Park carbon emission management method, equipment and medium

Abstract

The invention discloses a park carbon emission management method, equipment and a medium, and relates to the technical field of carbon emission intelligent monitoring, accounting and management. The method comprises the following steps: acquiring multi-source data related to carbon emission of a park; according to the multi-source data, calibrating a carbon emission factor corresponding to the park to obtain a dynamic calibration factor; performing carbon flow simulation on the carbon emission of the park based on the twins of the park and the dynamic calibration factor to obtain a simulation result of at least one emission reduction scheme; according to a simulation result, generating a zero-carbon path plan corresponding to the park; and under the framework of zero carbon path planning, based on a reinforcement learning algorithm, generating a carbon emission strategy corresponding to the park, and executing the carbon emission strategy. Therefore, precise park carbon emission management can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring, accounting and management of carbon emissions, and in particular to a park carbon emission management method, device and medium. BACKGROUND

[0002] With the implementation of the "double carbon" policy, the park as a core carrier of carbon emissions, the accuracy of its carbon emission accounting directly affects the fairness of carbon trading, the scientificity of ESG rating and the effectiveness of emission reduction strategy.

[0003] With the deepening of the "double carbon" policy and the expansion of the global carbon market, the park as a core carrier of carbon emissions, its carbon management needs have been upgraded from "passive compliance accounting" to "active performance optimization". The existing technology has defects and cannot meet the precise carbon management needs of the park. SUMMARY

[0004] The embodiments of the present application provide a park carbon emission management method, device and medium to solve the technical problem of how to meet the precise carbon management needs of the park.

[0005] In a first aspect, the embodiments of the present application provide a park carbon emission management method, which comprises: acquiring multi-source data related to carbon emissions of a park; calibrating carbon emission factors corresponding to the park according to the multi-source data to obtain dynamic calibration factors; simulating carbon flow of carbon emissions of the park based on a twin body of the park and the dynamic calibration factors to obtain simulation results of at least one emission reduction scheme; generating a zero-carbon path plan corresponding to the park according to the simulation results; and generating a carbon emission strategy corresponding to the park based on a reinforcement learning algorithm under the framework of the zero-carbon path plan and executing the carbon emission strategy.

[0006] In a second aspect, the embodiments of the present application further provide a park carbon emission management device, which comprises: at least one processor; and a memory in communication connection with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the park carbon emission management method of the first aspect.

[0007] In a third aspect, the embodiments of the present application further provide a computer storage medium storing computer executable instructions, which when executed, implement the park carbon emission management method of the first aspect.

[0008] The park carbon emission management method, device and medium provided by the embodiments of the present application have the following beneficial effects: In the embodiments of the present application, the carbon emission factor corresponding to the park can be calibrated according to multi-source data related to the carbon emission of the park to obtain a dynamic calibration factor, so that a more accurate carbon emission factor corresponding to the park can be obtained. Then, based on the twin of the park and the dynamic calibration factor, carbon flow simulation of the carbon emission of the park is performed to obtain a simulation result of at least one emission reduction scheme, and then a zero-carbon path planning corresponding to the park is generated; finally, under the framework of the zero-carbon path planning, a carbon emission strategy corresponding to the park is generated based on the above data and the reinforcement learning algorithm, and the carbon emission strategy is executed. In this way, the above method is developed around the carbon emission management of the park, and through data collection, accurate carbon emission factor calibration, scientific zero-carbon planning and intelligent decision execution, precise park carbon emission management can be realized, the park can be promoted to the direction of green, low-carbon and sustainable development, and the scientificity and efficiency of the park in carbon emission management can be improved, thereby helping the park to achieve the carbon emission target. BRIEF DESCRIPTION OF DRAWINGS

[0009] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation on the present application. In the drawings: Figure 1 A flowchart of a park carbon emission management method provided by an embodiment of the present application; Figure 2 An architecture diagram of a carbon data intelligent hub system provided by an embodiment of the present application; Figure 3 An architecture diagram of an intelligent gateway deployment provided by an embodiment of the present application; Figure 4 A structural schematic diagram of the inside of a park carbon emission management device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0010] To make the purpose, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below in combination with the specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts fall within the scope of protection of the present application.

[0011] With the deepening of the "double carbon" policy and the expansion of the global carbon market, the park, as the core carrier of carbon emission, has upgraded its carbon management demand from "passive compliance accounting" to "active performance optimization + risk control + ecological synergy". The existing technology has five core defects, which cannot meet the carbon management demand of the whole cycle, high adaptation and low threshold: Carbon factor static, poor adaptation to extreme scenarios Traditional methods use fixed industry factors (such as 0.6101 tCO2 / MWh for power grids), without considering extreme weather (cold waves leading to gas surges), sudden production disturbances (equipment failures), and policy mutations (temporary adjustment of carbon prices), resulting in accounting errors of 25-30%; some dynamic solutions (such as CN202310567890.2) only optimize power grid factors, without covering multiple types of factors and extreme scenarios, and have insufficient adaptability.

[0012] High cost of data collection, difficult to renovate old parks Existing solutions rely on wired intelligent devices. Old parks (built more than 10 years ago) have difficulty in wiring due to the lack of intelligent interfaces, and the renovation cost is more than 100 million yuan per park, which is difficult for small and medium-sized parks to bear. Data coverage is only energy consumption (coverage ≤60%), missing key data such as process and supply chain.

[0013] Accounting and management are disconnected, and risk control is missing Existing technologies stop at carbon emission statistics (such as CN202210897654.3), without correlating carbon policy compliance risks, carbon asset impairment risks, and supply chain carbon risks; lack of "risk early warning-response" mechanism, parks often suffer millions of losses due to policy misjudgment and asset depreciation.

[0014] Long-term planning is missing, and cross-domain collaboration is insufficient No 5-10 year zero-carbon path planning tool is provided, and parks are prone to blindly investing in high-cost emission reduction technologies; limited to single park management, without using cross-park carbon resource sharing (quota surplus / deficit complementation) and cross-industry technology migration (such as data center liquid cooling technology) to improve emission reduction efficiency.

[0015] High operation threshold, low third-party trust Core functions (such as model training, twin adjustment) require professional technical personnel to operate, and non-professional managers are difficult to use; the accounting method has not been authorized, and the third-party verification period is as long as 2-3 months, and the results are difficult to directly used for carbon trading and ESG rating.

[0016] After searching, existing technologies (such as CN202210897654.3, CN202310567890.2) only solve partial problems, lack of "multi-scenario adaptation + full-cycle management + cross-domain collaboration + low threshold landing" integrated solution, and there is an urgent need for a carbon data intelligent central system covering "data-algorithm-application-ecosystem".

[0017] Embodiments of the present application provide a park carbon emission management scheme, which will be described in detail below with reference to the accompanying drawings.

[0018] Figure 1A flowchart of a park carbon emission management method provided by an embodiment of the present application. As shown in Figure 1 The park carbon emission management method provided by an embodiment of the present application specifically includes the following steps: Step 101, obtaining multi-source data related to carbon emissions of the park.

[0019] In the embodiment of the present application, various multi-source data of different structures related to carbon emissions of the park can be obtained. The above multi-source data can comprehensively and accurately show the carbon emission situation of the park. In this way, a comprehensive, accurate and complete data basis can be provided for subsequent analysis and decision-making, ensuring reliable data quality and avoiding the influence of data problems on the accuracy and effectiveness of subsequent links.

[0020] Step 102, calibrating the carbon emission factor corresponding to the park according to the multi-source data to obtain a dynamic calibration factor.

[0021] In the embodiment of the present application, the carbon emission factor corresponding to the park can be calibrated according to the above multi-source data to obtain a dynamic calibration factor. The dynamic calibration factor can more truly reflect the real-time changes of the carbon emissions of the park. Traditional static factors are often based on annual or long-term average data, and it is difficult to capture the influence of seasonal fluctuations, energy structure adjustment or production activity changes on carbon emissions. In this way, the carbon emission situation of different energy and process links can be reflected in real time, providing accurate parameter basis for subsequent carbon emission calculation, analysis and decision-making.

[0022] Step 103, performing carbon flow simulation on the carbon emissions of the park based on the twin of the park and the dynamic calibration factor to obtain simulation results of at least one emission reduction scheme.

[0023] In actual application, a twin is a virtual model constructed by digital means, which is highly consistent with a physical entity or system. That is, the twin is a “mirror image” of the physical entity in the digital space, which not only replicates the geometric structure (such as shape, size) of the physical entity, but also simulates its dynamic behavior (such as running logic, performance changes) and environmental interaction (such as temperature, pressure, user operation).

[0024] In actual application, carbon flow simulation can simulate the complete path of carbon elements from energy input, production process to product output or waste treatment, realizing “end-to-end” tracking of carbon emissions. For example, in a manufacturing park, the carbon emission contributions of raw material mining, transportation, processing, assembly and other links can be clearly identified, providing a basis for precise emission reduction.

[0025] In the embodiment of the present application, the carbon flow simulation of the carbon emissions of the park can be performed based on the twin of the park and the dynamic calibration factor, and the simulation result of at least one emission reduction scheme is obtained. In this way, in the virtual park, the future carbon emission trajectory is simulated based on the dynamic calibration factor, and the effect of various intervention measures is evaluated. For example, the control variable method can be used to independently and accurately simulate discrete technical schemes such as new photovoltaic, deployment of energy storage, and implementation of energy-saving reconstruction. In this way, the independent emission reduction potential and implementation complexity can be accurately quantified. Through the above-mentioned manner, it is helpful to accurately understand the source and distribution of carbon emissions, and intuitively understand the carbon emission dynamics of the park through digital twinning technology, plan the path to achieve the zero-carbon target in advance, and evaluate the risks that different emission reduction schemes may face, which can provide support for formulating a scientific and reasonable zero-carbon path planning.

[0026] Step 104, generating the corresponding zero-carbon path planning of the park according to the simulation result.

[0027] In actual application, the zero-carbon path planning is the overall and strategic arrangement of the park to achieve zero carbon emission in the future. It clearly defines the vision, stage division and final zero-carbon state of the zero-carbon target, and provides a macro direction for the whole carbon emission management work.

[0028] In the embodiment of the present application, the zero-carbon path planning corresponding to the park can be generated according to the simulation result. Each verified emission reduction scheme is regarded as a "strategic resource library" that can be dispatched, and the zero-carbon path planning can realize the appropriate configuration of the above-mentioned emission reduction schemes in the time sequence and spatial dimension. For example, by using the operational research optimization algorithm and comprehensively considering the technical maturity curve, the time value of funds, the synergy and exclusion effect between projects and other long-term dynamic constraints, a globally optimal trajectory that smoothly and economically transitions from the current state to the carbon neutral target can be generated. In this way, global optimization can be achieved to avoid local optimization and investment waste, the technical evolution path can be clearly defined to avoid the risk of "technology lock-in", and carbon management can be upgraded from a passive and scattered behavior to an active, systematic and quantitative behavior.

[0029] Step 105, generating the carbon emission strategy corresponding to the park based on the reinforcement learning algorithm under the framework of the zero-carbon path planning and executing the carbon emission strategy.

[0030] In actual application, the carbon emission strategy is a specific action plan and measure formulated to achieve the goal of zero-carbon path planning, aiming to directly reduce carbon emissions or improve carbon utilization efficiency. It focuses on the specific operation in the present and near future to ensure that every step of action moves towards the zero-carbon target. For example, specific strategies such as improving energy utilization efficiency and adopting renewable energy are formulated to reduce carbon emissions.

[0031] In the embodiments of the present application, based on the reinforcement learning algorithm, the carbon emission strategy corresponding to the park can be generated and the carbon emission strategy is executed under the framework of the zero-carbon path planning. In this way, intelligent decision-making of carbon emission control is realized, ensuring that the decision can be quickly and accurately conveyed and executed. In actual application, the reinforcement learning algorithm can be continuously optimized through a feedback mechanism to improve the scientificity and effectiveness of the decision.

[0032] In the embodiments of the present application, the carbon emission factor corresponding to the park can be calibrated according to multi-source data related to carbon emission of the park to obtain a dynamic calibration factor, so that a more accurate carbon emission factor corresponding to the park can be obtained. Then, based on the twin of the park and the dynamic calibration factor, carbon flow simulation of carbon emission of the park is performed to obtain simulation results of at least one emission reduction scheme, and then the zero-carbon path planning corresponding to the park is generated. Finally, based on the above data and the reinforcement learning algorithm, the carbon emission strategy corresponding to the park is generated and the carbon emission strategy is executed under the framework of the zero-carbon path planning. In this way, the above method is developed around the carbon emission management of the park, and through data collection, accurate carbon emission factor calibration, scientific zero-carbon planning and intelligent decision execution, precise park carbon emission management can be realized, the park can be promoted to the direction of green, low-carbon and sustainable development, and the scientificity and efficiency of the park in carbon emission management can be improved to help the park achieve the goal of carbon emission.

[0033] In one possible implementation, the multi-source data related to carbon emission of the park is obtained, including: The numerical data related to carbon emission of the park is obtained through a carbon-energy integrated edge intelligent gateway. The multi-modal data related to the carbon emission data of the park is obtained through a mobile device and an interface. The numerical data and multi-modal data are preprocessed to obtain the multi-source data of the park.

[0034] In practical applications, the carbon integrated edge intelligent gateway can be deployed at key nodes in the park, support Modbus RTU (baud rate 9600 bps), OPC UA (sampling frequency 1 Hz), and MQTT (QoS level 2) multi-protocol access, and can synchronously collect power (0.5-level intelligent power meter DTZ341, 15 minutes / once), gas (1.0-level intelligent gas meter G2.5, 30 minutes / once), process tail gas (Testo 350 flue gas analyzer, real-time), and production work order (ERP system interface) data. The built-in ARM Cortex-A53 lightweight AI chip can complete real-time carbon flow estimation (error ≤3%) and initial screening of abnormal data on site. The carbon integrated edge intelligent gateway supports solar power supply (endurance ≥72 hours), LoRa wireless communication (transmission distance ≥5 km), and is suitable for Southeast Asian high-temperature and high-humidity environments (working temperature -10℃-60℃, humidity ≤95%) and wiring-free scenarios in old parks. In this way, no on-site hardware deployment is required in the park, and only deployment at key nodes (to read water meters, power meters, etc. in the park) is required, which can save costs. The specific numerical data content is not limited.

[0035] Secondly, multi-modal data can be supplemented and collected. For old equipment, a portable infrared meter (non-contact reading, accuracy ±1%) and a LoRa wireless sensor (battery endurance 1 year) can be provided. Through OCR / NLP technology, gas composition reports, policy news, and satellite remote sensing data can be accessed. The system can be connected to power grid enterprises (real-time power supply structure), banks (green credit data), and third-party certification agencies (test reports) to achieve full coverage of internal and external data. In this way, the high cost of upgrading old parks and the incomplete data coverage problem can be solved, the upgrading cost can be reduced by 70%, and energy-process-supply chain data coverage can be achieved by 100%. When preprocessing the above numerical data and multi-modal data, an autoencoder anomaly detection method (robustness improved by 40% compared to the 3σ criterion) can be used to remove extreme values. A spatio-temporal graph convolution network can be used to fill in missing values (accuracy ≥98%). Finally, Z-score standardization can be used to unify the data format, and then multi-source data can be obtained. The data integrity after preprocessing is ≥99.5%. In practical applications, other preprocessing methods can also be used, and the specific method is not limited. In one possible implementation, the calibration of the carbon emission factor corresponding to the park based on the multi-source data to obtain a dynamic calibration factor includes: The historical factor data corresponding to the carbon emission factor and the multi-source data are input into a multi-modal carbon emission model to obtain a predicted factor corresponding to the carbon emission factor, wherein the multi-modal carbon emission model is used to predict the trend of the carbon emission factor in a preset event period. The base factor of at least one carbon emission factor is corrected based on the multi-source data, to obtain a correction factor corresponding to the carbon emission factor.

[0036] In practical applications, the carbon emission factor is a key parameter for measuring the carbon emissions generated by different activities or processes, such as the carbon emissions corresponding to unit energy consumption (e.g., per kilowatt-hour of electricity, per cubic meter of natural gas) or unit product production (e.g., per ton of steel, per cubic meter of concrete). In practical applications, dynamic calibration of carbon emission factors can obtain high-precision and traceable carbon emission calculation benchmarks for the current and near future.

[0037] In the above embodiments, a multi-dimensional base factor library can be constructed. The national / industry base factors (such as the Provincial Greenhouse Gas Inventory), regional-specific factors (North China Power Grid winter factors), and historical factors of the park (measured in the past three years) are integrated, covering 15 types of factors such as electricity, gas, and process. The above factors can be regarded as base factors of carbon emission factors. Then, a multi-modal carbon emission large model can be used to predict the prediction factors at future times, which can adopt a hybrid architecture of Transformer+GNN. When used, the time series data (historical factors, power structure), text data (policy news), image data (satellite cloud map), and graph data (supply chain relationship) are input, and the features are fused through an attention mechanism. Adversarial training (including historical data sets of extreme weather, equipment failure, and policy mutation) and transfer learning (chemical industry / commercial / data center scenario sub-library) can also be introduced. Moreover, the final factor prediction error is ≤2%, and the causal traceability accuracy is ≥90% (such as identifying the cause of the factor fluctuation of the thermal power unit failure). In practical applications, other models can also be used for prediction, and the specific limitations are not made.

[0038] In practical applications, when the prediction factor fluctuation amplitude is greater than a preset threshold, the data acquisition frequency of the multi-source data can be increased, so that more accurate information of the carbon emission factor can be obtained.

[0039] In the above embodiments, the carbon emission factors can be corrected in real time and the uncertainty quantified. For example, the grid factor is calculated by weighting the wind power proportion multiplied by the wind power factor, the photovoltaic proportion multiplied by the photovoltaic factor, and the thermal power proportion multiplied by the thermal power factor. By monitoring the power generation of wind power, photovoltaic power and thermal power in real time, the proportion of their total power generation can be calculated, and then combined with the pre-determined baseline factors of the carbon emission factors of wind power, photovoltaic power and thermal power, the correction factor of the current grid carbon emission can be calculated in real time. This dynamic adjustment can more accurately reflect the carbon emission in the actual power generation process, because the energy structure of the grid may vary at different times and in different regions. The gas factor is corrected according to the methane content (e.g. 0.578 tCO2 / m³ for 98% methane). By detecting the methane content of the gas in real time, the gas factor is corrected in real time according to the pre-established corresponding relationship between the methane content and the gas factor. In this way, the accuracy of the gas carbon emission accounting can be ensured, and errors caused by changes in gas composition can be avoided. The process factor can be corrected in combination with the tail gas concentration and the production account. The tail gas concentration reflects the actual situation of carbon emission in the production process, and the production account records various parameters and operation information in the production process, which helps to more comprehensively understand the influence of the production process on carbon emission.

[0040] In practical applications, a confidence interval (e.g. 95%) can also be provided for each factor to quantify the uncertainty. The confidence interval refers to the range in which the true value of the parameter may fall under a certain confidence level. Through statistical analysis and modeling methods, the variability of various influencing factors can be considered, such as the fluctuation of energy generation proportion, the measurement error of gas methane content, the change of tail gas concentration in the process, etc., to calculate the 95% confidence interval of each carbon emission factor. The 95% confidence interval provides a range, indicating that the probability of the true carbon emission factor falling within this range is relatively high. This helps decision-makers understand the degree of uncertainty of the carbon emission accounting results, and makes more scientific and reasonable decisions in formulating emission reduction strategies, evaluating emission reduction effects, and conducting carbon emission trading.

[0041] In practical applications, the calibration frequency can be 15 minutes per time for electricity, 1 hour per time for gas, 4 hours per time for process, and 1 day per time for waste.

[0042] In one possible implementation, based on the twin of the park and the dynamic calibration factor, carbon flow simulation is performed on the carbon emission of the park to obtain simulation results of at least one emission reduction scheme, including: A preset industrial park template is used to construct the twin of the park; Based on the dynamic calibration factor, carbon flow simulation is performed on the carbon emission of the park to obtain the flow of carbon elements in the park; A baseline scenario is simulated to obtain simulation results; In the case that the simulation result does not meet the expectation, at least one emission reduction scheme is generated; An independent What-If simulation is started for the at least one emission reduction scheme, and a simulation result is obtained.

[0043] In actual application, lightweight modeling can be performed, an industrial / commercial / complex park template is preset, and a simplified twin body can be automatically generated by a user uploading a plan view + equipment list (modeling cycle: 1-2 weeks). For a large park, a slicing technology can be used to split into energy / production / carbon sink subsystems, an edge end (Huawei Atlas 500) calculates a local slice (delay ≤ 30 ms), and a cloud end synchronizes global data. Then, high-fidelity carbon flow simulation can be performed, based on BIM / GIS and physical mechanism models, real-time simulation of carbon elements flowing in an energy-process-building-waste system (in actual application, the accuracy of simulation ≥ 97%). At this time, a baseline model simulation can be performed, that is, an evaluation baseline is established, and it is determined how the future carbon emission will be if everything remains unchanged. In actual application, a time range for deduction can be determined, for example, carbon emission in the future one year is deducted. Based on current various data and historical rules, based on physical mechanism models (such as thermodynamic equations, chemical reaction equations), the conversion efficiency of energy in the equipment and the flow path of carbon elements are simulated, and a simulation result, that is, a baseline carbon emission trajectory line, is obtained. In actual application, the simulation result can be explicitly displayed, and the predicted value of carbon emission of the park in a future period of time is predicted if it develops in this way. If the predicted value does not meet the expectation, at least one emission reduction scheme can be generated at this time, and the emission reduction scheme supports “what-if” analysis (such as the influence of adding photovoltaic on carbon emission). In this way, a simulation result corresponding to the emission reduction scheme can be obtained. In actual application, for each simulation result, carbon emission influence, such as total emission reduction amount and peak time, can be calculated. Economic influence, such as initial investment cost and operation and maintenance cost, can also be calculated. In this way, the simulation result can clearly show the emission reduction potential, economic cost and implementation complexity of the emission reduction scheme, which facilitates laying a foundation for subsequent zero-carbon path planning. Moreover, it is also helpful to build a carbon risk “early warning-response” mechanism, to avoid three types of core risks of policy, asset and supply chain, and to reduce potential loss by 50%.

[0044] It should be noted that, in actual application, carbon flow simulation and emission reduction scheme simulation can also be directly performed, and specific limitations are not made.

[0045] In one possible implementation, the generating, according to the simulation result, of the zero-carbon path planning corresponding to the park comprises: Fusing a scenario analysis model, a baseline target, an optimization target and a zero-carbon target corresponding to the park are generated; According to the simulation results of at least one emission reduction scheme, a short-term path, a medium-term path, and a long-term path are generated based on the benchmark target, the optimization target, and the zero-carbon target.

[0046] In the above embodiments, the scenario analysis model can be used to generate the benchmark target, the optimization target, and the zero-carbon target corresponding to the park; in actual applications, the scenario analysis model can be fused to generate three types of targets (such as reaching the peak in 2028 and carbon neutralization in 2050) of the benchmark target, the optimization target, and the zero-carbon target; for example, the scenario analysis model can be used to input key driving factors (such as carbon price, green electricity cost, green hydrogen cost, policy strength, etc.) to generate the three types of targets. The benchmark target is to continue the current emission reduction measures without significant new policies or investments. The energy efficiency is slowly improved, and the energy structure is slightly optimized. The target result can be that the carbon emissions will peak around 2028, with a peak of A million tons of CO2 equivalent. By 2050, carbon emissions will slowly decline to about 10 million tons, and carbon neutralization cannot be achieved. The optimization target considers the maturity and cost reduction curve of the technology, and selects the technology combination with the highest cost performance. The core principle is to maximize the return on investment. The target result is that through energy efficiency improvement, energy substitution (natural gas transition), circular economy, etc., carbon emissions can peak in 2026, and by 2040, carbon emissions will be about 70% lower than the peak. This is a pragmatic and economic target. The zero-carbon target (1.5°C path) is to take the most aggressive emission reduction measures to respond to climate change, without considering short-term costs, and deploying all feasible technologies. The target result can be that it must reach the peak in 2028 and achieve carbon neutralization in 2050. This is an ultimate strategic goal, and all path planning will be guided by it. In the above embodiments, the role of the scenario analysis model is to simulate different future scenarios corresponding to the three types of targets by adjusting key variables (such as energy prices, technology breakthrough speed, policy strength, and consumer behavior changes), and quantifying their carbon emission trajectories, energy structures, and economic costs. In actual applications, the above targets can also be determined by other means, and the specific implementation is not limited.

[0047] After obtaining the three types of targets, short-term, medium-term and long-term paths can be generated according to the simulation results of the above-mentioned emission reduction scheme, for example, can be decomposed into short (1-3 years), medium (3-5 years), long (5-10 years) period path, and determine the emission reduction measures, investment cost, income estimation of each stage. In this way, the long-term carbon neutral vision is divided into executable phased tasks. In the above process, the simulation results of the emission reduction scheme need to be considered, and various emission reduction schemes are combined optimally in time and space on the basis of considering the technology maturity and investment cost decline curve. For example, short-term path may prefer to implement roof photovoltaic, lighting LED transformation and other "low cost high return" projects. The medium-term path may deploy large-scale energy storage and upgrade process equipment. The long-term path can introduce frontier technologies such as green hydrogen, biomass energy, carbon capture and storage. In this way, a zero-carbon park full-cycle planning tool can be provided to achieve 5-10 year goal decomposition and cost-benefit estimation, which can reduce long-term transformation cost by 30%.

[0048] In one possible implementation, under the framework of the zero-carbon path planning, generating the carbon emission strategy corresponding to the park based on the reinforcement learning algorithm and executing the carbon emission strategy, includes: Under the framework of the zero-carbon path planning, generating the carbon emission strategy corresponding to the park based on the reinforcement learning algorithm; Monte Carlo simulation is performed on the carbon emission strategy to obtain the success probability and risk level of the carbon emission strategy; In the case where the success probability and risk level meet the preset conditions, the carbon emission strategy is issued to the building automation system or micro-grid management system of the park.

[0049] In actual application, a reinforcement learning algorithm can be constructed for multi-agent game reinforcement learning, for example, with the goal of "park emission reduction + enterprise cost reduction + supplier profit", the reward function is set as R=0.4x emission reduction (unit: tons of CO2) +0.3x cost saving (unit: ten thousand yuan) -0.3x production disturbance, where production disturbance=(actual device downtime / monthly planned downtime threshold) x 100%, monthly planned downtime threshold≤1 hour, when the actual downtime exceeds the threshold, the disturbance value increases in proportion (such as 2 hours of downtime corresponding to a disturbance value of 200%); and then training the agent to generate an optimal or near-optimal carbon emission strategy (such as energy storage charging and discharging plan, air conditioning group control parameters).

[0050] In the above embodiment, after determining the zero-carbon path planning, the multi-source data of the park and the prediction factors of the carbon emission factors can be input into the above reinforcement learning algorithm, at this time, a carbon emission strategy can be obtained which makes the value of the reward function satisfy the condition, then Monte Carlo simulation can be used to simulate 1000+ scenarios (photovoltaic output ±20% fluctuation, carbon price ±15% change), output the success probability of the strategy (such as 5% emission reduction in 90% scenarios) and the risk level (low / medium / high), in the case that the success probability and the risk level satisfy the preset conditions, for example, the success probability is higher than the preset probability threshold and the risk level is lower than the preset level, it can be determined that the carbon emission strategy can be executed and it is possible to successfully implement, at this time, the carbon emission strategy can be issued to facilitate execution, for example, the carbon emission strategy can be converted into a control instruction that can be read by equipment and the time point is clear; finally, the carbon emission strategy can be issued to the building automation system and the micro-grid management system through the OPC UA / MQTT standard interface, so that the park executes according to the carbon emission strategy.

[0051] In practical applications, after obtaining a clear zero-carbon path planning, generating daily or short-term carbon emission strategies is a process of decomposing strategies into tactics. Taking the example of a path planning that has clear stage goals (such as "reduce emissions by 8% in A year") and key measures (such as "complete 5MW photovoltaic construction and grid connection in A year"), the process is as follows. First, the medium and long-term goals in the zero-carbon path can be decomposed into operational short-term goals, such as monthly goals, weekly goals, and daily goals. Then, the prediction factors of carbon emission factors can be obtained, such as dynamic carbon emission factor predictions for the next 24 hours (such as grid factor curves), weather forecasts (which affect photovoltaic output), and the latest production plans and equipment maintenance plans of the park. Then, under the constraints of the zero-carbon path planning, a multi-objective optimization calculation can be performed using reinforcement learning algorithms. Reinforcement learning agents try countless device scheduling schemes in a virtual environment. Any generated strategy must first conform to the stage direction of the zero-carbon path. For example, the strategy will never suggest enabling a backward diesel generator to save money. Through the reward function, the agent learns how to balance emission reduction, cost savings, and production disturbances. For example, it can find a carbon emission strategy: "Tomorrow at noon, even if the grid price is not the highest, but due to the high predicted factor, priority should be given to using energy storage discharge to achieve the emission reduction target." The generated preferred carbon emission strategy is then stress tested. Using Monte Carlo simulation, uncertainties are injected (such as: photovoltaic output is 10% lower than predicted, load increases by 5%). In this way, the success probability of the strategy can be obtained. For example, in 95% of scenarios, the strategy can achieve the emission reduction target for tomorrow. Risk warnings can also be made, such as the main risk coming from load uncertainty, and emergency plans are recommended (such as slightly adjusting the air conditioning settings in non-critical areas). In the case of successful implementation of the carbon emission strategy to achieve the emission reduction target, the carbon emission strategy can be converted into specific and executable instructions and issued. After the strategy enters the execution phase, it can be continuously monitored. For example, actual carbon emission data and energy consumption data can be collected in real time. Compare "actual value" with "predicted value". If there is a deviation (such as sudden clear weather, large increase in photovoltaic output), a rapid optimization cycle can be started immediately to adjust the subsequent strategy (such as reducing energy storage discharge and prioritizing more green power). If there is a persistent and significant deviation (such as continuous multi-day emission reduction not meeting the target), it may trigger a re-evaluation and adjustment of the zero-carbon path planning, such as whether to start the next emission reduction project ahead of schedule.

[0052] In practical applications, under the framework of zero-carbon path planning, the steps to generate the corresponding carbon emission strategy of the park based on reinforcement learning algorithms can also be directly generated, without considering the uncertainty of the scheme for the time being, and then input into the twin body for simulation to obtain simulation results, based on which it can be determined whether the carbon emission strategy can be directly executed and what needs to be improved.

[0053] In practical applications, the carbon emission strategy is generated on the premise of existing zero-carbon paths, which has the greatest value in ensuring that short-term tactical decisions always align with long-term strategic goals. It changes the carbon management of the park from passive response to active planning, and every daily operation is for the ultimate carbon neutral vision. In one possible implementation, after the carbon emission strategy corresponding to the park is generated and executed, the method further includes: Monitoring the carbon key performance indicators of the park in real time; In response to a specific risk of carbon emissions of the park, a risk warning corresponding to the risk is performed; Generating a report according to a specific format.

[0054] In practical applications, carbon performance management and risk warning can be performed after the carbon emission strategy is executed. In the above embodiment, a multi-level KPI dashboard can be provided to display the total carbon, carbon intensity, and emission reduction progress of the park, building, and equipment in real time. In the above embodiment, when a specific risk occurs, a corresponding risk warning can be triggered, for example, a policy compliance warning (calculate compliance gap and push response suggestions), a carbon asset impairment warning (trigger realization reminder when carbon price falls by ≥20%), and a supply chain carbon risk warning (monitor supplier carbon emissions and push a replacement list). In practical applications, automatic report generation can also be performed, for example, generating accounting reports and ESG reports according to ISO14064, GRI, and CBAM standards, and supporting one-key export. In this way, the carbon emission performance of the park can be comprehensively controlled to ensure that the carbon emissions of the park meet relevant standards and requirements.

[0055] In practical applications, carbon asset financialization services can be provided, and dynamic carbon asset valuation can be performed, that is, the quota surplus / deficit value is calculated in real time by combining carbon price prediction and park emission data. In this way, enterprises can real-time master the carbon quota surplus or deficit value. For example, a certain steel enterprise uses the platform to realize monthly carbon emission fluctuation automatic analysis, and early warning of quota deficit to avoid penalties due to over-emission (such as facing 3-5 times market price penalties for non-compliance). Combined with carbon price prediction, enterprises can develop cross-market arbitrage strategies. Financial interface docking can also be performed, and RSA-2048 encrypted API can be connected to energy exchanges (quota trading), banks (green credit), and SGS (certification) to support one-key completion of carbon asset mortgage, compliance declaration, and green project financing. Through RSA-2048 encrypted API, enterprises can complete carbon asset mortgage application one-key. Moreover, by connecting energy exchanges and third-party certification agencies (such as SGS), enterprises can quickly complete project certification and financing applications. For example, new energy project developers can generate carbon footprint reports that meet international standards through the platform to attract green fund investment.

[0056] In one possible implementation, after the generation of the carbon emission strategy corresponding to the park and the execution of the carbon emission strategy, the method further comprises: Building a regional carbon resource pool to realize point-to-point trading between surplus parks and gap parks; Building a carbon technology knowledge base, connecting technology suppliers and calculating investment returns according to the skill needs of the park.

[0057] In practical applications, cross-domain carbon collaborative services can be provided for parks, cross-park carbon trading, construction of a regional carbon resource pool, and realization of point-to-point trading between surplus parks and gap parks with a transaction fee of ≤2% (deducted from the transaction amount) and blockchain evidence (non-tamperable); Example: Park A has a surplus of 2000 tons of quota, which is sold to Park B (which needs 1800 tons) at 95 yuan / ton (5 yuan lower than the market price). Park A's actual income = 1800 tons x 95 yuan / ton x (1-2%) = 168540 yuan.

[0058] In practical applications, cross-industry technology sharing can also be carried out, a carbon technology knowledge base (integrating liquid cooling, photovoltaic curtain wall, carbon capture, etc.) can be built, intelligent matching of park needs (such as recommending photovoltaic curtain walls for commercial parks) can be carried out, and technology suppliers can be connected and investment returns can be calculated. In this way, cross-park carbon resource collaboration and cross-industry technology sharing can be realized, and regional emission reduction efficiency can be improved by at least 25%.

[0059] In practical applications, through carbon trading and cross-industry technology matching, more efficient emission reduction can be achieved while meeting compliance requirements, and the overall level of carbon emission management in the park can be improved.

[0060] In practical applications, in order to ensure data security, a cloud-edge collaborative architecture can be used, with the edge end responsible for real-time collection, local calibration, and emergency calculation; the cloud end uses a MySQL+InfluxDB hybrid database (structured data+time series data) that supports 5000 data transactions per second, meets the concurrency of 10 10km² parks, and synchronizes data through 5G / optical fiber. At the same time, security and permission management are carried out, with AES-256 data encryption, three levels of permission (administrator / accountant / viewer), 1 year of operation log retention, support for carbon trading regulatory departments and third-party institutions to access permissions. In this way, the operation threshold for non-professional users and the third-party verification period can be reduced, the operation independence rate can be improved to 90%, and the verification period can be shortened to 1 week.

[0061] In practical applications, the above-mentioned manner can bring various beneficial effects. In terms of technical benefits, the factor calibration accuracy is ≤2% (improved by more than 40% compared to the traditional fixed factor method), the accounting error is ≤3%, and the carbon flow simulation accuracy is ≥97%; the cross-park adaptation period is shortened from 3 months to 1 week, and the cost of old park transformation is reduced by 70% (from 1 million yuan / park to 30-50 million yuan / park). In terms of economic benefits, the emission reduction cost is reduced by 20%-30% (such as a park that reduces 10,000 tons of carbon emissions per year can save 2-3 million yuan); the risk of carbon asset impairment is reduced by 50%, the carbon cost of the supply chain is saved by 15%, and the cross-park carbon trading cost is reduced by 40% (the service fee is reduced from the industry average of 5% to 2%). In terms of social benefits, the application of 100 typical parks can reduce carbon emissions by 500-1,000 tons per year (equivalent to the annual carbon sequestration of 2.5-5 million mature trees).

[0062] In practical applications, it is especially suitable for various parks (industrial parks, commercial complexes, data center parks, and old park transformation parks) that require accurate carbon accounting, real-time risk management, long-term zero-carbon planning, and cross-domain collaboration. It is also suitable for overseas double-carbon core regions such as the European Union (CBAM policy) and Southeast Asia (low-cost demand).

[0063] To make the above technical solutions clearer, the following will explain the implementation process in detail based on a "10 km² comprehensive industrial park (including chemical zone, old manufacturing zone, supporting commercial zone, and distributed photovoltaic power station)" and the European Union CBAM compliance scenario. The hardware used includes: carbon energy integrated edge intelligent gate (supporting LoRa + solar), smart electricity meter (DTZ341, 0.5 level), portable infrared meter (accuracy ±1%), flue gas analyzer (Testo 350), edge server (Huawei Atlas 500), cloud server (Ali Cloud 8-core 16G x 15 nodes), carbon performance management platform (Web / APP end).

[0064] Example 1: Multimodal data acquisition and old park transformation 1. Hardware deployment: deploy carbon energy integrated gate + smart electricity meter DTZ341 in 10 power distribution rooms, deploy smart gas meter G2.5 in 3 gas stations, and deploy Testo 350 flue gas analyzer in 5 chemical workshops; configure 10 portable infrared meters and 20 LoRa wireless sensors (with a battery life of 1 year) in the old manufacturing area; interface with the North China Power Grid enterprise data platform (real-time power supply structure), local gas companies (methane content data), park ERP systems (production work orders), and the European Union CBAM declaration system.

[0065] 2. Data collection: From 8:00 to 18:00 on June 1, 2025, the carbon energy integration gateway collects power data every 15 minutes (e.g., at 10:00, a certain chemical plant consumes 1200 kWh of electricity), and collects gas data every 30 minutes (at 10:30, the gas consumption is 50 m³); the portable infrared meter reads the old workshop meter data (at 11:00, the reading is 850 kWh); the OCR technology identifies the gas composition report (methane content 98%), and the NLP technology analyzes the EU CBAM new policy ("steel product carbon tax will increase by 10% in 2026").

[0066] 3. Preprocessing: Remove the 12:00 smart meter false alarm of 10000 kWh abnormal data by self-encoder; use spatio-temporal graph convolution network to supplement the missing data of gas meter at 14:00 (based on the interpolation of 48 m³ at 13:30 and 52 m³ at 14:30, the value is 50 m³); after Z-score standardization, the data integrity reaches 99.6%, and is pushed to the carbon data intelligent hub layer.

[0067] Example 2: Multimodal large model calibration and extreme scenario test 1. Initialization of basic factor library: Import the power benchmark factor 0.6101 tCO2 / MWh and the gas benchmark factor 0.561 tCO2 / m³ in the Provincial Greenhouse Gas Inventory Compilation Guide, and the average factor of the park chemical process in the past three years is 0.85 tCO2 / ton product; import the EU CBAM steel industry special factor 0.72 tCO2 / ton.

[0068] 2. Real-time correction: On June 1, 10:00, North China Power Grid pushes the real-time power structure: wind power 30% (emission factor 0.03 tCO2 / MWh), photovoltaic 20% (emission factor 0.01 tCO2 / MWh), thermal power 50% (emission factor 0.82 tCO2 / MWh), calculate the real-time grid factor = 0.3 x 0.03 + 0.2 x 0.01 + 0.5 x 0.82 = 0.421 tCO2 / MWh; combined with the methane content of gas 98%, correct the gas factor = 0.561 x (98 / 95) = 0.578 tCO2 / m³; the Testo 350 measured tail gas CO2 concentration of the chemical plant is 2500 ppm (the benchmark concentration is 2000 ppm), combined with the production account (output on June 1 is 5 tons), correct the process factor = 0.85 x (2500 / 2000) = 1.06 tCO2 / ton product.

[0069] 3. Extreme scenario test: simulate "a cold wave causes a 30% increase in gas consumption in the park", multi-modal carbon emission model input for nearly 5 years of cold weather energy consumption data, June 2 weather forecast (temperature -5°C), predict that on June 2, the gas factor is 0.585 tCO2 / m³ (error 2.8%≤3%); simulate "a 10% increase in EU CBAM carbon tax", model traceability analysis "will lead to a 7.2% increase in carbon costs for steel products in the park (0.72 tCO2 / ton x 10% x carbon price 100 yuan / ton = 7.2 yuan / ton)", push response suggestions "add 5MW photovoltaic in 2 months, which can offset 60% of the cost increase".

[0070] Example 3: Digital twin zero-carbon planning and cross-park collaboration 1. Zero-carbon path planning: input the park's current carbon emission base of 80,000 tons / year into the carbon digital twin, generate a zero-carbon goal: carbon peak in 2028 (peak 80,000 tons / year), achieve carbon neutrality in 2050; decomposed into three-stage path: (1) Short term (2025-2027): add 10MW photovoltaic (investment 50 million, annual emission reduction 10,000 tons, 5 years to recover cost), old motor retrofit (investment 30 million, annual emission reduction 8,000 tons, 6 years to recover cost); (2) Medium term (2028-2030): process green hydrogen replacement (investment 100 million, annual emission reduction 30,000 tons, 8 years to recover cost); (3) Long term (2031-2040): carbon capture system (investment 150 million, annual emission reduction 32,000 tons, 12 years to recover cost).

[0071] 2. Cross-park collaboration: on June 10, the system monitors the park's quota surplus of 2000 tons, through regional carbon resource pool to dock quota gap park B (need 1800 tons), to 95 yuan / ton (lower than Shanghai environmental energy exchange daily quoted price 100 yuan / ton) to reach a transaction, transaction fee 2%, the actual income of the park = 1800x95x(1-2%)=168540 yuan, transaction data through blockchain notarization (hash value: 0x7aF3...2Dc9); match "chemical park carbon capture system transformation technology" from carbon technology knowledge base, interface with supplier to estimate investment return: 150 million investment, annual emission reduction 32,000 tons, carbon price 100 yuan / ton, annual income 3200,000, 12 years to recover cost.

[0072] Example 4: Risk early warning and strategy closed-loop execution 1. Risk warning: On June 15, the system monitored the national carbon market price falling from 100 yuan / ton to 78 yuan / ton (a drop of 22%), triggering a carbon asset realization reminder: "The current quota value has decreased by 22% compared to last week, suggesting selling 1500 tons of quota to avoid a further decrease of 33,000 yuan (1500 tons x 22 yuan / ton)"; Through the blockchain carbon ledger, the upstream supplier C (steel raw materials) carbon intensity reached 1.2 tCO2 / ton (industry average 0.88 tCO2 / ton, 35% above standard), automatically pushing 3 alternative suppliers (carbon intensity ≤0.9 tCO2 / ton, price difference ≤3%).

[0073] 2. Strategy execution: Simulate two schemes in the carbon digital twin, "central air conditioning temperature from 24℃ to 26℃ + energy storage system peak-valley charging and discharging optimization", after Monte Carlo simulation of 1000+ scenarios, the success probability of the combination strategy is 92% (low risk, emission reduction 4%, cost saving 80,000 yuan / month); Through the OPC UA interface, the strategy parameters (air conditioner set temperature 26℃, energy storage charging period 23:00-7:00, discharging period 10:00-16:00) are sent to the building automation system and microgrid management system of the park; After 1 week of execution, the actual measurement data shows that the power consumption is reduced by 8%, the carbon emissions are reduced by 4%, and the deviation from the simulation result is ≤2%.

[0074] Based on the same inventive concept, the embodiments of the present application also provide a carbon data intelligent hub system, the overall three-layer architecture diagram of the system is as shown in Figure 2 Figure 2 ​In the middle, the hierarchical relationship of "data collection and edge processing layer-carbon data intelligent hub layer-carbon application and service layer" is displayed from top to bottom, and the core components and data flow direction of each layer are marked: data collection and edge processing layer: including carbon energy integrated edge intelligent gateway (protocol analysis, edge computing, AI chip), portable collection equipment (infrared meter reading instrument, LoRa sensor), multi-system interface (power grid, gas company, ERP, CBAM system), intelligent preprocessing unit (abnormality detection, missing value repair, standardization); carbon data intelligent hub layer: including carbon emission factor dynamic calibration engine (multimodal carbon emission large model, real-time correction unit, uncertainty quantification module), carbon digital twin engine (slicing modeling unit, carbon flow simulation unit, zero-carbon path planning unit), intelligent decision engine (multi-agent game reinforcement learning unit, Monte Carlo scenario simulation unit, strategy issuing unit); carbon application and service layer: including carbon performance management platform (multi-level KPI dashboard, carbon risk early warning module, automatic report generation unit), carbon asset financialization service (dynamic valuation unit, financial system interface), cross-domain carbon collaboration service (carbon resource pool transaction unit, carbon technology knowledge base unit); bottom support: cloud-edge collaborative architecture (Huawei Atlas 500 edge, Ali cloud cluster cloud), security and permission management module (AES-256 encryption, three-level permission control, 1 year operation log). The intelligent gateway deployment diagram is shown in Figure 3 .

[0075] The above is the method embodiment and system proposed by the present application. Based on the same inventive concept, the park carbon emission management device is also provided by the embodiments of the present application, and the structure is shown in Figure 4 .

[0076] Figure 4 The internal structure diagram of the park carbon emission management device provided by the embodiments of the present application is shown in Figure 4 . at least one processor 401; and the memory 402 in communication connection with the at least one processor; Wherein, the memory 402 stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor 401 to enable the at least one processor 401 to execute the above-mentioned park carbon emission management method.

[0077] The non-volatile computer storage medium corresponding to Figure 1 provided by some embodiments of the present application stores computer executable instructions, and the computer executable instructions are set to execute the above-mentioned park carbon emission management method.

[0078] The various embodiments in the present application are described in a progressive manner, and the same or similar parts among the various embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, the IoT device and medium embodiments are described simply because they are basically similar to the method embodiments.

[0079] The system and medium provided by the embodiments of the present application are one-to-one corresponding to the method, and therefore, the system and medium also have similar beneficial technical effects to the method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the system and medium will not be described here.

[0080] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. In addition, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0081] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device implemented in the flowcharts and / or block diagrams. Figure 1 The device that implements the function specified in one flow or multiple flows and / or blocks. Figure 1 The device that implements the function specified in one flow or multiple flows and / or blocks.

[0082] These computer program instructions can also be stored in a computer-readable memory that can guide the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction devices that implement the flowcharts and / or block diagrams. Figure 1 The device that implements the function specified in one flow or multiple flows and / or blocks. Figure 1 The device that implements the function specified in one flow or multiple flows and / or blocks.

[0083] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 ​

[0084] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0085] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the computer stores information about an operating system, application software, and / or the like. Memory is an example of computer readable media.

[0086] Computer readable media includes permanent and non-permanent, moveable and non- moveable media that can be implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules 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 technology, compact disc read-only memory (CD-ROM), digital versatile disks (DVDs) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that is accessible to a computing device. According to the definition provided herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0087] It is also important to note that the terms "comprises", "comprising", or other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0088] ​​The above merely provides an example of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the scope of claims of the present application.

Claims

1. A method for managing carbon emissions in a park, characterized in that, include: Obtain multi-source data related to carbon emissions in the park; Based on the multi-source data, the carbon emission factor corresponding to the park is calibrated to obtain a dynamic calibration factor; Based on the twin of the park and the dynamic calibration factor, carbon flow simulation of the park's carbon emissions is performed to obtain simulation results for at least one emission reduction scheme. Based on the simulation results, a zero-carbon pathway plan corresponding to the park is generated; Within the framework of the zero-carbon pathway planning, a carbon emission strategy corresponding to the park is generated and executed based on a reinforcement learning algorithm.

2. The method according to claim 1, characterized in that, Obtain multi-source data related to the park's carbon emissions, including: Through the carbon energy integrated edge smart gateway, numerical data related to the carbon emissions of the park are obtained; Multimodal data related to the carbon emission data of the park are acquired through mobile devices and interfaces; The numerical data and multimodal data are preprocessed to obtain multi-source data of the park.

3. The method according to claim 1, characterized in that, Based on the multi-source data, the carbon emission factor corresponding to the park is calibrated to obtain a dynamic calibration factor, including: The historical factor data corresponding to the carbon emission factor and the multi-source data are input into the multimodal carbon emission model to obtain the prediction factor corresponding to the carbon emission factor. The multimodal carbon emission model is used to predict the change trend of the carbon emission factor in a preset event period. Based on the multi-source data, the baseline factor of at least one carbon emission factor is corrected to obtain the correction factor corresponding to the carbon emission factor.

4. The method according to claim 1, characterized in that, Based on the twin of the park and the dynamic calibration factor, carbon flow simulation of the park's carbon emissions is performed to obtain simulation results for at least one emission reduction scheme, including: Using a pre-set industrial park template, construct a twin of the park; Based on the dynamic calibration factor, carbon flow simulation is performed on the carbon emissions of the park to obtain the flow of carbon elements in the park. Simulate the baseline scenario and obtain simulation results; If the simulation results do not meet expectations, at least one emission reduction scheme should be generated; Initiate independent What-If simulations for at least one of the emission reduction schemes and obtain simulation results.

5. The method according to claim 1, characterized in that, The step of generating a zero-carbon pathway plan for the park based on the simulation results includes: By integrating scenario analysis models, baseline targets, optimization targets, and zero-carbon targets corresponding to the park are generated. Based on the simulation results of at least one emission reduction scheme, short-term, medium-term, and long-term paths are generated based on the baseline target, the optimization target, and the zero-carbon target.

6. The method according to claim 1, characterized in that, Within the framework of the zero-carbon pathway planning, the carbon emission strategy corresponding to the park is generated and executed based on a reinforcement learning algorithm, including: Within the framework of the zero-carbon pathway planning, a carbon emission strategy corresponding to the park is generated based on a reinforcement learning algorithm. Monte Carlo simulations were performed on the carbon emission strategy to obtain the success probability and risk level of the carbon emission strategy. If the success probability and risk level meet the preset conditions, the carbon emission strategy will be distributed to the building automation system or microgrid management system of the park.

7. The method according to claim 1, characterized in that, After generating and executing the carbon emission strategy corresponding to the park, the method further includes: Real-time monitoring of the park's key carbon performance indicators; In response to specific risks to carbon emissions in the park, a risk warning corresponding to the risk will be issued; Generate reports according to a specific format.

8. The method according to claim 1, characterized in that, After generating and executing the carbon emission strategy corresponding to the park, the method further includes: Establish a regional carbon resource pool to enable point-to-point trading between surplus and deficit industrial parks; Establish a carbon technology knowledge base, connect with technology suppliers and calculate return on investment based on the skill requirements of the park.

9. A carbon emission management device for industrial parks, characterized in that, The device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform a park carbon emission management method as described in any one of claims 1-8.

10. A computer storage medium storing computer-executable instructions, characterized in that, When the computer-executable instructions are executed, a carbon emission management method for a park as described in any one of claims 1-8 is implemented.

Images