, and the exchange power repair value. The repair cost function includes at least the unit output deviation penalty term, the city load deviation penalty term, and the exchange power deviation penalty term. The constraint states include at least one of the following: node power balance constraint, unit output upper and lower bound constraints, exchange power upper and lower bound constraints, and rate of change constraint. The rate of change constraint includes at least one of the following: unit ramp-up constraint and load rate of change constraint.

[0139] For example, data quality assessment of streaming data can identify missing, outlier, noisy, and inconsistent records, as well as power balance residuals and exchange consistency residuals, generating a quality score (qt). Repair can then be triggered for the streaming data within the time slice corresponding to low-quality scores. When repairing target outlier data, decision variables can be constructed. (Generator output data) (Urban load data) (Exchange power data) is used as the minimum repair cost function. With the minimum repair cost function as the objective, the target abnormal data is solved under the constraints of power balance, upper and lower bounds and rate of change to obtain the repair data.

[0140] The detailed calculation is as follows:

[0141]

[0142] In one implementation, after obtaining the repaired data, the confidence level q of the repaired data can also be output simultaneously. i,t .

[0143] The aforementioned technical solution enables more refined and accurate anomaly identification and repair of streaming data from power generation systems. By distinguishing between time-series anomalies and residual anomalies, and combining this with data constraints for comprehensive judgment, it avoids the missed or false alarms that might occur with a single anomaly detection method. Simultaneously, by minimizing repair costs, the repaired data ensures that it conforms to the physical laws and operational characteristics of the power system while preserving the effective information of the original data to the maximum extent, thus significantly improving the quality and reliability of the repaired data. This lays a solid data foundation for subsequent accurate construction of time-varying power flow matrices, accurate allocation of carbon emission injection, and calculation of average carbon intensity and marginal emission factors, ultimately enhancing the overall accuracy and reliability of electricity consumption carbon emission monitoring.

[0144] In some of the embodiments described above in this application, a method for constructing a time-varying power flow matrix based on repaired data to allocate carbon emission injection is proposed. However, in actual power system operation, the sources and types of streaming data may be diverse, such as cross-regional data, intra-regional data, or topologyless data. If different processing is not carried out for the characteristics of different types of data, the accuracy and applicability of the time-varying power flow matrix construction may be limited, thereby affecting the accuracy of carbon emission allocation.

[0145] Based on this, such as Figure 4 As shown, step 23, based on the repaired data, constructs a time-varying power flow matrix. Using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, the carbon emission injection of the power generation system is allocated to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity at the consumption end. This specifically includes steps 231 to 235, where:

[0146] Step 231: Construct a time-varying power flow matrix based on the repaired data.

[0147] Here, streaming data includes at least one of the first type of data, the second type of data, and the third type of data. Based on the data type of the repaired data, constructing the time-varying power flow matrix can include the following three cases:

[0148] In the first scenario, when the streaming data is of the first type, including cross-regional data, a time-varying power flow matrix is ​​constructed based on the exchanged power between cross-regional data.

[0149] The first type of data may include power data collected from interconnecting lines or interfaces between different power regions, which is used to reflect the power exchange between regions.

[0150] It should be understood that the first type of data has a high degree of aggregation and clear physical boundaries. For example, streaming data includes electricity exchange data across regions or cities, or power measurements on regional interconnection lines between regions or cities.

[0151] Possible implementation methods include, but are not limited to, obtaining inter-regional tie-line power data through a wide-area measurement system, or obtaining cross-regional transaction settlement data through a data exchange protocol between regional dispatch centers, to construct a time-varying power flow matrix.

[0152] For example, the streaming data of the first type of data can be used to construct an equivalent flow matrix based on the exchange of electricity between cities in the repair data. The positive net exchange amount of the exchange edge of the equivalent flow matrix is ​​taken as the first effective net power flow between cities. The first effective net power flow is used as the matrix element to construct a time-varying power flow matrix.

[0153] It can be explained that when the streaming data is the first type of data, including cross-regional data, a time-varying power flow matrix is ​​constructed based on the exchanged electricity between cross-regional data. When the streaming data mainly reflects the power exchange between regions, the key to constructing the time-varying power flow matrix lies in accurately quantifying the electricity flow between each region. The exchanged electricity is a direct manifestation of power transmission between regions and can effectively reflect the power coupling relationship between regions. By using the exchanged electricity as the basis for constructing the matrix, it can be ensured that the matrix can accurately capture the power flow characteristics at the regional level. Specifically, this can be achieved by treating each region as a node and the exchanged electricity between regions as the connection weight between nodes, or by constructing a simplified inter-regional power flow model.

[0154] In the second scenario, when the streaming data is of the second type, including data from the same region, the line power flow in the data from the same region is aggregated and combined with the network loss term to construct a time-varying power flow matrix.

[0155] The second type of data can be power data collected from various nodes, lines, or equipment within the same power area of ​​the transmission and distribution network. This second type of data typically contains more granular network operation information, such as line power flow data, cross-sectional power flow data, or grid topology version data, or power flow data on substation outgoing lines, feeders, or main lines.

[0156] Possible implementation methods include, but are not limited to, substation bus voltage, line current, and active / reactive power data obtained through monitoring and data acquisition systems, or distribution line operation data obtained through distribution automation systems.

[0157] For example, streaming data of the second type of data can calculate the power flow between nodes based on the line power flow or cross-sectional power flow in the repair data, aggregate the power flow between nodes into the power flow between cities according to the node-to-city mapping, estimate the network loss term and allocate the network loss term to the corresponding inter-city power flow according to the power supply side rule or the power consumption side rule to obtain the second equivalent net power flow between cities, and use the second equivalent net power flow as matrix elements to construct a time-varying power flow matrix.

[0158] When streaming data is second-type data including data from the same region, line power flows within the same region are aggregated and combined with network loss terms to construct a time-varying power flow matrix. When streaming data consists of line power flow data from the same region, due to the potentially large data volume and fine granularity, appropriate aggregation processing is required to simplify the model, while also considering unavoidable network losses during power grid operation. Line power flow aggregation refers to summarizing or equivalently processing power flow data from multiple lines or nodes to obtain a more macroscopic representation of power flow. Combining network loss terms ensures that the power flow matrix satisfies the power balance principle, i.e., input power minus output power should equal network loss, thereby improving the physical accuracy of the matrix. Aggregation can employ topology-based partitioned aggregation or cluster analysis based on power flow similarity. Network loss terms can be obtained through historical data statistics, real-time calculation, or estimation based on line parameters.

[0159] In the third scenario, when the streaming data is of the third type, including topologyless data, the equivalent transmission coefficients are obtained by constraining and optimizing the topologyless data, and a time-varying power flow matrix is ​​constructed based on the equivalent transmission coefficients.

[0160] The third type of data can be power data that lacks clear network topology information or has incomplete topology information. This third type of data may originate from distributed power sources, smart electricity terminals, or some devices not connected to the monitoring system. Streaming data may include urban load data and exchanged power data, but it lacks topology parameters.

[0161] Possible implementation methods include, but are not limited to, obtaining electricity consumption / generation data from devices such as smart meters, distributed photovoltaic inverters, or electric vehicle charging piles. This data may only contain timestamps and power values, but lacks the precise physical location and connection relationship within the power grid.

[0162] For example, the streaming data of the first type of data can be used to construct an equivalent transmission coefficient matrix. The equivalent transmission coefficients are obtained through multi-source data and constraint optimization inversion. The equivalent transmission coefficients are multiplied by the corresponding city load data to obtain the third equivalent power flow between cities. The third equivalent power flow is used as matrix elements to construct a time-varying power flow matrix.

[0163] It should be noted that when streaming data is a third type of data, including topologyless data, constraint optimization can be performed on the topologyless data to obtain equivalent transmission coefficients, and a time-varying power flow matrix can be constructed based on these equivalent transmission coefficients. When streaming data lacks explicit topological information, directly constructing the power flow matrix presents challenges. In this case, constraint optimization methods are needed to infer the equivalent transmission relationships between data points using known physical laws (such as power balance, voltage constraints, etc.) and statistical relationships between data. The equivalent transmission coefficients reflect the "strength" or "path" of power transfer between different data points; even without explicit physical connections, virtual power flow relationships can be established through optimization methods. Constraint optimization can employ linear programming, quadratic programming, or machine learning-based methods, aiming to minimize errors or satisfy specific constraints to solve for the equivalent transmission coefficients that best reflect actual operating conditions.

[0164] In this embodiment, by adaptively selecting the construction method of the time-varying power flow matrix based on different types of streaming data, it is ensured that the power flow of the power system can be accurately reflected under various data conditions. Specifically, when the acquired streaming data is of the first type, i.e., containing cross-regional data, the time-varying power flow matrix is ​​directly constructed using the exchanged electricity between regions. This allows the matrix to directly capture the macroscopic power exchange relationship between regions, making it suitable for large-scale carbon emission allocation. When the streaming data is of the second type, i.e., containing data within the same region, the line power flow data within the region is aggregated and combined with network loss terms to construct the matrix. While preserving the details of power flow within the region, the aggregation reduces computational complexity, and the consideration of network losses ensures the physical accuracy of power balance. When the streaming data is of the third type, i.e., containing topologyless data, the equivalent transmission coefficient is derived from these seemingly discrete data using constrained optimization methods, thereby constructing a time-varying power flow matrix that reflects the potential power flow relationship. This allows for effective carbon emission allocation even in the absence of complete topology information, making the construction process of the time-varying power flow matrix more flexible and robust, and adaptable to the increasingly complex and diverse data environment of the power system.

[0165] As an example, the construction method for the time-varying power flow matrix F_t can be determined by choosing one or a combination of the following methods based on the data format of the streaming data. Specifically:

[0166] (1) Streaming data of exchanged electricity: For each exchanged edge (i,j), let , Both equations are true.

[0167] (2) Streaming data of power flow data aggregation: calculate the power flow between nodes from line power flow or cross-sectional power flow, and then aggregate it to the city layer by mapping; estimate network loss if necessary. tAnd allocated to Loss according to the rules of the power supply side or the power consumption side. i,t .

[0168] (3) Streaming data used for inverse derivation of equivalent coefficients: When complete topology parameters are lacking, construct the equivalent transmission coefficient matrix T, obtain T through multi-source data and constraint optimization inversion, and construct the inverse derivation formula of equivalent coefficients accordingly. It also outputs the confidence level of the inversion error.

[0169] Step 232: Calculate the carbon emission injection of the power generation system.

[0170] Among them, carbon emission injection can be used to quantify the total amount of carbon emissions injected into the grid by the power generation system within a specific operating cycle.

[0171] For example, total carbon emissions can be determined based on the fuel consumption of the power generation system and a preset fuel carbon emission factor. This can be achieved, for instance, by monitoring the coal consumption and carbon emission factor of coal-fired power plants. Alternatively, it can be estimated based on the real-time power generation of the power generation system and the carbon emission intensity per unit of power generation, for example, by extrapolating from real-time electricity production data and historical carbon emission intensity data.

[0172] For example, the unit output restoration value and unit emission factor of each unit can be obtained from the restoration data. The carbon emissions of each unit can be calculated based on the product of the unit output restoration value and the unit emission factor. The carbon emissions of each unit can be summed to obtain the carbon emission injection on the power generation side.

[0173] Alternatively, if unit data is missing in the repair data, the regional power structure, fuel consumption, and power generation can be obtained, and the carbon emission injection on the power generation side can be calculated based on the regional power structure, fuel consumption, and power generation.

[0174] Step 233: The carbon emission injection amount is split using the effective net power flow of the time-varying power flow matrix as the allocation coefficient to obtain the electricity consumption parameters corresponding to multiple consumer ends.

[0175] Among them, effective net power flow can be used to reflect the actual path and amount of electrical energy flowing from the generation side to the consumption side.

[0176] Based on the actual power flow in the power grid, represented by the effective net power flow, the "contribution" or "share" of each consumer to the total carbon emission injection can be determined.

[0177] For example, the proportion of the effective net power flow received by each consumer to the total effective net power flow can be used as an allocation coefficient to proportionally divide the total carbon emission injection, thereby obtaining the preliminary carbon emission allocation for each consumer and deriving the electricity consumption parameters.

[0178] As another example, the carbon emission injection can be distributed to each consumer node based on the net power injection / outflow of each consumer node in the time-varying power flow matrix, combined with its electrical distance or topological relationship with the power generation side, through weighted averaging or other allocation algorithms, to obtain its corresponding electricity consumption parameters.

[0179] Step 234: Construct a carbon tracking linear equation based on carbon emission injection and electricity consumption parameters from multiple consumer ends.

[0180] Among them, the carbon tracking linear equation satisfies the carbon conservation law.

[0181] The carbon tracking linear equation can be used to describe the flow and distribution of carbon emissions from the power generation side to the consumption side, and ensure that the entire distribution process conforms to the physical principle of carbon conservation.

[0182] It is understandable that the carbon conservation law states that the total amount of carbon emissions remains unchanged before and after distribution, and a system of linear equations is constructed with the average carbon intensity of each consumption end as the unknown quantity.

[0183] It should be noted that the coefficients of the carbon tracking linear equation are determined by the amount of carbon emissions injected into the power generation system and the electricity consumption parameters (such as electricity consumption and power flow) at each consumer end. Virtual nodes or balance equations are introduced to ensure that the total inflow and outflow of carbon emissions are equal. Alternatively, the node injection method or branch tracking method can be used, treating each node in the power grid (including generation nodes, load nodes, and transmission nodes) as a sink or source of carbon emissions. Based on the equivalent form of carbon emissions according to Kirchhoff's current law, a set of linear equations describing the carbon emission balance relationship at each node is established, thereby achieving accurate tracking and conservation of carbon emissions.

[0184] For example, the electricity consumption of multiple consumer terminals can be obtained; where the electricity consumption is the sum of the power generation attributable to the consumer terminal and the effective net power flow flowing into the consumer terminal; based on the average carbon intensity of the consumer terminal and the conservation relationship between the total supply of electricity on the consumer side and the carbon emission injection of the consumer terminal, plus the power flow flowing into the consumer terminal and the average carbon intensity of the upstream consumer terminal, an initial carbon tracking linear equation corresponding to the consumer terminal can be constructed.

[0185] Step 235: Solve the carbon tracking linear equation to obtain the average carbon intensity at the consumer end.

[0186] After constructing the carbon tracking linear equation, the final average carbon intensity for each consumer can be derived from the constructed carbon tracking linear equation.

[0187] Specifically, direct numerical methods such as Gaussian elimination and conjugate gradient method can be used to solve the linear equations and directly obtain the average carbon intensity value of each consumption end. Alternatively, iterative methods, such as Jacobi iteration and Gauss-Seidel iteration, can be used to obtain the average carbon intensity of each consumption end through a step-by-step approximation method after satisfying the preset convergence conditions.

[0188] The following example illustrates how, after constructing a time-varying power flow matrix and determining the amount of carbon emission injection, the amount of carbon emission injection can be mapped and allocated to multiple consumer ends (multiple cities or nodes) on the electricity consumption side according to the city or node to which the generator sets belong, as described by the effective net power flow in the time-varying power flow matrix, to obtain the corresponding electricity consumption parameters.

[0189] Specifically, through the allocation formula Distribute the allocation when ef g When using model parameters, electricity consumption parameters corresponding to multiple consumer terminals can be calculated. .

[0190] Furthermore, the total supply S belonging to city i can be defined. i,t =G i,t +Σ j F t (j,i), constructing a linear carbon tracking equation A based on carbon conservation. t ×CI t =b t And solve for CI t .

[0191] The linear equation for carbon tracing is: .

[0192] In another approach, when there are zero supply, isolated nodes, or ill-conditioned matrices, strategies such as backoff rules, adding slight regularization terms, or outputting range results can be used to ensure numerical stability and usability of results.

[0193] The aforementioned technical solution enables precise and physically conserved allocation of carbon emission injection from the power generation system to multiple consumption ends on the electricity consumption side. By explicitly calculating the carbon emission injection and performing preliminary decomposition based on the effective net power flow of the time-varying power flow matrix, a foundation is laid for subsequent refined allocation. Crucially, by constructing and solving a carbon tracking linear equation that satisfies the carbon conservation law, the problem of carbon emission imbalance or accumulated allocation errors that may occur in traditional allocation methods is effectively avoided. This ensures the accuracy and reliability of the carbon emission allocation results, making the final average carbon intensity at the consumption end more scientific and credible, thereby significantly improving the accuracy and practicality of carbon emission monitoring information.

[0194] In some embodiments described above in this application, the marginal emission factor is determined by applying load change to the repaired data and calculating the corresponding emission increment. However, the operating modes of power generation systems are complex and varied, and the response mechanisms of generator units to load changes differ significantly under different operating types. If the marginal emission factor is calculated without differentiation, the accuracy of the calculation results may be insufficient, thereby affecting the reliability of carbon emission monitoring information.

[0195] It should be understood that streaming data includes streaming data collected from multiple generator sets in a power generation system, and streaming data includes generator set processing data corresponding to multiple generator sets.

[0196] It should be noted that generator output data refers to real-time or historical operating data collected from multiple generator sets in the power generation system, such as the actual power generation, fuel consumption rate, start-up and shutdown status, available capacity, and corresponding carbon emission coefficient of the generator sets.

[0197] The operation type of a power generation system refers to the typical operating mode or dispatch strategy exhibited by the system under different time periods or load conditions. Operation types can be classified based on factors such as system load level, adjustable unit combination, and dispatch objectives (e.g., economic efficiency, safety). For example, different operation types can be distinguished based on the peak-valley characteristics of the system load, the penetration rate of renewable energy, and the presence of specific dispatch instructions.

[0198] Due to the diversity of multiple generator sets, the output data of the generator sets are different. Therefore, the operation mode of the power generation system is also different, which can be divided into three types: the first operation type, the second operation type, and the third operation type.

[0199] First, the first operating type can be that multiple generator sets in the power generation system adjust their output according to a pre-set proportional relationship in response to load changes.

[0200] The pre-set ratio can be determined based on factors such as unit capacity, fuel type, or historical operating experience.

[0201] Under this operating type, the contribution of each generator set to the total load change is relatively fixed. The marginal participation factor can be the proportion of the change in output of each generator set to the total load change when the system load undergoes a small change. The marginal participation factor corresponding to the first operating type can be used to indicate which generator sets will respond to load changes and the extent of their response under specific operating conditions.

[0202] It should be understood that the marginal participation factor is usually related to factors such as the operating cost of the generator set, capacity limitations, and scheduling strategies.

[0203] Secondly, the second operating type can be the output adjustment of multiple generator units in the power generation system, which is allocated based on a real-time scheduling optimization algorithm.

[0204] Dispatch and allocation typically consider multiple factors such as generation costs, transmission losses, grid security constraints, and environmental requirements to optimize system operation. Under this type of operation, the output allocation of generator units is dynamic and complex, requiring the solution of an optimization problem to determine its optimality.

[0205] Dispatch constraints refer to various technical and operational limitations that must be met during power system dispatching. These constraints include, but are not limited to, minimum / maximum output limits for generator units, ramp rate limits, power flow limitations for transmission lines, voltage stability constraints, reserve capacity requirements, and environmental emission restrictions. Optimization must be performed while satisfying these constraints during actual dispatching and allocation.

[0206] Finally, the third operating type refers to the power distribution pattern of generator units in the power generation system, which is obtained through statistical analysis and modeling based on historical operating data.

[0207] When real-time scheduling information is incomplete or the system operation mode is difficult to describe by an accurate model, the statistical relationship between historical load data and corresponding generator output data and marginal emission factor data (such as linear regression, machine learning model) can be used to predict or infer the current marginal emission factor.

[0208] The linear relationship between load and marginal factor in historical data refers to the mathematical relationship discovered and established through statistical analysis of changes in power system load over a past period and the corresponding marginal emission factor. This relationship can be a simple linear regression model or a more complex nonlinear model, used to predict the marginal emission factor under specific load conditions.

[0209] After determining the operating type of the power generation system based on the generator output data, the corresponding marginal emission factor can be calculated. By applying load change to the repair data, the emission increment corresponding to the load change is calculated. The marginal emission factor is calculated based on the emission increment and load change in the following three ways.

[0210] Method 1: When the operation type is the first operation type in which multiple generator sets are allocated according to a preset ratio, the marginal participation factors corresponding to the load change of multiple generator sets are integrated to obtain the marginal emission factor.

[0211] For example, the marginal participation factor of each unit can be determined and its sum can be normalized. The load change is allocated to each unit according to the marginal participation factor to obtain the unit output change. The output change of each unit is multiplied by the corresponding unit emission factor and then summed to obtain the emission increment. The ratio of the emission increment to the load change is used as the marginal emission factor.

[0212] Method 2: When the operation type is the second operation type with multiple generator sets based on actual scheduling allocation, the marginal emission factor is obtained by solving the scheduling constraints based on the load change.

[0213] For example, a simplified scheduling model can be constructed that includes unit ramp-up constraints, minimum output constraints, and standby constraints. The model can be solved before and after the load change to obtain the baseline emissions and the emissions after the disturbance. The ratio of the difference between the two to the load change is used as the marginal emission factor.

[0214] Method 3: When the operating type is the third operating type allocated by multiple generator sets based on historical data, the marginal emission factor is determined based on the linear relationship between load and marginal factor in the historical data.

[0215] For example, a local linear relationship can be fitted using emission increments and load changes within a historical time window, and parameters can be updated online. The marginal emission factor can then be corrected by combining this with the current power structure to obtain the marginal emission factor.

[0216] The above technical solution improves the accuracy of marginal emission factor calculation by introducing the concept of power generation system operation types and employing targeted marginal emission factor calculation methods based on different operation types. Specifically, after acquiring streaming data from multiple generator units in the power generation system, including the processing data of the corresponding generator units, the system first determines the current operation type of the power generation system based on the output data of these generator units. For example, when the system is operating stably and the generator unit output adjustment follows preset rules, it can be identified as the first operation type. When the system faces sudden load changes or needs real-time economic dispatch, it can be identified as the second operation type. When the system operation mode is complex and historical data is abundant, it can be identified as the third operation type. Once the operation type is determined, subsequent marginal emission factor calculations will be based on that type. If it is determined to be the first operation type, i.e., multiple generator units allocate output according to a preset ratio, the marginal emission factor is calculated by integrating the marginal participation factors corresponding to the load changes of each generator unit. This means that the system considers the emission characteristics of each generator unit and its share in the load response. If the system is identified as operating under the second type, where multiple generator sets allocate power based on actual dispatch, the system will perform optimization based on load changes and generator set dispatch constraints (such as output limits, ramp rates, fuel costs, etc.) to accurately determine the marginal generator sets and their emission factors under all constraints. If the system is identified as operating under the third type, where multiple generator sets allocate power based on historical data, the system will directly infer the current marginal emission factor using pre-established linear relationships between load and marginal factors in historical data or other statistical models. Through this categorized and refined calculation method, this application can more accurately reflect the power system's carbon emission response to load changes under different operating scenarios, thus providing a more reliable marginal emission factor for monitoring carbon emissions from electricity consumption.

[0217] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0218] Based on the same inventive concept, this application also provides a carbon emission monitoring system for electricity consumption to implement the above-mentioned method for real-time monitoring of urban-level electricity consumption carbon emissions based on power system streaming data. The solution provided by this system is similar to the implementation described in the above method. Therefore, the specific limitations of one or more embodiments of the carbon emission monitoring system for electricity consumption provided below can be found in the limitations of the method for real-time monitoring of urban-level electricity consumption carbon emissions based on power system streaming data described above, and will not be repeated here.

[0219] In one exemplary embodiment, such as Figure 5 As shown, a carbon emission monitoring system for electricity consumption is provided, comprising: an acquisition module, an assessment and remediation module, a first calculation module, a second calculation module, and an information output module, wherein:

[0220] The acquisition module is used to acquire streaming data of the power generation system operation;

[0221] The evaluation and repair module is used to identify target anomalous data in streaming data, repair the target anomalous data based on constraint state estimation, and generate repaired data.

[0222] The first calculation module is used to construct a time-varying power flow matrix based on the repaired data, and use the effective net power flow of the time-varying power flow matrix as the allocation coefficient to allocate the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity of the consumption end.

[0223] The second calculation module is used to calculate the emission increment corresponding to the load change by applying the load change to the repair data, and to calculate the marginal emission factor based on the emission increment and the load change.

[0224] The information output module is used to output carbon emission monitoring information based on average carbon intensity and marginal emission factor.

[0225] In one embodiment, the evaluation and repair module may include a first identification unit, a second identification unit, and a repair unit, wherein:

[0226] The first identification unit is used to identify time-series anomalous data in streaming data based on time-series consistency, and to identify residual anomalous data in streaming data based on physical consistency.

[0227] The second identification unit is used to identify target abnormal data in streaming data based on time-series abnormal data and residual abnormal data, as well as data constraints for streaming data.

[0228] The repair unit is used to repair the target abnormal data with the goal of minimizing the repair cost, and obtain repaired data.

[0229] In one embodiment, the streaming data includes at least one of a first type of data, a second type of data, and a third type of data. Based on the repaired data, the first computing module includes a first building unit, a second building unit, and a third building unit, wherein:

[0230] The first building unit is used to construct a time-varying power flow matrix based on the exchanged power between cross-regional data when the streaming data is a first type of data including cross-regional data.

[0231] The second building unit is used to aggregate the line power flow in the same area data and combine it with the network loss term to construct a time-varying power flow matrix when the streaming data is a second type of data including data in the same area.

[0232] The third building unit is used to perform constraint optimization on the topologyless data to obtain the equivalent transmission coefficient when the streaming data is a third type of data including topologyless data, and to construct the time-varying power flow matrix based on the equivalent transmission coefficient.

[0233] In one embodiment, the first computing module further includes a third computing unit, a splitting unit, a carbon tracing unit, and a solving unit, wherein:

[0234] The third calculation unit is used to calculate the carbon emission injection of the power generation system;

[0235] The splitting unit is used to split the carbon emission injection amount using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, so as to obtain the electricity consumption parameters corresponding to multiple consumer ends;

[0236] The tracking unit is used to construct a carbon tracking linear equation based on carbon emission injection and electricity consumption parameters from multiple consumer terminals; wherein the carbon tracking linear equation satisfies the carbon conservation law;

[0237] The solver unit is used to solve the carbon tracing linear equation to obtain the average carbon intensity at the consumer end.

[0238] In one embodiment, the streaming data includes streaming data collected from multiple generator sets in a power generation system, and the streaming data includes generator set processing data corresponding to the multiple generator sets. The second calculation module includes a judgment unit, a first marginal factor calculation unit, a second marginal factor calculation unit, and a third marginal factor calculation unit, wherein:

[0239] The judgment unit is used to determine the operating type of the power generation system based on the generator set output data; wherein the operating type includes a first operating type, a second operating type, and a third operating type;

[0240] The first marginal factor calculation unit is used to integrate the marginal participation factors corresponding to the load change of multiple generator sets when the operation type is the first operation type in which multiple generator sets are allocated according to a preset ratio, to obtain the marginal emission factor.

[0241] The second marginal factor calculation unit is used to solve the scheduling constraints based on the load change when the operation type is the second operation type with multiple generator sets based on actual scheduling allocation, and obtain the marginal emission factor.

[0242] The third marginal factor calculation unit is used to determine the marginal emission factor based on the linear relationship between load and marginal factor in historical data when the operating type is the third operating type allocated to multiple generator sets according to historical data.

[0243] In one embodiment, the carbon emission monitoring system for electricity consumption further includes a monitoring module comprising a fourth computing unit, an anomaly monitoring unit, and an anomaly location unit, wherein:

[0244] The fourth calculation unit is used to calculate the confidence intervals for average carbon intensity and marginal emission factor;

[0245] The anomaly monitoring unit is used to monitor the changes in average carbon intensity and marginal emission factor. When the changes exceed the confidence interval, it outputs an anomaly alarm event.

[0246] Anomaly location unit is used to locate abnormal locations that affect average carbon intensity and marginal emission factor based on abnormal alarm events.

[0247] Each module in the aforementioned carbon emission monitoring system for electricity consumption can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the device in hardware form or independent of it, or stored in the device's memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0248] In one exemplary embodiment, a device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6As shown, the device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores streaming and monitoring data from the power system's operation. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for real-time monitoring of urban-level electricity consumption carbon emissions based on streaming power system data.

[0249] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the device to which the present application is applied. Specific devices may include more or fewer components than those shown in the figure, or may combine certain components, or may have different component arrangements.

[0250] In one exemplary embodiment, an apparatus is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, implements the steps of a method for real-time monitoring of urban-level electricity consumption carbon emissions based on streaming data from an electricity system.

[0251] In one embodiment, a readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for real-time monitoring of carbon emissions from city-level electricity consumption based on streaming data from a power system.

[0252] In one embodiment, a program product is provided, including a computer program that, when executed by a processor, implements the steps of a method for real-time monitoring of city-level electricity consumption carbon emissions based on streaming data from a power system.

[0253] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0254] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0255] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this 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 real-time monitoring of carbon emissions from urban-level electricity consumption based on streaming data of the power system, characterized in that, The method includes: Acquire streaming data of the power generation system operation; Identify target anomalous data in the streaming data, repair the target anomalous data based on constraint state estimation, and generate repaired data; Based on the repaired data, a time-varying power flow matrix is ​​constructed. Using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, the carbon emission injection of the power generation system is allocated to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity of the consumption ends. By applying a load change to the repaired data, the emission increment corresponding to the load change is calculated, and the marginal emission factor is calculated based on the emission increment and the load change. Based on the average carbon intensity and the marginal emission factor, carbon emission monitoring information is output.

2. The method according to claim 1, characterized in that, The steps of identifying target anomalous data in the streaming data and repairing the target anomalous data based on constraint state estimation to generate repaired data include: Based on temporal consistency, identify temporal anomalies in the streaming data; and based on physical consistency, identify residual anomalies in the streaming data. Based on the time-series abnormal data and the residual abnormal data, as well as the data constraints on the streaming data, identify the target abnormal data in the streaming data; With the goal of minimizing the repair cost, the target abnormal data is repaired to obtain the repaired data.

3. The method according to claim 1, characterized in that, The streaming data includes at least one of a first type of data, a second type of data, and a third type of data. The step of constructing a time-varying power flow matrix based on the repaired data includes: When the streaming data is a first type of data that includes cross-regional data, the time-varying power flow matrix is ​​constructed based on the exchanged power between the cross-regional data. When the streaming data is a second type of data that includes data from the same region, the line power flow in the data from the same region is aggregated and combined with the network loss term to construct the time-varying power flow matrix; When the streaming data is a third type of data including topologyless data, the topologyless data is constrained and optimized to obtain equivalent transmission coefficients, and the time-varying power flow matrix is ​​constructed based on the equivalent transmission coefficients.

4. The method according to claim 1, characterized in that, The step of allocating the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side using the effective net power flow of the time-varying power flow matrix as the allocation coefficient, and obtaining the average carbon intensity of the consumption ends, includes: Calculate the carbon emission injection of the power generation system; The carbon emission injection amount is split using the effective net power flow of the time-varying power flow matrix as the allocation coefficient to obtain multiple electricity consumption parameters corresponding to the consumer end. A carbon tracking linear equation is constructed based on the carbon emission injection amount and the electricity consumption parameters of multiple consumer terminals; wherein the carbon tracking linear equation satisfies the carbon conservation law; Solve the carbon tracking linear equation to obtain the average carbon intensity at the consumer end.

5. The method according to claim 1, characterized in that, The streaming data includes streaming data collected from multiple generator sets in the power generation system, and the streaming data includes generator set processing data corresponding to the multiple generator sets. The step of applying a load change to the repaired data, calculating the emission increment corresponding to the load change, and calculating the marginal emission factor based on the emission increment and the load change includes: Based on the generator set output data, the operating type of the power generation system is determined; wherein, the operating type includes a first operating type, a second operating type, and a third operating type; When the operating type is a first operating type in which multiple generator sets are allocated according to a preset ratio, the marginal participation factors corresponding to multiple generator sets and the load change amount are integrated to obtain the marginal emission factor; When the operating type is the second operating type based on the actual scheduling allocation of multiple generator sets, the marginal emission factor is obtained by solving the scheduling constraints according to the load change. When the operating type is the third operating type assigned by multiple generator sets based on historical data, the marginal emission factor is determined based on the linear relationship between load and marginal factor in the historical data.

6. The method according to claim 1, characterized in that, After the step of outputting carbon emission monitoring information based on the average carbon intensity and the marginal emission factor, the method further includes: Calculate the confidence intervals for the average carbon intensity and the marginal emission factor; Monitor the changes in the average carbon intensity and the marginal emission factor, and output an abnormal alarm event when the changes exceed the confidence interval; Based on the abnormal alarm events, locate the abnormal locations that affect the average carbon intensity and the marginal emission factor.

7. A carbon emission monitoring system for electricity consumption, characterized in that, The system includes: The acquisition module is used to acquire streaming data of the power generation system operation; An evaluation and repair module is used to identify target anomalous data in the streaming data, repair the target anomalous data based on constraint state estimation, and generate repaired data. The first calculation module is used to construct a time-varying power flow matrix based on the repaired data, and use the effective net power flow of the time-varying power flow matrix as the allocation coefficient to allocate the carbon emission injection of the power generation system to multiple consumption ends on the electricity consumption side to obtain the average carbon intensity of the consumption ends. The second calculation module is used to calculate the emission increment corresponding to the load change by applying the load change to the repair data, and to calculate the marginal emission factor based on the emission increment and the load change. The information output module is used to output carbon emission monitoring information based on the average carbon intensity and the marginal emission factor.

8. A device comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A program product comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images