Carbon emission prediction method and device based on combination of GIS-LCA and deep learning
By combining GIS-LCA and deep learning, a carbon emission prediction model for coastal cities was constructed, which solved the problems of accuracy and implementation difficulty in carbon emission prediction in coastal cities, and achieved multi-dimensional carbon emission prediction and planning convenience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-13
- Publication Date
- 2026-04-03
AI Technical Summary
Existing carbon emission prediction methods suffer from low accuracy and difficulty in implementation in coastal cities, especially due to the insufficient consideration of the impact of industrial structure.
By combining GIS-LCA and deep learning technologies, a spatialized carbon footprint map is constructed by acquiring spatiotemporal data on industries and carbon emissions in coastal areas. A deep learning model is then used to predict carbon emissions, and a multi-dimensional analysis is conducted in conjunction with industrial policies and technological scenarios.
It enables accurate prediction of carbon emissions in coastal cities, improves the convenience and accuracy of carbon emission reduction planning, and can display various carbon emission prediction data under different industry scenarios.
Smart Images
Figure CN121787672A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of carbon emission prediction, specifically to a carbon emission prediction method and apparatus based on a combination of GIS-LCA and deep learning. Background Technology
[0002] The scientific validity and rationality of carbon emission accounting methods directly affect the reliability of carbon emission data, which is crucial for formulating emission reduction policies, assessing environmental impacts, and promoting green development. Therefore, choosing an appropriate carbon emission calculation method is particularly important. Currently, carbon emission calculation methods and models used in academia and practice can be broadly classified into two categories based on their design principles: macroscopic and microscopic. Macroscopic estimation models primarily take a holistic perspective, providing a conceptual explanation of carbon emissions on a large scale and offering corresponding accounting frameworks and methods. Microscopic estimation models, on the other hand, are more specific, directly analyzing different types of emission sources to estimate specific carbon emissions.
[0003] Among the many existing methods, three techniques that combine macroscopic and microscopic characteristics are widely used: the emission factor method, the mass balance method, and the measurement method. The basic idea of the emission factor method is to identify and list all possible emission sources based on the contents of the carbon emission inventory, and then construct corresponding activity data and emission factors for each emission source. Based on this, multiplying the activity data by the emission factors yields an estimated carbon emission from the project. This method is highly operable, but its accuracy is highly dependent on the precision of the activity data and emission factors. The mass balance method estimates carbon emissions through a different logical path: it calculates the share of new chemical substances consumed to meet the capacity requirements of these new devices or the removal of gases when replacing old devices, based on the quantity of new chemical substances and equipment used annually for production and daily life, combined with the capacity requirements of these new devices or the removal of gases when replacing old equipment, thus deriving the relevant carbon emissions. In contrast, the measurement method focuses more on the actual measurement process. It uses on-site measured data from emission sources as its core, systematically collecting, organizing, and summarizing this data to obtain a more accurate carbon emission result.
[0004] However, while the aforementioned traditional carbon emission estimation methods can meet current needs to some extent, they still have some significant technical shortcomings. First, these methods generally rely too heavily on acquiring data from various energy sources, which typically requires substantial human and material resources for detailed recording and reporting. Furthermore, some methods (such as the measurement method) require high-precision and expensive measuring equipment, undoubtedly increasing the difficulty and economic burden of implementation. More importantly, for coastal cities, their industrial structure directly impacts carbon emission data; therefore, ignoring industrial structure in carbon emission prediction also results in low prediction accuracy. Summary of the Invention
[0005] The purpose of this application is to overcome the shortcomings and deficiencies of the existing technology and provide a carbon emission prediction method and device based on the combination of GIS-LCA and deep learning.
[0006] The first aspect of this application provides a carbon emission prediction method based on a combination of GIS-LCA and deep learning, including: Using GIS-LCA technology, historical spatiotemporal data of target industries and historical spatiotemporal data of carbon emissions in the target coastal area are obtained, and a historical spatialized carbon footprint map is constructed based on the historical carbon emission spatiotemporal data. A first deep learning model is trained based on the historical spatialized carbon footprint map and the historical remote sensing images of the target coastal area to obtain a spatial prediction model for carbon emissions. A generative adversarial network is trained based on historical spatiotemporal data of industries, historical industrial policy data, and historical industrial technology data of the target coastal area to obtain a scenario generation model that simulates industrial scenario data based on industrial policy data and industrial technology data. A second deep learning model is trained based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model. The actual remote sensing images of the target coastal area are input into the carbon emission spatial prediction model to obtain the actual spatialized carbon footprint map of the target coastal area. Data analysis is performed based on the actual industrial data of the target coastal area and the actual spatialized carbon footprint map to obtain the actual carbon emission data; Based on the scenario generation model, the following are obtained: first industry scenario data corresponding to the actual industrial policy data and actual industrial technology of the target coastal area; second simulated industry scenario data corresponding to the simulated industrial policy data of the target industry; and third simulated industry scenario data corresponding to the simulated industrial technology data. The actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data are input into the carbon emission time series prediction model to obtain the first carbon emission prediction data, the second carbon emission prediction data affected by industrial policies, and the third carbon emission prediction data affected by industrial technologies. Data analysis is performed based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry.
[0007] As one implementation method, the carbon emission prediction and analysis data includes a carbon emission reduction score; The step of performing data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry includes: The first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data are divided into periods according to a preset time period. Based on the data differences between the second and third carbon emission prediction data and the first carbon emission prediction data in each period, a comprehensive carbon emission reduction score for each period corresponding to the second and third carbon emission prediction data is obtained.
[0008] As one implementation method, the step of obtaining the comprehensive carbon emission reduction score for each period corresponding to the second and third carbon emission prediction data based on the data differences between the second and third carbon emission prediction data and the first carbon emission prediction data in each period includes: Based on the preset correspondence between data difference and cycle score, and the data difference between the second carbon emission prediction data and the third carbon emission prediction data and the first carbon emission prediction data in each cycle, the carbon emission reduction cycle score of the second carbon emission prediction data and the third carbon emission prediction data for each cycle is obtained. The overall carbon emission reduction score for each cycle is obtained by combining the carbon emission reduction cycle score of each cycle with the carbon emission reduction cycle scores of all preceding cycles.
[0009] As one implementation method, the step of obtaining the comprehensive carbon emission reduction score for each cycle based on the carbon emission reduction cycle score of each cycle and the carbon emission reduction cycle scores of all preceding cycles includes: The overall carbon emission reduction score for each cycle can be obtained using the following formula:
[0010] in, The overall carbon emission reduction score for the k-th cycle is... Score the carbon emission reduction cycle for the i-th cycle. This is the preset time compensation coefficient.
[0011] As one implementation method, the step of performing data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry includes: The output includes a visualization of the first carbon emission prediction data, the second carbon emission prediction data, the third carbon emission prediction data, and a decision support chart of the carbon emission prediction analysis data.
[0012] As one implementation method, the step of constructing a historical spatialized carbon footprint map based on the historical carbon emission spatiotemporal data includes: Based on the historical carbon emission spatiotemporal data, the target industry's full life-cycle carbon footprint is calculated, and the calculation results are spatialized and visualized to generate a spatialized carbon footprint map. The spatialized grid allocation includes: allocating the unit process carbon emissions obtained from LCA calculation to a unified geographic grid system based on their spatial attributes or impact range through GIS spatial overlay analysis and area weight allocation method.
[0013] As one implementation, the input features of the carbon emission time series prediction model include coded policy text quantification features, energy structure features, industry scale features, and meteorological features. The carbon emission time series prediction model uses a self-attention mechanism to capture the nonlinear coupling relationship of multiple input features.
[0014] The step of training a second deep learning model based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model includes: Based on the historical industry spatiotemporal data and the historical carbon emission spatiotemporal data, multiple training data samples are generated; each training data sample includes historical carbon emission spatiotemporal data corresponding to a first time period as input, historical industry spatiotemporal data corresponding to a first time period and a second time period as output, and historical emission spatiotemporal data corresponding to a second time period as output; wherein, the second time period is later than the first time period; A second deep learning model is trained based on the multiple training data samples to obtain a carbon emission time series prediction model.
[0015] Compared to existing technologies, the carbon emission prediction method of this application obtains the actual spatial carbon footprint map of the target coastal area through a carbon emission spatial prediction model, and then performs data analysis based on the actual industrial data of the target coastal area and the actual spatial carbon footprint map to obtain actual carbon emission data. Through a scenario generation model, it obtains first industry scenario data corresponding to the actual industrial policy data and actual industrial technology of the target coastal area, second simulated industry scenario data corresponding to the simulated industrial policy data of the target industry, and third simulated industry scenario data corresponding to the simulated industrial technology data. Then, it inputs the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data into the carbon emission time-series prediction model to obtain first carbon emission prediction data, second carbon emission prediction data affected by industrial policies, and third carbon emission prediction data affected by industrial technologies. This method combines multiple dimensions such as space, industry, and time to predict carbon emissions, and can predict multiple emission prediction data corresponding to different industry scenarios for data display and analysis. It can accurately and clearly display the analysis of multiple carbon emission prediction data corresponding to different industry scenarios, improving the convenience of carbon reduction planning for target coastal cities.
[0016] The second embodiment of this application provides a carbon emission prediction device based on a combination of GIS-LCA and deep learning, comprising: The data acquisition module is used to acquire historical spatiotemporal data of the target industry and historical carbon emission spatiotemporal data of the target coastal area through GIS-LCA technology, and to construct a historical spatialized carbon footprint map based on the historical carbon emission spatiotemporal data. The first model training module is used to train a first deep learning model based on the historical spatialized carbon footprint map and the historical remote sensing images of the target coastal area to obtain a carbon emission spatial prediction model. The second model training module is used to train a generative adversarial network based on the historical spatiotemporal data of industries, historical industrial policy data, and historical industrial technology data of the target coastal area, so as to obtain a scenario generation model. The third model training module is used to train a second deep learning model based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model. The actual carbon footprint map acquisition module is used to input the actual remote sensing images of the target coastal area into the carbon emission spatial prediction model to obtain the actual spatialized carbon footprint map of the target coastal area. The actual carbon emission data acquisition module is used to perform data analysis based on the actual industrial data of the target coastal area and the actual spatialized carbon footprint map to obtain actual carbon emission data. The industry scenario data acquisition module is used to acquire, based on the scenario generation model, first industry scenario data corresponding to the actual industry policy data and actual industry technology of the target coastal area, second simulated industry scenario data corresponding to the simulated industry policy data of the target industry, and third simulated industry scenario data corresponding to the simulated industry technology data. The carbon emission prediction module is used to input the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data into the carbon emission time series prediction model to obtain the first carbon emission prediction data, the second carbon emission prediction data affected by industrial policies, and the third carbon emission prediction data affected by industrial technologies. The carbon emission prediction and analysis module is used to perform data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction and analysis data corresponding to the target coastal area and the target industry.
[0017] Compared to existing technologies, the carbon emission prediction device of this application obtains the actual spatial carbon footprint map of the target coastal area through a carbon emission spatial prediction model, and then performs data analysis based on the actual industrial data of the target coastal area and the actual spatial carbon footprint map to obtain actual carbon emission data. Through a scenario generation model, it obtains first industry scenario data corresponding to the actual industrial policy data and actual industrial technology of the target coastal area, second simulated industry scenario data corresponding to the simulated industrial policy data of the target industry, and third simulated industry scenario data corresponding to the simulated industrial technology data. Then, it inputs the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data into the carbon emission time series prediction model to obtain first carbon emission prediction data, second carbon emission prediction data affected by industrial policies, and third carbon emission prediction data affected by industrial technologies. It combines multiple dimensions such as space, industry, and time to predict carbon emissions, and can predict multiple emission prediction data corresponding to different industry scenarios for data display and data analysis. It can accurately and clearly display the analysis of multiple carbon emission prediction data corresponding to different industry scenarios, improving the convenience of carbon emission reduction planning for target coastal cities.
[0018] To provide a clearer understanding of this application, the specific embodiments of this application will be described below in conjunction with the accompanying drawings. Attached Figure Description
[0019] Figure 1 This is a flowchart illustrating a carbon emission prediction method based on a combination of GIS-LCA and deep learning, according to one embodiment of this application.
[0020] Figure 2This is a schematic diagram illustrating data analysis based on a combination of GIS-LCA and deep learning, according to one embodiment of this application.
[0021] Figure 3 This is a module connection diagram of a carbon emission prediction device based on the combination of GIS-LCA and deep learning according to an embodiment of this application.
[0022] 100. Carbon emission prediction device; 101. Data acquisition module; 102. First model training module; 103. Second model training module; 104. Third model training module; 105. Actual carbon footprint map acquisition module; 106. Actual carbon emission data acquisition module; 107. Industry scenario data acquisition module; 108. Carbon emission prediction module; 109. Carbon emission prediction and analysis module. Detailed Implementation
[0023] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.
[0024] It should be understood that the described embodiments are merely some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.
[0025] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances. The singular forms "a," "the," and "the" used in this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. The word "if" as used herein can be interpreted as "when," "when," or "in response to determination."
[0026] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.
[0027] Please see Figure 1This is a flowchart of a carbon emission prediction method based on a combination of GIS-LCA and deep learning, according to the first embodiment of this application. The method includes: S1: Using GIS-LCA technology, acquire historical spatiotemporal data of the target industry and historical spatiotemporal data of carbon emissions in the target coastal area, and construct a historical spatialized carbon footprint map based on the historical carbon emission spatiotemporal data. The target industries are those that account for ≥5% of carbon emissions in the target coastal areas, including but not limited to energy production, steel industry, port transportation, mariculture, and offshore wind power.
[0028] S2: Train the first deep learning model based on the historical spatialized carbon footprint map and the historical remote sensing images of the target coastal area to obtain a carbon emission spatial prediction model; Among them, the carbon emission spatial prediction model is used to predict the carbon footprint based on remote sensing imagery, so as to predict the carbon emission data of the target coastal area based on spatial dimensions.
[0029] S3: Train a generative adversarial network based on the historical spatiotemporal data of industries, historical industrial policy data, and historical industrial technology data of the target coastal area to obtain a scenario generation model that simulates industrial scenario data based on industrial policy data and industrial technology data; Industrial policy data can be either vectorized or numerical. Taking numerical data as an example, the "subsidy amount" of industrial policy is converted according to the proportion of subsidy amount to industrial cost (10 points for a proportion ≥20%, 8 points for 10%-20%, and 5 points for ≤10%). The "mandatory standards" of industrial policy are converted according to the proportion of enterprises covered (10 points for coverage ≥90%, and 8 points for 70%-90%).
[0030] Industrial technology data refers to quantitative or numerical data, where industrial technology includes low-carbon technologies that can be applied to the target industry.
[0031] S4: Train a second deep learning model based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model; Step S4 includes: S41: Based on the historical industry spatiotemporal data and the historical carbon emission spatiotemporal data, generate multiple training data samples; each training data sample includes historical carbon emission spatiotemporal data corresponding to the first time period as input, historical industry spatiotemporal data corresponding to the first time period and the second time period as output, and historical emission spatiotemporal data corresponding to the second time period as output; wherein, the second time period is later than the first time period; S42: Train a second deep learning model based on the multiple training data samples to obtain a carbon emission time series prediction model.
[0032] S5: Input the actual remote sensing image of the target coastal area into the carbon emission spatial prediction model to obtain the actual spatialized carbon footprint map of the target coastal area; S6: Based on the actual industrial data of the target coastal area and the actual spatialized carbon footprint map, data analysis is performed to obtain the actual carbon emission data; The actual carbon emission data includes current carbon emission data as well as carbon emission data from previous periods, such as carbon emission data for the past 3 months, carbon emission data for the past 6 months, carbon emission data for the past 12 months, etc.
[0033] S7: Based on the scenario generation model, obtain the first industrial scenario data corresponding to the actual industrial policy data and actual industrial technology of the target coastal area, the second simulated industrial scenario data corresponding to the simulated industrial policy data of the target industry, and the third simulated industrial scenario data corresponding to the simulated industrial technology data; Among them, the primary industry scenario data is the industry data corresponding to the baseline scenario. The corresponding conditions are the actual industry policy data and the actual industry technology. The condition vector obtained by converting the actual industry policy data is [current proportion, current quota], and the condition vector obtained by converting the actual industry technology is [current technology carbon emission data], thus generating the "development as usual" path.
[0034] The second simulated industry scenario data consists of industry data corresponding to policy-driven scenarios. For example, policy-driven scenario A: the condition vector is [the proportion is increased to 80%, and the subsidy amount is increased by 50%], generating a "strengthened incentive" path; policy-driven scenario B: the condition vector is [the proportion is increased to 80%, and the subsidy amount remains unchanged], generating a "mandatory regulation" path, etc.
[0035] The third simulated industry scenario data point corresponds to the industry data of the technology update scenario. For example, in the technology update scenario, the condition vector is [new technology carbon emission data].
[0036] S8: Input the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data into the carbon emission time series prediction model to obtain the first carbon emission prediction data, the second carbon emission prediction data affected by industrial policies, and the third carbon emission prediction data affected by industrial technologies. S9: Based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data, perform data analysis to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry.
[0037] Please see Figure 2Using the baseline scenario, policy-driven scenario A, and policy-driven scenario B as examples for data analysis, policy-driven scenario A represents an optimal path with a comparable emission reduction of approximately 28% by 2030 compared to the baseline scenario. If the analysis incorporates the technology update scenario, the emission reduction data from the technology update scenario in 2030 compared to the baseline scenario are analyzed together to determine the optimal carbon emission reduction scenario with 2030 as the limiting factor.
[0038] In one feasible embodiment, the carbon emission prediction and analysis data includes a carbon emission reduction score; S9: The step of performing data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry includes: S91: Divide the first carbon emission prediction data, the second carbon emission prediction data and the third carbon emission prediction data into periods according to a preset time period; S92: Based on the data differences between the second carbon emission prediction data and the third carbon emission prediction data and the first carbon emission prediction data in each period, obtain the comprehensive carbon emission reduction score for each period corresponding to the second carbon emission prediction data and the third carbon emission prediction data.
[0039] In a feasible embodiment, S92: the step of obtaining the comprehensive carbon emission reduction score for each period corresponding to the second carbon emission prediction data and the third carbon emission prediction data based on the data difference between the second carbon emission prediction data and the third carbon emission prediction data and the first carbon emission prediction data in each period includes: S921: Based on the preset correspondence between data difference and cycle score, and the data difference between the second carbon emission prediction data and the third carbon emission prediction data and the first carbon emission prediction data in each cycle, obtain the carbon emission reduction cycle score of the second carbon emission prediction data and the third carbon emission prediction data for each cycle. S922: Obtain the comprehensive carbon emission reduction score for each cycle based on the carbon emission reduction cycle score of each cycle and the carbon emission reduction cycle scores of all preceding cycles.
[0040] In a feasible embodiment, step S922: obtaining the comprehensive carbon emission reduction score for each cycle based on the carbon emission reduction cycle score of each cycle and the carbon emission reduction cycle scores of all preceding cycles, includes: The overall carbon emission reduction score for each cycle can be obtained using the following formula:
[0041] in, The overall carbon emission reduction score for the k-th cycle is... Score the carbon emission reduction cycle for the i-th cycle. This is the preset time compensation coefficient.
[0042] The time compensation coefficient represents the increase in the time dimension corresponding to the beneficial effect of carbon emission reduction on the environment. For example, it includes periods a, b, c, d, e, and f. In this case, period a of the second carbon emission prediction data can achieve a relatively high carbon emission reduction, while period d of the third carbon emission prediction data is required to achieve a relatively high carbon emission reduction. Therefore, period a of the second carbon emission prediction data can significantly bring about changes in carbon emission reduction from the beginning. When the carbon emission reduction period scores of period a of the second carbon emission prediction data and period d of the third carbon emission prediction data are the same, period a of the second carbon emission prediction data is superior to period d of the third carbon emission prediction data in terms of time. For example, it can be set to 1.05, 1.1, 1.12, 1.2, etc.
[0043] In a feasible embodiment, the step S9, after performing data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry, includes: S93: Output a visual decision support chart that includes the first carbon emission prediction data, the second carbon emission prediction data, the third carbon emission prediction data, and the carbon emission prediction analysis data.
[0044] In one feasible embodiment, the step of constructing a historical spatialized carbon footprint map based on the historical carbon emission spatiotemporal data includes: Based on the historical carbon emission spatiotemporal data, the target industry's full life-cycle carbon footprint is calculated, and the calculation results are spatialized and visualized to generate a spatialized carbon footprint map. The spatialized grid allocation includes: allocating the unit process carbon emissions obtained from LCA calculation to a unified geographic grid system based on their spatial attributes or impact range through GIS spatial overlay analysis and area weight allocation method.
[0045] In one feasible embodiment, the input features of the carbon emission time series prediction model include coded policy text quantification features, energy structure features, industry scale features, and meteorological features. The carbon emission time series prediction model uses a self-attention mechanism to capture the nonlinear coupling relationship of multiple input features.
[0046] Compared to existing technologies, the carbon emission prediction method of this application obtains the actual spatial carbon footprint map of the target coastal area through a carbon emission spatial prediction model, and then performs data analysis based on the actual industrial data of the target coastal area and the actual spatial carbon footprint map to obtain actual carbon emission data. Through a scenario generation model, it obtains first industry scenario data corresponding to the actual industrial policy data and actual industrial technology of the target coastal area, second simulated industry scenario data corresponding to the simulated industrial policy data of the target industry, and third simulated industry scenario data corresponding to the simulated industrial technology data. Then, it inputs the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data into the carbon emission time-series prediction model to obtain first carbon emission prediction data, second carbon emission prediction data affected by industrial policies, and third carbon emission prediction data affected by industrial technologies. This method combines multiple dimensions such as space, industry, and time to predict carbon emissions, and can predict multiple emission prediction data corresponding to different industry scenarios for data display and analysis. It can accurately and clearly display the analysis of multiple carbon emission prediction data corresponding to different industry scenarios, improving the convenience of carbon reduction planning for target coastal cities.
[0047] Please see Figure 3 The second embodiment of this application provides a carbon emission prediction device 100 based on a combination of GIS-LCA and deep learning, comprising: The data acquisition module 101 is used to acquire historical spatiotemporal data of the target industry and historical carbon emission spatiotemporal data of the target coastal area through GIS-LCA technology, and to construct a historical spatialized carbon footprint map based on the historical carbon emission spatiotemporal data. The first model training module 102 is used to train a first deep learning model based on the historical spatialized carbon footprint map and the historical remote sensing images of the target coastal area to obtain a carbon emission spatial prediction model. The second model training module 103 is used to train a generative adversarial network based on the historical spatiotemporal data of the target coastal area, historical industrial policy data, and historical industrial technology data to obtain a scenario generation model. The third model training module 104 is used to train a second deep learning model based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model. The actual carbon footprint map acquisition module 105 is used to input the actual remote sensing image of the target coastal area into the carbon emission spatial prediction model to obtain the actual spatialized carbon footprint map of the target coastal area. The actual carbon emission data acquisition module 106 is used to perform data analysis based on the actual industrial data of the target coastal area and the actual spatialized carbon footprint map to obtain actual carbon emission data. The industry scenario data acquisition module 107 is used to acquire, according to the scenario generation model, first industry scenario data corresponding to the actual industry policy data and actual industry technology of the target coastal area, second simulated industry scenario data corresponding to the simulated industry policy data of the target industry, and third simulated industry scenario data corresponding to the simulated industry technology data. The carbon emission prediction module 108 is used to input the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data and the third simulated industry scenario data into the carbon emission time series prediction model to obtain the first carbon emission prediction data, the second carbon emission prediction data affected by industrial policies and the third carbon emission prediction data affected by industrial technologies. The carbon emission prediction and analysis module 109 is used to perform data analysis based on the first carbon emission prediction data, the second carbon emission prediction data and the third carbon emission prediction data to obtain carbon emission prediction and analysis data corresponding to the target coastal area and the target industry.
[0048] It should be noted that the carbon emission prediction device 100 provided in the second embodiment of this application is only illustrated by the above-described division of functional modules when executing the carbon emission prediction method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the carbon emission prediction device 100 provided in the second embodiment of this application and the carbon emission prediction method of the first embodiment of this application belong to the same concept, and its implementation process is detailed in the method embodiment, which will not be repeated here.
[0049] The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without any inventive effort.
[0050] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0051] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function selected in one or more boxes.
[0052] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function selected in one or more boxes.
[0053] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0054] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0055] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0056] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0057] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A carbon emission prediction method based on a combination of GIS-LCA and deep learning, characterized in that, include: Using GIS-LCA technology, historical spatiotemporal data of target industries and historical spatiotemporal data of carbon emissions in the target coastal area are obtained, and a historical spatialized carbon footprint map is constructed based on the historical spatiotemporal data of carbon emissions. A first deep learning model is trained based on the historical spatialized carbon footprint map and the historical remote sensing images of the target coastal area to obtain a spatial prediction model for carbon emissions. A generative adversarial network is trained based on historical spatiotemporal industrial data, historical industrial policy data, and historical industrial technology data of the target coastal area to obtain a scenario generation model that simulates industrial scenario data based on industrial policy data and industrial technology data. A second deep learning model is trained based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model. The actual remote sensing images of the target coastal area are input into the carbon emission spatial prediction model to obtain the actual spatialized carbon footprint map of the target coastal area. Data analysis is performed based on the actual industrial data of the target coastal area and the actual spatialized carbon footprint map to obtain the actual carbon emission data; Based on the scenario generation model, the following are obtained: first industry scenario data corresponding to the actual industrial policy data and actual industrial technology of the target coastal area; second simulated industry scenario data corresponding to the simulated industrial policy data of the target industry; and third simulated industry scenario data corresponding to the simulated industrial technology data. The actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data are input into the carbon emission time series prediction model to obtain the first carbon emission prediction data, the second carbon emission prediction data affected by industrial policies, and the third carbon emission prediction data affected by industrial technologies. Data analysis is performed based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry.
2. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to claim 1, characterized in that, The carbon emission prediction and analysis data includes carbon emission reduction scores; The step of performing data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry includes: The first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data are divided into periods according to a preset time period. Based on the data differences between the second and third carbon emission prediction data and the first carbon emission prediction data in each period, a comprehensive carbon emission reduction score for each period corresponding to the second and third carbon emission prediction data is obtained.
3. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to claim 2, characterized in that, The step of obtaining the comprehensive carbon emission reduction score for each period based on the data difference between the second and third carbon emission prediction data and the first carbon emission prediction data in each period includes: Based on the preset correspondence between data difference and cycle score, and the data difference between the second carbon emission prediction data and the third carbon emission prediction data and the first carbon emission prediction data in each cycle, the carbon emission reduction cycle score of the second carbon emission prediction data and the third carbon emission prediction data for each cycle is obtained. The comprehensive carbon reduction score for each cycle is obtained by combining the carbon reduction cycle score of each cycle with the carbon reduction cycle scores of all preceding cycles.
4. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to claim 3, characterized in that, The step of obtaining the comprehensive carbon emission reduction score for each cycle based on the carbon emission reduction cycle score of each cycle and the carbon emission reduction cycle scores of all preceding cycles includes: The comprehensive carbon emission reduction score for each cycle can be obtained using the following formula: in, The overall carbon emission reduction score for the k-th cycle is... Score the carbon emission reduction cycle for the i-th cycle. This is the preset time compensation coefficient.
5. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to any one of claims 1-4, characterized in that, After performing data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction analysis data corresponding to the target coastal area and the target industry, the process includes: The output includes a visualization of the first carbon emission prediction data, the second carbon emission prediction data, the third carbon emission prediction data, and a decision support chart of the carbon emission prediction analysis data.
6. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to claim 1, characterized in that, The steps for constructing a historical spatialized carbon footprint map based on the historical spatiotemporal data of carbon emissions include: Based on the historical carbon emission spatiotemporal data, the target industry's full life-cycle carbon footprint is calculated, and the calculation results are spatialized and visualized to generate a spatialized carbon footprint map. The spatialized grid allocation includes: allocating the unit process carbon emissions obtained from LCA calculation to a unified geographic grid system based on their spatial attributes or impact range through GIS spatial overlay analysis and area weight allocation method.
7. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to claim 1, characterized in that, The input features of the carbon emission time series prediction model include coded policy text quantification features, energy structure features, industry scale features, and meteorological features. The carbon emission time series prediction model uses a self-attention mechanism to capture the nonlinear coupling relationship of multiple input features.
8. The carbon emission prediction method based on the combination of GIS-LCA and deep learning according to claim 1, characterized in that, The step of training a second deep learning model based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model includes: Based on the historical industry spatiotemporal data and the historical carbon emission spatiotemporal data, multiple training data samples are generated; each training data sample includes historical carbon emission spatiotemporal data corresponding to a first time period as input, historical industry spatiotemporal data corresponding to a first time period and a second time period as output, and historical emission spatiotemporal data corresponding to a second time period as output; wherein, the second time period is later than the first time period; A second deep learning model is trained based on the multiple training data samples to obtain a carbon emission time series prediction model.
9. A carbon emission prediction device based on a combination of GIS-LCA and deep learning, characterized in that, include: The data acquisition module is used to acquire historical spatiotemporal data of the target industry and historical carbon emission spatiotemporal data of the target coastal area through GIS-LCA technology, and to construct a historical spatialized carbon footprint map based on the historical carbon emission spatiotemporal data. The first model training module is used to train a first deep learning model based on the historical spatialized carbon footprint map and the historical remote sensing images of the target coastal area to obtain a carbon emission spatial prediction model. The second model training module is used to train a generative adversarial network based on the historical spatiotemporal data of industries, historical industrial policy data, and historical industrial technology data of the target coastal area, so as to obtain a scenario generation model. The third model training module is used to train a second deep learning model based on the historical industrial spatiotemporal data and the historical carbon emission spatiotemporal data to obtain a carbon emission time series prediction model. The actual carbon footprint map acquisition module is used to input the actual remote sensing images of the target coastal area into the carbon emission spatial prediction model to obtain the actual spatialized carbon footprint map of the target coastal area. The actual carbon emission data acquisition module is used to perform data analysis based on the actual industrial data of the target coastal area and the actual spatialized carbon footprint map to obtain actual carbon emission data. The industry scenario data acquisition module is used to acquire, based on the scenario generation model, first industry scenario data corresponding to the actual industry policy data and actual industry technology of the target coastal area, second simulated industry scenario data corresponding to the simulated industry policy data of the target industry, and third simulated industry scenario data corresponding to the simulated industry technology data. The carbon emission prediction module is used to input the actual carbon emission data, the first industry scenario data, the second simulated industry scenario data, and the third simulated industry scenario data into the carbon emission time series prediction model to obtain the first carbon emission prediction data, the second carbon emission prediction data affected by industrial policies, and the third carbon emission prediction data affected by industrial technologies. The carbon emission prediction and analysis module is used to perform data analysis based on the first carbon emission prediction data, the second carbon emission prediction data, and the third carbon emission prediction data to obtain carbon emission prediction and analysis data corresponding to the target coastal area and the target industry.