An information processing system for regional carbon emission intelligent deduction and early warning

Abstract

The present application relates to carbon emission intelligent management technical field, specifically to a kind of information processing system for regional carbon emission intelligent deduction and early warning, it includes carbon source coupling identification extraction abnormal point, evolution path determination screening suitable section, migration cause calculation output leading factor, early warning response switching identification lagging area, model reconstruction trial calculation optimization configuration graph group.The present application integrates multi-source data such as energy consumption, industrial structure and land use, constructs time-space coupling mutation analysis mechanism, identifies the collaborative deviation and continuous migration section of carbon emission in grid, introduces traffic flow, energy consumption and carbon flow density gradient linkage calculation, extracts leading factor sorting, combines response delay and mutation duration interval check deduction result, dynamically matches model space level and time step, ensures deduction stable under complex scenario, supports early warning triggering and model linkage adjustment.

Description

Technical Field

[0001] This invention relates to the field of intelligent carbon emission management technology, and in particular to an information processing system for intelligent prediction and early warning of regional carbon emissions. Background Technology

[0002] Currently, there is an urgent need to build an information processing system capable of integrating multi-source heterogeneous data, supporting dynamic evolution simulation, and possessing intelligent early warning capabilities to enhance the scientific rigor and foresight of regional carbon governance. However, existing carbon emission information processing technologies still have significant shortcomings in terms of regional-scale extrapolation capabilities, multi-factor coupled modeling, and proactive early warning mechanisms. A search revealed a carbon emission information processing method, apparatus, and equipment with publication number CN118504424B. This patent constructs a carbon emission monitoring model by acquiring scanned images of a target scene and trains a prediction model based on historical carbon emission data to improve the accuracy of carbon source emission data. However, this technical solution mainly focuses on the static identification and location prediction of a single scene or specific carbon source, lacking systematic integration and dynamic interaction modeling of multiple industries and factors of carbon emission sources within a region, making it difficult to support regional-scale carbon emission evolution extrapolation. Furthermore, its prediction model lacks an embedded early warning mechanism, failing to automatically trigger tiered early warnings based on extrapolation results, thus limiting its proactive prevention and control capabilities in regional carbon management.

[0003] A search revealed a patent, CN120106597A, which discloses a method and system for long-term prediction of regional net carbon emissions using a multi-factor hybrid judgment approach. This patent utilizes a Convolutional Long Short-Term Memory (ConvLSTM) network combined with an improved particle swarm optimization algorithm to predict regional net carbon emissions over the long term, aiming to improve the accuracy of long-term predictions. However, this approach focuses on predicting the long-term trend of macro-level total emissions. Its model input relies on pre-screened influencing factors and does not fully consider the spatial heterogeneity within the region and the dynamic migration characteristics of carbon flows. Furthermore, the system only provides prediction results and lacks real-time monitoring, causal tracing, and intelligent early warning functions for abnormal emission scenarios, failing to meet the actual needs of closed-loop management of regional carbon emissions through "monitoring-deduction-early warning-response."

[0004] The aforementioned problems indicate that existing technologies for regional carbon emission information processing generally suffer from limitations such as limited scenario coverage, weak dynamic simulation capabilities, lack of spatial coupling relationships, and inadequate early warning mechanisms. Therefore, this invention proposes an information processing system for intelligent simulation and early warning of regional carbon emissions. The system aims to construct a comprehensive platform that integrates multi-source data, supports spatiotemporal dynamic simulation, and possesses intelligent early warning and decision support capabilities. This platform will enable accurate perception, scientific simulation, and proactive early warning of regional carbon emissions, providing strong technical support for regional low-carbon governance. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an information processing system for intelligent extrapolation and early warning of regional carbon emissions.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an information processing system for intelligent regional carbon emission prediction and early warning, the system comprising:

[0007] The carbon source coupling identification module acquires energy consumption data streams, dynamic indicators of industrial structure and land use change maps within the region, calculates the frequency of coordinated mutations of the three in the spatiotemporal dimension and cross-compares them, extracts high-frequency coupling offset points, and generates a list of carbon emission coupling anomaly trigger points.

[0008] The evolution path determination module analyzes the synchronicity between the rate of change of traffic flow intensity and the sudden change point of industrial energy consumption based on the points with multiple periodic repeat records in the list of carbon emission coupling anomaly trigger points, determines whether the dynamic migration process of carbon flow is continuous, and generates a set of spatiotemporal segments suitable for deduction and modeling.

[0009] The migration motive calculation module calls the segments marked as high carbon flow interaction regions in the set of spatiotemporal segments suitable for modeling and inference, measures their carbon flow density gradient, industrial linkage intensity slope and energy structure change curve, and calculates the overlapping area of ​​the three change rates on the spatiotemporal grid, and outputs the carbon flow dynamic migration dominant factor ranking map.

[0010] The early warning response switching module checks the current carbon emission response delay duration and carbon flow mutation duration in the corresponding region based on the first-rank factor in the carbon flow dynamic migration dominant factor ranking map, extracts the region identification code, confirms whether the graded early warning triggering conditions are met, and generates a list of region segments for the optimization and simulation model.

[0011] As a further aspect of the present invention, the list of abnormal carbon emission coupling trigger points includes energy consumption surge points, critical segments of industrial structure transition, and high-frequency areas of land use conversion; the set of spatiotemporal segments suitable for modeling includes abnormal carbon flow density growth zones, synchronous areas of industrial energy consumption abrupt changes, and segments of multi-source factor response mismatch; the ranking map of dominant factors for dynamic carbon flow migration includes transportation network driving factors, industrial linkage transmission factors, and energy structure transformation factors; and the list of regional segments of the model to be optimized includes response delay regions, periodic out-of-step grids, and action lag units.

[0012] As a further aspect of the present invention, the carbon source coupling identification module includes:

[0013] The multi-source data extraction submodule acquires continuous periodic datasets of regional power grid load time series data, industrial enterprise output dynamic sequences, and land cover type change layers interpreted from remote sensing images. It calls the differences of each type of data within adjacent time steps to construct energy consumption difference sequences, industrial structure change interval sets, and land use conversion trend sequences. For the time steps corresponding to the industrial structure change intervals, it extracts spatiotemporal segments in the energy consumption difference sequences and land use conversion trend sequences to generate three sets of multi-source element change interval sets.

[0014] The spatiotemporal collaborative offset submodule calls the energy consumption difference sequence and the corresponding land use conversion trend sequence in all spatiotemporal units of the three sets of multi-source element change segment sets, calculates the deviation between the land use conversion rate of adjacent units before and after the peak energy consumption difference unit and the average conversion rate in that segment, and constructs a coupling offset judgment index based on the spatiotemporal unit of the intersection of the deviation and the industrial structure change interval to obtain the spatiotemporal collaborative offset frequency sequence.

[0015] The coupling trigger point identification submodule calls all frequency data in the spatiotemporal coordinated offset frequency sequence, calculates the product of the increase ratio and the directional stability ratio based on the difference between the change trend of each frequency and the change in its adjacent spatiotemporal units, records the spatiotemporal units whose product exceeds the coupling offset trigger threshold, extracts the synchronous fluctuation values ​​of the three elements of energy consumption, industrial structure and land use in the spatiotemporal unit, and generates a list of carbon emission coupling anomaly trigger points.

[0016] As a further aspect of the present invention, the evolution path determination module includes:

[0017] The repeating anomaly extraction submodule obtains the cycle number and anomaly marking status of all points in the carbon emission coupling anomaly triggering point list, filters point numbers with repeating anomaly records, extracts the marking frequency and distribution interval of the corresponding point in each cycle, calculates the anomaly interval difference in consecutive cycles and compares it with the cycle judgment threshold, establishes a point index set that meets the cycle frequency condition, and generates a multi-cycle anomaly point set.

[0018] The multi-dimensional sequence interleaving submodule calls the traffic flow sequence and industrial energy consumption sequence corresponding to all point numbers in the multi-period anomaly point set, calculates the rate of change of the difference between the two sequences, extracts the fluctuation frequency change sequence, compares the peak position of the two types of sequence changes with the extreme period of fluctuation, identifies the location index of the interleaving and records the interleaving frequency interval, and obtains the interleaving location interval value.

[0019] The deduction segment selection submodule calls the frequency interlacing time periods recorded in the interlacing position interval values, compares the occurrence positions of abrupt change segments in the carbon flux density sequence, selects the time periods in which the intersection position and density abrupt change point are successfully matched, calculates the time interval value between two abrupt change points in each time period and compares it with the deduction coherence identification threshold, selects the numbered segments that meet the interval conditions and establishes a complete path index, and generates a set of spatiotemporal segments suitable for deduction modeling.

[0020] As a further aspect of the present invention, the migration motive calculation module includes:

[0021] The intensity slope extraction submodule calls the path number marked as a high carbon flow interaction area in the set of spatiotemporal segments suitable for modeling and extrapolation, extracts the continuous time series value of the industrial association intensity corresponding to each path number, calculates the intensity change slope between each two adjacent sampling points, constructs a slope sequence in time order, extracts the segment number and its change amount corresponding to the maximum slope change amplitude, and generates the association intensity slope change value.

[0022] The gradient synchronization identification submodule calls the region number recorded in the associated intensity slope change value, extracts the corresponding traffic network flow change sequence and energy structure change sequence, calculates the synchronization segment number between the carbon flow density decrease value and the energy structure value, extracts the peak interval of the carbon flow density gradient and the segment number with the largest continuous change in the energy structure curve, and obtains the synchronization gradient change interval.

[0023] The overlapping area sorting submodule calls the carbon flow density gradient segment, energy structure change segment, and slope segment in the correlation intensity slope change value in the synchronous gradient change interval. It performs an intersection operation on the relative positions of the three segments on the spatiotemporal grid, calculates the multidimensional overlapping area value formed by the change rate of the intersection segment in three dimensions, and rearranges the segment numbers according to the area value to generate a sorting map of the dominant factors of carbon flow dynamic migration.

[0024] As a further aspect of the present invention, the early warning response switching module includes:

[0025] The first-leader identification submodule calls the sorting first-leader number in the sorting map of the dominant factors of carbon flow dynamic migration, extracts the regional early warning response time series and carbon flow mutation time series data corresponding to the factor, calculates the time difference between the early warning signal issuance time and the carbon flow mutation start point, and compares the difference with the regional matching early warning response benchmark value to determine whether it is in a response delay state and generates an early warning response deviation interval value.

[0026] The response delay determination submodule calls the region number recorded in the warning response deviation interval value, extracts the warning action setting time value and the actual response start time value in the corresponding carbon emission cycle, calculates the setting response start lag time and compares it with the response lag determination threshold, marks the region identification code with lag problem, and obtains the response lag identification list.

[0027] The early warning potential assessment submodule calls all area identification codes in the response lag identification list, compares their frequency of occurrence in the early warning control record with the response intensity sequence of the corresponding segment, filters out area numbers whose frequency exceeds the repeated identification threshold and whose response intensity value is lower than the response benchmark value, extracts the spatiotemporal segment information to which the corresponding area number belongs and summarizes it into an index list, and generates a list of area segments to be optimized inference models.

[0028] As a further aspect of the present invention, the system further includes:

[0029] The model reconstruction trial module calls the region name in the list of region segments of the model to be optimized and replaces the spatial resolution level and time step granularity in the model accordingly. It tests the stability of the model combination under the carbon flow mutation scenario, selects compliant combinations, and obtains the carbon emission inference model optimization configuration map group.

[0030] The optimized configuration group of the carbon emission extrapolation model includes a spatial granularity adjustment unit, a time step configuration module, and a coupling mechanism integrated structure.

[0031] As a further aspect of the present invention, the model reconstruction trial calculation module includes:

[0032] The model parameter replacement submodule calls the region name recorded in the region segment list of the model to be optimized, extracts the replacement model number and spatial resolution setting value corresponding to each region, obtains the time step level and initial offset parameter in the region's inference process, performs mapping verification on the combination of resolution setting value and time step and reconstructs its replacement model list, and generates the model replacement parameter group.

[0033] The mutation scenario verification submodule extracts the inference error sequence and peak deviation position recorded in the carbon flow mutation event for the corresponding region based on the model number already labeled in the model replacement parameter group. It compares the peak error value with the upper and lower limits of the inference accuracy set by the model to which the region belongs, filters the model numbers within the accuracy range, and obtains a list of accuracy-matching models.

[0034] The configuration integration judgment submodule calls the spatial resolution setting value and time step level corresponding to each model number in the precision matching model list, constructs a combination mapping relationship, and performs deduplication processing on the repetitive offset combinations. It extracts the effective model units under each combination and compares them with the original model distribution map for configuration alignment, establishes a usable model combination library, and generates a carbon emission extrapolation model optimization configuration map group.

[0035] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0036] In this invention, by integrating multi-source information such as regional energy consumption, industrial structure changes, and land use evolution, a spatiotemporal coupled abrupt change analysis mechanism is constructed. This mechanism can identify the collaborative offset characteristics of carbon emission elements in a spatial grid and capture continuous migration segments. In this process, the linkage calculation of traffic flow intensity, industrial energy consumption changes, and carbon flow density gradient is introduced to form the ranking results of the dominant driving factors. The extrapolation results are then verified by combining the carbon emission response delay and the abrupt change duration interval. The spatial hierarchy and time step of the model are dynamically matched to ensure that the regional carbon emission evolution extrapolation maintains stable output under complex scenarios, while supporting the linkage operation of early warning triggering and model adjustment. Attached Figure Description

[0037] Figure 1 This is a system flowchart of the present invention;

[0038] Figure 2 This is a flowchart illustrating the acquisition process of the carbon source coupling identification module of the present invention.

[0039] Figure 3 This is a flowchart illustrating the acquisition process of the evolution path determination module of the present invention.

[0040] Figure 4 This is a flowchart illustrating the acquisition process of the migration motive calculation module of the present invention.

[0041] Figure 5 This is a flowchart illustrating the acquisition process of the early warning response switching module of the present invention.

[0042] Figure 6 This is a flowchart illustrating the acquisition process of the model reconstruction trial calculation module of this invention. Detailed Implementation

[0043] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0044] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0045] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0046] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0047] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0048] Please see Figure 1 This invention provides a technical solution: an information processing system for intelligent prediction and early warning of regional carbon emissions, the system comprising:

[0049] The carbon source coupling identification module acquires energy consumption data streams, dynamic indicators of industrial structure and land use change maps within the region, calculates the frequency of coordinated mutations of the three in the spatiotemporal dimension and cross-compares them, extracts high-frequency coupling offset points, and generates a list of carbon emission coupling anomaly trigger points.

[0050] The evolution path determination module analyzes the synchronicity between the rate of change of traffic flow intensity and the sudden change point of industrial energy consumption based on the points with multiple periodic repeat records in the list of carbon emission coupling anomaly trigger points, determines whether the dynamic migration process of carbon flow is continuous, and generates a set of spatiotemporal segments suitable for deduction and modeling.

[0051] The migration dynamics calculation module calls the segments marked as high carbon flow interaction regions from the set of spatiotemporal segments suitable for simulation and modeling, measures their carbon flow density gradient, industrial linkage intensity slope and energy structure change curve, and calculates the overlapping area of ​​the three change rates on the spatiotemporal grid, outputting the carbon flow dynamic migration dominant factor ranking map.

[0052] The early warning response switching module checks the current carbon emission response delay and carbon flow mutation duration in the region corresponding to the first factor in the carbon flow dynamic migration dominant factor ranking map, determines the lag in early warning action, extracts the region identification code, confirms whether the tiered early warning triggering conditions are met, identifies whether there is a disconnect between the output of the simulation model and the actual monitoring data, identifies the response lag section by comparing the early warning triggering frequency and the carbon flow rate change interval, and generates a list of region segments for the simulation model to be optimized.

[0053] The model reconstruction trial module calls the region names in the list of region segments of the model to be optimized, replaces the spatial resolution level and time step granularity in the model accordingly, tests the stability of the model combination under the scenario of sudden carbon flow, selects compliant combinations, and obtains the optimized configuration map group of carbon emission simulation models.

[0054] The list of abnormal carbon emission coupling trigger points includes energy consumption surge points, critical segments of industrial structure transition, and high-frequency areas of land use conversion. The set of spatiotemporal segments suitable for modeling includes abnormal carbon flow density growth zones, synchronous areas of industrial energy consumption abrupt changes, and segments of multi-source factor response mismatch. The ranking map of dominant factors in carbon flow dynamic migration includes transportation network driving factors, industrial linkage transmission factors, and energy structure transformation factors. The list of regional segments for the model to be optimized includes response delay regions, periodic out-of-step grids, and action lag units. The optimized configuration map set for the carbon emission model includes spatial granularity adjustment units, time step configuration modules, and coupling mechanism integrated structures.

[0055] Please see Figure 2 The carbon source coupling identification module includes:

[0056] The multi-source data extraction submodule acquires continuous periodic datasets of regional power grid load time series data, industrial enterprise output dynamic sequences, and land cover type change layers interpreted from remote sensing images. It calls the differences of each type of data within adjacent time steps to construct energy consumption difference sequences, industrial structure change interval sets, and land use conversion trend sequences. For the time steps corresponding to the industrial structure change intervals, it extracts spatiotemporal segments in the energy consumption difference sequences and land use conversion trend sequences to generate three sets of multi-source element change interval sets.

[0057] The system acquires time-series data (unit: MW) of regional power grid load within a city's administrative area, comprising five time steps; dynamic sequences of industrial enterprise output value (unit: 100 million yuan) within the same time period; and continuous periodic data (unit: km²) of land cover type change layers obtained from remote sensing image interpretation. The processing unit calls adjacent time steps from each data type and constructs three sequences by subtracting the previous time step from the next. For the power grid load data, the first difference is... The second difference is Based on this, the energy consumption difference sequence is calculated. Similarly, the land use conversion trend sequence can be obtained. For the output value sequence, the processing unit compares adjacent values ​​point by point, identifies all continuous intervals where the subsequent point is greater than the previous point, and performs this process in the sequence. middle, , , ,and Therefore, the time step index of the first increasing interval is The second one is By traversing the entire output value sequence in this way, we obtain the set of industrial structure change intervals, which in this example is the set. For the time steps corresponding to these variation intervals, spatiotemporal segments are extracted from the energy consumption difference sequence and the land use conversion trend sequence, specifically by extracting values ​​from index 0 to 1, and thus extracting the energy consumption difference sequence. By extracting the land use conversion trend sequence, we can obtain... By combining these two extracted sequences with the original output value increasing segments, three sets of multi-source factor change segments are generated.

[0058] The spatiotemporal collaborative offset submodule calls the energy consumption difference sequence and the corresponding land use conversion trend sequence in all spatiotemporal units of the three sets of multi-source element change segment sets. It calculates the deviation between the land use conversion rate of the adjacent units before and after the peak energy consumption difference unit and the average conversion rate in that segment. Based on the intersection spatiotemporal unit of the deviation and the industrial structure change interval, it constructs a coupled offset judgment index to obtain the spatiotemporal collaborative offset frequency sequence.

[0059] The energy consumption difference sequence and the corresponding land use conversion trend sequence within all spatiotemporal units of three sets of multi-source factor change segments are retrieved, specifically for energy consumption difference segments. and land use conversion trend sections The processing unit identifies peak units within the energy consumption difference range. Here, a peak unit is defined as the maximum absolute value within the range. In the above, the absolute maximum value is 20, and its corresponding value is 20 with an index of 0. Then, the processing unit calculates the land use conversion rate of the adjacent units before and after this peak unit, namely the units with indices 0 (no predecessor) and 1. In the land use conversion trend segment, the land use conversion rate corresponding to index 1 is... Simultaneously, the average conversion rate within the land use conversion trend segment is calculated by summing all values ​​within the segment and dividing by the number of values. Calculate the deviation between the land use conversion rate of the adjacent units after the aforementioned peak unit and the average conversion rate. For the unit at index 1, the deviation is: The processing unit is based on this deviation and the range of industrial structure changes, i.e., the time step index is... The intersection of spatiotemporal units is used to construct a coupling offset judgment index. This process compares each deviation with a coupling offset threshold, which is set based on the standard deviation of land use conversion rate under stable conditions in historical data. For example, if the standard deviation is 0.15, then the threshold is... ,because If a significant offset is detected, the frequency is recorded as 1. If the offset does not exceed the threshold, the frequency is recorded as 0. All judgment results are combined in chronological order to obtain the spatiotemporal coordinated offset frequency sequence.

[0060] The coupling trigger point identification submodule calls all frequency data in the spatiotemporal collaborative offset frequency sequence. Based on the change trend of each frequency and the difference in change within its adjacent spatiotemporal units, it calculates the product of the increase ratio and the directional stability ratio. Spatiotemporal units where the product exceeds the coupling offset trigger threshold are recorded. The synchronous fluctuation values ​​of the corresponding three elements—energy consumption, industrial structure, and land use—are extracted in the spatiotemporal unit to generate a list of carbon emission coupling anomaly trigger points.

[0061] All frequency data from the spatiotemporal cooperative migration frequency sequence are retrieved. To demonstrate the calculation process, a longer sequence is introduced here. The processing unit calculates based on the difference between the frequency variation trend and the variation in its adjacent spatiotemporal units. This process starts from the second point in the sequence. For the point at index 1 (value 1), the variation compared to the previous point is... The change between it and the next point is The processing unit calculates the increase ratio, defined as the ratio of the current point's value to the average of the two points before and after it. If the average is zero, the ratio is a large number. For the point at index 1, the increase ratio is... Then calculate its directional stability ratio, defined as Then calculate their product. The processing unit compares the product with a preset coupling offset trigger threshold. This threshold is set based on the statistical distribution of the product value under a large number of normal and abnormal operating conditions, selecting a quantile that can identify 95% of abnormal operating conditions as the threshold, for example, set to 1.5. Since 0 does not exceed 1.5, it is not recorded in the spatiotemporal unit of index 1, and the processing continues to the point of index 4 (value 2), whose preceding point is 0 and the following point is 3, with the changes being respectively and The growth rate ratio is The directional stability ratio is The product is 0. When processing the point at index 5 (value 3), its preceding and following points are 2 and 1, respectively, with a change of [missing value]. and The growth rate ratio is The directional stability ratio is (The absolute value is 2, but the directional stability ratio is defined as non-negative, so it is 0). The product is 0. After the calculation and comparison of the entire sequence are completed, all spatiotemporal units whose products exceed the coupling offset trigger threshold are recorded, and the synchronous fluctuation values ​​of the three elements of the corresponding regional power grid load, industrial enterprise output value and land cover type in the spatiotemporal unit are extracted to generate a list of carbon emission coupling anomaly trigger points.

[0062] Please see Figure 3 The evolution path determination module includes:

[0063] The repeating anomaly extraction submodule obtains the cycle number and anomaly marking status of all points in the carbon emission coupling anomaly triggering point list, filters point numbers with repeating anomaly records, extracts the marking frequency and distribution interval of the corresponding point in each cycle, calculates the anomaly interval difference in consecutive cycles and compares it with the cycle judgment threshold, establishes a point index set that meets the cycle frequency condition, and generates a multi-cycle anomaly point set.

[0064] The cycle number and anomaly marking status of all points in the carbon emission coupling anomaly trigger point list are obtained, as shown in Table 1. The processing unit filters out point numbers with repeated anomaly records. In Table 1, point numbers A3 and B7 are marked as anomaly in multiple cycles and are therefore selected. The number of times these points are marked and their distribution intervals in each cycle are extracted. For point A3, it is marked in cycles 1, 3, and 5, with a marking frequency of 3 and distribution intervals of 2 cycles and 2 cycles respectively. For point B7, it is marked in cycles 2 and 4, with a marking frequency of 2 and a distribution interval of 2 cycles. Then, the difference in anomaly intervals within consecutive cycles is calculated. For point A3, its interval sequence is as follows: The interval difference is For point B7, there is only one interval and no difference can be calculated. These interval differences are compared with a period judgment threshold. The threshold is set to identify the regularity of anomaly occurrence. Based on the study of the evolution cycle of regional carbon emission anomalies, the threshold is set to 0, which means that only points with completely consistent anomaly intervals (i.e., periodic stability) are considered to meet the conditions. Since the interval difference of point A3 is 0, it meets the condition of being less than or equal to the period judgment threshold. Therefore, it is included in the point index set that meets the periodicity frequent occurrence condition. Point B7 cannot be judged due to insufficient data and is not included. A multi-period anomaly point set is generated.

[0065] Table 1: Trigger Point Anomaly Marking Status Table

[0066] A3 1 0 1 0 1 B7 0 1 0 1 0 C2 1 0 0 0 0

[0067] As shown in Table 1, this table records the abnormal marking status of each trigger point in different periods, where the mark "1" indicates an abnormality.

[0068] The multidimensional sequence interleaving submodule calls the traffic flow sequence and industrial energy consumption sequence corresponding to all point numbers in the multi-period anomaly point set, calculates the rate of change of the difference between the two sequences, extracts the fluctuation frequency change sequence, compares the peak position of the two types of sequence changes with the extreme period of fluctuation, identifies the location index of the interleaving and records the interleaving frequency interval, and obtains the interleaving location interval value.

[0069] The system retrieves the traffic flow sequence and industrial energy consumption sequence corresponding to all point numbers in the multi-period anomaly point set, specifically the traffic flow sequence collected from point A3 over three consecutive periods. (Unit: veh / h) and industrial energy consumption series (Unit: GWh) The processing unit calculates the rate of change of the difference between the two sequences, i.e., the rate of change of traffic-energy consumption coupling. First, it calculates the point-by-point difference sequence. Then calculate the rate of change sequence of adjacent points of the difference sequence. The process involves extracting the fluctuation frequency variation sequence of the rate sequence. This is achieved by performing a Fourier transform on the rate sequence and identifying the changes in the dominant frequency components in the frequency domain. It is assumed that the extracted fluctuation frequency variation sequence is... The processing unit compares the peak positions and extreme fluctuation periods of these two types of sequences, namely the difference rate of change sequence and the fluctuation frequency change sequence. In the difference rate of change sequence, the peak position (maximum absolute value) is at index 2 with a value of 285. In the fluctuation frequency change sequence, the peak position is also at index 1 with a value of 1.2. The processing unit identifies that these two peak positions do not coincide in terms of time index, but at index 1, the difference rate of change is -190 and the fluctuation frequency is 1.2. The two change in opposite directions and with significant amplitudes. This phenomenon is determined to be an interleaving, and the position index 1 where the interleaving occurs and the corresponding frequency interval, i.e., 1.2 Hz, are recorded to obtain the interleaving position interval value.

[0070] The deduction segment selection submodule calls the frequency interlacing time periods recorded in the interlacing position interval values, compares the occurrence positions of the abrupt change segments in the carbon flux density sequence, selects the time periods in which the intersection position and density abrupt point are successfully matched, calculates the time interval value between two abrupt points in each time period and compares it with the deduction coherence identification threshold, selects the numbered segments that meet the interval conditions and establishes a complete path index, and generates a set of spatiotemporal segments suitable for deduction modeling.

[0071] The processing unit retrieves the frequency interlacing time intervals recorded in the interlacing position interval values, specifically index 1 and its associated frequency interval of 1.2 Hz. It then compares these intervals with the locations of abrupt change segments in an independently acquired regional carbon flux density sequence. This carbon flux density sequence is... The abrupt change segments are identified by calculating the rate of change between adjacent points. For example, at index 2, the change is... The change is much greater than the changes in other locations (such as...). Therefore, index 2 is marked as a mutation point. The processing unit filters out the time intervals where the crossover position (index 1) and the density mutation point (index 2) successfully match. Here, a successful match is defined as the difference between their time indices being within a preset small range, such as one time unit. If the conditions are met, the match is considered successful. The matched time period is the segment from index 1 to 2. Then, the time interval between two mutation points within this time period is calculated. If there are other mutation points before and after this segment, such as a small mutation point at index 0, the time interval between index 0 and index 2 is calculated to be 2 time units. This interval value is compared with a deduction coherence identification threshold. The threshold is set based on the minimum stable observation window required for regional carbon flow dynamic deduction. The value is determined to be 2 time units through experiments. Since it is equal to 2, the segment meets the interval condition. The processing unit uses the segment number, such as R12, to create a complete path index containing start and end indices, generating a set of spatiotemporal segments suitable for deduction modeling.

[0072] Please see Figure 4 The migration motive calculation module includes:

[0073] The intensity slope extraction submodule calls the path number marked as a high carbon flow interaction area in the set of spatiotemporal segments suitable for inference and modeling, extracts the continuous time series value of the industrial association intensity corresponding to each path number, calculates the intensity change slope between each two adjacent sampling points, constructs the slope sequence in time order, extracts the segment number and its change amount corresponding to the maximum slope change amplitude, and generates the association intensity slope change value.

[0074] The path numbers marked as high-carbon flow interaction regions in the set of suitable spatiotemporal segments for modeling are invoked, specifically the continuous time-series values ​​of industrial association intensity for segment R12 within the time period of index 1-2. (Dimensionless), the processing unit calculates the slope of the intensity change between every two adjacent sampling points. For the point pair of 0.75 and 0.82, the slope is... Slope sequence is constructed in chronological order. Extract the segment number and its change amount corresponding to the maximum slope change amplitude in the slope sequence. The maximum amplitude is the maximum absolute value, which is 0.07. The original intensity change segment corresponding to it is the segment with index 1 to 2. The change amount is the slope value. Record the segment number (R12 in this case) and its corresponding change amount to generate the associated intensity slope change value.

[0075] The gradient synchronization identification submodule calls the region number recorded in the associated intensity slope change value, extracts the corresponding traffic network flow change sequence and energy structure change sequence, calculates the synchronization segment number between the carbon flow density decrease value and the energy structure value, extracts the peak interval of carbon flow density gradient and the segment number of the largest continuous change in the energy structure curve, and obtains the synchronization gradient change interval.

[0076] Retrieve the region number R12 recorded in the correlation strength slope change value, and extract the traffic network flow change sequence corresponding to the same time period, i.e., index 1-2. veh / h and energy structure change series (Fossil energy share) The processing unit calculates the synchronous segment number between the carbon flux density decrease value and the energy structure value. This process is accomplished by identifying continuous segments with the same trend (increase or decrease) in the two sequences. In the carbon flux density sequence, index 1 to 2 is the rising segment (0.48→0.62), and in the energy structure sequence, index 1 to 2 is the rising segment (0.35→0.40). The two trends are synchronous, so their sequence number is recorded as [1,2]. The peak interval of the carbon flux density gradient is extracted, that is, the point with the largest absolute value in the carbon flux density difference sequence, which is the segment corresponding to index 2. The segment number with the largest continuous change in the energy structure curve is extracted, that is, the point with the largest absolute value in the energy structure change sequence, which is the segment corresponding to index 2. The two segment numbers are combined to obtain the synchronous gradient change interval.

[0077] The overlapping area sorting submodule calls the carbon flow density gradient segment, energy structure change segment, and slope segment in the slope change value of the correlation intensity in the synchronous gradient change interval. It performs an intersection operation on the relative positions of the three segments on the spatiotemporal grid, calculates the multidimensional overlapping area value formed by the change rate of the intersection segment in three dimensions, and rearranges the segment numbers according to the size of the area value to generate the carbon flow dynamic migration dominant factor sorting map.

[0078] The processing unit calls upon the carbon flux density gradient segment (index 2), the energy structure change segment (index 2), and the slope segment (index 1-2) recorded in the correlation strength slope change values ​​identified within the synchronous gradient change interval. The processing unit performs an intersection operation on the relative positions of these three segments on the spatiotemporal grid. Since the time indices of the three segments all cover index 2, their intersection is the time unit represented by index 2. Then, it calculates the multidimensional overlapping area value formed by the change rate values ​​of the intersection segment in three dimensions: carbon flux density gradient, energy structure change rate, and industry correlation strength slope. Specifically, the value is 0.14 for the carbon flux density difference. The energy structure change rate is 0.05, and the industrial linkage intensity slope is 0.07. The area value is calculated as the product of the absolute values ​​of the three change rates, i.e. If multiple such overlapping segments exist, for example, if the area value of another segment is calculated to be 0.00062, the processing unit pairs all segment numbers with their corresponding area values ​​to obtain... These segment numbers were then reordered according to their area values ​​from largest to smallest, and the sorting result is as follows: A ranking map of the dominant factors in the dynamic migration of carbon flows was generated.

[0079] Please see Figure 5 The early warning response switching module includes:

[0080] The first factor identification submodule calls the sorting first factor number in the carbon flow dynamic migration dominant factor sorting map, extracts the regional early warning response time series and carbon flow mutation time series data corresponding to the factor, calculates the time difference between the early warning signal issuance time point and the carbon flow mutation start point, and compares the difference with the regional matching early warning response benchmark value to determine whether it is in a response delay state and generates the early warning response deviation interval value.

[0081] The processing unit retrieves the region corresponding to R23, the first-ranked factor in the carbon flow dynamic migration dominant factor ordination map, and extracts the warning response time series and carbon flow mutation time series data for that region within the corresponding time period, as shown in Table 2. The processing unit calculates the time difference between the warning signal issuance time and the carbon flow mutation initiation point. The warning signal is issued at T=3, and the carbon flow begins to mutate at T=1 (density jumps from 0.50 to 0.65). The time difference between the two is... The processing unit calculates the difference between the calculated difference of 2 hours and compares it with a baseline value for early warning response matching the region. This baseline value is the average response delay obtained through statistical analysis of the region's response time in historical emergency drills, set at 1.5 hours, with an additional 10% tolerance, i.e., 1.65 hours. Since the calculated difference of 2 hours is greater than the baseline value, the processing unit determines that the region is currently in a state of response delay, records the region's number and the calculated response deviation, and generates an early warning response deviation range value.

[0082] Table 2: Early Warning Response and Carbon Flow Sudden Change Data

[0083] 0 0 0.50 1 0 0.65 2 0 0.68 3 1 0.70 4 1 0.65

[0084] As shown in Table 2, this table records the data on the changes in the warning signal status and the corresponding carbon flow density over time.

[0085] The response delay determination submodule calls the area number recorded in the early warning response deviation interval value, extracts the early warning action setting time value and the actual response start time value in the corresponding carbon emission cycle, calculates the lag time of the set response start and compares it with the response lag determination threshold, marks the area identification code with lag problem, and obtains the response lag identification list.

[0086] The system retrieves the region number R23 recorded in the early warning response deviation interval value and extracts the corresponding early warning action set time value within the carbon emission cycle from the emergency control parameter table for that region. For example, 2 hours represents the maximum permissible time from monitoring a sudden change to issuing an early warning, as specified in the emergency plan. Simultaneously, it retrieves the actual response start time value, i.e., the early warning signal issuance time T=3 identified in the previous steps. The processing unit calculates a set response start lag time, which is obtained by modulo the actual response start time against the early warning action set time. The result indicates that the actual response start time lags behind the theoretical response period by an integer multiple of 1 hour. Then, the processing unit compares this lag time with a preset response lag judgment threshold. This threshold is set to identify small but continuous response lags caused by data transmission delays or decision process redundancy. Its value is set to 30% of the warning action setting time, i.e., 0.6 hours, based on the performance analysis of the information system in this area. Since the calculated lag time of 1 hour is greater than the threshold, the processing unit marks the identification code R23 of this area as having a lag problem, and obtains the response lag identification list.

[0087] The early warning potential assessment submodule calls all area identification codes in the response lag identification list, compares their frequency of occurrence in the early warning control record with the response intensity sequence of the corresponding segment, filters out area numbers whose frequency exceeds the repeated identification threshold and whose response intensity value is lower than the response benchmark value, extracts the spatiotemporal segment information to which the corresponding area number belongs and summarizes it into an index list, and generates a list of area segments to be optimized inference models.

[0088] The system retrieves the identification codes (R23) for all regions in the response lag identification list and calculates the frequency with which a region has been marked as lagging over the past 10 operating cycles from its long-term early warning control records (e.g., 4 times). Simultaneously, it extracts the response intensity sequence for these lagging segments, e.g., an average intensity value of 0.65 (normalized value, 1 for complete response). The processing unit then filters out regions whose frequency exceeds the repeat identification threshold and whose response intensity value is lower than the response baseline value. The repeat identification threshold is set to filter out occasional lag phenomena; its value is based on statistical process control theory and is set to [value missing]. The total observation period is 30%, or 3 times. The response baseline value represents the minimum intensity of the effective response in the region, which is set to 0.70. The processing unit judges that the frequency of R23 occurring 4 times exceeds the threshold of 3 times. Secondly, the response intensity value of R23, 0.65, is lower than the baseline value of 0.70. Both conditions are met, so the region R23 is selected. The spatiotemporal segment information to which the region number belongs is extracted. For example, it belongs to the urban "industrial-transportation" composite functional area. This information is summarized to form an index list, generating a list of regional segments to be optimized inference models.

[0089] Please see Figure 6 The model reconstruction trial calculation module includes:

[0090] The model parameter replacement submodule calls the region name recorded in the region segment list of the model to be optimized, extracts the replacement model number and spatial resolution setting value corresponding to each region, obtains the time step level and initial offset parameter in the region's inference process, performs mapping verification on the combination of resolution setting value and time step, reconstructs its replacement model list, and generates the model replacement parameter group.

[0091] The system calls upon the region name recorded in the region segment list of the model to be optimized, namely the "Industrial-Transportation" composite functional area R23, and extracts multiple alternative model numbers corresponding to this region from the model library, such as M1, M2, and M3, along with their respective spatial resolution settings of 500m, 300m, and 100m, respectively. Simultaneously, it obtains the number of time step levels in the original model, such as three levels (day, week, month), and the initial offset parameter, such as 0.5 hours. The processing unit performs a combined mapping verification of the resolution settings and time step levels for each candidate model. This verification process involves calculating the spatiotemporal coupling index of each model at the three levels, which is the sum of the reciprocal of the resolution and the reciprocal of the time step. For M1 (500m, day), the index is... For M2 (300m, day), the index is For M3 (100m, hour), the index is Then, these indices are compared with the optimal coupling index range required by the original model, such as 4.0–5.0. Only models with indices falling within this range are retained. In this example, M2's 4.33 meets the requirement, while M1 and M3 do not. Therefore, the list of alternative models after reconstruction is as follows: Generate a model to replace the parameter set.

[0092] The mutation scenario verification submodule extracts the inference error sequence and peak deviation position recorded in the carbon flow mutation event for the corresponding region based on the model number already labeled in the model replacement parameter group. It compares the peak error value with the upper and lower limits of the inference accuracy set by the model to which the region belongs, filters the model numbers within the accuracy range, and obtains a list of accuracy-matching models.

[0093] Based on the model number already labeled in the model replacement parameter group, i.e., M2, extract the inference error sequence recorded in historical carbon flow abrupt events for the region associated with this model, for example... The system identifies the location of peak deviations and their error values; the peak error in this sequence is 0.15. The processing unit compares this peak error value with the upper and lower limits of the extrapolation accuracy set for the model to which this area belongs, namely the "industrial-transportation" composite functional zone. This accuracy range is determined based on the measurement errors of the regional carbon monitoring equipment and the physical constraints of the model, with the lower limit set at 0.05. The upper limit is set at 0.20. The processing unit performs numerical comparisons. If the peak error is determined to be within the range of the inference accuracy, then model M2 passes the test and is added to a new list. If other alternative models also pass the test, they are added together to obtain the accuracy-matched model list.

[0094] Configure the integrated judgment submodule, call the spatial resolution setting value and time step level corresponding to each model number in the precision matching model list, construct the combination mapping relationship and deduplicate the repetitive offset combination, extract the effective model unit under each combination and compare the configuration alignment with the original model distribution map, establish a usable model combination library, and generate carbon emission extrapolation model optimization configuration map group.

[0095] The system calls upon the spatial resolution settings and time step levels corresponding to each model number in the precision matching model list. Assuming the list is #imgpt69#, the processing unit constructs a combination mapping relationship for these parameters. That is, for model M2, it generates its specific spatiotemporal configuration at different time step levels. For example, for the daily level, the resolution is 300m, and for the weekly level, the resolution is 500m. Then, all these configurations are aggregated, and duplicate or overly similar offset combinations are deduplicated. For example, if the resolution difference between two models at the daily level is less than the preset 50m resolution threshold, only one is retained. The effective model units under each deduplicated combination are extracted, and these units are compared with the original model distribution map describing the physical space of the region for configuration alignment. This comparison checks whether the computational resource requirements of the replaced model match the capabilities of the existing server cluster. Only combinations that do not cause resource overload are considered usable. The information of all usable model combinations is stored to establish a usable model combination library. Based on the information in this library, a carbon emission extrapolation model optimization configuration map group containing multiple optimization schemes is generated.

[0096] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An information processing system for intelligent regional carbon emission projection and early warning, characterized in that, The system includes: The carbon source coupling identification module acquires energy consumption data streams, dynamic indicators of industrial structure and land use change maps within the region, calculates the frequency of coordinated mutations of the three in the spatiotemporal dimension and cross-compares them, extracts high-frequency coupling offset points, and generates a list of carbon emission coupling anomaly trigger points. The evolution path determination module analyzes the synchronicity between the rate of change of traffic flow intensity and the sudden change point of industrial energy consumption based on the points with multiple periodic repeat records in the list of carbon emission coupling anomaly trigger points, determines whether the dynamic migration process of carbon flow is continuous, and generates a set of spatiotemporal segments suitable for deduction and modeling. The migration motive calculation module calls the segments marked as high carbon flow interaction regions in the set of spatiotemporal segments suitable for modeling and inference, measures their carbon flow density gradient, industrial linkage intensity slope and energy structure change curve, and calculates the overlapping area of ​​the three change rates on the spatiotemporal grid, and outputs the carbon flow dynamic migration dominant factor ranking map. The early warning response switching module checks the current carbon emission response delay duration and carbon flow mutation duration in the corresponding region based on the first-rank factor in the carbon flow dynamic migration dominant factor ranking map, extracts the region identification code, confirms whether the graded early warning triggering conditions are met, and generates a list of region segments for the optimization and simulation model.

2. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 1, characterized in that: The list of abnormal carbon emission coupling trigger points includes energy consumption surge points, critical segments of industrial structure transition, and high-frequency areas of land use conversion. The set of spatiotemporal segments suitable for modeling includes abnormal carbon flow density growth zones, synchronous areas of industrial energy consumption abrupt changes, and segments of multi-source factor response mismatch. The ranking map of dominant factors for dynamic carbon flow migration includes transportation network driving factors, industrial linkage transmission factors, and energy structure transformation factors. The list of regional segments in the model to be optimized includes response delay regions, periodic out-of-step grids, and action lag units.

3. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 1, characterized in that, The carbon source coupling identification module includes: The multi-source data extraction submodule acquires continuous periodic datasets of regional power grid load time series data, industrial enterprise output dynamic sequences, and land cover type change layers interpreted from remote sensing images. It calls the differences of each type of data within adjacent time steps to construct energy consumption difference sequences, industrial structure change interval sets, and land use conversion trend sequences. For the time steps corresponding to the industrial structure change intervals, it extracts spatiotemporal segments in the energy consumption difference sequences and land use conversion trend sequences to generate three sets of multi-source element change interval sets. The spatiotemporal collaborative offset submodule calls the energy consumption difference sequence and the corresponding land use conversion trend sequence in all spatiotemporal units of the three sets of multi-source element change segment sets, calculates the deviation between the land use conversion rate of adjacent units before and after the peak energy consumption difference unit and the average conversion rate in that segment, and constructs a coupling offset judgment index based on the spatiotemporal unit of the intersection of the deviation and the industrial structure change interval to obtain the spatiotemporal collaborative offset frequency sequence. The coupling trigger point identification submodule calls all frequency data in the spatiotemporal coordinated offset frequency sequence, calculates the product of the increase ratio and the directional stability ratio based on the difference between the change trend of each frequency and the change in its adjacent spatiotemporal units, records the spatiotemporal units whose product exceeds the coupling offset trigger threshold, extracts the synchronous fluctuation values ​​of the three elements of energy consumption, industrial structure and land use in the spatiotemporal unit, and generates a list of carbon emission coupling anomaly trigger points.

4. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 1, characterized in that, The evolution path determination module includes: The repeating anomaly extraction submodule obtains the cycle number and anomaly marking status of all points in the carbon emission coupling anomaly triggering point list, filters point numbers with repeating anomaly records, extracts the marking frequency and distribution interval of the corresponding point in each cycle, calculates the anomaly interval difference in consecutive cycles and compares it with the cycle judgment threshold, establishes a point index set that meets the cycle frequency condition, and generates a multi-cycle anomaly point set. The multi-dimensional sequence interleaving submodule calls the traffic flow sequence and industrial energy consumption sequence corresponding to all point numbers in the multi-period anomaly point set, calculates the rate of change of the difference between the two sequences, extracts the fluctuation frequency change sequence, compares the peak position of the two types of sequence changes with the extreme period of fluctuation, identifies the location index of the interleaving and records the interleaving frequency interval, and obtains the interleaving location interval value. The deduction segment selection submodule calls the frequency interlacing time periods recorded in the interlacing position interval values, compares the occurrence positions of abrupt change segments in the carbon flux density sequence, selects the time periods in which the intersection position and density abrupt change point are successfully matched, calculates the time interval value between two abrupt change points in each time period and compares it with the deduction coherence identification threshold, selects the numbered segments that meet the interval conditions and establishes a complete path index, and generates a set of spatiotemporal segments suitable for deduction modeling.

5. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 1, characterized in that, The migration motivation calculation module includes: The intensity slope extraction submodule calls the path number marked as a high carbon flow interaction area in the set of spatiotemporal segments suitable for modeling and extrapolation, extracts the continuous time series value of the industrial association intensity corresponding to each path number, calculates the intensity change slope between each two adjacent sampling points, constructs a slope sequence in time order, extracts the segment number and its change amount corresponding to the maximum slope change amplitude, and generates the association intensity slope change value. The gradient synchronization identification submodule calls the region number recorded in the associated intensity slope change value, extracts the corresponding traffic network flow change sequence and energy structure change sequence, calculates the synchronization segment number between the carbon flow density decrease value and the energy structure value, extracts the peak interval of the carbon flow density gradient and the segment number with the largest continuous change in the energy structure curve, and obtains the synchronization gradient change interval. The overlapping area sorting submodule calls the carbon flow density gradient segment, energy structure change segment, and slope segment in the correlation intensity slope change value in the synchronous gradient change interval. It performs an intersection operation on the relative positions of the three segments on the spatiotemporal grid, calculates the multidimensional overlapping area value formed by the change rate of the intersection segment in three dimensions, and rearranges the segment numbers according to the area value to generate a sorting map of the dominant factors of carbon flow dynamic migration.

6. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 1, characterized in that, The early warning response switching module includes: The first-leader identification submodule calls the sorting first-leader number in the sorting map of the dominant factors of carbon flow dynamic migration, extracts the regional early warning response time series and carbon flow mutation time series data corresponding to the factor, calculates the time difference between the early warning signal issuance time and the carbon flow mutation start point, and compares the difference with the regional matching early warning response benchmark value to determine whether it is in a response delay state and generates an early warning response deviation interval value. The response delay determination submodule calls the region number recorded in the warning response deviation interval value, extracts the warning action setting time value and the actual response start time value in the corresponding carbon emission cycle, calculates the setting response start lag time and compares it with the response lag determination threshold, marks the region identification code with lag problem, and obtains the response lag identification list. The early warning potential assessment submodule calls all the region identification codes in the response lag identification list, compares their frequency of occurrence in the early warning control record with the response intensity sequence of the corresponding segment, filters out the region numbers whose frequency exceeds the repeated identification threshold and whose response intensity value is lower than the response benchmark value, extracts the spatiotemporal segment information to which the corresponding region number belongs and summarizes it into an index list, and generates a list of region segments to be optimized inference models.

7. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 1, characterized in that, The system also includes: The model reconstruction trial module calls the region name in the list of region segments of the model to be optimized and replaces the spatial resolution level and time step granularity in the model accordingly. It tests the stability of the model combination under the carbon flow mutation scenario, selects compliant combinations, and obtains the carbon emission inference model optimization configuration map group. The optimized configuration group of the carbon emission extrapolation model includes a spatial granularity adjustment unit, a time step configuration module, and a coupling mechanism integrated structure.

8. The information processing system for intelligent regional carbon emission simulation and early warning according to claim 7, characterized in that, The model reconstruction trial calculation module includes: The model parameter replacement submodule calls the region name recorded in the region segment list of the model to be optimized, extracts the replacement model number and spatial resolution setting value corresponding to each region, obtains the time step level and initial offset parameter in the region's inference process, performs mapping verification on the combination of resolution setting value and time step and reconstructs its replacement model list, and generates the model replacement parameter group. The mutation scenario verification submodule extracts the inference error sequence and peak deviation position recorded in the carbon flow mutation event for the corresponding region based on the model number already labeled in the model replacement parameter group. It compares the peak error value with the upper and lower limits of the inference accuracy set by the model to which the region belongs, filters the model numbers within the accuracy range, and obtains a list of accuracy-matching models. The configuration integration judgment submodule calls the spatial resolution setting value and time step level corresponding to each model number in the precision matching model list, constructs a combination mapping relationship, and performs deduplication processing on the repetitive offset combinations. It extracts the effective model units under each combination and compares them with the original model distribution map for configuration alignment, establishes a usable model combination library, and generates a carbon emission extrapolation model optimization configuration map group.

Images