A park real-time energy-carbon intelligent management system based on dynamic carbon emission factor driving
The park's real-time energy and carbon intelligent management system, driven by dynamic carbon emission factors, solves the problems of low energy utilization efficiency and difficulty in reducing carbon emissions in traditional park energy management. It achieves precise energy scheduling and carbon emission management, and optimizes the park's energy structure and carbon emission reduction effect.
Patent Information
- Application Number
- CN202510054674.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-14
- Publication Date
- 2026-11-06
Smart Images
Figure CN123050858A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of carbon emission reduction technology, and more specifically to a real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors. Background Technology
[0002] Currently, with the escalation of global climate change, reducing carbon emissions has become a major issue that countries urgently need to address. Industrial parks, as concentrated areas of energy consumption and carbon emissions, play a crucial role in achieving carbon reduction targets. Traditional energy and carbon emission management methods in industrial parks often have many shortcomings and are unable to meet the current demands for precise, real-time, and efficient energy and carbon management.
[0003] The main challenges of integrated energy and carbon reduction systems in industrial parks lie in system optimization and economic efficiency. On one hand, industrial parks exhibit diverse energy consumption patterns, involving fossil fuel combustion and various types of electricity consumption, each corresponding to different carbon emission levels. On the other hand, energy management lacks detailed analysis of the electricity consumption characteristics of different enterprises within the park (e.g., electricity consumption volume, time of day, and industry), and lacks real-time data-driven optimization scheduling strategies. Traditional energy supply structures are relatively singular, lacking systematic planning for the introduction and integration of clean energy. This hinders the effective utilization of clean energy sources such as photovoltaic and wind power to replace high-carbon energy sources, thus preventing the optimization and transformation of the energy structure and ultimately improving the overall energy efficiency and reducing carbon emissions of the park.
[0004] Therefore, how to improve the overall energy efficiency of the park and reduce carbon emissions is a problem that needs to be solved by those skilled in the art. Summary of the Invention
[0005] In view of this, the present invention provides a real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors, in order to solve the problems existing in the background technology.
[0006] To achieve the above objectives, the present invention adopts the following technical solution:
[0007] A real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors includes: a carbon information acquisition system, a carbon information processing system, a carbon data transmission system, a real-time carbon information feedback system, a deep reinforcement learning system, and an intelligent management system.
[0008] The carbon information acquisition system is used to collect energy consumption and carbon emission data within the park;
[0009] The carbon information processing system is used to preprocess the collected energy consumption and carbon emission data;
[0010] The carbon data transmission system is used to transmit preprocessed energy consumption and carbon emission data to other parts of the system.
[0011] The real-time carbon information feedback system is used to detect the real-time changes in dynamic carbon emission factors and carbon emissions, and obtain real-time feedback information.
[0012] The deep reinforcement learning system establishes a deep reinforcement learning model for carbon emissions in the park based on the preprocessed energy consumption and carbon emission data, determines carbon emission management strategies, and automatically adjusts the strategies based on real-time feedback information.
[0013] The intelligent management system integrates multiple types of information resources and optimizes management strategies, strategy execution, energy scheduling, and carbon emission management based on various factors.
[0014] Optionally, the carbon information acquisition system acquires energy consumption and carbon emission data generated by different carbon emission technologies within the park; on the other hand, it acquires relevant carbon emission information and energy consumption based on the park's historical carbon emission information. By integrating these two data sources, the carbon information acquisition system can comprehensively collect current and historical carbon emission data of the park, thereby providing real-time support for the park's carbon emission management and analysis.
[0015] Optionally, it also includes obtaining carbon emission factors using any one of the following methods: online monitoring, emission factor method, and dynamic accounting model, and being able to automatically update carbon emission factors for different carbon emission sources within the park.
[0016] Optionally, this also includes using a variety of technological means to obtain carbon emissions, including:
[0017] Regression analysis: Based on collected environmental and historical data, a statistical model is constructed to predict future emission trends;
[0018] Time series analysis: Conduct time series analysis on carbon emissions and energy consumption within the park to identify long-term trends and seasonal fluctuations;
[0019] Machine learning algorithms: Train energy and carbon models to accurately predict carbon emissions from various emission sources through classification and regression methods;
[0020] Real-time feedback: Calculate the park's real-time carbon emission factor based on carbon emissions and energy consumption, and dynamically transmit the analysis results to the intelligent management system to optimize energy allocation strategies.
[0021] Optional, fossil fuel combustion CO2 emissions: refers to CO2 emissions from the intentional oxidation process of fossil fuels for energy utilization purposes, and CO2 emissions implied by net purchased electricity: refers to CO2 emissions generated in the electricity production process corresponding to the net purchased electricity consumed by enterprises. The carbon emissions are calculated using the following formula:
[0022]
[0023] C 化石燃料 =C 化石燃料 +C 天然气 +C 其他
[0024] Where E represents total carbon dioxide emissions; C represents fossil fuel consumption; and j represents the j-th type of fossil fuel; e i Z represents the carbon dioxide emission factor corresponding to the i-th type of energy consumption; Z represents the energy consumed by electricity in the park.
[0025] Optionally, the formula for calculating the real-time carbon emission factor is:
[0026]
[0027] Where e i,t cp represents the carbon dioxide emission factor corresponding to the i-th energy consumption at time t; k,t The value represents the carbon potential of node k at time t; E represents carbon dioxide emissions; ESC represents the energy structure adjustment coefficient; EOC represents the equipment operating efficiency coefficient; A represents the node area; and TAC is the time adjustment coefficient.
[0028] Optionally, the deep reinforcement learning system is a deep reinforcement learning method based on dynamic carbon emission factors and real-time carbon measurement systems; specifically, it includes establishing a deep reinforcement learning model and a reinforcement learning algorithm.
[0029] In this process, establishing a deep reinforcement learning model requires defining the state space as the current state of the environment, including energy consumption level, carbon emission factor, and environmental variables, while the action space is defined as the actions to be performed; the reward function sets the carbon emission reduction target, and the reward is proportional to the reduction in carbon emissions.
[0030] Regarding reinforcement learning algorithms, the algorithm is selected, and optimization is performed using deep Q-networks or proximal policies. Experience data of past states, actions, rewards, and the next state is stored through experience replay for random sampling updates.
[0031] Optionally, the intelligent management system includes a model building unit and an optimization and adjustment unit:
[0032] The model building unit is used to build a pre-simulation model of future carbon emissions;
[0033] The optimization and adjustment unit is used to classify and manage various carbon emission sources in the park through the carbon emission management system, and to provide feedback and optimization control on carbon emissions based on real-time monitoring data.
[0034] Optionally, the method for constructing the model is as follows:
[0035] By converting dynamic carbon emission factors and real-time carbon emission data into a time series format, a time series matrix is created, and then a matrix autoregressive model is used to predict carbon emissions for future years.
[0036] Optionally, the optimization and adjustment unit performs optimization and adjustment by providing feedback within the industrial park through the following steps:
[0037] Prioritization strategies are developed based on real-time monitored energy data and dynamic carbon emission factors.
[0038] By introducing clean energy power generation facilities to replace traditional high-carbon energy, the energy structure can be optimized and transformed, and the proportion of clean energy in microgrid systems can be increased.
[0039] Adjustments will be made based on policy factors.
[0040] As can be seen from the above technical solution, compared with the prior art, this invention discloses a real-time intelligent management system for park energy and carbon based on dynamic carbon emission factors, including: a carbon information acquisition system, a carbon information processing system, a carbon data transmission system, a real-time carbon information feedback system, a deep reinforcement learning system, and an intelligent management system; the carbon information acquisition system collects energy consumption and carbon emission data within the park; the carbon information processing system preprocesses the collected data; the carbon data transmission system ensures timely data transmission; the real-time carbon information feedback system provides real-time carbon information feedback on the energy consumption of various departments within the park based on real-time carbon emission factors; the deep reinforcement learning system automatically adjusts management strategies and optimizes energy allocation based on the park's carbon emissions and energy usage; and the intelligent management system makes dynamic adjustments through comprehensive analysis. This invention accurately calculates carbon emissions based on real-time carbon emission factors, optimizes the energy structure, and integrates multiple energy sources to replace high-carbon energy sources to improve efficiency and reduce carbon emissions. Attached Figure Description
[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0042] Figure 1 This is a schematic diagram of the system structure provided by the present invention. Detailed Implementation
[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0044] This invention discloses a real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors, such as... Figure 1 As shown, it includes: carbon information acquisition system, carbon information processing system, carbon data transmission system, real-time carbon information feedback system, deep reinforcement learning system, and intelligent management system;
[0045] A carbon information acquisition system is used to collect energy consumption and carbon emission data within the park;
[0046] A carbon information processing system is used to preprocess the collected energy consumption and carbon emission data;
[0047] A carbon data transmission system is used to transmit pre-processed energy consumption and carbon emission data to other parts of the system.
[0048] A real-time carbon information feedback system is used to detect the real-time changes in dynamic carbon emission factors and carbon emissions, and obtain real-time feedback information.
[0049] The deep reinforcement learning system establishes a deep reinforcement learning model for carbon emissions in the park based on the preprocessed energy consumption and carbon emission data, determines carbon emission management strategies, and automatically adjusts the strategies based on real-time feedback information.
[0050] The intelligent management system integrates multiple information resources and optimizes management strategies, strategy execution, energy scheduling, and carbon emission management based on various factors.
[0051] The specific implementation method of this embodiment is as follows:
[0052] In this embodiment, the carbon information acquisition system within the industrial park acquires carbon information. This system includes an integrated data acquisition module, which obtains various data related to carbon emissions in the park through two methods. First, the module installs flow meters and fuel consumption monitoring equipment within the park to record different energy consumption levels (e.g., coal consumption, natural gas usage) in real time and transmits this data to the carbon information acquisition system. Second, it accesses the park's existing carbon emission record database to collect carbon emissions and related energy consumption data over a past period (e.g., the past five years), and cleans and organizes this data to ensure its accuracy and consistency. By integrating these two data sources, the integrated data acquisition module can comprehensively collect current and historical carbon emission data from the park, thereby providing real-time support for carbon emission management and analysis.
[0053] This embodiment obtains carbon emission factors through various technical means, including online monitoring, emission factor method and dynamic accounting model-based method, and can automatically update carbon emission factors for different carbon emission sources in the park.
[0054] By installing online sensors within the park, the gas emission concentration and emission volume of various emission sources are monitored in real time, and carbon emissions are estimated using internationally recognized emission factors based on the energy consumption types of various industrial equipment within the park.
[0055] In this embodiment, the system uses regression analysis, time series analysis, and machine learning algorithms to obtain the carbon emissions within the industrial park.
[0056] The system regularly performs regression analysis, using linear regression, nonlinear regression and other methods to establish mathematical models of energy consumption and carbon emissions based on the park's energy consumption data, and calculates the carbon emission factor for each type of energy.
[0057] The system periodically analyzes the carbon emission data of the park, and uses other time series analysis methods such as the ARIMA (Autoregressive Integral Moving Average) model to identify the time patterns of carbon emissions, and then adjusts the estimation of carbon emission factors.
[0058] Meanwhile, by training on a large amount of historical carbon emission data, a machine learning model is built to automatically identify carbon emission patterns, calculate and update carbon emission factors for various energy sources, and regularly optimize the model according to the characteristics of different processes and production activities.
[0059] Based on the energy consumption and carbon emission factors obtained from the carbon information acquisition system, the carbon emissions are displayed to the park management personnel in the form of charts, curves and other methods through a visual dashboard, and the carbon emissions in the industrial park are calculated through feedback data.
[0060] The carbon emission accounting statistics for the industrial park mainly include the following parameters (the energy consumption of the industrial park mainly refers to the energy consumption of the power system):
[0061] CO2 emissions from fossil fuel combustion: refers to CO2 emissions generated by the intentional oxidation process of fossil fuels for energy utilization purposes, and CO2 emissions implied by net purchased electricity: refers to CO2 emissions generated during the electricity production process corresponding to the net purchased electricity consumed by enterprises. Carbon emissions are calculated using the following formula:
[0062]
[0063] C 化石燃料 =C 化石燃料 +C 天然气 +C 其他
[0064] Where E represents total carbon dioxide emissions; C represents fossil fuel consumption; and j represents the j-th type of fossil fuel; e i Z represents the carbon dioxide emission factor corresponding to the i-th type of energy consumption; Z represents the energy consumed by electricity in the park.
[0065] The formula for calculating the real-time carbon emission factor is:
[0066]
[0067] Where e i,t cp represents the carbon dioxide emission factor corresponding to the i-th energy consumption at time t; k,t The value represents the carbon potential of node k at time t; E represents carbon dioxide emissions; ESC represents the energy structure adjustment coefficient; EOC represents the equipment operating efficiency coefficient; A represents the node area; and TAC is the time adjustment coefficient.
[0068] The coefficients in the formula were obtained through data collection within the industrial park, as well as through the compilation of historical data. Carbon emission coefficients and other coefficients may vary at different times and points in time. The TAC coefficient also differs between day and night. Therefore, when processing data to calculate the implementation carbon emission factor, it is best to use reliable coefficient data tailored to local conditions.
[0069] In this embodiment, the carbon emissions and dynamic carbon emission factors calculated by the carbon information processing system are first processed through the carbon data transmission system. The collected data is then classified, filtered, and transmitted, ensuring that the data can be transmitted to all parts of the system in real time, thus ensuring the accuracy and timeliness of the data.
[0070] The data is transmitted to a real-time carbon information feedback system, which detects the real-time energy consumption of various parts of the industrial park based on real-time carbon emission factors; and the real-time carbon emissions are obtained based on the real-time consumption and real-time carbon emission factors.
[0071] The calculation formula is as follows:
[0072] E i,t =∑(Q i ×e i,t )
[0073] In the formula, E i,t Q represents the carbon emissions corresponding to the i-th energy consumption at time t. i e represents the consumption of the i-th type of energy. i,t Let represent the carbon dioxide emission factor corresponding to the i-th type of energy consumption at time t.
[0074] Based on the real-time carbon emissions obtained above, the foundation of the deep reinforcement learning system is transmitted in real time and accurately to facilitate subsequent deep reinforcement.
[0075] The deep reinforcement learning system is a deep reinforcement learning method based on dynamic carbon emission factors and real-time carbon measurement systems; specifically, it includes the establishment of deep reinforcement learning models and reinforcement learning algorithms.
[0076] In this process, establishing a deep reinforcement learning model requires defining the state space as the current state of the environment, including energy consumption level, carbon emission factor, and environmental variables, while the action space is defined as the actions to be performed; the reward function sets the carbon emission reduction target, and the reward is proportional to the reduction in carbon emissions.
[0077] Regarding reinforcement learning algorithms, the algorithm is selected, and optimization is performed using deep Q-networks or proximal policies. Experience data of past states, actions, rewards, and the next state is stored through experience replay for random sampling updates.
[0078] Constructing a state-space network first involves identifying several key carbon emission-related factors, including energy consumption data such as the usage and consumption time of various energy sources; production activity indicators such as production line operating rates, product types and quantities; equipment operating status, including the operating time of high-energy-consuming equipment; and environmental monitoring data. Based on the monitoring data, an appropriate state representation format is selected, with some data discretized or presented using a combination of continuous and vector representations. Finally, carbon information processing is carried out, including normalization to unify the data scale and encoding conversion to transform non-numerical factors into numerical forms. Simultaneously, state-space boundaries can be defined as needed. After repeated verification and adjustments, the constructed state-space is ensured to accurately reflect the actual carbon emission situation in the industrial park, effectively promoting the achievement of carbon emission reduction goals.
[0079] Action space networks are used to generate a series of selectable action strategies, such as adjusting energy allocation, optimizing equipment operation modes, and changing production plans. Through continuous trial and error and learning, the optimal action plan that effectively reduces carbon emissions and achieves carbon reduction targets is found. Action space networks include energy regulation actions, equipment operation adjustment actions, and production plan optimization actions. The steps for obtaining the optimal carbon reduction plan based on the action space network of an industrial park are as follows: Initialize the environment and model; then, in a cyclical interaction, the action space network generates action plans based on the current state; after the agent executes the action plan, it receives a new state and reward. Then, a reinforcement learning algorithm is used to update the policy and adjust the action selection probability. Finally, continuous evaluation is performed, and when convergence conditions such as no significant increase in reward value are met, a relatively optimal carbon reduction plan is obtained.
[0080] The reward function is defined in this example through the following steps: First, clarify the target orientation, focusing on carbon emission reduction while also considering production efficiency and cost control; then quantify key indicators, obtaining carbon emission reduction data through carbon emission monitoring equipment, measuring production efficiency by statistically analyzing product output and production cycle, and reflecting cost control by considering the degree of energy and equipment maintenance cost savings; second, set indicator weights according to the park's priorities, increasing the weight of indicators that prioritize carbon emission reduction and allocating them reasonably for balanced development; finally, set the reward range and thresholds, clarifying the range of values, and providing additional rewards for achieving specific carbon emission reduction standards.
[0081] The deep reinforcement learning model is trained based on Deep Q-Network (DNQ) and Proximal Policy Optimization (PPO). The training process includes:
[0082] Step 1: Experience Accumulation and Random Sampling Update. Initiate the data collection process, accumulating experience data through operation and monitoring in the actual industrial park environment. After each action is performed, record the current state (including energy consumption, equipment status, production status, etc.), the action taken, the reward received (calculated based on a preset reward function, such as rewards for carbon emission reduction, production efficiency improvements, etc.), and the next state generated after the action. Store this "state-action-reward-next state" experience data to form an ExperienceReplay pool. Then, randomly sample a certain number of data points from this ExperienceReplay pool for subsequent model updates, ensuring the model can learn from various past experiences and avoid getting trapped in local optima.
[0083] Step 2: Model Training and Parameter Update Process. The model is trained using sample data randomly sampled from the experience replay pool. Based on the state and action information in the samples, the predicted Q-value is calculated using a neural network. This predicted Q-value is then combined with the actual reward and the maximum Q-value estimate for the next state. The parameters of the Q-value network are updated according to a specific update formula, making the predicted Q-value closer to the actual expected reward.
[0084] Step 3: Dynamic Carbon Emission Factor Fusion and Adjustment. Dynamic carbon emission factors within the park are continuously acquired using a real-time carbon information feedback system. These factors reflect the actual carbon emission situation in the park in real time. Based on the acquired dynamic carbon emission factors, the model's state input is adjusted and optimized. This is specifically achieved through the following formula:
[0085] R t = -α×Emission t +β×Performance t
[0086] Among them, R t The reward for the current time step, Emission t For current carbon emissions, Performance t For production efficiency, α and β are weighting coefficients.
[0087] In this way, the model can be closely integrated with the current carbon emission level to optimize decision-making and better adapt to the dynamic changes in the park's carbon emission situation.
[0088] Step 4: Real-time Monitoring and Dynamic Strategy Adjustment. Online sensors continuously monitor the operating status and energy consumption data of various devices within the park, ensuring timely access to the latest information on energy utilization and equipment operation. Simultaneously, the intelligent management system dynamically adjusts operational strategies based on action selection suggestions output by the reinforcement learning model and the actual state of the park. For example, it may adjust the operating power of a device, change the energy allocation scheme, or optimize production planning based on model suggestions, thereby achieving carbon emission reduction goals and making the park's energy utilization more efficient and rational, and production activities more environmentally friendly.
[0089] Step 5: Model Evaluation and Iterative Optimization Process. Regularly conduct a comprehensive evaluation of the reinforcement learning model's performance. Analyze historical data (including states, actions, and rewards at each past time step) to measure the model's actual effectiveness in areas such as carbon reduction and production efficiency improvement. Based on the evaluation results, if shortcomings are found, such as carbon reduction not meeting expectations or production efficiency declining in certain situations, make targeted adjustments and optimizations to the model. This may involve redefining the state space and action space, modifying and improving the reward function. The final model can be represented using a Markov Decision Process (MDP).
[0090] MDP=(S,A,P,R,γ)
[0091] in:
[0092] S: State space
[0093] A: Action Space
[0094] P: State transition probability
[0095] R: Reward function
[0096] γ: Discount factor
[0097] Furthermore, the process of steps 1 to 5 is repeated again. Through continuous iteration and optimization, a complete and efficient carbon management system for the park is gradually formed, ultimately achieving the ideal emission reduction target and realizing the sustainable development of the park.
[0098] The aforementioned intelligent management system includes model building units and optimization and adjustment units. Based on feature learning, the intelligent agent makes decisions according to conditions such as carbon emissions and energy usage. The agent then adjusts its strategy by receiving environmental feedback to achieve the optimal carbon reduction target.
[0099] Model building unit, used to build a simulation model of future carbon emissions.
[0100] The optimization and adjustment unit is used to classify and manage various carbon emission sources in the park through the carbon emission management system, and to provide feedback and optimize control of carbon emissions based on real-time monitoring data.
[0101] The process of building a predictive model of future carbon emissions includes six parts: data preparation, conversion to time series format, creation of time series matrix, matrix autoregression model, parameter estimation, and prediction.
[0102] The data preparation process includes collecting:
[0103] Dynamic carbon factors (such as carbon emissions per unit of product); real-time carbon metering data (such as carbon emissions per industrial unit); timestamps (such as annual or monthly data).
[0104] The data is then used to transform it into a time series format.
[0105] Furthermore, it is converted into a time series format in the following manner.
[0106] Create a DataFrame where each row represents a point in time, and columns can include:
[0107] Time (year or month), carbon factor, carbon emissions, and other relevant indicators that may affect carbon emissions (such as energy consumption, production, etc.).
[0108] Furthermore, a time series matrix can be constructed from a structured data frame using the following formula:
[0109] y t =[C1,C2,…,C t ]
[0110] Wherein, C t This represents the values of various indicators at time t.
[0111] The matrix autoregressive model is constructed using the following formula:
[0112] Y t =A1Y t-1 +A2Y t-2 +…+A p Y t-p +∈ t
[0113] In the formula, p is used to determine the appropriate lag number, which can be selected using information criteria; ∈ t The error term represents the portion that cannot be predicted using historical data; parameter estimation requires historical data to estimate the model parameters A1, A2, ..., A. p Methods such as least squares or maximum likelihood estimation can be used.
[0114] Predictions for future years based on estimated model parameters:
[0115] Y t+h =A1Y t+h-1 +A2Y t+h-2 +…+A p Y t+h-p
[0116] In the formula, h is the number of future time steps (e.g., 1 year, 2 years, etc.).
[0117] The optimization and adjustment of the unit specifically includes the following three methods:
[0118] The first approach is to develop a priority adjustment strategy based on real-time monitored energy data and dynamic carbon emission factors.
[0119] A comprehensive analysis of energy sources (fossil fuels, renewable energy, etc. and their proportions), consumption patterns (total amounts, peaks and troughs, etc. for different enterprises, regions, and time periods), and dynamic carbon emission factors (the changing patterns under different conditions).
[0120] Measure the electricity consumption of each enterprise at different times, clarify the patterns of electricity consumption periods (peak, valley, and flat), and determine the industry attributes of electricity consumption based on the industry (such as the electricity consumption characteristics of manufacturing and service industries).
[0121] Dynamic strategies are developed based on real-time carbon information and analysis results, and adjusted in a timely manner according to changes in energy and carbon emissions in the park, incentivizing enterprises to optimize energy management, improve utilization rates, and reduce carbon emissions.
[0122] The second approach is to introduce clean energy power generation facilities to replace traditional high-carbon energy sources.
[0123] Analyze the current energy demand in the park to gain a comprehensive understanding of the energy needs of various enterprises and production processes at different times, including energy types, demand quantities, demand time distribution, and quality requirements, so as to provide a basis for subsequent selection.
[0124] This includes analyzing and assessing the current demand situation and conducting a comprehensive evaluation of the energy microgrid system, including the existing energy structure, load demand, energy storage capacity, and grid connection conditions.
[0125] Consider the applicability of clean energy technologies. For example, for photovoltaic power generation, attention should be paid to the local solar resources (annual sunshine hours, intensity changes, etc.) and land resources (area, flatness, orientation, obstructions, etc.) in the park. For wind power generation, the local wind speed conditions (average wind speed, fluctuations, etc.) and site conditions (openness, obstacles, etc.) should be understood to determine its applicability in the park.
[0126] By comparing the differences in power generation costs, efficiency, equipment investment costs, and operation and maintenance costs of various new energy sources within the park, we can understand the overall benefits.
[0127] Based on the actual conditions of the park, select appropriate types and scales of clean energy, either single or multiple, to achieve diversified energy supply, promote the optimization and transformation of the park's energy structure, and reduce carbon emissions.
[0128] The third approach is to measure the management system in real time based on policy factors and carbon emission factors to ensure the operation of the carbon management system, improve energy efficiency, and reduce carbon emissions.
[0129] Based on local peak and off-peak electricity pricing, and according to the local policy's division of peak, off-peak, and normal periods, prices are appropriately increased during peak hours to guide enterprises to reduce electricity consumption; prices are reduced during off-peak hours to encourage enterprises to increase electricity consumption, thereby achieving rational allocation of power resources.
[0130] In accordance with the national carbon emissions trading policy, this mechanism has been introduced into the industrial park, using carbon emission factors as a guiding signal. Enterprises with high carbon emissions are required to purchase carbon emission rights, while those with low emissions can sell them, thereby encouraging enterprises to reduce carbon emissions, improve efficiency, and increase energy utilization.
[0131] By managing the system through real-time measurement of policy factors and carbon emission factors, we can keep abreast of policy implementation and carbon emission dynamics, so as to adjust strategies in a timely manner, ensure system operation, reduce carbon emissions, and improve energy efficiency.
[0132] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0133] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors, characterized in that, include: Carbon information acquisition system, carbon information processing system, carbon data transmission system, real-time carbon information feedback system, deep reinforcement learning system, and intelligent management system; The carbon information acquisition system is used to collect energy consumption and carbon emission data within the park; The carbon information processing system is used to preprocess the collected energy consumption and carbon emission data; The carbon data transmission system is used to transmit preprocessed energy consumption and carbon emission data to other parts of the system. The real-time carbon information feedback system is used to detect the real-time changes in dynamic carbon emission factors and carbon emissions, and obtain real-time feedback information. The deep reinforcement learning system establishes a deep reinforcement learning model for carbon emissions in the park based on the preprocessed energy consumption and carbon emission data, and determines carbon emission management strategies. And automatically adjust the strategy based on real-time feedback information; The intelligent management system integrates multiple types of information resources and optimizes management strategies, strategy execution, energy scheduling, and carbon emission management based on various factors.
2. The park real-time energy and carbon intelligent management system driven by dynamic carbon emission factors according to claim 1, characterized in that, The carbon information acquisition system acquires energy consumption and carbon emission data generated by different carbon emission technologies within the park. On the other hand, it acquires relevant carbon emission information and energy consumption based on the park's historical carbon emission information. By integrating these two data sources, the carbon information acquisition system can comprehensively collect current and historical carbon emission data of the park, thereby providing real-time support for the park's carbon emission management and analysis.
3. The park real-time energy and carbon intelligent management system driven by dynamic carbon emission factors according to claim 1, characterized in that, It also includes obtaining carbon emission factors using any one of the following methods: online monitoring, emission factor method, and dynamic accounting model, and is able to automatically update carbon emission factors for different carbon emission sources within the park.
4. The park real-time energy and carbon intelligent management system based on dynamic carbon emission factors as described in claim 1, characterized in that, This also includes using a variety of technological means to obtain carbon emissions, including: Regression analysis: Based on collected environmental and historical data, a statistical model is constructed to predict future emission trends; Time series analysis: Conduct time series analysis on carbon emissions and energy consumption within the park to identify long-term trends and seasonal fluctuations; Machine learning algorithms: Train energy and carbon models to accurately predict carbon emissions from various emission sources through classification and regression methods; Real-time feedback: Calculate the park's real-time carbon emission factor based on carbon emissions and energy consumption, and dynamically transmit the analysis results to the intelligent management system to optimize energy allocation strategies.
5. The park real-time energy and carbon intelligent management system driven by dynamic carbon emission factors according to claim 1, characterized in that, CO2 emissions from fossil fuel combustion: refers to CO2 emissions generated by the intentional oxidation process of fossil fuels for energy utilization purposes, and CO2 emissions implied by net purchased electricity: refers to CO2 emissions generated during the electricity production process corresponding to the net purchased electricity consumed by enterprises. Carbon emissions are calculated using the following formula: C 化石燃料 =C 化石燃料 +C 天然气 +C 其他 Where E represents total carbon dioxide emissions; C represents fossil fuel consumption; and j represents the j-th type of fossil fuel; e i Z represents the carbon dioxide emission factor corresponding to the i-th type of energy consumption; Z represents the energy consumed by electricity in the park.
6. The park real-time energy and carbon intelligent management system driven by dynamic carbon emission factors according to claim 1, characterized in that, The formula for calculating the real-time carbon emission factor is: Where e i,t Let represent the carbon dioxide emission factor corresponding to the i-th type of energy consumption at time t; cp k,t The value represents the carbon potential of node k at time t; E represents carbon dioxide emissions; ESC represents the energy structure adjustment coefficient; EOC represents the equipment operating efficiency coefficient; A represents the node area; and TAC is the time adjustment coefficient.
7. The park real-time energy and carbon intelligent management system driven by dynamic carbon emission factors according to claim 1, characterized in that, The deep reinforcement learning system is a deep reinforcement learning method based on dynamic carbon emission factors and real-time carbon measurement systems; specifically, it includes establishing a deep reinforcement learning model and a reinforcement learning algorithm. In this process, establishing a deep reinforcement learning model requires defining the state space as the current state of the environment, including energy consumption level, carbon emission factor, and environmental variables, while the action space is defined as the actions to be performed; the reward function sets the carbon emission reduction target, and the reward is proportional to the reduction in carbon emissions. Regarding reinforcement learning algorithms, the algorithm is selected, and optimization is performed using deep Q-networks or proximal policies. Experience data of past states, actions, rewards, and the next state is stored through experience replay for random sampling updates.
8. The park real-time energy and carbon intelligent management system based on dynamic carbon emission factors as described in claim 1, characterized in that, The intelligent management system includes a model building unit and an optimization and adjustment unit: The model building unit is used to build a pre-simulation model of future carbon emissions; The optimization and adjustment unit is used to classify and manage various carbon emission sources in the park through the carbon emission management system, and to provide feedback and optimization control on carbon emissions based on real-time monitoring data.
9. A real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors as described in claim 8, characterized in that, The method for constructing the model is as follows: By converting dynamic carbon emission factors and real-time carbon emission data into a time series format, a time series matrix is created, and then a matrix autoregressive model is used to predict carbon emissions for future years.
10. A real-time energy and carbon intelligent management system for industrial parks driven by dynamic carbon emission factors according to claim 8, characterized in that, The optimization and adjustment of the unit involves the following steps to provide feedback within the industrial park: Prioritization strategies are developed based on real-time monitored energy data and dynamic carbon emission factors. By introducing clean energy power generation facilities to replace traditional high-carbon energy, the energy structure can be optimized and transformed, and the proportion of clean energy in microgrid systems can be increased. Adjustments will be made based on policy factors.