Full-link electricity consumption monitoring method, system and equipment based on intelligent internet of things

Abstract

The invention provides a full-link electricity consumption monitoring method, system and equipment based on intelligent Internet of Things, and relates to the technical field of power system management. The method comprises the following steps: acquiring operation data acquired by an intelligent sensor group deployed at a power network node in real time, and executing localized carbon flow accounting according to the operation data through an edge computing node; constructing a joint probability model based on Monte Carlo simulation and a deep belief network, and generating a node carbon flow density matrix under multiple scenes; dynamically updating an electricity-carbon conversion coefficient according to real-time energy structure data, and calculating node-level carbon emission: optimizing a power grid topology and an energy storage scheduling strategy by adopting a double-delay depth deterministic strategy gradient algorithm TD3 and taking a carbon flow density matrix as a constraint condition; and generating an intelligent report including a carbon footprint thermodynamic diagram, emission reduction potential evaluation and block chain evidence storage. According to the invention, the capability of multi-target collaborative optimization of the security and economy of the power grid can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system management, and in particular to a full-link electricity energy consumption monitoring method, system and device based on intelligent Internet of Things. BACKGROUND

[0002] With the acceleration of the global energy structure towards low carbonization, the power system as a key field of carbon emissions urgently needs to establish a precise and dynamic carbon footprint management mechanism. The traditional carbon footprint tracking method (such as the comparative document CN118863923A) is mostly based on deterministic power flow calculation, and the node carbon emission is calculated by using a static electric carbon conversion coefficient. Such method has defects: first, the load fluctuation and uncertainty of new energy grid connection are not considered, resulting in inaccurate carbon emission evaluation; second, the fixed electric carbon coefficient is used, which cannot reflect the dynamic changes of the mixed proportion of wind, light and storage in the regional power grid, limiting the accuracy of carbon flow mapping; third, the optimization strategy mainly focuses on a single emission reduction target, ignoring key constraints such as economic operation and voltage safety of the power grid, making it difficult to achieve multi-dimensional collaborative optimization.

[0003] In view of the above problems, the industry tries to introduce probabilistic power flow analysis and edge computing technology, but the existing scheme of the traditional power system carbon footprint tracking method relies on deterministic power flow calculation and static electric carbon conversion model, which cannot adapt to the multi-energy dynamic fluctuation and uncertain load characteristics of the power grid in the new energy high penetration scenario, resulting in insufficient carbon emission evaluation accuracy, and lacking the ability of multi-objective collaborative optimization of power grid safety and economy. SUMMARY

[0004] The embodiment of the present application provides a full-link electricity energy consumption monitoring method, system and device based on intelligent Internet of Things, to solve the problem of how to realize the multi-objective collaborative optimization ability of power grid safety and economy.

[0005] In the first aspect, the embodiment of the present application provides a full-link electricity energy consumption monitoring method based on intelligent Internet of Things, comprising:

[0006] Obtaining the operation data collected by the intelligent sensor group deployed in the power network node in real time, and performing localized carbon flow accounting according to the operation data through the edge computing node; wherein the operation data includes voltage, current, power and device energy consumption data;

[0007] Based on Monte Carlo simulation and deep belief network, a joint probability model is constructed to generate a node carbon flow density matrix under multiple scenarios;

[0008] According to the real-time energy structure data, the electric carbon conversion coefficient is dynamically updated, and the node-level carbon emission is calculated:

[0009] A double-delay deep deterministic policy gradient algorithm (TD3, Twin Delayed Deep Deterministic Policy Gradient) is used to optimize the grid topology and energy storage scheduling strategy with the carbon flow density matrix as the constraint condition.

[0010] An intelligent report containing a carbon footprint heat map, emission reduction potential assessment, and blockchain storage is generated.

[0011] In one possible implementation, the joint probability model is constructed based on Monte Carlo simulation and deep belief network to generate node carbon flow density matrices in multiple scenarios, including:

[0012] Based on meteorological factors and historical load data, the load probability distribution in the next 24 hours is predicted.

[0013] An improved Nataf transformation is used to process non-Gaussian correlated variables to generate multiple scenario power flow samples.

[0014] According to the multiple scenario power flow samples, the reverse power flow method is used to identify the carbon emission source propagation path and calculate the node carbon flow contribution degree.

[0015] According to the point carbon flow contribution degree, the node carbon flow density matrix in multiple scenarios is generated.

[0016] In one possible implementation, the dynamic updating of the electricity-carbon conversion coefficient based on real-time energy structure data includes:

[0017] Real-time grid-connected energy type and proportion data are obtained, and a dynamic correction factor is calculated.

[0018] The electricity-carbon conversion parameters of the edge node and the cloud are updated synchronously using a federated learning framework.

[0019] In one possible implementation, it further includes:

[0020] Abnormal data is subjected to credibility weighting, and low-confidence data is filtered out for calculation.

[0021] In one possible implementation, the node-level carbon emission calculation formula is:

[0022] node=∑i=1 n (P i ·α i (t)+Q i ·β i (t))·γ(t)]

[0023] Where γ(t) is calculated by real-time grid-connected energy data, satisfying:

[0024]

[0025] wherein P i is the active power of node i, Q i is the reactive power of node i, α i (t) is the time-varying active power-carbon conversion coefficient of node i, β i (t) is the time-varying reactive power-carbon conversion coefficient of node i, γ(t) is a dynamic correction factor; EF k (t) is the real-time carbon emission factor of the kth type of energy; w k is the real-time proportion weight of the kth type of energy in the power grid.

[0026] In one possible implementation, the reward function of the TD3 algorithm is defined as:

[0027] R = -λ1·∑C node -λ2·ΔLoss + λ3·V margin

[0028] wherein λ1, λ2, λ3 are dynamic weight coefficients, C node is the real-time carbon emission of the node, ΔLoss is the change in network loss, V margin is the voltage safety margin.

[0029] In one possible implementation, before the optimization of the power grid topology and the energy storage scheduling strategy, the method further includes:

[0030] constructing a carbon-electricity coupling sensitivity matrix to identify nodes with high emission reduction potential;

[0031] initializing a population using a chaotic particle swarm algorithm to accelerate convergence to a feasible solution domain;

[0032] embedding a carbon trading cost factor in the objective function to define economic constraints.

[0033] In one possible implementation, the local carbon flow accounting performed by the edge computing node includes:

[0034] calculating the local carbon emission intensity based on equipment nameplate parameters and real-time operating condition data;

[0035] updating the electricity-carbon conversion coefficient using a lightweight federated learning model.

[0036] In one possible implementation, after the electricity-carbon conversion coefficient is updated using the lightweight federated learning model, the method further includes uploading the sensitive data to the cloud after differential privacy encryption.

[0037] In a second aspect, an embodiment of the present application provides a full-link electricity consumption monitoring system based on intelligent Internet of Things, which includes:

[0038] The Internet of Things perception layer, the probabilistic power flow analysis module, the carbon flow dynamic modeling module, the adaptive optimization engine and the carbon footprint analysis platform;

[0039] The Internet of Things perception layer is configured to acquire operation data collected by intelligent sensor groups deployed at power network nodes in real time, and perform localized carbon flow accounting based on the operation data through edge computing nodes.

[0040] The probabilistic power flow analysis module is configured to construct a joint probability model based on Monte Carlo simulation and deep belief networks, and generate a node carbon flow density matrix under multiple scenarios.

[0041] The carbon flow dynamic modeling module is configured to dynamically update the electric-carbon conversion coefficient based on real-time energy structure data, and calculate the node-level carbon emissions.

[0042] The adaptive optimization engine is configured to use TD3 to optimize the power grid topology and energy storage scheduling strategy with the carbon flow density matrix as a constraint condition.

[0043] The carbon footprint analysis platform is configured to generate an intelligent report including a carbon footprint heat map, emission reduction potential evaluation and blockchain storage.

[0044] In a third aspect, an electronic device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the method of the first aspect or any possible implementation manner of the first aspect when executing the computer program.

[0045] In the embodiments of the present application, the comprehensive optimization model constructed by combining the intelligent Internet of Things technology, big data analysis and artificial intelligence technology can effectively realize the accurate quantification and collaborative optimization of carbon footprint in the power system. By combining probabilistic power flow analysis calculation and carbon footprint optimization, the scheme can globally model and dynamically adjust carbon emissions in a multi-dimensional scenario, fully considering uncertainty factors to improve optimization accuracy. Based on the "electricity-energy-carbon" adaptive conversion algorithm, online monitoring technology can collect, share and process carbon footprint data of different nodes in real time, and realize data credibility and traceability by combining blockchain technology. With the aid of artificial intelligence technology, the scheme can self-learn and optimize iteration of carbon footprint indicators, dynamically adjust the power grid topology structure and energy storage configuration, thereby maximizing the carbon reduction benefit. At the same time, the intelligent carbon footprint report provides a transparent carbon emission evaluation tool for managers. Overall, the embodiments of the present application enhance the multi-objective collaborative optimization capability of power grid safety and economy, improve the low-carbon level and resource utilization efficiency of power grid operation, and realize the transition from local optimization to global collaboration. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1The implementation flowchart of the whole-link electricity energy consumption monitoring method based on smart Internet of Things is provided in the embodiments of the present application.

[0047] Figure 2 The structural schematic diagram of the whole-link electricity energy consumption monitoring device based on smart Internet of Things is provided in the embodiments of the present application.

[0048] Figure 3 The schematic diagram of the electronic device is provided in the embodiments of the present application. DETAILED DESCRIPTION

[0049] The embodiments of the present application will be described in detail below with reference to the drawings.

[0050] Figure 1 The implementation flowchart of the whole-link electricity energy consumption monitoring method based on smart Internet of Things is provided in the embodiments of the present application. As shown in the figure, Figure 1 the method comprises the following steps:

[0051] S101, acquiring operation data collected in real time by a smart sensor group deployed at a power network node, and performing localized carbon flow accounting according to the operation data through an edge computing node; wherein the operation data comprises voltage, current, power and device energy consumption data.

[0052] The smart sensor group deployed at the power network node is used to collect operation data in real time. These sensors include but are not limited to voltage transformers, current transformers, power sensors and energy consumption monitoring devices, which can accurately capture key parameters such as voltage, current, power and device energy consumption in the power grid. The acquired raw data is transmitted to the edge computing node through wired or wireless communication mode to realize real-time data processing and analysis.

[0053] The current system relies on centralized computing architecture, and uploading a large amount of data to the cloud may cause transmission delay and privacy leakage problems, which cannot meet the requirements of real-time and security in distributed energy scenarios. In the present embodiment, a lightweight carbon flow accounting model is built in the edge node, and the localized carbon flow accounting is performed according to the operation data through the edge computing node, which combines the characteristics of the power network to efficiently process the voltage and current waveform data and calculate the carbon emissions in real time. Compared with the traditional cloud computing mode, the edge computing node is closer to the data source, significantly reducing the delay of data transmission and improving the real-time performance of data processing. At the same time, since the data is processed locally at the edge node, it is not necessary to upload to the cloud, effectively avoiding the risk of privacy leakage and enhancing the security of the system.

[0054] S102, constructing a joint probability model based on Monte Carlo simulation and deep belief network to generate a node carbon flow density matrix under multiple scenarios.

[0055] In order to comprehensively analyze the carbon emission characteristics of each node in the power network under different operating states, the invention adopts a method combining Monte Carlo simulation and deep belief network to construct a joint probability model. The specific steps are as follows:

[0056] Monte Carlo simulation: generate a variety of possible scenario parameter combinations through random sampling techniques, covering uncertainty factors such as power load, energy structure changes, and equipment operating states;

[0057] Deep belief network: use a multi-layer neural network to learn the non-linear relationships between variables in the system, capturing complex probability distribution characteristics;

[0058] Joint probability model: combine the probability density functions generated by Monte Carlo simulation and DBN to construct a probability model that can reflect the complexity of the system;

[0059] Node carbon flow density matrix: based on the above model, simulate the power flow under multiple scenarios, calculate the carbon emission intensity distribution of each node, and generate the corresponding carbon flow density matrix, providing input basis for subsequent optimization.

[0060] S103, dynamically update the electric-carbon conversion coefficient according to the real-time energy structure data, and calculate the node-level carbon emission.

[0061] In the case of changing energy structure, a single electric-carbon conversion coefficient cannot accurately reflect the carbon emission characteristics of the system. Therefore, the invention proposes a dynamic updating mechanism.

[0062] Among them, the real-time energy structure data includes the proportion information of each power source type (such as coal-fired, gas-fired, wind and solar power, etc.). According to the real-time power generation composition, the corresponding electric-carbon conversion coefficient is calculated, and the updated electric-carbon conversion coefficient is applied to the calculation of node-level carbon emission, combined with the previous carbon flow density matrix, to generate a more accurate carbon emission distribution map.

[0063] In the specific implementation process, it also includes: when the manually adjusted electric-carbon conversion coefficient is obtained, the node-level carbon emission is calculated according to the manually adjusted electric-carbon conversion coefficient.

[0064] Among them, the system allows manual adjustment of the electric-carbon conversion coefficient according to seasonal changes or policy adjustments and other external factors. The manually set electric-carbon conversion coefficient has a higher priority than the coefficient calculated based on real-time energy structure data.

[0065] The node-level carbon emissions are calculated by manually adjusting the electric-carbon conversion coefficient, aiming to adapt to the complex and variable external environment, and ensure the accuracy and flexibility of carbon emission calculation. For example, during the winter heating period, the coal-fired power generation may increase significantly, which in turn causes the electric-carbon conversion coefficient to rise. In this case, manually adjusting the electric-carbon conversion coefficient can quickly respond to this change and prevent errors in carbon emission calculation due to data update delays. Similarly, when policy changes affect the proportion of certain power types, manual adjustment can also ensure the accuracy of the calculation results. This mechanism not only ensures the automatic operation of the system, but also provides flexibility for human intervention, making the monitoring and management of carbon emissions more scientific and efficient.

[0066] S104, TD3 is used to optimize the grid topology and energy storage scheduling strategy with the carbon flow density matrix as the constraint condition.

[0067] For the complex power network structure and multi-objective optimization requirements, the improved TD3 algorithm is introduced to optimize the grid topology structure and energy storage device scheduling strategy. Compared with traditional reinforcement learning methods, TD3 algorithm has higher stability and convergence speed, and can effectively deal with high-dimensional, nonlinear and other problems in the power system.

[0068] Among them, the grid topology and energy storage scheduling strategy are optimized with the carbon flow density matrix as the constraint condition, which can realize global reduction of carbon emissions, improve system operation efficiency and economy, and enhance power supply stability and resilience; at the same time, it promotes energy consumption, optimizes the utilization rate of energy storage devices, and supports the realization of low-carbon energy transformation goals.

[0069] S105, generate an intelligent report containing a carbon footprint thermal map, emission reduction potential assessment and blockchain storage.

[0070] Among them, the carbon emission analysis results are displayed in the form of intelligent report. Through the regional carbon flow thermal map, the carbon emission intensity distribution of each grid node is intuitively displayed. Combined with historical data and current operating state, the carbon emission trend in the future period is predicted, and the carbon emission prediction curve is generated. Based on this curve, the carbon emission changes before and after the implementation of the optimization strategy are compared and analyzed, and the potential emission reduction benefits are calculated to complete the emission reduction potential assessment. In addition, in terms of data privacy protection, traditional methods rely on uploading in the original data set, which is difficult to meet the requirements of laws and regulations such as the Data Security Law, and restricts the compliance of cross-regional carbon management. In order to ensure the authenticity and traceability of data, all results are encrypted and stored by blockchain technology to effectively prevent tampering and forgery.

[0071] In this embodiment, the comprehensive optimization model constructed by combining smart internet technology, big data analysis and artificial intelligence technology can effectively realize the accurate quantification and collaborative optimization of carbon footprint in the power system. By combining probabilistic power flow analysis and calculation with carbon footprint optimization, this scheme can globally model and dynamically adjust carbon emissions in a multi-dimensional scenario, fully considering uncertainty factors to improve optimization accuracy. Based on the "electricity-energy-carbon" adaptive conversion algorithm, online monitoring technology can collect, share and process carbon footprint data of different nodes in real time, and realize data credibility and traceability combined with blockchain technology. With the help of artificial intelligence technology, this scheme can self-learn and optimize iteration of carbon footprint indicators, dynamically adjust the power grid topology and energy storage configuration, so as to maximize the carbon reduction benefit. At the same time, the intelligent carbon footprint report provides a transparent carbon emission evaluation tool for managers. Overall, the embodiment of the application enhances the multi-objective collaborative optimization capability of power grid safety and economy, improves the low-carbon level and resource utilization efficiency of power grid operation, and realizes the transition from local optimization to global collaboration.

[0072] In a possible implementation, the generating the node carbon flow density matrix in multiple scenarios based on the joint probability model constructed by the Monte Carlo simulation and the deep belief network comprises:

[0073] predicting a load probability distribution in the next 24 hours based on meteorological factors and historical load data;

[0074] generating multiple scenario power flow samples by using an improved Nataf transformation to process non-Gaussian correlated variables;

[0075] identifying carbon emission source propagation paths and calculating node carbon flow contribution degrees according to the multiple scenario power flow samples by using a reverse power flow method;

[0076] generating the node carbon flow density matrix in multiple scenarios according to the node carbon flow contribution degrees.

[0077] In the implementation process, a multi-dimensional feature input vector is constructed by integrating meteorological factors (including temperature, wind speed and solar intensity) and historical load time series data; a deep belief network is used to extract hierarchical features of the input data, and capture the nonlinear correlation between the load and the meteorological conditions; based on a time sliding window mechanism, the probability density function of the load in each period in the next 24 hours is predicted, and an intervalized load distribution characterized by mean and variance is output.

[0078] For non-Gaussian correlated variables such as load and new energy output, an improved Nataf transformation is used to process multiple scenario power flow samples, including:

[0079] K-L transformation is used to orthogonally decompose non-Gaussian variables to eliminate high-order correlation between variables;

[0080] Reconstructing the nonlinear correlation between variables based on Copula function;

[0081] Generate multi-scenario sample set conforming to target distribution through Cholesky decomposition.

[0082] Subsequently, reverse power flow analysis is performed on each scenario sample, and a node-branch carbon flow correlation network is constructed using multi-scenario power flow samples. Through this network, the carbon flow transmission coefficients of each power generation node to the load node are determined, and the carbon emission source-load propagation path topology is established. Based on the carbon emission intensity of the power generation node and the transmission path weight, the carbon flow contribution degree of each node in a specific scenario is calculated.

[0083] When generating the node carbon flow density matrix in multiple scenarios according to the point carbon flow contribution degree, the matrix is constructed with nodes as rows and scenarios as columns, and the matrix elements are the normalized carbon flow density values of each node in different scenarios. By analyzing and processing the multi-scenario matrix, the expected carbon flow density and confidence interval are extracted, forming a density matrix representing the spatiotemporal distribution characteristics of the power grid carbon flow.

[0084] In the specific implementation process, the matrix is coupled with the geographic information system (GIS) to visually display the carbon emission hotspots and key propagation paths, providing spatial dimension decision basis for subsequent optimization.

[0085] This embodiment realizes the probabilistic modeling of the carbon footprint of the power system through meteorological-load joint prediction, non-Gaussian scenario generation, carbon flow reverse tracing, and matrix representation, solves the defects of traditional deterministic models that cannot depict the volatility of new energy and the influence of multi-energy coupling, and provides reliable data support for carbon flow dynamic tracking and optimization.

[0086] In one possible implementation, the electric carbon conversion coefficient is dynamically updated according to real-time energy structure data, including:

[0087] Real-time acquisition of grid-connected energy type and proportion data, and calculation of dynamic correction factor;

[0088] Synchronously updating the electric carbon conversion parameters of the edge node and the cloud using the federated learning framework.

[0089] Through the interface with the power grid dispatching system or related data platform, the output power and proportion information of each type of power source in the current power system can be obtained. These data reflect the real-time operating state and energy structure characteristics of the system, and are the basis for calculating the dynamic correction factor. Since different types of power sources (such as coal-fired, gas-fired, wind, and solar) have different carbon emission intensities, it is necessary to adjust the electric carbon conversion coefficient according to the real-time energy structure changes.

[0090] The electric-carbon conversion parameters of the edge nodes and the cloud are synchronously updated by using a federal learning framework. By distributing the calculation tasks on multiple edge nodes and combining the collaborative optimization of the central server (the cloud), the efficient synchronization and optimization of the parameters can be realized while protecting the data privacy.

[0091] In the embodiment, the dynamic correction factor quantifies the influence of the difference in carbon emission intensity of different types of power sources on the overall system carbon emission, thereby realizing dynamic updating of the electric-carbon conversion parameters. The distributed collaborative mechanism of the edge nodes and the cloud not only improves the scalability and flexibility of the system, but also avoids the risks caused by single point failure or data leakage in the traditional centralized calculation mode.

[0092] In a possible implementation manner, the method further includes:

[0093] The abnormal data is subjected to credibility weighting, and the low-confidence data is filtered and then participates in the calculation.

[0094] In the process of acquiring the energy type and proportion data in real time, abnormal data or low-confidence data (such as error information caused by sensor failure, communication delay, etc.) may be encountered. Specifically, the credibility of the collected data is evaluated, and the reliability of the data is judged by historical data comparison, statistical analysis or other machine learning methods. For low-confidence data, it is weighted and adjusted or directly filtered to avoid its interference on the subsequent calculation results.

[0095] In the embodiment, the abnormal data is subjected to credibility weighting, and the low-confidence data is filtered and then participates in the calculation, thereby ensuring the accuracy and reliability of the dynamic correction factor and the electric-carbon conversion parameters.

[0096] In a possible implementation manner, the node-level carbon emission calculation formula is:

[0097] node=∑i=1 n (P i ·α i (t)+Q i ·β i (t))·γ(t)]

[0098] Wherein, γ(t) is calculated by real-time grid-connected energy data, and satisfies:

[0099]

[0100] Wherein, P i is the active power of node i, Q i is the reactive power of node i, α i (t) is the time-varying active electric-carbon conversion coefficient of node i, β i (t) is the time-varying reactive electric-carbon conversion coefficient of node i, and γ(t) is the dynamic correction factor; EFk (t) is the real-time carbon emission factor of the kth type of energy; w k is the real-time proportion weight of the kth type of energy in the power grid.

[0101] In one possible implementation, the reward function of the TD3 algorithm is defined as:

[0102] R = -λ1·∑C node -λ2·ΔLoss + λ3·V margin

[0103] where λ1, λ2, and λ3 are dynamic weight coefficients, C node is the real-time carbon emission of the node, ΔLoss is the change in network loss, and V margin is the voltage safety margin.

[0104] In one possible implementation, before optimizing the power grid topology and energy storage scheduling strategy, the following steps are further included:

[0105] Constructing a carbon-electric coupling sensitivity matrix to identify nodes with high emission reduction potential;

[0106] Initializing the population using a chaotic particle swarm algorithm to accelerate convergence to the feasible solution domain;

[0107] Embedding a carbon trading cost factor in the objective function to define economic constraints.

[0108] Obtaining real-time operation data collected by intelligent sensor groups deployed at nodes in the power network, including node voltage, current, power, and device energy consumption data, as well as carbon emission data related to the nodes. For each node, calculate the degree of influence of power change on carbon emission change. For example, when the power of a node increases by a certain percentage, analyze the corresponding increase in carbon emissions to quantify the coupling relationship between the two. By performing such calculations for all nodes, a complete matrix is formed.

[0109] In this matrix, through specific threshold setting and data analysis methods, nodes with high emission reduction potential can be identified. Nodes whose power changes have a greater impact on carbon emissions, i.e., nodes with high carbon-electric coupling sensitivity, are identified as nodes with high emission reduction potential. These nodes become the focus of subsequent optimization work because targeted adjustments to them are expected to achieve significant results in carbon emission reduction.

[0110] The chaotic particle swarm optimization algorithm is used to initialize the initial population of the optimization algorithm, improving the optimization efficiency. First, set the algorithm parameters, such as the number of particles and the search range, balance the problem complexity and computing resources. Select the appropriate number of particles to find high-quality solutions within a reasonable time. Use chaotic models such as the Logistic map to generate chaotic sequences that are random and ergodic, mapping to the search space to form the initial particle positions. Uniformly distributed initial particle positions help cover a wider search space and avoid local optima, which is better than traditional random initialization. After generating the initial particles, the particle swarm optimization algorithm begins iteration. Particles adjust their speed and position based on historical best and global best, and chaotic factors fine-tune particle positions to improve search efficiency. Iteration brings particles closer to the feasible solution domain, accelerating convergence to solutions that meet engineering constraints, providing good initial solutions for grid topology and energy storage scheduling strategy optimization.

[0111] When optimizing the grid and energy storage scheduling, economic considerations are crucial. To this end, a carbon trading cost factor is added to the objective function. First, understand the rules and prices of the carbon trading market and obtain real-time carbon trading data. These data can be obtained from market platforms or authoritative agencies. Based on the carbon trading price and the carbon emissions of each node, a carbon trading cost term is constructed. For example, the carbon emissions of each node are multiplied by the carbon trading price to obtain the carbon trading cost of that node. Accumulate the carbon trading costs of all nodes to form the total cost term. Incorporate the carbon trading cost term into the objective function, which usually includes device operating costs, energy storage charging and discharging costs, etc. By setting cost weights, the objective function can comprehensively reflect the economic cost of grid operation. At the same time, according to the demand and economic constraints, define the constraint conditions, such as limiting carbon emissions, power balance of the grid, and energy storage charging and discharging limits, to ensure that the optimization process meets the requirements of carbon emission reduction, economic feasibility, and system stability.

[0112] In this embodiment, by constructing a carbon-electricity coupling sensitivity matrix to locate high emission reduction potential nodes, using a chaotic particle swarm optimization algorithm to accelerate the convergence of the population to the feasible solution domain, and embedding a carbon trading cost factor in the objective function to define economic constraints, the foundation is laid for subsequent optimization of grid topology and energy storage scheduling strategies, improving optimization efficiency, identifying key emission reduction nodes, and considering economic costs.

[0113] In one possible implementation, the localization of carbon flow accounting is performed by an edge computing node, including:

[0114] Calculate the local carbon intensity based on the equipment nameplate parameters and real-time operating data;

[0115] Update the electricity-carbon conversion coefficient using a lightweight federated learning model.

[0116] The equipment nameplate details the basic information of the equipment, such as for a transformer, the nameplate will mark its rated capacity, no-load loss, short-circuit loss and other key parameters; for a motor, the rated power, efficiency, etc. will be marked. These parameters are the performance indicators of the equipment under standard working conditions, and are important basic data for calculating carbon intensity.

[0117] Real-time working condition data includes voltage, current, power and equipment energy consumption data when the equipment is currently running. Taking a transformer as an example, the voltage and current data under real-time working conditions can be used to calculate the load rate during actual operation, and combined with the no-load loss and short-circuit loss parameters on the nameplate, the actual loss power of the transformer at the current time can be accurately calculated.

[0118] After obtaining the equipment nameplate parameters and real-time data, the local carbon intensity is calculated. For different types of equipment, the corresponding model is used. For example, according to the power and running time, combined with the carbon emission factor (obtained from the standard database), the carbon emission per unit time is calculated. Then, the carbon emission is divided by the product quantity or service quantity to obtain the local carbon intensity.

[0119] To build a lightweight federated learning model, participants such as edge computing nodes and central servers in the power network need to be determined. The edge nodes have local power data, including real-time energy structure and carbon emission data. During model training, the edge nodes use local data to preliminarily train the model, and then upload the model parameters to the central server. The server aggregates the parameters and uses encryption algorithms to protect data privacy. For example, homomorphic encryption technology allows aggregation operations to be performed without decryption, ensuring data security. The aggregated model parameters are returned to the edge nodes, which fine-tune the model with local data. After multiple rounds of training and fine-tuning, the model converges and can accurately update the electricity-carbon conversion coefficient. This makes the carbon flow accounting system more accurate and real-time.

[0120] In this embodiment, through the edge computing node, the local carbon intensity is accurately calculated using the equipment nameplate and real-time working condition data, and the electricity-carbon conversion coefficient is safely and efficiently updated using the lightweight federated learning model, realizing accurate accounting of local carbon flow and improving the real-time and accuracy of the carbon flow accounting system.

[0121] In one possible implementation, after updating the electricity-carbon conversion coefficient using the lightweight federated learning model, the sensitive data is encrypted using differential privacy and uploaded to the cloud.

[0122] In this embodiment, after updating the electricity-carbon conversion coefficient using the lightweight federated learning model, the sensitive data is encrypted using differential privacy and uploaded to the cloud to ensure the security and privacy of the data during transmission and storage.

[0123] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiment of the application.

[0124] The following is an embodiment of the device of the application, and for details not described in detail, reference can be made to the corresponding method embodiments described above.

[0125] Figure 2 The structure of the full-link electricity consumption monitoring system based on smart internet provided by the embodiment of the application is shown, only the part related to the embodiment of the application is shown for convenience of description, and the details are as follows:

[0126] As Figure 2 shown, the full-link electricity consumption monitoring system based on smart internet includes: an internet sensing layer 201, a probabilistic power flow analysis module 202, a carbon flow dynamic modeling module 203, an adaptive optimization engine 204 and a carbon footprint analysis platform 205;

[0127] The internet sensing layer 201 is used to acquire the operation data collected by the intelligent sensor group deployed in the power network node in real time, and perform local carbon flow accounting according to the operation data through the edge computing node; wherein the operation data includes: voltage, current, power and device energy consumption data;

[0128] The probabilistic power flow analysis module 202 is used to construct a joint probability model based on Monte Carlo simulation and deep belief network, and generate a node carbon flow density matrix under multiple scenarios;

[0129] The carbon flow dynamic modeling module 203 is used to dynamically update the electric carbon conversion coefficient according to the real-time energy structure data, and calculate the node-level carbon emission;

[0130] The adaptive optimization engine 204 is used to adopt TD3, and optimize the power grid topology and energy storage scheduling strategy with the carbon flow density matrix as the constraint condition;

[0131] The carbon footprint analysis platform 205 is used to generate an intelligent report containing a carbon footprint thermal map, emission reduction potential evaluation and blockchain storage.

[0132] In one possible implementation, the probabilistic power flow analysis module 202 is specifically used for:

[0133] Based on meteorological factors and historical load data, the future 24-hour load probability distribution is predicted;

[0134] An improved Nataf transformation is used to process non-Gaussian correlated variables to generate multi-scenario power flow samples;

[0135] According to the multi-scenario power flow sample, a carbon emission source propagation path is identified by a reverse power flow method, and a node carbon flow contribution degree is calculated.

[0136] According to the point carbon flow contribution degree, a node carbon flow density matrix under multi-scenarios is generated.

[0137] In a possible implementation, the carbon flow dynamic modeling module 203 is specifically configured to:

[0138] Real-time grid-connected energy type and proportion data are acquired, and a dynamic correction factor is calculated.

[0139] The electric-carbon conversion parameters of the edge node and the cloud are synchronously updated by using a federated learning framework.

[0140] In a possible implementation, the carbon flow dynamic modeling module 203 is further configured to:

[0141] Abnormal data are subjected to credibility weighting, and low-confidence data are filtered and then participate in calculation.

[0142] In a possible implementation, the adaptive optimization engine 204 is further configured to:

[0143] A carbon-electricity coupling sensitivity matrix is constructed, and a high emission reduction potential node is identified.

[0144] A chaotic particle swarm optimization algorithm is used to initialize a population, and convergence to a feasible solution domain is accelerated.

[0145] A carbon trading cost factor is embedded in a target function, and an economic constraint condition is defined.

[0146] In a possible implementation, the Internet of Things perception layer 201 is specifically configured to:

[0147] Local carbon emission intensity is calculated based on equipment nameplate parameters and real-time working condition data.

[0148] A lightweight federated learning model is used to update electric-carbon conversion coefficients.

[0149] In a possible implementation, the Internet of Things perception layer 201 is further configured to upload sensitive data to the cloud after implementing differential privacy encryption.

[0150] In the embodiment, the comprehensive optimization model constructed by combining wisdom Internet technology, big data analysis and artificial intelligence technology can effectively realize accurate quantification and collaborative optimization of carbon footprint in the power system. By combining probabilistic power flow analysis and calculation with carbon footprint optimization, the scheme can globally model and dynamically adjust carbon emissions in a multi-dimensional scenario, fully considering uncertainty factors to improve optimization accuracy. Based on the "electricity-energy-carbon" adaptive conversion algorithm, the online monitoring technology can collect, share and process carbon footprint data of different nodes in real time, and realize data credibility and traceability by combining blockchain technology. With the aid of artificial intelligence technology, the scheme can self-learn and optimize iteration of carbon footprint indicators, dynamically adjust the power grid topology and energy storage configuration, so as to maximize the carbon reduction benefit. At the same time, the intelligent carbon footprint report provides a transparent carbon emission evaluation tool for managers. Overall, the embodiment of the application enhances the multi-objective collaborative optimization capability of power grid safety and economy, improves the low-carbon level and resource utilization efficiency of power grid operation, and realizes global collaboration from local optimization.

[0151] Figure 3 is a schematic diagram of an electronic device provided by an embodiment of the application. As shown in Figure 3 the electronic device 3 of the embodiment includes a processor 30 and a memory 31. The memory 31 stores a computer program 32. The processor 30 implements the steps in each of the method embodiments described above when executing the computer program 32. Alternatively, the processor 30 implements the functions of each module / unit in each of the device embodiments described above when executing the computer program 32.

[0152] For example, the computer program 32 can be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 30 to complete the application. The one or more modules / units can be a series of computer program instruction segments that can complete a specific function, which are used to describe the execution process of the computer program 32 in the electronic device 3.

[0153] The electronic device 3 can include, but is not limited to, the processor 30, the memory 31. Those skilled in the art can understand that Figure 3 The electronic device 3 is only an example and does not constitute a limitation on the electronic device 3, and can include more or fewer components than shown, or combine certain components, or different components, for example, the electronic device 3 can also include an input / output device, a network access device, a bus, etc.

[0154] For the convenience and brevity of description, only the above-mentioned division of each functional module / unit is exemplified, and in actual application, the above-mentioned functions can be completed by different functional modules / units according to needs. The above-mentioned modules / units can be realized in the form of hardware, software or a combination of hardware and software.

[0155] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in a certain embodiment can be referred to the relevant description of other embodiments. If there is no special description and logical conflict, the terms and / or descriptions of different embodiments are consistent and can be mutually referred to, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0156] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for monitoring energy consumption across all stages based on the Internet of Things, characterized in that, include: The system acquires real-time operational data from intelligent sensor arrays deployed at power grid nodes and performs localized carbon flow calculations based on this data via edge computing nodes. The operational data includes voltage, current, power, and equipment energy consumption data. A joint probability model is constructed based on Monte Carlo simulation and deep belief network to generate node carbon flow density matrices in multiple scenarios. The electricity-carbon conversion coefficient is dynamically updated based on real-time energy structure data, and node-level carbon emissions are calculated: The dual-delay deep deterministic strategy gradient algorithm TD3 is adopted to optimize the grid topology and energy storage scheduling strategy with the carbon flux density matrix as a constraint. Generate intelligent reports that include carbon footprint heatmaps, emission reduction potential assessments, and blockchain-based evidence.

2. The method for monitoring energy consumption across all stages based on the Internet of Things as described in claim 1, characterized in that, The joint probability model constructed based on Monte Carlo simulation and deep belief network to generate node carbon flow density matrices under multiple scenarios includes: Predicting the probability distribution of load for the next 24 hours based on meteorological factors and historical load data; An improved Nataf transform is used to process non-Gaussian correlated variables to generate power flow samples for multiple scenarios; Based on the multi-scenario current flow samples, the carbon emission source propagation path is identified using the reverse current flow method, and the node carbon flow contribution is calculated. Based on the point carbon flow contribution, a node carbon flow density matrix for multiple scenarios is generated.

3. The method for monitoring energy consumption across all stages based on the Internet of Things as described in claim 1, characterized in that, The dynamic updating of the electro-carbon conversion coefficient based on real-time energy structure data includes: Real-time acquisition of grid-connected energy type and proportion data, and calculation of dynamic correction factors; The federated learning framework is used to synchronously update the carbon conversion parameters between edge nodes and the cloud.

4. The method for monitoring energy consumption across all stages based on the Internet of Things according to claim 3, characterized in that, Also includes: Abnormal data is weighted by confidence level, and low-confidence data is filtered out before being included in the calculation.

5. The method for monitoring energy consumption across all stages based on the Internet of Things as described in claim 1, characterized in that, The formula for calculating node-level carbon emissions is: node=∑i=1 n (P i ·a i (t)+Q i ·b i (t))·γ(t)] Wherein, γ(t) is calculated using real-time grid-connected energy data and satisfies: Among them, P i Let Q be the active power at node i. i Let α be the reactive power of node i. i (t) represents the time-varying active carbon conversion coefficient at node i, β i γ(t) is the time-varying reactive carbon conversion coefficient at node i, and γ(t) is the dynamic correction factor; EF k (t) represents the real-time carbon emission factor for the k-th energy type; w k represents the real-time weight of the k-th energy type in the power grid.

6. The method for monitoring energy consumption across all stages based on smart IoT as described in claim 1, characterized in that, According to the method described in claim 1, the reward function of the TD3 algorithm is defined as: R=-λ1·∑C node -λ2·ΔLoss+λ3·V margin Where λ1, λ2, and λ3 are dynamic weighting coefficients, and C node V represents the real-time carbon emissions of the node, ΔLoss represents the change in network loss, and V represents the network loss. margin This is for voltage safety margin.

7. The method for monitoring energy consumption across all stages based on smart IoT as described in claim 1, characterized in that, Before the optimization of grid topology and energy storage dispatch strategy, the following is also included: Construct a carbon-electric coupling sensitivity matrix to identify nodes with high emission reduction potential; The chaotic particle swarm optimization algorithm is used to initialize the population, which accelerates convergence to the feasible solution domain. By embedding a carbon trading cost factor into the objective function, economic constraints are defined.

8. The method for monitoring energy consumption across all stages based on the Internet of Things as described in claim 1, characterized in that, The process of performing localized carbon flow accounting via edge computing nodes includes: Calculate local carbon emission intensity based on equipment nameplate parameters and real-time operating data; A lightweight federated learning model is used to update the electrocarbon conversion coefficient.

9. A comprehensive carbon footprint optimization system based on the Internet of Things, characterized in that, include: IoT sensing layer, probabilistic power flow analysis module, carbon flow dynamic modeling module, adaptive optimization engine, and carbon footprint analysis platform; The IoT sensing layer is used to acquire real-time operational data collected by intelligent sensor groups deployed on power network nodes, and to perform localized carbon flow calculations based on the operational data through edge computing nodes; wherein, the operational data includes: voltage, current, power and equipment energy consumption data; The probabilistic power flow analysis module is used to construct a joint probabilistic model based on Monte Carlo simulation and deep belief network, and generate node carbon flow density matrices under multiple scenarios. The carbon flow dynamic modeling module is used to dynamically update the electro-carbon conversion coefficient based on real-time energy structure data and calculate node-level carbon emissions. The adaptive optimization engine is used to optimize the grid topology and energy storage scheduling strategy using TD3 with the carbon flow density matrix as a constraint. The carbon footprint analysis platform is used to generate intelligent reports that include carbon footprint heatmaps, emission reduction potential assessments, and blockchain-based evidence.

10. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 8.

Images