Typical city carbon reduction method and system based on spatio-temporal evolution law of power carbon emission

Abstract

The application discloses a typical city carbon reduction method and system based on power carbon emission space-time evolution law, relates to the technical field of carbon emission, analyzes the space-time pattern of power carbon emission of a target city based on power carbon emission accounting data, and obtains spatial correlation characteristics; a geographic space is introduced, a multi-dimensional driving factor evaluation index system is constructed, and then a nonlinear carbon emission analysis model is constructed in combination with a random forest algorithm to obtain an analysis result; a target variable and a specific characteristic variable are decoupled based on a partial dependence plot, a power carbon reduction collaborative strategy analysis model is constructed in combination with a partial least squares structural equation model; a power carbon emission theoretical hypothesis is constructed, a causal path between variables is defined, path coefficients of the causal path are estimated based on the power carbon reduction collaborative strategy analysis model, and a differentiated power carbon reduction strategy facing spatial constraints is determined by comprehensively integrating all results. The nonlinear law of driving factors is deeply mined, and a theoretical basis is provided for analysis of a city power industry carbon reduction strategy.

Description

Technical Field

[0001] This invention relates to the field of carbon emission technology, and in particular to typical urban carbon reduction methods and systems based on the spatiotemporal evolution of electricity carbon emissions. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] In recent years, with the deepening of urban complex systems theory, research has gradually broken through the limitations of a single energy perspective and begun to explore the construction of a comprehensive evaluation framework covering multiple dimensions such as environment, society, and economy. Existing technologies have constructed a comprehensive evaluation system for urban sustainable development from five dimensions: building form, infrastructure, environment, society, and economy. Research has found that the economic dimension has a dominant influence on urban development performance. Existing technologies, based on the triple bottom line model, have constructed an evaluation index system for urban sustainable development from three dimensions: economy, society, and environment, and have also innovatively introduced static interaction effects and dynamic evolution trends among indicators. Although existing technologies have established multi-dimensional evaluation paradigms covering the economic, social, and environmental aspects, they tend to focus on a macro perspective, neglecting the decisive impact of the power industry's unique attributes on carbon emissions. Some existing technologies, targeting power grid operation scenarios, have constructed multi-dimensional evaluation index systems from four dimensions: safety, economy, fairness, and environmental protection, achieving a comprehensive quantitative assessment of carbon emission levels and environmental performance during power grid dispatching and operation. However, existing technologies often suffer from the dual limitations of coarse macro-scale and lack of industry-specific characteristics, making it difficult to adapt to the spatial heterogeneity of carbon emissions.

[0004] Existing technologies have yielded significant results in the field of carbon emissions, with various quantitative analysis models applied to reveal emission-driving mechanisms, predict long-term evolution trends, and simulate differentiated emission reduction strategies. Multivariate grey optimization models incorporating association rules can effectively identify the joint effects of various driving factors on carbon emissions, achieving high accuracy in carbon emission prediction under limited information, but they struggle to characterize the deep nonlinear mechanisms of carbon emission evolution. While the logarithmic average Dichroic index decomposition method meticulously deconstructs the deep driving effects of carbon emissions from a production theory perspective, enhancing the depth of attribution, it remains confined to a static analysis paradigm and cannot adapt to the adaptive expansion of multidimensional variables. Furthermore, as an environmental factor with spatial externalities, the evolution of carbon emissions is often deeply constrained by geographical location. Summary of the Invention

[0005] To overcome the shortcomings of the existing technologies, this invention provides a typical urban carbon reduction method and system based on the spatiotemporal evolution of electricity carbon emissions, providing effective support for typical cities to formulate scientific electricity emission reduction plans.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, this invention provides a typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions, including: Obtain electricity consumption data for the target city and combine it with implicit carbon emissions to obtain electricity carbon emission accounting data for the target city; By introducing spatial statistics theory, the spatiotemporal pattern of electricity carbon emissions in the target city is analyzed based on the electricity carbon emission accounting data to obtain spatial correlation characteristics; Based on the spatial correlation characteristics constraining electricity carbon emissions, a multi-dimensional driving factor evaluation index system is constructed by introducing geographic space. Based on the multidimensional driving factors in the multidimensional driving factor evaluation index system, a nonlinear carbon emission analysis model is constructed by combining the random forest algorithm to obtain the nonlinear response law of multidimensional driving factors to electricity carbon emissions. Based on the partial dependency graph, the target variable and specific feature variables are decoupled from the nonlinear carbon emission analysis model, the marginal effect and nonlinear threshold characteristics of the feature variables are obtained, and a power carbon reduction collaborative strategy analysis model is constructed by combining the partial least squares structural equation model. Theoretical assumptions about carbon emissions from electricity are constructed, causal paths between variables are defined, and path coefficients of the causal paths are estimated based on a collaborative strategy analysis model for carbon reduction in electricity. By combining the spatial correlation characteristics, nonlinear threshold characteristics, and path coefficients of the causal paths, differentiated carbon reduction strategies for electricity oriented towards spatial constraints are determined.

[0007] Further technical solutions involve analyzing the spatiotemporal patterns of electricity carbon emissions in target cities based on electricity carbon emission accounting data, specifically including: Based on electricity carbon emission accounting data, the global Moran index is calculated to quantify the overall spatial correlation attributes of electricity carbon emissions, and the global spatial correlation characteristics are obtained. Based on electricity carbon emission accounting data, the local Moran index is calculated to obtain local spatial clustering characteristics.

[0008] A further technical solution is that the multi-dimensional driving factor evaluation index system includes primary indicators and secondary indicators. The primary indicators include economic development level, urbanization process, power energy characteristics and geographic space. The secondary indicators include GDP, urbanization rate, total population at the end of the year, power generation, power consumption, renewable energy ratio and urban centripetal force.

[0009] Further technical solutions include using geospatial factors as a distance constraint variable when constructing a multi-dimensional driving factor evaluation index system, and taking into account the characteristics of power energy and the level of socio-economic development.

[0010] A further technical solution is that the nonlinear carbon emission analysis model is expressed as follows:

[0011] in, The final output of the model is the predicted carbon emissions from electricity. As multidimensional driving factor characteristic variables, For the first The regression results of the decision trees, The total number of decision trees.

[0012] A further technical solution is that the power carbon reduction collaborative strategy analysis model uses the year-end total population, urbanization rate, GDP, and urban centripetal force as independent variables, and the characteristics of power energy as potential independent variables.

[0013] In a further technical solution, the relationship between each variable in the power carbon reduction collaborative strategy analysis model is represented by path coefficients, as follows:

[0014] in, For total carbon emissions, 、 、 、 、 These are the total population at the end of the year, electricity and energy characteristics, urbanization rate, GDP, and urban centripetal force. 、 、 、 、 These are the path coefficients for carbon emissions for each independent variable. This is the error term.

[0015] Secondly, this invention provides a typical urban carbon reduction system based on the spatiotemporal evolution of electricity carbon emissions, including: The data acquisition module is configured to: acquire electricity energy consumption data of the target city, and combine it with implicit carbon emissions to obtain electricity carbon emission accounting data of the target city; The spatiotemporal analysis module is configured to: introduce spatial statistics theory, analyze the spatiotemporal pattern of electricity carbon emissions in the target city based on the electricity carbon emission accounting data, and obtain spatial correlation characteristics; The indicator system construction module is configured to: based on the spatial correlation characteristics of the constraint on carbon emissions from electricity, introduce geographic space, and construct a multi-dimensional driving factor evaluation indicator system; The multidimensional factor analysis module is configured to: construct a nonlinear carbon emission analysis model based on the multidimensional driving factors in the multidimensional driving factor evaluation index system and the random forest algorithm, and obtain the nonlinear response law of multidimensional driving factors to electricity carbon emissions. The collaborative analysis module is configured to: decouple the target variable and specific feature variables from the nonlinear carbon emission analysis model based on the partial dependency graph, obtain the marginal effect and nonlinear threshold characteristics of the feature variables, and construct a collaborative strategy analysis model for power carbon reduction by combining the partial least squares structural equation model. The strategy determination module is configured to: construct theoretical assumptions about electricity carbon emissions, define causal paths between variables, estimate the path coefficients of the causal paths based on the electricity carbon reduction collaborative strategy analysis model, and determine differentiated electricity carbon reduction strategies oriented towards spatial constraints by integrating the spatial correlation characteristics, nonlinear threshold characteristics and path coefficients of the causal paths.

[0016] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in the first aspect.

[0017] Fourthly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in the first aspect.

[0018] The above one or more technical solutions have the following beneficial effects: This invention first establishes an urban electricity carbon emission index system encompassing the characteristics of electric energy and socio-economic factors, achieving a synergistic consideration of electricity constraints and urban development indicators; secondly, it applies spatial statistical models to characterize the spatiotemporal evolution of carbon emissions, analyzing the spatial dependence and agglomeration effect of urban electricity carbon emissions; finally, it combines machine learning and structural equation modeling to deeply explore the nonlinear laws of driving factors, providing a theoretical basis for the analysis of carbon reduction strategies in the urban power industry.

[0019] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1This is a flowchart of a typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions according to an embodiment of the present invention; Figure 2 This is a PLS-SEM model of the carbon emission influencing factor pathway in this embodiment of the invention; Figure 3 This is a weighted diagram of key driving factors for carbon emissions from electricity in target cities from XX19 to XX23, as described in this embodiment of the invention. Figure 4 This is an example of the impact intensity diagram of various influencing factors on the target city's electricity carbon emissions from XX19 to XX23. Among them, (a) is the impact curve of the year-end total population on electricity carbon emissions, (b) is the impact curve of power generation on electricity carbon emissions, (c) is the impact curve of GDP on electricity carbon emissions, (d) is the impact curve of urban centripetal intensity on electricity carbon emissions, (e) is the impact curve of urbanization rate on electricity carbon emissions, (f) is the impact curve of electricity consumption on electricity carbon emissions, and (g) is the impact curve of the proportion of renewable energy on electricity carbon emissions. Figure 5 This is an analysis diagram of the relationship between electricity generation and urban centripetal force in an embodiment of the present invention; Figure 6 This is an analysis diagram of the spatial urbanization rate and urban centripetal force of electricity carbon emissions according to an embodiment of the present invention; Figure 7 This is an analysis diagram of the spatial proportion of renewable energy in electricity carbon emissions and urban centripetal force in an embodiment of the present invention; Figure 8 This is a graph showing the spatial analysis of electricity carbon emissions, year-end total population, and urban centripetal force in an embodiment of the present invention. Figure 9 This is an analysis diagram of spatial electricity consumption and urban centripetal force in an embodiment of the present invention regarding electricity carbon emissions; Figure 10 This is an embodiment of the present invention showing the relationship between electricity carbon emission space GDP and urban centripetal force analysis. Figure 11 This is the PLS-SEM model result of an embodiment of the present invention. Detailed Implementation

[0022] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0023] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0024] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0025] Example 1 This invention defines large cities characterized by both high energy density and low local supply as typical cities. During rapid urbanization, these cities experience both high economic output and high electricity load. However, constrained by local resource shortages and environmental carrying capacity limits, power generation development lags behind load growth, resulting in a typical receiving-end power grid operation pattern. Their carbon emissions exhibit strong import and external dependence, making their low-carbon evolution path significantly different from resource-exporting cities. They must balance local clean energy substitution with the green proportion of electricity imported from outside the region. Selecting these cities as research subjects has significant demonstrative value in revealing how to break down resource barriers and construct new power systems by utilizing the interaction between power generation, grid, load, and storage under local supply-side bottlenecks.

[0026] Urban electricity carbon emissions are influenced by a combination of multiple driving factors, resulting in complex and variable emission patterns and a lack of refined carbon reduction mechanisms and pathways. Therefore, this invention proposes a typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions. First, a city carbon emission indicator system with characteristics specific to the power industry is constructed to provide a foundation for quantifying the power industry's impact on carbon emissions. Second, spatial autocorrelation and local LISA clustering indices are used to characterize the evolution of carbon emissions. Finally, combining the advantages of random forest and PLS-SEM models, the importance of carbon emission influencing factors and their path indices are analyzed, leading to the proposal of scientific electricity carbon reduction strategies. The innovation of this invention is mainly reflected in the following three aspects: 1) An urban electricity carbon emission index system integrating electricity supply and demand characteristics and geographical spatial constraints was constructed, realizing the unified quantification of multi-dimensional nonlinear driving factors and power industry operation characteristics, and reducing the systematic evaluation bias caused by ignoring spatial heterogeneity.

[0027] 2) A method for analyzing the evolution of typical urban electricity carbon emissions by integrating spatiotemporal dimensions was proposed, revealing the spatial interaction mechanism within the city driven by electricity load characteristics, and indicating the optimization direction for implementing differentiated carbon reduction strategies for urban electricity under spatial constraints.

[0028] 3) A method for weighted analysis of driving factors and analysis of power carbon reduction strategies incorporating spatial distance variables was proposed, clarifying the nonlinear threshold boundary and complex action chain of power carbon reduction strategy implementation, and providing a scientific quantitative basis for establishing priority control areas and key node management of power carbon reduction.

[0029] like Figure 1 As shown in the figure, this embodiment discloses a typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions. The method includes the following steps: Obtain electricity consumption data for the target city, and combine the carbon emissions from the combustion of fossil fuels in the city's thermal power plants with the implicit carbon emissions corresponding to external power reception to calculate the electricity carbon emission accounting data for the target city. By introducing spatial statistics theory, we analyze the spatiotemporal pattern of electricity carbon emissions in the target city based on electricity carbon emission accounting data, extract spatial dependence and agglomeration effect, and obtain spatial correlation characteristics. Based on the spatial correlation characteristics constraining electricity carbon emissions, this paper introduces geographic space as a distance constraint variable, and considers the characteristics of electricity energy and the level of socio-economic development to construct a multi-dimensional driving factor evaluation index system. Based on the multidimensional driving factors in the evaluation index system, a nonlinear carbon emission analysis model is constructed by combining the random forest algorithm. The feature importance metric of the multidimensional driving factors is calculated to identify key driving factors, and the nonlinear response law of the multidimensional driving factors to electricity carbon emissions is extracted as the analysis result. Based on the partial dependency graph, the target variable and specific feature variables are decoupled from the nonlinear carbon emission analysis model to obtain the marginal effect and nonlinear threshold characteristics of the feature variables. At the same time, in order to analyze the interaction between multidimensional driving factors, a power carbon reduction synergistic strategy analysis model is constructed by combining the partial least squares structural equation model. We construct theoretical assumptions about carbon emissions from electricity, define causal paths between variables, and estimate path coefficients of causal paths based on a collaborative strategy analysis model for carbon reduction in electricity. By integrating spatial correlation characteristics, nonlinear threshold characteristics, and path coefficients of causal paths, we determine differentiated carbon reduction strategies for electricity that are oriented towards spatial constraints.

[0030] In this embodiment, the present invention constructs an integrated quantitative analysis framework for electricity carbon emissions adapted to urban characteristics. First, an electricity carbon emission accounting model is established based on the IPCC inventory method, clarifying the research boundaries and data accounting foundation. Then, spatial statistics theory is introduced, and global and local spatial autocorrelation models are used to analyze the spatial dependence and clustering evolution characteristics of carbon emissions, verifying the necessity of introducing a spatial dimension. Furthermore, a multi-dimensional driving factor evaluation index system integrating spatial distance dimension is constructed, and a quantitative analysis model of driving factors is established based on the nonlinear fitting advantages of the random forest algorithm. This framework achieves a logical closed loop from identifying macroscopic spatiotemporal evolution patterns to analyzing nonlinear driving mechanisms, providing theoretical basis and data support for subsequent analysis of urban electricity carbon reduction strategies.

[0031] (I) Electricity Carbon Emission Accounting Model According to the internal operation of the power industry, carbon dioxide emissions from the power sector mainly originate from thermal power generation. Considering the typical "high load - low local supply" operating pattern in receiving cities, their power supply is significantly dependent on external power sources, resulting in strong input and external dependence on carbon emissions. Therefore, this invention, based on the IPCC inventory method for calculating carbon emissions from local thermal power fossil fuel combustion, further incorporates the implicit carbon emissions corresponding to external power sources into the power carbon emission accounting boundary to obtain power carbon emission accounting data for the target city. The calculation formula is as follows:

[0032] in, This indicates the carbon dioxide emissions from the power industry; This indicates the types of fossil fuels. The eight main types of energy include coal, coke, crude oil, gasoline, kerosene, diesel, fuel oil, and natural gas. Indicates energy consumption; Indicates the carbon emission factor; This indicates the average lower heating value; Indicates average calorific value and carbon content; This indicates the carbon oxidation rate; 44 / 12 is the molecular weight ratio of carbon dioxide. This indicates the amount of electricity received from external sources during the accounting period; This represents the average carbon emission factor of the power grid in the region where the purchased electricity is located.

[0033] (II) Spatiotemporal Pattern Analysis Model of Electricity Carbon Emissions To quantify the spatial correlation characteristics of urban electricity carbon emissions, this invention constructs a spatial statistical analysis framework based on typical time segments. Years with both data completeness and timeliness are selected as research objects. The Global Moran's I index is used to study the overall spatial correlation attributes of the region, and the Local Indicators of Spatial Association (LISA) map is introduced to visualize and analyze the spatial topology of high-value and low-value clustering areas. The static analysis intuitively reveals the spatial distribution structure of urban electricity carbon emissions, confirming that different areas of the city do not operate in isolation, but rather exhibit spatial dependence and clustering effects. This conclusion verifies the objective existence of spatial correlation and lays the foundation for further expanding the full-time-domain dimension and deeply analyzing the spatiotemporal evolution of electricity carbon emissions.

[0034] Spatial statistical analysis includes global spatial correlation feature analysis and local spatial clustering feature analysis. Specifically, based on electricity carbon emission accounting data, a global Moran index is calculated to quantify the overall spatial correlation attributes of electricity carbon emissions, thus obtaining global spatial correlation features; based on electricity carbon emission accounting data, a local Moran index is calculated to obtain local spatial clustering features.

[0035] Furthermore, global spatial autocorrelation analysis aims to quantify the overall spatial correlation attributes of urban electricity carbon emissions. It uses the Global Moran's I index to reveal the spatial correlation characteristics of regional carbon emissions, and its mathematical model is constructed as follows:

[0036] in, This is the global Moran index. For the total number of studies, This is the spatial weight matrix. and They are respectively and Carbon emissions from electricity This represents the average carbon emissions from all electricity sources within the study area. The range of values ​​for is ( 1,1). A value close to 0 indicates that carbon emissions from electricity exhibit positive spatial autocorrelation, while a value close to 0 indicates that carbon emissions from electricity exhibit negative spatial autocorrelation. A value close to 0 indicates that carbon emissions from electricity exhibit random distribution. and It refers to different spatial units, such as specific administrative counties and districts.

[0037] To verify To assess statistical validity, the normality statistic is introduced. Z Value and significance levelP Hypothesis testing is performed on the values. Z The value reflects the significance factor of the spatial clustering degree. P The value represents the probability that the observed outcome is a random event. If | Z |>1.96 and P If the value is less than 0.05, it confirms that there is a significant spatial autocorrelation characteristic in urban electricity carbon emissions; Z The larger the value, the stronger the spatial dependence and clustering effect. The mathematical model is shown below:

[0038]

[0039] in, The Z-value represents the global Moran exponent. For random expectations, For variance, It is the integral variable.

[0040] Furthermore, given that global autocorrelation analysis can only characterize the overall spatial correlation trend of a region and is insufficient to depict the spatial dependence characteristics at the local scale, this invention further introduces LISA for in-depth analysis. As the core statistic of LISA, the local Moran index can accurately characterize the correlation strength and significance level between a specific unit and its neighborhood in carbon emission distribution. This method overcomes the masking effect of global statistics on local non-stationary characteristics by revealing the patterns of local spatial heterogeneity. Its computational model is constructed as follows:

[0041] in, This refers to Local Moran's I index. This indicates that the neighboring species exhibit similar characteristics, indicating spatial homogeneity (high-high clustering or low-low clustering). This indicates that the product exhibits different characteristics from its neighbors, demonstrating spatial heterogeneity (high-low clustering or low-high clustering). This indicates that there is no spatial correlation.

[0042] (III) Multidimensional Driving Factor Analysis Model for Carbon Emissions from Electricity Given the complex multidimensional coupling of carbon emissions from the power sector, this invention addresses the limitations of traditional indicator systems that neglect spatial factors, focusing instead on the constraining effect of geographical location on resource allocation. Furthermore, a carbon emission evaluation indicator system for the power sector, as shown in Table 1, is constructed, laying the foundation for accurately identifying the nonlinear responses of carbon emission drivers.

[0043] Table 1. Evaluation Index System for Carbon Emissions from Electricity

[0044] The "urban centripetal force" is defined as the Euclidean spatial distance from the geometric center of each district / county to the absolute center of the main urban area. This term was chosen because: firstly, the spatial differences in carbon emissions from electricity in typical cities mainly stem from the spatial mismatch between power generation and consumption. This mismatch between load and power supply exhibits a high degree of concentric correlation with the physical spatial distance from the main urban area; secondly, compared to the underlying micro-level power grid topology, Euclidean spatial distance can be better integrated with multi-dimensional macro-level factors such as GDP, population, and urbanization rate for unified quantification and multi-dimensional interactive analysis within the same geographical coordinate system, thus more intuitively representing the comprehensive constraint effect of urbanization expansion and economic evolution gradients on carbon emissions from electricity.

[0045] Following the principles of scientific rigor and data availability, this invention constructs a driving factor evaluation system comprising four dimensions and seven indicators. The dimensions of economic development level and urbanization process primarily characterize the demand-side growth momentum of electricity carbon emissions; the dimension of electricity energy characteristics focuses on depicting the energy dependence of the power industry's supply-side operation; and the dimension of geospatial characteristics aims to quantify the constraining effect of spatial dependence and agglomeration effects on electricity carbon emissions. This system achieves comprehensive coverage from endogenous driving forces to external constraints.

[0046] Traditional linear regression models are often inadequate to address the multicollinearity and complex nonlinear interactions among variables present in the aforementioned indicator system. In contrast, the random forest model, with its superior nonlinear fitting capabilities, has become the preferred tool for analyzing the complex driving mechanisms of carbon emissions: firstly, its built-in feature importance assessment mechanism can accurately quantify and screen key driving factors such as industrial structure and population from multidimensional heterogeneous data; secondly, it overcomes the limitations of traditional linear regression, effectively revealing the nonlinear response patterns of various factors to carbon emission intensity, providing in-depth quantitative support for differentiated policy formulation. Therefore, this invention uses the random forest (RF) algorithm to construct a driving factor analysis model, namely, a nonlinear carbon emission analysis model based on random forests. This algorithm is based on the Bagging ensemble learning concept, constructing multiple independent decision trees for parallel computation, and finally taking the average as the output result. Its mathematical expression is:

[0047] in, The predicted carbon emissions from electricity (dependent variable) are the final output of the model. As multidimensional driving factor characteristic variables, For the first The regression results of the decision trees, This represents the total number of decision trees. The model possesses strong noise resistance and nonlinear fitting advantages, effectively overcoming the overfitting problem.

[0048] To quantitatively identify the contribution weight of each driving factor to carbon emissions from electricity generation, this invention uses the IncMSE (Indicative Significance of Emissions) measure as an evaluation index. By randomly shuffling the order of observations of a driving variable, the decrease in model prediction accuracy (OOB error) is calculated. If the random permutation of a variable leads to a significant increase in the model's MSE, it indicates that the variable is more important. The calculation formula is as follows:

[0049] in, Indicates the first The importance of each variable This represents the total number of decision trees in the RF model. This represents a specific decision tree. This represents the baseline accuracy of the model under normal conditions. Indicates the first The new error calculated after randomly shuffling the data of each variable.

[0050] To further reveal the marginal effect and nonlinear impact threshold of driving factors on carbon emissions, a partial dependency plot (PDP) is introduced. PDP decouples the functional relationship between the target variable and specific characteristic variables from complex multidimensional models, visually demonstrating the independent trajectory of a single variable's change on carbon emissions while keeping the average levels of other variables constant. This analysis overcomes the limitation of simple importance ranking, which fails to reflect positive and negative correlations and inflection point characteristics. The mathematical definition function of the partial dependency plot (PDP) is:

[0051] in, Represents target feature variables Partial dependency function, This represents the constructed nonlinear carbon emission analysis model. This represents the total number of samples in the training dataset. This represents the target feature variable in the current analysis. Indicates the first Each sample in the complementary feature set The specific value that can be taken.

[0052] Based on the above technical features, the nonlinear carbon emission analysis model (RF model) derives the contribution weight of each multidimensional driving factor to carbon emissions, i.e., the importance score, as well as the nonlinear response boundary and threshold inflection point of each factor to carbon emissions, such as the threshold of 80% urbanization rate.

[0053] (iv) Analysis Model of Coordinated Strategy for Carbon Reduction in the Power Industry While the RF model effectively reveals the nonlinear threshold and spatial heterogeneity of individual factors, it is difficult to intuitively characterize the intricate interactions between these factors. Therefore, this invention also introduces the partial least squares structural equation model (PLS-SEM) to further deconstruct the direct and indirect effects of the "population-economy-energy" system on carbon emissions from a structural perspective, thus complementing the nonlinear analysis for verification.

[0054] Compared to traditional covariance-based structural equation modeling, PLS-SEM demonstrates significant advantages in handling small sample sizes and non-normally distributed data. It exhibits greater robustness against multicollinearity and can simultaneously handle multiple causal paths and latent variables. Furthermore, with limited sample size, PLS-SEM demonstrates higher statistical power and convergence, obtaining robust path coefficient estimates even with smaller sample sizes. Given that this invention focuses on carbon emission data at the level of population, the sample size is relatively limited and it is difficult to satisfy the stringent assumption of a multivariate normal distribution; therefore, choosing PLS-SEM is the most appropriate analytical strategy. In this invention, PLS-SEM is used to construct a causal model incorporating variables such as population, electricity consumption, power generation, and relative distance, aiming to assess the impact of these electricity factors on carbon emissions and their interrelationships.

[0055] (1) Theoretical assumptions about carbon emissions from electricity When using the PLS-SEM model for analysis, the first step is to establish hypotheses and define the paths and relationships between latent variables, thus providing a basis for subsequent path analysis. Based on the PDP plot results, the year-end total population, urbanization rate, GDP, and urban centripetal force are selected as independent variables, and electricity energy characteristics are selected as latent independent variables. The relationship between each variable in the model is expressed by the path coefficient formula, as shown below:

[0056] in, Represents total carbon emissions; 、 、 、 、 These represent the total population at the end of the year, electricity energy characteristics, urbanization rate, GDP, and urban centripetal force, respectively. Among them, electricity energy characteristics are jointly affected by electricity consumption, power generation, and the proportion of renewable energy. 、 、 、 、 These are the path coefficients of each independent variable for carbon emissions; The error term represents other influencing factors not captured by the model. The path coefficients mentioned above are estimated by inputting the actual data of each evaluation index (independent variable) and the electricity carbon emission accounting data (dependent variable) into the PLS-SEM model, and then iteratively fitting the model using a sampling algorithm.

[0057] Subsequently, the present invention proposes the following hypothesis: Hypothesis H1: The total population at the end of the year has a positive impact on the characteristics of electricity energy. ).

[0058] Hypothesis H2: Year-end total population has a positive impact on GDP ( ).

[0059] Hypothesis H3: Urban centripetal force has a negative impact on urbanization rate. ).

[0060] Hypothesis H4: Urbanization rate has a positive impact on GDP ( ).

[0061] Hypothesis H5: The characteristics of electricity energy have a positive impact on carbon emissions ( ).

[0062] Hypothesis H6: Urbanization rate has a negative impact on carbon emissions ( ).

[0063] Assumption H7: GDP has a positive impact on carbon emissions ( ) To verify the above hypotheses, each hypothesis is represented as a causal path between variables. Specifically, the study consists of six variables: year-end total population, electricity energy characteristics, urbanization rate, GDP, urban centripetal force, and carbon emissions. These variables are reflected through a series of observable parameters. The PLS-SEM model of influencing factors established based on this hypothesis is shown below. Figure 2 As shown.

[0064] The power sector carbon reduction synergy strategy analysis model ultimately outputs the quantitative strength of the impact of each driving factor on carbon emissions, and then determines differentiated carbon reduction strategies based on this strength. In this model, variables such as year-end total population, power energy characteristics, urbanization rate, GDP, and urban centripetal force jointly affect carbon emissions through path relationships. These variables are connected by path coefficients, forming a causal relationship model. The arrows indicate the causal relationship between variables. The model quantifies the direct and indirect impacts of each factor on carbon emissions by estimating the path coefficients, helping to identify key driving factors for emission reduction optimization.

[0065] The Power Sector Carbon Reduction Collaborative Strategy Analysis Model (PLS-SEM Model) clarifies the causal paths of the interactions between variables, while the nonlinear carbon emission analysis model constructed using the random forest algorithm provides the numerical boundaries and spatial characteristics of the impact. Combining these two approaches, the control targets are identified macroscopically, and the spatial range and numerical intervals of control are determined microscopically, ultimately leading to the derivation of the final carbon reduction strategy.

[0066] The example analysis of this embodiment is as follows: (1) Analysis of the spatiotemporal evolution of carbon emissions from electricity 1) Electricity carbon emission accounting data To verify the effectiveness and universality of the multidimensional driving factor analysis model and the collaborative carbon reduction strategy analysis framework constructed above, it is necessary to select typical cities with high electricity consumption and low local power support characteristics for empirical research. This invention incorporates the implicit emissions from externally received electricity as an important component in electricity carbon emission accounting to avoid underestimating the carbon emissions on the electricity consumption side of the receiving city by only accounting for local thermal power emissions, and to more accurately reflect the effect of external clean energy substitution. Based on the electricity carbon emission accounting data of the target city from XX19 to XX23, a heat map of spatial distribution evolution is drawn. Its spatiotemporal differentiation characteristics are analyzed from the following two aspects: Spatial Distribution Pattern: The target city's carbon emissions from electricity exhibit a gradient distribution pattern, with higher emissions in the east and lower emissions in the west. The peak emission area has remained stable in the eastern industrial load center, driven by high-density manufacturing loads, and its emissions have consistently remained at the highest level. The western ecological conservation area, constrained by strict industrial access policies, has maintained stable low levels. The central urban area lies between the two, exhibiting an overall hollowed-out topology structure with high-value agglomeration in the peripheral industrial belt and low-value collapse in the central service sector, demonstrating significant spatial imbalance.

[0067] Temporal Evolution Characteristics: From XX19 to XX23, regional carbon emissions generally showed a trend of "rapid growth followed by high-level fluctuations." Initially, influenced by the capacity expansion of the Eastern New Area, the emission center shifted eastward; later, although affected by the external environment, emissions from major industries remained at a high level. In particular, in XX23, with the recovery of overall electricity load, emissions from key industrial clusters such as electronics and information began to rise, reflecting the strong dependence and inertia of key industrial regions on fossil fuel consumption.

[0068] Furthermore, the research period of this invention includes the objective disruption caused by the pandemic. However, the rigid electricity consumption of energy-intensive industries means that short-term fluctuations have not masked the long-term trend of spatial accumulation of carbon emissions. Therefore, even in the face of strong external shocks, the spatiotemporal analysis framework of this invention can still effectively filter out short-term disturbances, accurately capture stage-specific characteristics and evolution patterns, and verify the robustness of the model.

[0069] 2) Spatiotemporal correlation analysis of electricity carbon emission data Table 2 shows that the global Moran's I index for target city electricity carbon emissions was positive from 2019 to 2023, and the Z-value consistently exceeded the 1% significance threshold, confirming a significant positive spatial autocorrelation characteristic in electricity carbon emissions. This indicates that electricity carbon emissions are not isolated but exhibit strong neighborhood homogeneity in space, meaning high-emission areas tend to be adjacent to other high-emission areas, and low-emission areas are interdependent with other low-emission areas. From a temporal evolution perspective, the Moran's I index generally showed a fluctuating upward trend over the past five years, consistently remaining above the significance level. This reveals that the spatial agglomeration effect of target city electricity carbon emissions is strengthening year by year. This increased spatial dependence essentially reflects the increasing concentration of energy-intensive industries in geographical space as regional strategies such as the "Eastward Expansion" deepen, leading to a gradual shift from a discrete to a solidified spatial pattern of carbon emissions, and enhancing the explanatory power of inter-regional spatial spillover effects on local carbon emission levels.

[0070] Table 2 Moran's I Index of Global Electricity Carbon Emissions, XX19-XX23

[0071] However, the global Moran's I index can only characterize the overall spatial correlation trend of the study area at a macro level. Essentially, it is an average measurement of spatial differences across the entire region, which can easily mask potential spatial anomalies in local areas. To further accurately identify the local spatial heterogeneity of adjacent electricity carbon emissions and clarify the specific spatial locations of high-value and low-value clusters, this invention introduces LISA for in-depth analysis to reveal the correlation patterns between local spatial units and their neighbors. Based on this, a LISA clustering evolution map of target city electricity carbon emissions from XX19 to XX23 was calculated and plotted.

[0072] Based on local spatial autocorrelation properties, the study area is divided into four spatial clustering patterns: HH (high-high) clustering: local and adjacent power generation carbon emission levels are relatively high, showing a high emission hotspot; LH (low-high) clustering: local emissions are relatively low but surrounded by high emission areas; LL (low-low) clustering: local and adjacent power generation carbon emission levels are relatively low, forming a low-carbon defense zone; HL (high-low) clustering: local emissions are relatively high but surrounding areas are relatively low.

[0073] The HH and HL clusters highly overlap with typical load centers such as the northern industrial heavy load area and Shuangliu District, exhibiting a synergistic effect of high load density and high emission intensity. Due to the concentration of energy-intensive industries such as petrochemicals and advanced manufacturing in this region, the marginal emission reduction cost of electricity substitution is high. Furthermore, the strong operational inertia and rigid energy consumption of industrial loads lead to the spatial solidification of high carbon emissions and significant positive spillover. Conversely, the expansion of the LL cluster towards the central urban area confirms the effectiveness of structural carbon reduction on the load side. As the central urban area's industries transform from manufacturing to modern services, high-energy-consuming loads are relocated in an orderly manner, and the nature of the terminal power grid load has successfully shifted from industrial dominance to low-carbon commercial, residential, and comprehensive energy-consuming loads. This optimized load structure improves the economic output efficiency of end-user energy consumption, prompting the central urban area and the western hydropower-rich area to spatially connect and construct a stable low-carbon electricity consumption region, achieving a relative decoupling between regional electricity consumption growth and carbon emissions.

[0074] (2) Analysis of key drivers of carbon emissions from electricity Key influencing factors identification and analysis.

[0075] To enhance the interpretability of the RF model, the importance of driving factors was first ranked, and then a partial dependency plot (PDP) was used to visualize the marginal impact of these factors on the model output. IncMSE was employed as a metric to quantify and rank the contribution of each driving factor. This metric reflects the decrease in model prediction accuracy after a variable is randomly rearranged; a higher value indicates a stronger explanatory power of that variable for electricity carbon emissions.

[0076] Figure 3 For the variable importance scores of the RF model, such as Figure 3 , Figure 4 As shown in (a)-(g), the driving mechanism of carbon emissions from electricity in target cities is jointly dominated by the year-end total population, the proportion of renewable energy, and urban centripetal force. Among them, the year-end total population ranks first in importance and is the most critical factor causing spatial differentiation of carbon emissions; the influence of the proportion of renewable energy and urban centripetal force follows closely behind, forming a secondary driving echelon; while the importance of power generation, GDP, urbanization rate, and electricity consumption is relatively weak, belonging to basic influencing factors. The reason for this is that the year-end total population, as the primary driving factor, establishes a rigid base load scale, and its explanatory power for spatial differentiation of carbon emissions from electricity significantly exceeds that of total economic output; the proportion of renewable energy, as the core variable for source-side emission reduction, utilizes the differences in attributes between hydropower in the west and thermal power in the east, and its influence weight is far higher than that of simple load-side electricity consumption indicators; urban centripetal force, by representing the gradient evolution of industrial load, profoundly reshapes the regional carbon emission intensity. This indicates that the economy or electricity consumption volume alone is no longer the only core factor; the cleanliness of electricity sources and geographical spatial layout are also deep variables determining the current carbon emission pattern.

[0077] While the aforementioned importance scores clearly identify the dominant roles of factors such as population size and power structure, they do not reveal the specific forms and directions of their impact on carbon emissions. To compensate for the shortcomings of single indicators in detail characterization, this invention utilizes a Power-Phase Analytical Process (PDP) to separate the marginal effects of each driving factor from the model. Through single-factor response curve analysis of seven key variables, this study further explores the nonlinear driving patterns and inflection point characteristics of each factor on electricity carbon emissions within different numerical ranges.

[0078] Analysis based on the partial dependency graph reveals that the impact of various driving factors on carbon emissions from electricity exhibits significant nonlinear threshold characteristics and a phased evolutionary pattern. Specifically, the year-end total population and GDP drive carbon emissions through a dynamic game between scale effects and congestion effects. Within a moderate scale range, intensive resource utilization can effectively suppress emissions. However, when the population exceeds 2.4 million or GDP surpasses 250 billion yuan, the rigidity of energy consumption brought about by ultra-large scale leads to a significant rebound in emissions. Electricity consumption also shows a similar leapfrog growth characteristic after a certain threshold. Meanwhile, the emission reduction effects of urbanization rate and renewable energy share exhibit clear threshold characteristics. Only when the urbanization rate exceeds 80% and enters the intensive development stage, or when the renewable energy share exceeds the critical point of 50%, can the system regulation constraints be overcome to establish a dominant substitution advantage, thereby significantly suppressing emission intensity. Furthermore, urban centripetal intensity reveals a spatial gradient pattern of decreasing emission intensity from the center to the periphery. The 40km mark becomes a key geographical boundary for the transition from a high-energy-consumption core area to a low-value stable area, verifying the spatial heterogeneity of resource allocation efficiency between the central urban area and the suburbs.

[0079] Given that single-factor analysis struggles to capture the heterogeneous emission characteristics of factors across different geographical regions and neglects the spatial constraints of urban centripetal force on resource allocation efficiency, this invention constructs a three-dimensional interactive model to deeply analyze how geographical distance reshapes the marginal impact of different driving factors on electricity carbon emissions, aiming to reconstruct the nonlinear driving mechanism at scale from a spatial coupling perspective. Table 3 shows the correlation analysis data between urban centripetal force and other indicators.

[0080] Table 3. Correlation data of city centripetal force

[0081] Table 3 shows that urbanization rate, renewable energy share, and power generation are strongly correlated with urban centripetal force, indicating that geographical region is a key factor restricting their distribution and exhibits a strict concentric gradient characteristic. Conversely, the correlation between year-end total population, GDP, and electricity consumption is weak, reflecting that the development of target cities has broken the single distance decay law, and the distribution of factors has a high degree of spatial complexity.

[0082] like Figure 5 , Figure 6 , Figure 7 The figure shows a spatial correlation analysis of carbon emissions from electricity generation. The power generation surface indicates that emissions are mainly concentrated in load centers. Contrary to the common perception of power generation relocation, the data shows that carbon emissions in the core area on the right are not reduced due to proximity but are actually very high. This is because the huge electricity demand in the central area leads to persistently high emission levels. As the transition to the far suburbs on the left progresses, emissions naturally decrease with increasing distance, reflecting the mitigating effect of distance on emissions. The urbanization rate surface shows a clear characteristic of high at the center and low at the periphery. In the central urban area on the right, the high urbanization rate combined with high population density keeps carbon emissions high, indicating extremely high resource consumption intensity. As the curve extends to the far suburbs on the left, it drops rapidly, indicating that the driving effect of urbanization on carbon emissions weakens in areas far from the center. This change confirms that geographical location affects the carbon emission effect of urbanization. The renewable energy share surface forms a clear low-emission zone in the far suburbs on the left. In the central urban area on the right, the emission reduction effect of increasing the green electricity share is limited due to the huge energy consumption base. However, in the far suburbs on the left, as distance increases, the substitution effect of green electricity is amplified by spatial advantages, and the emission value drops rapidly to its lowest point. This proves that developing green electricity in the suburbs is the best combination for achieving deep decarbonization.

[0083] like Figure 8 , Figure 9 , Figure 10 The graph shows the spatial weak correlation of carbon emissions from electricity. The graphs for year-end total population, GDP, and electricity consumption are highly similar, all exhibiting a pattern of high emissions in the center and a gradual decline towards the periphery. On the right side, the central urban area corresponds to high emissions for all three, reflecting the environmental pressure brought about by factor agglomeration. Extending to the left, the curves decline gently without complex fluctuations. This indicates that the driving force of these three factors on carbon emissions mainly depends on scale, with spatial distance primarily playing a simple dilution role and not forming a complex regulatory mechanism.

[0084] In summary, the impact of each driving factor on carbon emissions exhibits a core-clustering and periphery-decaying characteristic, revealing the current status of central urban areas as high-emission centers. Urbanization rate, power generation, and the proportion of renewable energy show spatial heterogeneity; that is, emission reduction is limited in central areas due to their fixed energy consumption base, while suburban areas, due to their spatial advantages, become the best regions for deep decarbonization through green electricity. In contrast, the emission drivers of population, GDP, and electricity consumption are mainly determined by scale, with spatial distance only playing a linear dilution role. This indicates that emission reduction strategies need to focus on total emission control in central areas and structural optimization in suburban areas.

[0085] (3) Analysis of carbon emission reduction strategies for the power industry Major drivers of carbon emissions and their impact on emission reduction strategies, such as Figure 11 As shown, the values ​​on the line represent path coefficients, and the values ​​inside the circles represent... The value represents the degree to which the independent variable explains the dependent variable. This invention uses the Bootstrapping algorithm, selecting 5000 repeated samplings at a 5% significance level to test the significance of the model parameters. Among them, The values ​​all remained above 0.6, indicating that the model has a strong explanatory power for changes in carbon emissions.

[0086] The path coefficient shows that the characteristics of electricity energy have a significant positive effect on carbon emissions, at 0.662, meaning that for every 1% increase in the electricity energy characteristic index, the carbon emission index increases by 0.662%. Among the three branches of the comprehensive influencing factors of electricity, the promoting effect of electricity consumption is the strongest, while the proportion of renewable energy generation has a weakening effect. The effect of electricity generation is relatively low because the target city itself has a small power generation capacity.

[0087] Interestingly, urbanization rate has a significant negative impact on carbon emissions, with a path coefficient of -0.348. This means that for every 1% increase in urbanization rate, carbon emissions decrease by 0.348%, indicating that increasing urbanization rate can mitigate carbon emissions to some extent. This is because as urbanization increases, energy consumption in rural areas, such as coal-fired heating and cooking, is replaced by electricity. Household energy consumption shifts from high-carbon scattered coal use to municipal electricity and centralized heating. Therefore, intuitively, urbanization rate reduces carbon emissions, and the carbon emissions from alternative energy sources are reflected in electricity consumption. Urbanization rate has a significant positive impact on GDP, with a path coefficient of 0.469, indicating that increasing urbanization rate can lead to rapid GDP growth.

[0088] The negative impact of urban centripetal intensity on urbanization rate was verified in this model, with a path coefficient of -0.887. This means that for every 1% increase in urban centripetal intensity, the urbanization rate decreases by 0.887%. This indicates that as the distance from the city center increases, the urbanization rate tends to decrease, and the impact is significant, which is consistent with the current situation of urban development radiating outward from the center.

[0089] From other perspectives, the increase in the total population at the end of the year has a strong positive impact on both GDP and the comprehensive factors related to electricity, with path coefficients of 0.949 and 0.379, respectively. For the same reason as the urbanization rate, the city's pillar industry is the tertiary sector, and the energy consumption resulting from GDP growth is primarily reflected in electricity consumption; therefore, GDP's impact on carbon emissions is not significant.

[0090] Based on the above path analysis, and considering the city's unique situation of abundant hydropower resources but limited utilization of photovoltaic and wind power resources, this invention proposes the following feasible emission reduction paths. First, accelerate the urbanization process in remote areas, using electricity to replace fossil fuels such as coal, enhancing energy utilization efficiency, reducing the carbon emission coefficient per unit of energy, and simultaneously boosting GDP growth, achieving a relative decoupling of economic development from carbon emissions. Second, the city should increase some local power generation, fully utilize local hydropower resources as baseload, strengthen the utilization of wind and solar resources, and continue to vigorously develop renewable energy, mitigating the indirect carbon emission impact of purchasing electricity from external sources. Finally, optimize the demographic dividend, reduce the population growth rate, cultivate a highly skilled population, and while achieving high-quality GDP development, offset the environmental pressure brought by population growth through improved population quality. Therefore, based on the city's unique characteristics, a comprehensive emission reduction policy is formulated.

[0091] In summary, this invention analyzes the spatiotemporal evolution of urban electricity carbon emissions, providing full-chain quantitative support from pattern identification to strategy analysis. First, it confirms the spatial correlation of electricity carbon emissions, clarifying the distribution pattern of high-emission clusters and low-carbon periphery areas. Second, it overcomes the limitations of single linear analysis, accurately quantifying the nonlinear marginal effects of different driving factors within specific numerical ranges and geographical areas. Finally, by deconstructing the complex causal transmission chain among multidimensional factors, it identifies differentiated emission reduction paths centered on controlling population size, breaking through the inflection point of clean energy penetration, and utilizing spatial gradient advantages.

[0092] Example 2 This embodiment discloses a typical urban carbon reduction system based on the spatiotemporal evolution of electricity carbon emissions, including: The data acquisition module is configured to: acquire electricity energy consumption data of the target city, and combine it with implicit carbon emissions to obtain electricity carbon emission accounting data of the target city; The spatiotemporal analysis module is configured to: introduce spatial statistics theory, analyze the spatiotemporal pattern of electricity carbon emissions in the target city based on the electricity carbon emission accounting data, and obtain spatial correlation characteristics; The indicator system construction module is configured to: based on the spatial correlation characteristics of the constraint on carbon emissions from electricity, introduce geographic space, and construct a multi-dimensional driving factor evaluation indicator system; The multidimensional factor analysis module is configured to: construct a nonlinear carbon emission analysis model based on the multidimensional driving factors in the multidimensional driving factor evaluation index system and the random forest algorithm, and obtain the nonlinear response law of multidimensional driving factors to electricity carbon emissions. The collaborative analysis module is configured to: decouple the target variable and specific feature variables from the nonlinear carbon emission analysis model based on the partial dependency graph, obtain the marginal effect and nonlinear threshold characteristics of the feature variables, and construct a collaborative strategy analysis model for power carbon reduction by combining the partial least squares structural equation model. The strategy determination module is configured to: construct theoretical assumptions about electricity carbon emissions, define causal paths between variables, estimate the path coefficients of the causal paths based on the electricity carbon reduction collaborative strategy analysis model, and determine differentiated electricity carbon reduction strategies oriented towards spatial constraints by integrating the spatial correlation characteristics, nonlinear threshold characteristics and path coefficients of the causal paths.

[0093] Example 3 The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method of Embodiment 1.

[0094] Example 4 The purpose of this embodiment is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method of Embodiment 1.

[0095] The steps and methods involved in the apparatuses of Embodiments 3 and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0096] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0097] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0098] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions, characterized in that, include: Obtain electricity consumption data for the target city and combine it with implicit carbon emissions to obtain electricity carbon emission accounting data for the target city; By introducing spatial statistics theory, the spatiotemporal pattern of electricity carbon emissions in the target city is analyzed based on the electricity carbon emission accounting data to obtain spatial correlation characteristics; Based on the spatial correlation characteristics constraining electricity carbon emissions, a multi-dimensional driving factor evaluation index system is constructed by introducing geographic space. Based on the multidimensional driving factors in the multidimensional driving factor evaluation index system, a nonlinear carbon emission analysis model is constructed by combining the random forest algorithm to obtain the nonlinear response law of multidimensional driving factors to electricity carbon emissions. Based on the partial dependency graph, the target variable and specific feature variables are decoupled from the nonlinear carbon emission analysis model, the marginal effect and nonlinear threshold characteristics of the feature variables are obtained, and a power carbon reduction collaborative strategy analysis model is constructed by combining the partial least squares structural equation model. Theoretical assumptions about carbon emissions from electricity are constructed, causal paths between variables are defined, and path coefficients of the causal paths are estimated based on an analytical model of coordinated carbon reduction strategies for electricity. By combining the spatial correlation characteristics, nonlinear threshold characteristics, and path coefficients of causal paths, a differentiated power carbon reduction strategy oriented towards spatial constraints is determined.

2. The typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in claim 1, characterized in that, This analysis examines the spatiotemporal patterns of electricity carbon emissions in the target city based on electricity carbon emission accounting data, specifically including: Based on electricity carbon emission accounting data, the global Moran index is calculated to quantify the overall spatial correlation attributes of electricity carbon emissions, and the global spatial correlation characteristics are obtained. Based on electricity carbon emission accounting data, the local Moran index is calculated to obtain local spatial clustering characteristics.

3. The typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in claim 1, characterized in that, The multidimensional driving factor evaluation index system includes primary indicators and secondary indicators. The primary indicators include economic development level, urbanization process, power energy characteristics and geographical space. The secondary indicators include GDP, urbanization rate, total population at the end of the year, power generation, power consumption, proportion of renewable energy and urban centripetal force.

4. The typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in claim 1, characterized in that, When constructing a multi-dimensional driving factor evaluation index system, geospatial space is used as a distance constraint variable, and the characteristics of power energy and the level of socio-economic development are considered together.

5. The typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in claim 1, characterized in that, The nonlinear carbon emission analysis model is expressed as follows: in, The final output of the model is the predicted carbon emissions from electricity. As multidimensional driving factor characteristic variables, For the first The regression results of the decision trees, The total number of decision trees.

6. The typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in claim 1, characterized in that, The analytical model for the coordinated strategy of reducing carbon emissions in the power sector uses the year-end total population, urbanization rate, GDP, and urban centripetal force as independent variables, and the characteristics of power energy as potential independent variables.

7. The typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in claim 6, characterized in that, In the power sector carbon reduction collaborative strategy analysis model, the relationship between each variable is represented by path coefficients, as follows: in, For total carbon emissions, 、 、 、 、 These are the total population at the end of the year, electricity and energy characteristics, urbanization rate, GDP, and urban centripetal force. 、 、 、 、 These are the path coefficients for carbon emissions for each independent variable. This is the error term.

8. A typical urban carbon reduction system based on the spatiotemporal evolution of carbon emissions from electricity, characterized in that... include: The data acquisition module is configured to: acquire electricity energy consumption data of the target city, and combine it with implicit carbon emissions to obtain electricity carbon emission accounting data of the target city; The spatiotemporal analysis module is configured to: introduce spatial statistics theory, analyze the spatiotemporal pattern of electricity carbon emissions in the target city based on the electricity carbon emission accounting data, and obtain spatial correlation characteristics; The indicator system construction module is configured to: based on the spatial correlation characteristics of the constraint on carbon emissions from electricity, introduce geographic space, and construct a multi-dimensional driving factor evaluation indicator system; The multidimensional factor analysis module is configured to: construct a nonlinear carbon emission analysis model based on the multidimensional driving factors in the multidimensional driving factor evaluation index system and the random forest algorithm, and obtain the nonlinear response law of multidimensional driving factors to electricity carbon emissions. The collaborative analysis module is configured to: decouple the target variable and specific feature variables from the nonlinear carbon emission analysis model based on the partial dependency graph, obtain the marginal effect and nonlinear threshold characteristics of the feature variables, and construct a collaborative strategy analysis model for power carbon reduction by combining the partial least squares structural equation model. The strategy determination module is configured to: construct theoretical assumptions about electricity carbon emissions, define causal paths between variables, estimate the path coefficients of the causal paths based on the electricity carbon reduction collaborative strategy analysis model, and determine differentiated electricity carbon reduction strategies oriented towards spatial constraints by integrating the spatial correlation characteristics, nonlinear threshold characteristics and path coefficients of the causal paths.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the typical urban carbon reduction method based on the spatiotemporal evolution of electricity carbon emissions as described in any one of claims 1-7.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the typical urban carbon reduction method based on the spatiotemporal evolution law of electricity carbon emissions as described in any one of claims 1-7.

Images