Virtual power plant operation data intelligent optimization method and system based on internet of things

Abstract

The application discloses a virtual power plant operation data intelligent optimization method and system based on the Internet of Things, relates to the technical field of Internet of Things energy monitoring, and acquires energy transmission data sets and energy load data sets after preprocessing and maps to each corresponding node of the virtual power plant through real-time collection of energy operation data of each distributed node layer by arranging power sensors on the distributed node layer of the virtual power plant; energy flow state evaluation is carried out through the construction of an energy flow level deviation index Neq; when there is an anomaly in the energy flow level, a spatial distance matrix Dl(x, i) of the virtual power plant topology is extracted and a comprehensive operation coordination index Eop is constructed to evaluate the residual error distribution state. When the energy flow of the virtual power plant cannot converge naturally, the energy flow Φadj is calculated and distributed to dynamically correct the energy flow of each node, and the energy storage equipment is controlled to switch between the charging and discharging modes. The method realizes dynamic coordination and abnormal self-recovery control of the virtual power plant operation data.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) energy monitoring technology, specifically to a method and system for intelligent optimization of virtual power plant operation data based on IoT. Background Technology

[0002] With the widespread integration of distributed and renewable energy sources, virtual power plants (VPS), as a scheduling and management method integrating distributed power sources, energy storage, and controllable loads, are gradually becoming an important component of the power system. VPS utilizes IoT technology to achieve data interconnection and real-time monitoring between nodes, enabling decentralized energy units to have centralized scheduling capabilities at the logical level. However, in multi-level energy interaction structures, the dynamic characteristics of energy flow between nodes exhibit strong time-varying and nonlinear properties, making it difficult for traditional centralized control models to accurately characterize real-time energy distribution. Therefore, intelligent optimization methods for VPS operation data should focus on energy flow status, achieving dynamic adjustment and adaptive coordination of energy flow through multi-dimensional data analysis of energy transmission, load, and spatial distribution.

[0003] Current energy flow adjustments in virtual power plants primarily rely on centralized scheduling or empirical models, lacking a real-time response mechanism for the energy interaction states of multiple nodes. Due to the complexity of energy transmission paths between nodes, and influenced by factors such as geographical topology, equipment response lag, and data latency, existing scheduling methods often fail to accurately reflect the energy conservation relationships between nodes, leading to mismatches between power output and input at some nodes. This adjustment method overly depends on preset power balance formulas and linear allocation strategies, lacking the ability to comprehensively analyze spatial distribution, temporal dynamics, and multi-node correlations. This easily results in lagging energy flow state assessments and prevents control commands from synchronizing with the actual operating state, thus affecting the overall coordinated operation of the virtual power plant. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for intelligent optimization of virtual power plant operation data based on the Internet of Things, thus solving the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent optimization method for virtual power plant operation data based on the Internet of Things, comprising the following steps:

[0006] S1. By deploying IoT sensors in the distributed node layer managed by the virtual power plant, the energy operation data of each distributed node layer is collected in real time. After preprocessing, the energy transmission dataset and energy load dataset are obtained and mapped to the corresponding nodes of the virtual power plant.

[0007] S2. Perform energy flow status analysis on the virtual power plant based on the energy transmission dataset, generate an energy flow status assessment based on the analysis results, and trigger S3 when there are anomalies at the energy flow level.

[0008] S3. Extract the spatial distance matrix Dl(x,i) of the IoT topology modeling layer of the virtual power plant, and perform spatial field residual analysis on the virtual power plant by fitting it with the energy load dataset.

[0009] S4. Fit the energy flow state analysis results with the spatial field residual analysis results, perform operation coordination analysis on the virtual power plant, generate residual distribution state assessment based on the analysis results, and trigger S5 when the energy flow cannot converge naturally.

[0010] S5. Adjust the energy flow of each node in a state where it cannot converge naturally by calculating and allocating the energy flow Φadj.

[0011] Preferably, S1 includes S11;

[0012] S11. By deploying IoT sensors in the distributed node layer managed by the virtual power plant, energy operation data of each distributed node layer is collected in real time.

[0013] The IoT sensor includes a power sensor;

[0014] The input power Pin received by the lower-level equipment nodes from the upper-level power grid is obtained in real time by power sensors installed on the input side of the substation bus at the distributed node layer.

[0015] The output power Pou released by the power grid to the lower-level device node in the last time is obtained in real time by a power sensor placed between the node output end and the grid bus.

[0016] The measured power Ps of each distributed node layer is collected in real time by power sensors installed at the output bus of the entire power flow path of each distributed node.

[0017] S12. During the data acquisition process, the energy operation data is preprocessed by the data preprocessing unit built into the IoT sensor to obtain the energy transmission dataset and the energy load dataset.

[0018] The preprocessing includes outlier removal and noise reduction;

[0019] The outlier removal is based on an anomaly detection algorithm to automatically identify and remove abnormal signals caused by communication delays, environmental interference, and transient fluctuations. The denoising is performed on the energy operation data through multi-scale filtering and median smoothing.

[0020] The energy transfer dataset includes input power Pin and output power Pou;

[0021] The energy load dataset includes measured power Ps.

[0022] Preferably, S1 further includes S13;

[0023] S13. Based on the hierarchical topology and energy dispatch configuration table of the virtual power plant, establish a mapping matrix M between the distributed node layer and the virtual nodes. map (p, v), where p represents the physical acquisition point index and v represents the virtual power plant logical node. The distributed node layer is logically bound to the virtual power plant node, and the energy transmission dataset and energy load dataset are mapped to the corresponding nodes of the virtual power plant according to the logical binding relationship.

[0024] Preferably, S2 includes S21;

[0025] S21. After completing the data mapping, the energy transfer dataset is fitted, the energy flow state of the virtual power plant is analyzed, and the energy flow hierarchy deviation index Neq is constructed to describe the energy conservation state of the virtual power plant within a given scheduling cycle, reflecting the overall balance of energy transfer between different levels, as detailed below:

[0026] ;

[0027] Where Neq(t) represents the energy flow hierarchy deviation index at time t, N represents the number of child nodes participating in scheduling, and Pin i (t) represents the input power received by node i from the upper-level power grid at time t, Pin j (t) represents the output power released by node j to the lower-level node at time t.

[0028] Preferably, S2 further includes S22;

[0029] S22. Calculate the mean value of the energy flow level deviation index Neq when the energy flow consistency is compliant using statistical methods, and set a preset energy flow deviation judgment threshold Py based on the mean value. Then, evaluate the energy flow status with the real-time energy flow level deviation index Neq. The specific evaluation scheme is as follows.

[0030] When the energy flow level deviation index Neq ≤ energy flow deviation judgment threshold Py, it means that the virtual power plant's energy flow consistency is compliant, and normal scheduling is maintained at this time;

[0031] When the energy flow level deviation index Neq > the energy flow deviation judgment threshold Py, it indicates that there is an anomaly at the energy flow level of the virtual power plant, and spatial residual analysis is triggered.

[0032] Preferably, S3 includes S31;

[0033] S31. When the energy flow status assessment indicates an anomaly at the energy flow level, extract the spatial distance matrix Dl(x, i) of the virtual power plant IoT topology modeling layer. This matrix represents the spatial distance between any spatial location node x within the virtual power plant area and the i-th node in the virtual power plant. Fit this matrix to the energy load dataset to perform spatial field residual analysis on the virtual power plant. Construct the spatial residual field exponent Rfc, representing the continuous mapping of the virtual power plant's operating residuals in spatial distribution. This exponent is used to describe the intensity and diffusion of prediction errors within the virtual power plant in the spatial dimension, as detailed below:

[0034] ;

[0035] Where Rfc(x, t) represents the spatial residual field exponent at spatial location x at time t, N represents the number of child nodes participating in the scheduling, and Ps i (t) represents the measured power of node i at time t, Pp i (t) represents the predicted power of node i at time t, which is set by the user according to the needs of the virtual power plant. e represents the exponential function, and λ represents the spatial diffusion scale, which is set according to the node density.

[0036] Preferably, S4 includes S41;

[0037] S41. The energy flow hierarchy deviation index Neq and the spatial residual field index Rfc are fitted to perform operational coordination analysis on the virtual power plant, and a comprehensive operational coordination index Eop is constructed to represent the overall consistency of the virtual power plant's operational data, reflecting the coordination between the energy layer and the spatial layer of the virtual power plant, as detailed below:

[0038] ;

[0039] Where Eop(t) represents the comprehensive operation and scheduling index at time t, Ω represents the geographical distribution range of all nodes of the virtual power plant, and dx represents the infinitesimal element of the spatial integral in the integral function.

[0040] Preferably, S4 further includes S42;

[0041] S42. Calculate the mean of the comprehensive operation scheduling index Eop when the residual spatial distribution is stable using statistical methods, and set a preset residual distribution critical threshold Cy based on the mean. Then, evaluate the residual distribution status with the real-time acquired comprehensive operation coordination index Eop. The specific evaluation scheme is as follows.

[0042] When the comprehensive operation coordination index Eop < the residual distribution critical threshold Cy, it indicates that the residual space distribution is stable and the energy flow of the virtual power plant can converge naturally. At this time, the monitoring frequency is increased by 50%, and the current state is recorded as a convergent state.

[0043] When the comprehensive operation coordination index Eop ≥ the residual distribution critical threshold Cy, it indicates that the residual spatial distribution is abnormal and the energy flow of the virtual power plant cannot converge naturally. At this time, energy re-regulation is performed.

[0044] Preferably, S5 includes S51 and S52;

[0045] S51. When the energy flow of the virtual power plant cannot converge naturally, the system is fitted based on the acquired energy transmission dataset, the energy flow hierarchy deviation index Neq, and the comprehensive operation scheduling index Eop. This is used to adjust the energy flow of each node in a state where it cannot converge naturally, and to calculate the allocated energy flow Φadj, which represents the power correction relationship between nodes under the energy conservation deviation condition and describes the dynamic adjustment ratio of each node under the energy consistency deviation, as detailed below:

[0046] ;

[0047] Where Φadj(i,j,t) represents the required energy flow between the i-th energy input node in the upper layer and the j-th energy output node in the lower layer at time t, ε is a small constant to prevent the denominator from approaching zero, with a value of 0.001, and Neq + Denotes the positive part operator, where Neq(t) ≤ Py, and Neq... + =0, when Neq(t) > Py, Neq + =Neq(t)−Py;

[0048] S52. Based on the acquired distributed energy flow Φadj, the virtual power grid is adjusted by the following:

[0049] When the energy flow Φadj≤1, ​​the energy storage device is switched to charging mode and the energy output of 20% of the distributed nodes is limited. Then, iterative analysis is performed through S1.

[0050] When the energy flow Φadj > 1, the energy storage device is switched to discharge mode, and the distributed nodes are adjusted to the rated energy output. Then, iterative analysis is performed through S1.

[0051] The IoT-based intelligent optimization system for virtual power plant operation data includes an IoT mapping module, an energy analysis module, a residual field analysis module, a convergence judgment module, and an energy flow balance module.

[0052] The IoT mapping module is used to deploy IoT sensors in the distributed node layer managed by the virtual power plant to collect energy operation data of each distributed node layer in real time, obtain energy transmission dataset and energy load dataset after preprocessing, and map them to the corresponding nodes of the virtual power plant.

[0053] The energy analysis module is used to perform energy flow state analysis on the virtual power plant based on the energy transmission dataset, generate an energy flow state assessment based on the analysis results, and trigger the residual field analysis module when there are anomalies at the energy flow level.

[0054] The residual field analysis module is used to extract the spatial distance matrix Dl(x,i) of the IoT topology modeling layer of the virtual power plant, and to perform spatial field residual analysis on the virtual power plant by fitting it with the energy load dataset.

[0055] The convergence judgment module is used to fit the energy flow state analysis results with the spatial field residual analysis results, perform operation coordination analysis on the virtual power plant, generate residual distribution state assessment based on the analysis results, and trigger the energy flow balance module when the energy flow cannot converge naturally.

[0056] The energy flow balancing module is used to calculate and allocate energy flow Φadj to adjust the energy flow of each node when it cannot converge naturally.

[0057] This invention provides a method and system for intelligent optimization of virtual power plant operation data based on the Internet of Things (IoT). It has the following beneficial effects:

[0058] (1) This method deploys IoT power sensors in the distributed node layer of the virtual power plant via S1 to collect energy operation data of each distributed node layer in real time, realizing full-link energy data acquisition from the input bus to the output bus. The collected energy operation data undergoes outlier removal and multi-scale filtering through a built-in algorithm, eliminating the influence of non-real factors such as communication delays and transient fluctuations, forming energy transmission datasets and energy load datasets. Furthermore, the physical acquisition points are bound to the virtual nodes through the topology mapping matrix Mmap(p,v), ensuring that the energy flow data of each node can be accurately located in both the spatial and logical layers. The entire energy monitoring system realizes the transformation from discrete measurement points to a structured network, providing an accurate, real-time, and hierarchically corresponding original data foundation for subsequent energy conservation analysis.

[0059] (2) Method S2 calculates the energy flow hierarchy deviation index Neq using the energy transmission dataset to quantitatively analyze the input-output power relationship of the virtual power plant, which is used to characterize the energy balance state within the scheduling cycle. When abnormal deviations occur in the energy flow, S3 automatically extracts the spatial distance matrix Dl(x, i) of the virtual power plant's IoT topology and fits it with the energy load dataset to calculate the spatial residual field index Rfc, revealing the diffusion intensity and concentration characteristics of energy anomalies in the spatial domain. Through joint modeling of energy flow and spatial residuals, it is possible to identify both the fluctuations in energy over time and the clustering of anomalies in geographical distribution, thus realizing multi-dimensional monitoring and positioning of the internal state of the virtual power plant. This mechanism can distinguish the source and nature of energy flow anomalies from two orthogonal dimensions: energy conservation and spatial consistency.

[0060] (3) In method S4, the energy level deviation index Neq and the spatial residual field index Rfc are fitted to construct the comprehensive operation coordination index Eop, which is used to measure the overall coordination degree between the energy flow and spatial distribution of the virtual power plant. When the comprehensive operation coordination index Eop is greater than or equal to the residual distribution critical threshold Cy, it indicates that the energy flow cannot converge naturally. The allocated energy flow Φadj is calculated through S5 to determine the power correction relationship between the upper and lower level nodes. Based on the magnitude of the allocated energy flow Φadj, the energy storage device is automatically triggered to switch between charging and discharging modes, while adjusting the output ratio of the distributed nodes. This process enables the virtual power plant to have the ability to self-correct and self-balance, realizing real-time redistribution and coordinated control of the energy flow. Through continuous iteration and feedback analysis, the stability and efficiency of energy transmission can be maintained in complex scheduling environments, ensuring that the virtual power plant continues to operate in the optimal balance state under multi-node and dynamic load conditions. Attached Figure Description

[0061] Figure 1 This is a schematic diagram of the intelligent optimization method for virtual power plant operation data based on the Internet of Things according to the present invention;

[0062] Figure 2 This is a schematic diagram illustrating the steps of the IoT-based intelligent optimization system for virtual power plant operation data in this invention.

[0063] Figure 3 This is a block diagram illustrating the principle of the intelligent optimization method for virtual power plant operation data based on the Internet of Things (IoT) of this invention. Detailed Implementation

[0064] 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.

[0065] Example 1

[0066] Please see Figure 1 This invention provides an intelligent optimization method for virtual power plant operation data based on the Internet of Things. To achieve the above objectives, this invention is implemented through the following technical solution, including the following steps:

[0067] S1. By deploying IoT sensors in the distributed node layer managed by the virtual power plant, the energy operation data of each distributed node layer is collected in real time. After preprocessing, the energy transmission dataset and energy load dataset are obtained and mapped to the corresponding nodes of the virtual power plant.

[0068] S2. Perform energy flow status analysis on the virtual power plant based on the energy transmission dataset, generate an energy flow status assessment based on the analysis results, and trigger S3 when there are anomalies at the energy flow level.

[0069] S3. Extract the spatial distance matrix Dl(x,i) of the IoT topology modeling layer of the virtual power plant, and perform spatial field residual analysis on the virtual power plant by fitting it with the energy load dataset.

[0070] S4. Fit the energy flow state analysis results with the spatial field residual analysis results, perform operation coordination analysis on the virtual power plant, generate residual distribution state assessment based on the analysis results, and trigger S5 when the energy flow cannot converge naturally.

[0071] S5. Adjust the energy flow of each node in a state where it cannot converge naturally by calculating and allocating the energy flow Φadj.

[0072] In this embodiment, in S1, IoT sensors are deployed at the distributed node layer to collect energy operation data of each distributed node layer in real time. Data preprocessing is performed to establish a real-time data channel for input power, output power, and measured power, forming an energy transmission dataset and an energy load dataset. These datasets are then bound to virtual nodes based on topological mapping relationships. Compared to traditional virtual power plants that rely on centralized collection or manual aggregation, this structure ensures data accuracy and real-time performance at the source, enabling energy operation information to be digitized and structured at the node layer, laying a reliable data foundation for subsequent analysis. In S2, an energy flow hierarchy deviation index Neq is constructed using the energy transmission dataset to assess the energy flow status during the scheduling cycle. When anomalies exist at the energy flow level, in S3, the spatial distance matrix Dl(x, i) of the virtual power plant's IoT topology modeling layer is extracted and fitted with the energy load dataset. Spatial field residual analysis is performed on the virtual power plant, and a spatial residual field index Rfc is constructed to analyze the spatial diffusion characteristics of power deviation within the virtual power plant. Unlike existing technologies that only perform single-dimensional fluctuation analysis on energy data, this invention achieves collaborative diagnosis in the time and spatial domains. It can simultaneously identify composite anomalies of energy flow imbalance and geographical distribution imbalance, providing quantifiable evidence for the precise positioning and operational optimization of virtual power plants. S4 uses a comprehensive operational coordination index, Eop, constructed by fitting the energy flow hierarchical deviation index Neq with the spatial residual field index Rfc to assess the residual distribution state and evaluate the overall energy coordination level of the virtual power plant. S5 When the energy flow of the virtual power plant cannot converge naturally, it automatically calculates and allocates the energy flow Φadj, dynamically corrects the power transmission relationship between nodes, and triggers adaptive adjustment of the energy storage device and power redistribution of distributed nodes. Compared with traditional methods relying on manual scheduling or fixed threshold control, this invention achieves real-time feedback and self-convergence capabilities for the energy flow of the virtual power plant, possessing the characteristics of automatic identification, dynamic adjustment, and closed-loop optimization. Through this mechanism, the virtual power plant can maintain the balance and stability of energy transmission under conditions of multi-node collaboration and dynamic load fluctuations, thereby improving the accuracy and operational coordination of energy scheduling and enhancing the overall reliability and economy of operation.

[0073] Example 2

[0074] Please refer to Figure 3 Specifically: S1 includes S11;

[0075] S11. By deploying IoT sensors in the distributed node layer managed by the virtual power plant, energy operation data of each distributed node layer is collected in real time.

[0076] The IoT sensor includes a power sensor;

[0077] The input power Pin received by the lower-level equipment nodes from the upper-level power grid is obtained in real time by power sensors installed on the input side of the substation bus at the distributed node layer.

[0078] The output power Pou released by the power grid to the lower-level device node in the last time is obtained in real time by a power sensor placed between the node output end and the grid bus.

[0079] The measured power Ps of each distributed node layer is collected in real time by power sensors installed at the output bus of the entire power flow path of each distributed node.

[0080] S12. During the data acquisition process, the energy operation data is preprocessed by the data preprocessing unit built into the IoT sensor to obtain the energy transmission dataset and the energy load dataset.

[0081] The preprocessing includes outlier removal and noise reduction;

[0082] The outlier removal is based on an anomaly detection algorithm to automatically identify and remove abnormal signals caused by communication delays, environmental interference, and transient fluctuations. The denoising is performed on the energy operation data through multi-scale filtering and median smoothing.

[0083] The energy transfer dataset includes input power Pin and output power Pou;

[0084] The energy load dataset includes measured power Ps.

[0085] S1 also includes S13;

[0086] S13. Based on the hierarchical topology and energy dispatch configuration table of the virtual power plant, establish a mapping matrix M between the distributed node layer and the virtual nodes. map (p, v), where p represents the physical acquisition point index and v represents the virtual power plant logical node. The distributed node layer is logically bound to the virtual power plant node, and the energy transmission dataset and energy load dataset are mapped to the corresponding nodes of the virtual power plant according to the logical binding relationship.

[0087] In this embodiment, S1 deploys IoT power sensors at the distributed node layer of the virtual power plant to achieve full-path energy monitoring from the substation bus input side to the node output bus, and collects energy operation data from each distributed node layer in real time. S2 utilizes the data preprocessing unit built into the sensors to perform outlier removal and multi-scale filtering to denoise the energy operation data, forming a stable and interference-free energy transmission dataset and energy load dataset. S13 establishes a mapping matrix M between the distributed node layer and virtual nodes based on the hierarchical topology and energy dispatch configuration table of the virtual power plant. map(p, v) enables a one-to-one correspondence between physical and virtual nodes, allowing for accurate location and association of collected data within the logical space. This implementation method achieves unified acquisition, purification, and structured mapping of multi-source energy data from the virtual power plant, avoiding problems such as isolated measuring points, data redundancy, and time asynchrony in traditional monitoring. It makes energy flow monitoring more accurate and the response more real-time, providing reliable data support for subsequent energy flow analysis and coordinated control, and overall improving the availability, accuracy, and consistency of virtual power plant operation data.

[0088] Example 3

[0089] Please refer to Figure 3 Specifically: S2 includes S21;

[0090] S21. After completing the data mapping, the energy transfer dataset is fitted, the energy flow state of the virtual power plant is analyzed, and the energy flow hierarchy deviation index Neq is constructed to describe the energy conservation state of the virtual power plant within a given scheduling cycle, reflecting the overall balance of energy transfer between different levels, as detailed below:

[0091] ;

[0092] Where Neq(t) represents the energy flow hierarchy deviation index at time t, N represents the number of child nodes participating in scheduling, and Pin i (t) represents the input power received by node i from the upper-level power grid at time t, Pin j (t) represents the output power released by node j to the lower-level node at time t.

[0093] S2 further includes S22;

[0094] S22. Calculate the mean value of the energy flow level deviation index Neq when the energy flow consistency is compliant using statistical methods, and set a preset energy flow deviation judgment threshold Py based on the mean value. Then, evaluate the energy flow status with the real-time energy flow level deviation index Neq. The specific evaluation scheme is as follows.

[0095] When the energy flow hierarchy deviation index Neq ≤ energy flow deviation judgment threshold Py, it indicates that the energy flow consistency of the virtual power plant is compliant, which means that the input energy and output energy under the scheduling plan are reasonably matched and the energy exchange process of each node is stable. At this time, normal scheduling is maintained.

[0096] When the energy flow hierarchy deviation index Neq > the energy flow deviation judgment threshold Py, it indicates that there is an anomaly in the energy flow level of the virtual power plant, namely uneven energy distribution at the hierarchy level, power transmission imbalance, and delayed response of some nodes. At this time, spatial residual analysis is triggered.

[0097] In this embodiment, after energy data mapping is completed in the virtual power plant, S21 performs fitting analysis on the energy transmission dataset to construct the energy flow hierarchy deviation index Neq. This index is used to quantitatively characterize the energy conservation state of the virtual power plant within a given scheduling cycle, reflecting the balance relationship of energy transfer between the station level, node level, and equipment level. The theory behind this formula is derived from the combination of two classical principles: the law of energy conservation and the normalized difference index. The basic principle is E... in (t)−E out (t)=ΔE(t), where E in E(t) represents the total energy absorbed during time t. out (t) represents the total energy output within time t, and ΔE(t) represents the energy difference. For a multi-layer node virtual power plant, the energy transfer at each layer can be discretized as the superposition of the input and output power of multiple nodes. The power is the time derivative of the energy (i.e., ...). Therefore, under discrete node conditions, the above formula can be transformed into In mathematics, to quantify the balance between input and output without being affected by power scale, a normalized difference form is often used. This expression uses the numerator to represent relative differences and the denominator to balance scale, achieving dimensionlessness, scale adaptability, and comparability. Applied to energy flow fields, by substituting the sum of input and output power into A and B respectively, the differential form of the energy flow at the virtual power plant level can be obtained. Considering that the positive and negative directions of the energy difference have no physical meaning, and only the degree of deviation has value for judgment, the absolute value is taken in the molecular part, and the energy flow level deviation index Neq is finally obtained. This represents the total input power received by all child nodes from the upper-level power grid during the scheduling cycle; its physical meaning is energy input intensity. This represents the total power output from all child nodes to the lower layer or load, physically signifying the intensity of energy release or consumption. The numerator term... This represents the energy non-conservation quantity of a virtual power plant within a given time period, i.e., the energy imbalance caused by node delays, line losses, or inconsistent equipment responses during energy transfer; denominator term This represents the total energy flow of the virtual power plant, serving as a reference benchmark for energy scale. S22 calculates the mean value of the energy flow hierarchy deviation index Neq under energy flow consistency conditions using statistical methods, establishing an energy flow deviation judgment threshold Py. This threshold serves as a dynamic judgment benchmark for assessing energy flow compliance in real-time operation. When the energy flow hierarchy deviation index Neq is less than or equal to the energy flow deviation judgment threshold Py, it automatically confirms that the current energy dispatch matching is reasonable and the energy exchange between nodes is stable. When the energy flow hierarchy deviation index Neq exceeds the energy flow deviation judgment threshold Py, it automatically triggers spatial residual analysis to identify the spatial distribution characteristics of energy flow anomalies. Through this step, the virtual power plant can achieve real-time identification and dynamic hierarchical judgment of energy flow deviations during operation. Compared to traditional methods relying on manual monitoring or fixed thresholds, this achieves higher adaptability and accuracy, thereby improving the balance of energy dispatch and the coordinated stability of virtual power plant operation. This allows the virtual power plant to maintain continuous and efficient energy transmission even in a multi-node distributed structure.

[0098] Example 4

[0099] Please refer to Figure 3 Specifically: S3 includes S31;

[0100] S31. When the energy flow status assessment indicates an anomaly at the energy flow level, extract the spatial distance matrix Dl(x, i) of the virtual power plant IoT topology modeling layer. This matrix represents the spatial distance between any spatial location node x within the virtual power plant area and the i-th node in the virtual power plant. Fit this matrix to the energy load dataset to perform spatial field residual analysis on the virtual power plant. Construct the spatial residual field exponent Rfc, representing the continuous mapping of the virtual power plant's operating residuals in spatial distribution. This exponent is used to describe the intensity and diffusion of prediction errors within the virtual power plant in the spatial dimension, as detailed below:

[0101] ;

[0102] Where Rfc(x, t) represents the spatial residual field exponent at spatial location x at time t, N represents the number of child nodes participating in the scheduling, and Ps i (t) represents the measured power of node i at time t, Pp i (t) represents the predicted power of node i at time t, which is set by the user according to the needs of the virtual power plant. e represents the exponential function, and λ represents the spatial diffusion scale, which is used to control the diffusion rate of the residual in the spatial domain and is set according to the node density.

[0103] In this embodiment, when the energy flow status assessment results of the virtual power plant indicate an anomaly at the energy flow level, the spatial residual analysis stage is automatically initiated. The spatial distance matrix Dl(x, i) of the virtual power plant's IoT topology modeling layer is extracted, quantifying the equivalent connectivity distance between any spatially located node x and the i-th node. This equivalent connectivity distance considers the geographic spatial relationship between nodes and comprehensively reflects the connection level of the two nodes in the power grid topology, as well as the operating load and remaining transmission margin of each line segment along the actual transmission path. This is then combined with the energy load dataset to analyze the measured power Ps. i (t) and predicted power Pp i Spatially fitting the deviation between (t) constructs a continuously distributed spatial residual field exponent Rfc(x,t), reflecting the diffusion trend and intensity distribution of prediction errors in the spatial dimension, and identifying spatial clustering regions and diffusion directions of energy anomalies. By introducing a spatial diffusion scale λ to adjust the node density, the analysis results maintain a dynamic balance between resolution and smoothness in different regions. The original idea of ​​this formula comes from the Gaussian kernel function and exponential spatial weight model in geophysics and heat conduction, with the classic form being... Where f(x) represents the electric field strength or weight at point x in space, and v i Let d(x, i) be the observation at node i, d(x, i) be the spatial distance, and λ be the diffusion scale. Mathematically, this is a form of exponential kernel interpolation used to form a continuous field from discrete samples through distance weighting. This application introduces this model into the energy residual field of a virtual power plant, replacing the node observations with power residuals Ps. i (t)-Pp i (t), and rewrite the distance term as the spatial distance matrix Dl(x,i) of the IoT topology modeling layer to match the spatial distribution of nodes. Meanwhile, λ is no longer a fixed physical constant but is dynamically set according to node density to reflect the diffusion capability of energy anomalies in the spatial domain, ultimately constructing the spatial residual field exponent Rfc. In this formula, Ps i (t)-Pp i (t) represents the power residual, which is the difference between the measured power and the predicted power of the i-th node. It reflects the degree of energy mismatch of the node within a given scheduling period and is a direct quantitative expression of the energy flow deviation. This is the spatial influence factor, used to describe how the influence of nodal residuals on spatial location x decays with distance; summation is used. This approach iterates through all participating child nodes, weights the local residual contributions of each node using an attenuation kernel, and then superimposes them. The discrete residual sources are then integrated in the spatial domain to generate a continuous residual distribution field. Compared to traditional energy anomaly analysis methods based solely on time series or local node errors, this implementation transforms discrete errors into a continuous spatial mapping. This allows the virtual power plant to accurately capture anomalous energy flow characteristics in three-dimensional space, providing spatial constraints and directional references for subsequent coordination analysis and energy readjustment. This improves the ability to locate energy anomaly sources and enhances the overall operational stability.

[0104] Example 5

[0105] Please refer to Figure 3 Specifically: S4 includes S41;

[0106] S41. The energy flow hierarchy deviation index Neq and the spatial residual field index Rfc are fitted to perform operational coordination analysis on the virtual power plant, and a comprehensive operational coordination index Eop is constructed to represent the overall consistency of the virtual power plant's operational data, reflecting the coordination between the energy layer and the spatial layer of the virtual power plant, as detailed below:

[0107] ;

[0108] Where Eop(t) represents the comprehensive operation and scheduling index at time t, Ω represents the geographical distribution range of all nodes of the virtual power plant, and dx represents the infinitesimal element of the spatial integral in the integral function.

[0109] S4 also includes S42;

[0110] S42. Calculate the mean of the comprehensive operation scheduling index Eop when the residual spatial distribution is stable using statistical methods, and set a preset residual distribution critical threshold Cy based on the mean. Then, evaluate the residual distribution status with the real-time acquired comprehensive operation coordination index Eop. The specific evaluation scheme is as follows.

[0111] When the comprehensive operation coordination index Eop < the residual distribution critical threshold Cy, it indicates that the residual space distribution is stable and the energy flow of the virtual power plant can converge naturally. At this time, the monitoring frequency is increased by 50%, and the current state is recorded as a convergent state.

[0112] When the comprehensive operation coordination index Eop ≥ the residual distribution critical threshold Cy, it indicates that the residual spatial distribution is abnormal and the energy flow of the virtual power plant cannot converge naturally. At this time, energy re-regulation is performed.

[0113] In this embodiment, S41 fits the energy flow hierarchy deviation index Neq with the spatial residual field index Rfc to construct the comprehensive operational coordination index Eop, thereby quantifying the overall coordination of the virtual power plant between the energy layer and the spatial layer. This formula belongs to a typical spatiotemporal coupled integral evaluation model. Its theoretical basis comes from the field integral principle in physics and the L1 norm integral measure in mathematics, and is derived by combining the coupling consistency evaluation theory in systems science. The core idea of ​​the formula is to integrate the combined effects of the spatial residual field and energy hierarchy deviation of the virtual power plant through integral form, so that the dynamic relationship between the energy layer (Neq) and the spatial layer (Rfc) can be uniformly expressed in the form of a continuous field. Field integral principle: In electromagnetism and fluid mechanics, the intensity or energy distribution of a continuous field is often measured by integrating over the spatial region Ω. Here, ρ(x) represents the distribution density of a physical quantity in space. This principle is used to describe the globalization process of local quantities, that is, to transform the characteristics of local points into an overall metric through integration. In this scheme, the "field" of the virtual power plant corresponds to the spatial distribution function Rfc(x,t) of the energy residual, reflecting the deviation between the predicted and measured energy at time t and spatial location x. Therefore, the overall coordination state can be obtained by integrating |Rfc(x,t)| over the spatial range Ω. L1 norm integral metric: The classic L1 norm is used to characterize the overall deviation of a function over its domain, i.e. This scheme uses |Rfc(x,t)| instead of Rfc(x,t), i.e., taking the absolute value, so that the integral result reflects the total intensity of the deviation rather than the average value of sign cancellation, ensuring the physical consistency of the deviation measurement. The original field integral form can only reflect the spatial layer deviation distribution and cannot reflect the hierarchical conservation characteristics of the energy layer. Neq(t) is the energy layer deviation index, which describes the overall deviation of energy conservation between each level of the virtual power plant. It is embedded in the integral in the form of multiplication, indicating that when the energy layer deviation is large, the overall weight of the spatial residual increases accordingly, realizing the coupled adjustment of spatiotemporal deviation, and finally obtaining the comprehensive operation coordination index Eop. When the comprehensive operation coordination index Eop is lower than the preset residual distribution critical threshold Cy, it is judged that the energy flow can converge naturally, and the monitoring frequency is automatically increased to continuously track the stable state; when the comprehensive operation coordination index Eop is greater than or equal to the threshold Cy, the residual distribution is judged to be abnormal, triggering the energy re-regulation process, and adaptively adjusting the node energy allocation and energy storage strategy. This mechanism enables real-time assessment and dynamic feedback of the virtual power plant's operating status, allowing energy flow analysis to go beyond mere time-based trend judgments and reflect the coupling characteristics of spatial distribution. Compared to traditional energy dispatching methods based on static indicators, this implementation can achieve accurate identification and rapid response in the early stages of energy deviation, ensuring the stability and self-convergence of energy allocation in the virtual power plant under complex topology and dynamic load conditions, thus improving the overall operational coordination and control accuracy.

[0114] Example 6

[0115] Please refer to Figure 3 Specifically: S5 includes S51 and S52;

[0116] S51. When the energy flow of the virtual power plant cannot converge naturally, the system is fitted based on the acquired energy transmission dataset, the energy flow hierarchy deviation index Neq, and the comprehensive operation scheduling index Eop. This is used to adjust the energy flow of each node in a state where it cannot converge naturally, and to calculate the allocated energy flow Φadj, which represents the power correction relationship between nodes under the energy conservation deviation condition and describes the dynamic adjustment ratio of each node under the energy consistency deviation, as detailed below:

[0117] ;

[0118] Where Φadj(i,j,t) represents the required energy flow between the i-th energy input node in the upper layer and the j-th energy output node in the lower layer at time t, ε is a small constant to prevent the denominator from approaching zero, with a value of 0.001, and Neq + Denotes the positive part operator, where Neq(t) ≤ Py, and Neq... + =0, when Neq(t) > Py, Neq + =Neq(t)−Py;

[0119] S52. Based on the acquired distributed energy flow Φadj, the virtual power grid is adjusted by the following:

[0120] When the energy flow Φadj≤1, ​​the energy storage device is switched to charging mode and the energy output of 20% of the distributed nodes is limited. Then, iterative analysis is performed through S1.

[0121] When the energy flow Φadj > 1, the energy storage device is switched to discharge mode, and the distributed nodes are adjusted to the rated energy output. Then, iterative analysis is performed through S1.

[0122] In this embodiment, S51 performs fitting analysis on the previously acquired energy transfer dataset, energy flow hierarchy deviation index Neq, and comprehensive operation scheduling index Eop to calculate the allocated energy flow Φadj, clarify the power correction ratio between upper and lower level nodes, and form a precise energy redistribution command. In traditional PID control, the proportional term is used to correct the output according to the relative proportion of the current deviation. This formula borrows this idea, using the ratio of input power to output power... Considered as an instantaneous adjustment coefficient for energy flow, it describes the energy exchange intensity between upper and lower level nodes. Since the virtual power plant is a spatially distributed system, local power correction must consider overall coordination; therefore, it is introduced into the formula. As a comprehensive correction term between the global energy state and local bias, the energy flow Φadj is constructed after fitting. In this formula, The energy transfer ratio represents the relative intensity of energy flow between nodes and is the basis for the local correction factor. A global correction term is constructed to balance the relationship between local deviations and the global coordination state. In S52, device-level adjustment is automatically executed based on the numerical range of the allocated energy flow Φadj. When the allocated energy flow Φadj≤1, ​​the energy storage device is controlled to enter charging mode, and the energy output of the distributed nodes is limited to reduce the load. When the allocated energy flow Φadj>1, the energy storage device is controlled to enter discharging mode, and the node output is restored to the rated level, completing the rapid restoration of energy balance. Through the above dynamic correction mechanism, adaptive adjustment and closed-loop optimization of energy flow within the virtual power plant are realized, enabling energy to be redistributed among hierarchical nodes and tend to stabilize. This avoids the lag and energy imbalance problems caused by traditional virtual power plants that rely on manual intervention or fixed scheduling models. Ultimately, the virtual power plant has continuous energy self-balancing capability and efficient dynamic control performance in complex multi-node operating environments.

[0123] Example 7

[0124] Please refer to Figure 2 The IoT-based intelligent optimization system for virtual power plant operation data includes an IoT mapping module, an energy analysis module, a residual field analysis module, a convergence judgment module, and an energy flow balance module.

[0125] The IoT mapping module is used to deploy IoT sensors in the distributed node layer managed by the virtual power plant to collect energy operation data of each distributed node layer in real time, obtain energy transmission dataset and energy load dataset after preprocessing, and map them to the corresponding nodes of the virtual power plant.

[0126] The energy analysis module is used to perform energy flow state analysis on the virtual power plant based on the energy transmission dataset, generate an energy flow state assessment based on the analysis results, and trigger the residual field analysis module when there are anomalies at the energy flow level.

[0127] The residual field analysis module is used to extract the spatial distance matrix Dl(x,i) of the IoT topology modeling layer of the virtual power plant, and to perform spatial field residual analysis on the virtual power plant by fitting it with the energy load dataset.

[0128] The convergence judgment module is used to fit the energy flow state analysis results with the spatial field residual analysis results, perform operation coordination analysis on the virtual power plant, generate residual distribution state assessment based on the analysis results, and trigger the energy flow balance module when the energy flow cannot converge naturally.

[0129] The energy flow balancing module is used to calculate and allocate energy flow Φadj to adjust the energy flow of each node when it cannot converge naturally.

[0130] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent optimization of virtual power plant operation data based on the Internet of Things, characterized in that: Includes the following steps: S1. By deploying IoT sensors in the distributed node layer managed by the virtual power plant, the energy operation data of each distributed node layer is collected in real time. After preprocessing, the energy transmission dataset and energy load dataset are obtained and mapped to the corresponding nodes of the virtual power plant. S2. Perform energy flow status analysis on the virtual power plant based on the energy transmission dataset, generate an energy flow status assessment based on the analysis results, and trigger S3 when there are anomalies at the energy flow level. S3. Extract the spatial distance matrix Dl(x,i) of the IoT topology modeling layer of the virtual power plant, and perform spatial field residual analysis on the virtual power plant by fitting it with the energy load dataset. S4. Fit the energy flow state analysis results with the spatial field residual analysis results, perform operation coordination analysis on the virtual power plant, generate residual distribution state assessment based on the analysis results, and trigger S5 when the energy flow cannot converge naturally. S5. Adjust the energy flow of each node in a state where it cannot converge naturally by calculating and allocating the energy flow Φadj.

2. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 1, characterized in that: S1 includes S11; S11. By deploying IoT sensors in the distributed node layer managed by the virtual power plant, energy operation data of each distributed node layer is collected in real time. The IoT sensor includes a power sensor; The input power Pin received by the lower-level equipment nodes from the upper-level power grid is obtained in real time by power sensors installed on the input side of the substation bus at the distributed node layer. The output power Pou released by the power grid to the lower-level device node in the last time is obtained in real time by a power sensor placed between the node output end and the grid bus. The measured power Ps of each distributed node layer is collected in real time by power sensors installed at the output bus of the entire power flow path of each distributed node. S12. During the data acquisition process, the energy operation data is preprocessed by the data preprocessing unit built into the IoT sensor to obtain the energy transmission dataset and the energy load dataset. The preprocessing includes outlier removal and noise reduction; The outlier removal is based on an anomaly detection algorithm to automatically identify and remove abnormal signals caused by communication delays, environmental interference, and transient fluctuations. The denoising is performed on the energy operation data through multi-scale filtering and median smoothing. The energy transfer dataset includes input power Pin and output power Pou; The energy load dataset includes measured power Ps.

3. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 2, characterized in that: S1 also includes S13; S13. Based on the hierarchical topology and energy dispatch configuration table of the virtual power plant, establish a mapping matrix M between the distributed node layer and the virtual nodes. map (p, v), where p represents the physical acquisition point index and v represents the virtual power plant logical node. The distributed node layer is logically bound to the virtual power plant node, and the energy transmission dataset and energy load dataset are mapped to the corresponding nodes of the virtual power plant according to the logical binding relationship.

4. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 3, characterized in that: S2 includes S21; S21. After completing the data mapping, the energy transfer dataset is fitted, the energy flow state of the virtual power plant is analyzed, and the energy flow hierarchy deviation index Neq is constructed to describe the energy conservation state of the virtual power plant within a given scheduling cycle, reflecting the overall balance of energy transfer between different levels, as detailed below: ； Where Neq(t) represents the energy flow hierarchy deviation index at time t, N represents the number of child nodes participating in scheduling, and Pin i (t) represents the input power received by node i from the upper-level power grid at time t, Pin j (t) represents the output power released by node j to the lower-level node at time t.

5. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 4, characterized in that: S2 further includes S22; S22. Calculate the mean value of the energy flow level deviation index Neq when the energy flow consistency is compliant using statistical methods, and set a preset energy flow deviation judgment threshold Py based on the mean value. Then, evaluate the energy flow status with the real-time energy flow level deviation index Neq. The specific evaluation scheme is as follows. When the energy flow level deviation index Neq ≤ energy flow deviation judgment threshold Py, it means that the virtual power plant's energy flow consistency is compliant, and normal scheduling is maintained at this time; When the energy flow level deviation index Neq > the energy flow deviation judgment threshold Py, it indicates that there is an anomaly at the energy flow level of the virtual power plant, and spatial residual analysis is triggered.

6. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 5, characterized in that: S3 includes S31; S31. When the energy flow status assessment indicates an anomaly at the energy flow level, extract the spatial distance matrix Dl(x, i) of the virtual power plant IoT topology modeling layer. This matrix represents the spatial distance between any spatial location node x within the virtual power plant area and the i-th node in the virtual power plant. Fit this matrix to the energy load dataset to perform spatial field residual analysis on the virtual power plant. Construct the spatial residual field exponent Rfc, representing the continuous mapping of the virtual power plant's operating residuals in spatial distribution. This exponent is used to describe the intensity and diffusion of prediction errors within the virtual power plant in the spatial dimension, as detailed below: ； Where Rfc(x, t) represents the spatial residual field exponent at spatial location x at time t, N represents the number of child nodes participating in the scheduling, and Ps i (t) represents the measured power of node i at time t, Pp i (t) represents the predicted power of node i at time t, which is set by the user according to the needs of the virtual power plant. e represents the exponential function, and λ represents the spatial diffusion scale, which is set according to the node density.

7. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 6, characterized in that: S4 includes S41; S41. The energy flow hierarchy deviation index Neq and the spatial residual field index Rfc are fitted to perform operational coordination analysis on the virtual power plant, and a comprehensive operational coordination index Eop is constructed to represent the overall consistency of the virtual power plant's operational data, reflecting the coordination between the energy layer and the spatial layer of the virtual power plant, as detailed below: ； Where Eop(t) represents the comprehensive operation and scheduling index at time t, Ω represents the geographical distribution range of all nodes of the virtual power plant, and dx represents the infinitesimal element of the spatial integral in the integral function.

8. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 7, characterized in that: S4 also includes S42; S42. Calculate the mean of the comprehensive operation scheduling index Eop when the residual spatial distribution is stable using statistical methods, and set a preset residual distribution critical threshold Cy based on the mean. Then, evaluate the residual distribution status with the real-time acquired comprehensive operation coordination index Eop. The specific evaluation scheme is as follows. When the comprehensive operation coordination index Eop < the residual distribution critical threshold Cy, it indicates that the residual space distribution is stable and the energy flow of the virtual power plant can converge naturally. At this time, the monitoring frequency is increased by 50%, and the current state is recorded as a convergent state. When the comprehensive operation coordination index Eop ≥ the residual distribution critical threshold Cy, it indicates that the residual spatial distribution is abnormal and the energy flow of the virtual power plant cannot converge naturally. At this time, energy re-regulation is performed.

9. The intelligent optimization method for virtual power plant operation data based on the Internet of Things according to claim 8, characterized in that: S5 includes S51 and S52; S51. When the energy flow of the virtual power plant cannot converge naturally, the system is fitted based on the acquired energy transmission dataset, the energy flow hierarchy deviation index Neq, and the comprehensive operation scheduling index Eop. This is used to adjust the energy flow of each node in a state where it cannot converge naturally, and to calculate the allocated energy flow Φadj, which represents the power correction relationship between nodes under the energy conservation deviation condition and describes the dynamic adjustment ratio of each node under the energy consistency deviation, as detailed below: ； Where Φadj(i,j,t) represents the required energy flow between the i-th energy input node in the upper layer and the j-th energy output node in the lower layer at time t, ε is a small constant to prevent the denominator from approaching zero, with a value of 0.001, and Neq + Denotes the positive part operator, where Neq(t) ≤ Py, and Neq... + =0, when Neq(t) > Py, Neq + =Neq(t)−Py; S52. Based on the acquired distributed energy flow Φadj, the virtual power grid is adjusted by the following: When the energy flow Φadj≤1, ​​the energy storage device is switched to charging mode and the energy output of 20% of the distributed nodes is limited. Then, iterative analysis is performed through S1. When the energy flow Φadj > 1, the energy storage device is switched to discharge mode, and the distributed nodes are adjusted to the rated energy output. Then, iterative analysis is performed through S1.

10. An intelligent optimization system for virtual power plant operation data based on the Internet of Things, applied to the intelligent optimization method for virtual power plant operation data based on the Internet of Things as described in any one of claims 1-9, characterized in that: It includes an IoT mapping module, an energy analysis module, a residual field analysis module, a convergence judgment module, and an energy flow balance module; The IoT mapping module is used to deploy IoT sensors in the distributed node layer managed by the virtual power plant to collect energy operation data of each distributed node layer in real time, obtain energy transmission dataset and energy load dataset after preprocessing, and map them to the corresponding nodes of the virtual power plant. The energy analysis module is used to perform energy flow state analysis on the virtual power plant based on the energy transmission dataset, generate an energy flow state assessment based on the analysis results, and trigger the residual field analysis module when there are anomalies at the energy flow level. The residual field analysis module is used to extract the spatial distance matrix Dl(x,i) of the IoT topology modeling layer of the virtual power plant, and to perform spatial field residual analysis on the virtual power plant by fitting it with the energy load dataset. The convergence judgment module is used to fit the energy flow state analysis results with the spatial field residual analysis results, perform operation coordination analysis on the virtual power plant, generate residual distribution state assessment based on the analysis results, and trigger the energy flow balance module when the energy flow cannot converge naturally. The energy flow balancing module is used to calculate and allocate energy flow Φadj to adjust the energy flow of each node when it cannot converge naturally.

Images