Charging load demand response method based on virtual power plant technology

Abstract

The invention provides a charging load demand response method based on a virtual power plant technology. The charging load demand response method comprises the steps of generating a virtual load node corresponding to a to-be-regulated charging load and an initial response curve sequence in a preset response scheduling layer through a virtual power plant coordination system; and performing time sequence rearrangement on the initial response curve sequence according to a time axis, executing a coupling operation through a topological coupling engine, and outputting a coupling result. And verifying whether the coupling result is adaptive to a preset response time condition, generating a first to-be-optimized response curve by adopting a node replacement strategy when the coupling result is adaptive, and constructing a second to-be-optimized response curve when the coupling result is not adaptive. And executing optimization operation according to the processing mode adaptive to the to-be-optimized response curve, generating a target response curve, and replacing the to-be-regulated charging load with the target response curve to realize optimized regulation of the charging load. The method can improve the operation stability of the power grid, reduce the load response delay, optimize the space-time distribution of the charging load, and improve the adaptability and adjustment efficiency of the power grid to the large-scale charging load.

Description

Technical Field

[0001] This invention relates to the field of smart grid and electric vehicle charging technology, and more specifically, to a charging load demand response method based on virtual power plant technology. Background Technology

[0002] With the widespread adoption of electric vehicles, the impact of charging load on the power grid is becoming increasingly significant. Traditional power grids often face problems such as uneven load distribution, decreased grid stability, and slow response times when dealing with large-scale charging loads. While some existing technologies exist for regulating charging loads, such as using simple timing control or price-based incentive mechanisms to guide users to adjust charging times, most of these methods lack the ability for refined management and dynamic optimization of charging loads. For example, some methods can only achieve basic load reduction and cannot effectively integrate distributed energy resources or dynamically adjust according to the real-time state of the power grid.

[0003] In terms of technical principles, most existing charging load management methods are based on static load forecasting and fixed scheduling strategies. While these methods can alleviate grid pressure to some extent, they are limited when facing complex grid operating environments and dynamically changing load demands. For example, when grid frequency fluctuates, existing methods struggle to quickly adjust charging loads to maintain grid stability. Furthermore, existing technologies often neglect the spatiotemporal distribution characteristics of charging loads when optimizing them, resulting in unsatisfactory optimization effects.

[0004] In implementing the embodiments of the present invention, the prior art has at least the following problems or defects: First, the prior art cannot achieve dynamic and refined management of charging load, and it is difficult to make rapid adjustments according to the real-time status of the power grid; second, the prior art lacks consideration of the spatiotemporal distribution characteristics of charging load when optimizing charging load, resulting in unsatisfactory optimization effect; finally, the prior art lacks an effective dynamic response mechanism when dealing with power grid frequency fluctuations, and cannot quickly adjust the charging load to maintain the stable operation of the power grid. Summary of the Invention

[0005] This invention provides a charging load demand response method based on virtual power plant technology, comprising:

[0006] Step S1: Perform load matching operation at the preset response scheduling layer through the virtual power plant coordination system: generate virtual load nodes corresponding to the charging load to be regulated, and synchronously generate the initial response curve sequence;

[0007] Step S2: Rearrange the initial response curve sequence according to the time axis; drive the topology coupling engine to perform coupling operation on the rearranged sequence and output the topology coupling result;

[0008] Step S3: Verify whether the topology coupling result matches the preset response time conditions:

[0009] During adaptation, a node replacement strategy is used to process the initial response curve sequence within the topological coupling result, and the first response curve to be optimized is output.

[0010] When there is a mismatch, a second response curve to be optimized is constructed based on the virtual load nodes of adjacent sequences;

[0011] The verification process includes:

[0012] Detect the response time gaps between sequences in the topological coupling results;

[0013] When the response time interval is less than the preset response time threshold: perform node replacement operation in the preset response scheduling layer to generate the replacement response curve, and implement node blanking processing to obtain the first response curve to be optimized.

[0014] When the response time interval is not less than the preset response time threshold: a connection curve is inserted between the virtual load nodes of each two sequences to form a supplementary response curve, and node blanking processing is performed to obtain a second response curve to be optimized; the second response curve to be optimized is a smooth and continuous curve.

[0015] Step S4: Perform optimization operations according to the processing mode adapted to the response curve to be optimized to generate the target response curve, and replace the charging load to be regulated with the target response curve; the processing mode includes aggregation mode and curve reconstruction mode.

[0016] Further, step S1 includes:

[0017] Analyze the charging load to be regulated to extract discrete load datasets;

[0018] The virtual power plant coordination system performs data transformation operations at the preset response scheduling layer: creating virtual load nodes and initial response curve sequences based on discrete load datasets;

[0019] Establish a first mapping relationship between virtual load nodes and charging load nodes to be regulated, and establish a second mapping relationship between the initial response curve sequence and the charging load to be regulated;

[0020] Alternatively, a parameterized construction operation can be performed at the preset response scheduling layer through a virtual power plant coordination system: the initial response curve sequence and the corresponding virtual load nodes can be directly created based on the preset curve generation parameters.

[0021] Furthermore, the optimization operation performed according to the processing mode in step S4 includes:

[0022] Identify the processing mode of the response curve to be optimized;

[0023] When the processing mode is aggregation mode: perform a sequence merging operation to fuse all initial response curve sequences in the topological coupling result to form the target response curve;

[0024] When the processing mode is curve reconstruction mode:

[0025] Extract the first virtual load node and its response vector from the first initial response curve sequence in the topological coupling results;

[0026] Extract the second virtual load node and its response vector from the last initial response curve sequence;

[0027] Vector integration is performed through a virtual power plant coordination system to construct a target response curve based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector.

[0028] Furthermore, the construction of the target response curve includes:

[0029] A transition curve generation operation is performed in the preset response scheduling layer: a transition response curve is output based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector;

[0030] Perform curve fusion operation: fuse the transition response curve with the initial response curve sequence to form a precoupled response curve;

[0031] Perform underlying binding operations: Connect the initial response curve sequence with the pre-coupled response curve through the topology binding engine, and output the target response curve.

[0032] Furthermore, after generating the target response curve, the process also includes:

[0033] Perform point set acquisition operation: acquire the first ordered load point set of the initial response curve sequence and the second ordered load point set of the precoupled response curve;

[0034] Setting the projection reference operation: Set the response normal of the first ordered load point set as the projection reference direction;

[0035] Perform point set mapping operation: Map the first ordered load point set to the second ordered load point set along the projection reference direction to generate the third ordered load point set as the representation point set of the target response curve;

[0036] Perform data encapsulation operations: Generate discrete load data representation based on the third ordered load point set and transmit it to the user interface.

[0037] Furthermore, after generating the discrete load data representation, the process also includes:

[0038] Perform a morphology comparison operation: compare the first load morphology of the response curve to be optimized with the second load morphology of the initial response curve sequence;

[0039] When the morphological consistency condition is not met: trigger curve correction operation, adjust the target response curve according to the control rules until the load adaptation requirement is met;

[0040] The control rules include: inserting control points, shifting control points, or correcting response vectors.

[0041] Furthermore, load adaptation requirements include:

[0042] Perform smoothness calculation: Calculate the load smoothness index of the target response curve through the virtual power plant coordination system. This index is obtained based on the integral of the absolute value of the second derivative of load power with respect to time.

[0043] Perform delay monitoring operations: monitor power grid response delay data in real time;

[0044] When the load smoothness index exceeds the preset smoothness threshold and the response delay is lower than the preset delay limit: generate a load adaptation confirmation signal;

[0045] When both conditions are not met simultaneously: Initiate density adaptive operation, increase the distribution density of control points in the section where the rate of change of load curvature exceeds the critical value, until the load smoothness index reaches the preset smoothness threshold.

[0046] Furthermore, the preset response time threshold setting includes:

[0047] Perform frequency acquisition operation: continuously acquire power grid frequency fluctuation data;

[0048] Perform offset calculation: Calculate the absolute offset between the actual frequency and the reference frequency per minute;

[0049] When the absolute offset exceeds the reference frequency deviation: activate the threshold scaling engine, multiply the base response time threshold by the product of the scaling factor and the absolute offset divided by the reference frequency deviation, and add one to the value to dynamically update the response time threshold.

[0050] Perform parameter synchronization operation: Write the updated response time threshold into the response constraint parameter library of all virtual load nodes.

[0051] Furthermore, constructing the target response curve also includes:

[0052] Perform parameter initialization: Set the optimized parameter set, including the response direction angle and amplitude;

[0053] Perform iterative optimization operations: use the stochastic gradient descent algorithm to perform the operation cyclically.

[0054] The sum of squares of the differences between the actual load and the target load is used as the objective function value;

[0055] The optimization parameter set is updated by multiplying the learning rate by the negative direction of the objective function gradient;

[0056] When the change in the objective function value is less than the convergence threshold for three consecutive iterations: lock the response vector and output the final optimized parameter set.

[0057] Furthermore, the user interface operations include:

[0058] Perform parameter input operation: Receive load priority parameters input by the operator;

[0059] Perform a weight transformation operation: convert the parameters into a weight sequence for each load node;

[0060] Perform dynamic visualization operations: display the distribution status of the third ordered load point set in the time power coordinate system in real time, and overlay a color scale map of grid frequency fluctuation intensity;

[0061] When the load change per unit time exceeds the safety threshold: activate the protection mechanism, trigger an audible and visual alarm signal, and suspend the execution of the current scheduling command.

[0062] The embodiments of the present invention have at least the following beneficial effects:

[0063] 1. By generating virtual load nodes and initial response curve sequences through a virtual power plant coordination system, and performing time-series rearrangement and topology coupling operations, dynamic matching and optimized scheduling of charging loads are achieved. This method can flexibly adjust the allocation of charging loads according to the real-time state of the power grid, effectively solving the problem of uneven load distribution in traditional methods, improving the power grid's adaptability to large-scale charging loads, and ensuring the stability of power grid operation.

[0064] 2. When verifying whether the topology coupling results are compatible with the preset response time conditions, depending on the different response time gaps, a node replacement strategy or a method of constructing supplementary response curves is used to generate the response curve to be optimized. This process not only ensures the smoothness and continuity of the response curve, but also optimizes the spatiotemporal distribution characteristics of the charging load, solves the problem of unsatisfactory optimization effects in existing technologies, and improves the level of precision in charging load management.

[0065] 3. The optimization operation is performed on the response curve to be optimized through aggregation mode or curve reconstruction mode, and the target response curve is constructed by combining steps such as transition curve generation, curve fusion and underlying binding. This method can flexibly adjust the optimization strategy according to different processing modes, further improving the regulation accuracy and response speed of charging load, effectively solving the problem of the lack of dynamic response mechanism in the existing technology when dealing with grid frequency fluctuations, and enhancing the stability and reliability of the grid. Attached Figure Description

[0066] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein:

[0067] Figure 1 This is a flowchart illustrating a charging load demand response method based on virtual power plant technology, provided in an embodiment of the present invention. Detailed Implementation

[0068] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0069] like Figure 1 As shown, this application proposes a charging load demand response method based on virtual power plant technology, comprising:

[0070] Step S1: Perform load matching operation at the preset response scheduling layer through the virtual power plant coordination system: generate virtual load nodes corresponding to the charging load to be regulated, and synchronously generate the initial response curve sequence;

[0071] Step S2: Rearrange the initial response curve sequence according to the time axis; drive the topology coupling engine to perform coupling operation on the rearranged sequence and output the topology coupling result;

[0072] Step S3: Verify whether the topology coupling result matches the preset response time conditions:

[0073] During adaptation, a node replacement strategy is used to process the initial response curve sequence within the topological coupling result, and the first response curve to be optimized is output.

[0074] When there is a mismatch, a second response curve to be optimized is constructed based on the virtual load nodes of adjacent sequences;

[0075] The verification process includes:

[0076] Detect the response time gaps between sequences in the topological coupling results;

[0077] When the response time interval is less than the preset response time threshold: perform node replacement operation in the preset response scheduling layer to generate the replacement response curve, and implement node blanking processing to obtain the first response curve to be optimized.

[0078] When the response time interval is not less than the preset response time threshold: a connection curve is inserted between the virtual load nodes of each two sequences to form a supplementary response curve, and node blanking processing is performed to obtain a second response curve to be optimized; the second response curve to be optimized is a smooth and continuous curve.

[0079] Step S4: Perform optimization operations according to the processing mode adapted to the response curve to be optimized to generate the target response curve, and replace the charging load to be regulated with the target response curve; the processing mode includes aggregation mode and curve reconstruction mode.

[0080] This application constructs a dynamic response curve generation framework through the collaborative mechanism of a virtual power plant coordination system and a topology coupling engine, achieving spatiotemporal coupling optimization of charging load. By dynamically combining time-series rearrangement and node replacement strategies, it addresses the issues of unbalanced load distribution and response lag in traditional methods. Employing supplementary response curves and a dual-mode optimization mechanism, it simultaneously meets the requirements of grid frequency fluctuation adaptability and load curve smoothness, effectively improving the real-time performance of charging load management and grid stability.

[0081] Specifically, the virtual power plant coordination system performs load matching operations at the preset response scheduling layer to generate virtual load nodes corresponding to the charging loads to be regulated, and simultaneously generates an initial response curve sequence. This step establishes a mapping relationship between charging loads and virtual nodes, laying the foundation for subsequent flexible regulation.

[0082] Next, the initial response curve sequence is rearranged along the time axis, driving the topology coupling engine to perform coupling operations on the rearranged sequence and output the topology coupling result. The time-series rearrangement ensures the temporal continuity of the load sequence, while the topology coupling operation takes into account the spatial relationships between load nodes.

[0083] Then, it is verified whether the topology coupling result is compatible with the preset response time conditions. The verification process includes detecting the response time gap between each sequence in the topology coupling result. When the response time gap is less than the preset response time threshold, a node replacement operation is performed at the preset response scheduling layer to generate a replacement response curve, and node blanking processing is implemented to obtain the first response curve to be optimized. When the response time gap is not less than the preset response time threshold, a connection curve is inserted between the virtual load nodes of each pair of sequences to form a supplementary response curve, and node blanking processing is implemented to obtain the second response curve to be optimized.

[0084] Finally, optimization operations are performed according to the processing mode adapted to the response curve to be optimized to generate the target response curve, and the charging load to be regulated is replaced with the target response curve. The processing modes include aggregation mode and curve reconstruction mode, and the appropriate mode is selected for optimization according to the actual situation.

[0085] A virtual power plant coordination system refers to an intelligent management system that integrates distributed energy resources and charging loads. It can be implemented using a combination of distributed control architecture and centralized scheduling algorithms to coordinate the dynamic balance between charging loads and grid conditions in real time. An initial response curve sequence is a set of power demand curves reflecting the time-varying power demand of the charging load. This can be generated through parametric modeling of discrete load datasets or learning from historical load patterns, and is used to characterize the dynamic load characteristics of different charging nodes. A topology coupling engine is a computational module that handles the spatiotemporal relationships between multiple sequences. This can be implemented using graph theory algorithms combined with time series analysis techniques to optimize the temporal arrangement and spatial connectivity between sequences. A node replacement strategy is a mechanism for adjusting the load distribution between nodes while maintaining a constant total load. This can be implemented through heuristic algorithms or integer programming models to eliminate time gap conflicts and maintain grid stability. Supplementary response curves are continuous load curves formed by inserting connection curves. This can be implemented using cubic spline interpolation or Bézier curve generation algorithms to ensure smooth transitions and temporal continuity of the load curves. Aggregation mode refers to the processing method of merging multiple load sequences into a single curve, which can be achieved through weighted superposition or convolution operations, and is used to improve the execution efficiency of power grid dispatching commands. Curve reconstruction mode refers to the processing method of reconstructing load curves based on the characteristics of the first and last nodes, which can be achieved by vector synthesis and transition curve fusion technology, and is used to adapt to the dynamic optimization needs under power grid frequency fluctuation conditions.

[0086] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0087] First, the virtual power plant coordination system receives the charging load data to be regulated. The system performs load matching operations at the preset response scheduling layer, creating a corresponding virtual load node for each charging pile and generating an initial response curve sequence. For example, for a charging station with 100 charging piles, the system will create 100 virtual load nodes, each node corresponding to an initial response curve.

[0088] Next, the system rearranges these 100 initial response curves in chronological order. Assuming the charging period is from 18:00 to 22:00, the system will sort the curves from earliest to latest according to their start time. Then, the topology coupling engine performs a coupling operation on the rearranged sequence, taking into account the physical location relationships of the charging stations, and outputs the topology coupling result.

[0089] Furthermore, the system verifies whether the topology coupling results are compatible with the preset response time conditions. The system detects the response time gap between each sequence, assuming a preset response time threshold of 5 minutes. For sequence pairs with a gap of less than 5 minutes, the system performs a node replacement operation at the preset response scheduling layer, generates a replacement response curve, and obtains the first response curve to be optimized through node blanking processing. For sequence pairs with a gap of not less than 5 minutes, the system inserts a connection curve between the virtual load nodes of the two sequences to form a supplementary response curve, and similarly obtains the second response curve to be optimized through node blanking processing.

[0090] Finally, based on the characteristics of the response curve to be optimized, the system selects an appropriate processing mode to perform the optimization operation. For example, for periods with relatively gentle load changes, an aggregation mode may be used; while for periods with drastic load changes, a curve reconstruction mode may be used. The system generates the final target response curve and replaces the original charging load to be regulated with this curve.

[0091] This application further proposes a scheme including two construction methods: 1) parsing the charging load to be regulated to extract discrete load datasets; 2) performing data transformation operations at the preset response scheduling layer through a virtual power plant coordination system to create virtual load nodes and initial response curve sequences based on the discrete load datasets; 3) establishing a first mapping relationship between virtual load nodes and charging load nodes to be regulated, and establishing a second mapping relationship between the initial response curve sequences and the charging loads to be regulated; or 4) performing parameterized construction operations at the preset response scheduling layer through a virtual power plant coordination system to directly create initial response curve sequences and corresponding virtual load nodes based on preset curve generation parameters.

[0092] The data transformation operation generates virtual nodes and response curves from a discrete load dataset, suitable for scenarios with clear historical load data. The parameterized construction operation generates initial curves based on preset parameters, suitable for scenarios with missing data or requiring rapid construction. Both operations share the same mapping mechanism, ensuring the traceability of the correspondence between the generated virtual nodes and the original load nodes. The discrete load dataset includes timestamps, power values, and load type fields; the preset curve generation parameters include initial power, slope, and duration.

[0093] Specifically, when the system detects discrete data in the charging load to be regulated, it performs a data conversion operation: discrete load data is collected from the charging piles, standardized in format, and input into the virtual power plant coordination system; at the response scheduling layer, the system converts each discrete data point into a virtual load node and generates an initial response curve sequence in chronological order; when establishing the first mapping relationship, the unique identifier of the virtual node is bound to the physical address of the original load node. When the system detects missing data or requires a rapid response, it performs a parameterization construction operation: it calls template parameters from the preset curve generation parameter library to generate an initial response curve sequence with a preset power change trend; the position coordinates of each virtual load node are dynamically calculated based on the start time and power value in the parameters. The virtual nodes generated by both operations are bidirectionally associated with the original load nodes through mapping relationships, ensuring that subsequent optimization operations can be traced back to the original load nodes.

[0094] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0095] The virtual power plant coordination system receives charging load data to be regulated. The system first parses the charging load to be regulated, extracting a discrete load dataset containing timestamps and corresponding power values. For example, the system may extract a dataset in the form of [(t1,P1),(t2,P2),...,(tn,Pn)], where ti represents the time point and Pi represents the corresponding power value.

[0096] Next, the system performs data transformation operations at the preset response scheduling layer. Based on the extracted discrete load dataset, the system creates virtual load nodes. Each virtual load node corresponds to a power value at a specific time point. Simultaneously, the system generates an initial response curve sequence, which consists of curve segments connecting adjacent virtual load nodes.

[0097] Furthermore, the system establishes a first mapping relationship between virtual load nodes and charging load nodes to be regulated. This mapping relationship can be implemented using a hash table, where the key is the unique identifier of the charging load node to be regulated, and the value is the corresponding virtual load node object. Simultaneously, the system establishes a second mapping relationship between the initial response curve sequence and the charging load to be regulated. This mapping relationship can be implemented using an array, where each element of the array contains the identifier of the initial response curve sequence and the corresponding identifier of the charging load to be regulated.

[0098] Alternatively, the system can directly create an initial response curve sequence and corresponding virtual load nodes using preset curve generation parameters. These parameters may include curve type, such as linear, quadratic, or exponential, start point, end point, critical control points, etc. The system generates mathematical function expressions based on these parameters, then samples virtual load nodes on the time axis and connects these nodes to form the initial response curve sequence.

[0099] This application further proposes a processing mode for identifying the response curve to be optimized. When the processing mode is aggregation mode, a sequence merging operation is performed to fuse all initial response curve sequences in the topology coupling result to form the target response curve. When the processing mode is curve reconstruction mode, the first virtual load node and its response vector of the first initial response curve sequence in the topology coupling result are extracted, and the second virtual load node and its response vector of the last initial response curve sequence are extracted. Vector synthesis operation is performed through the virtual power plant coordination system to construct the target response curve based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector.

[0100] The sequence merging operation superimposes the time power point sets of multiple curves using a weighted average algorithm, with the weight coefficients automatically allocated according to the node capacity ratio. The vector synthesis operation uses a cubic spline interpolation algorithm to connect the response vectors of the first and last nodes, with the interpolation step size set to 0.1-second precision. The response vector contains two dimensions: power change rate and phase offset, and is calculated by orthogonal projection in the vector space.

[0101] Specifically, when the processing mode is identified as curve reconstruction mode, the power data of the starting node of the first sequence is converted into a complex domain expression through coordinate transformation, and the data of the ending node of the last sequence undergoes Fourier transform processing simultaneously. The response vectors of the two nodes are convolved in the frequency domain, and the calculation result is inversely transformed to generate a transition function. This transition function is linearly superimposed with the original sequence in the time domain. During the superposition process, a Hanning window function is used to suppress boundary effects, and the final target response curve exhibits continuous differentiability at the junction of the first and last nodes. For example, in a scenario where the power of the first node is 5kW and the power of the last node is 8kW, the transition function achieves a smooth power ramp-up within a 2-second time window, with a ramp-up rate of 1.5kW / s, which meets the constraint requirements of power change rate for grid frequency adjustment.

[0102] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0103] Identify the processing mode of the response curve to be optimized. Processing modes include aggregation mode and curve reconstruction mode.

[0104] When the processing mode is aggregation mode, a sequence merging operation is performed. Specifically, all initial response curve sequences in the topology coupling result are merged to form the target response curve. For example, a weighted average method can be used to sum the load values ​​of each initial response curve sequence at each time point to obtain the corresponding point value of the target response curve.

[0105] When the processing mode is curve reconstruction mode, the first virtual load node and its response vector of the first initial response curve sequence in the topology coupling result are extracted first. The first virtual load node contains load value and timestamp information, and the response vector represents the load change trend of the node.

[0106] Furthermore, the second virtual load node and its response vector are extracted from the last initial response curve sequence. The second virtual load node also contains load value and timestamp information, and its response vector represents the load change trend of the last sequence.

[0107] Therefore, vector synthesis operations are performed through the virtual power plant coordination system. Specifically, a target response curve is constructed based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector. For example, an interpolation algorithm can be used to generate an intermediate point between the first and second virtual load nodes, and the position of the intermediate point can be adjusted according to the direction and magnitude of the two response vectors to form a smooth and continuous target response curve.

[0108] This application further proposes to perform a transition curve generation operation at the preset response scheduling layer, outputting a transition response curve based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector; perform a curve fusion operation to fuse the transition response curve with the initial response curve sequence to form a pre-coupled response curve; and perform a bottom-level binding operation to connect the initial response curve sequence with the pre-coupled response curve through the topology binding engine to output the target response curve.

[0109] The transition curve generation operation determines the curvature change direction of the transition interval by using the response vectors of the first and last virtual load nodes. The response vectors include the power change rate and phase offset. The curve fusion operation uses a weighted superposition algorithm to mix the power values ​​of the transition curve and the initial sequence in the time overlap region according to a preset ratio. The bottom-level binding operation establishes the connection constraint relationship between sequences through the topology engine. The constraint conditions include time continuity tolerance and power deviation threshold.

[0110] Specifically, the transition curve generation operation extracts the first response vector at the end of the first sequence and the second response vector at the beginning of the last sequence, calculates the angle and magnitude difference between the vectors, and generates a transition curve segment with continuous derivative changes. The curve fusion operation sets a weighting coefficient that linearly changes from 0 to 1 within the time window when the transition curve overlaps with the initial sequence, achieving a smooth transition of power values. The bottom-level binding operation detects the connection points of the pre-coupled curves through the topology engine, automatically establishing a binding relationship when the time deviation is less than 5 milliseconds and the power difference is less than 2% of the rated load. Thus, the target response curve forms a continuous power change trajectory without abrupt changes between the first and last sequences, eliminating step fluctuations in the load curve and improving grid frequency stability.

[0111] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0112] A transition curve generation operation is performed at the preset response scheduling layer: a transition response curve is output based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector. Specifically, a Bézier curve interpolation method can be used, with the first and second virtual load nodes as endpoints and the first and second response vectors as control points, to generate a smooth transition response curve.

[0113] Perform curve fusion: fuse the transition response curve with the initial response curve sequence to form a precoupled response curve. Further, a weighted averaging method can be used to fuse the transition response curve and the initial response curve sequence according to time weights to obtain the precoupled response curve.

[0114] Perform low-level binding operations: Connect the initial response curve sequence with the pre-coupled response curves using a topology binding engine to output the target response curve. The topology binding engine can employ the minimum spanning tree algorithm from graph theory to connect the nodes of the initial response curve sequence with the nodes of the pre-coupled response curves, forming the complete target response curve.

[0115] This application further proposes that after generating the target response curve, the process includes: performing a point set acquisition operation: acquiring a first ordered load point set of the initial response curve sequence and a second ordered load point set of the pre-coupled response curve; setting a projection reference operation: setting the response normal of the first ordered load point set as the projection reference direction; performing a point set mapping operation: mapping the first ordered load point set to the second ordered load point set along the projection reference direction to generate a third ordered load point set as the representation point set of the target response curve; and performing a data encapsulation operation: generating discrete load data representation based on the third ordered load point set and transmitting it to the user interface.

[0116] The point set acquisition operation synchronously extracts the discrete load coordinate information of the initial response curve sequence and the pre-coupled response curve through the data acquisition module, forming an ordered point set with timestamps. The projection reference direction setting employs a vector normalization algorithm to convert the response normal of the first ordered load point set into a unit vector as the mapping reference axis. The point set mapping operation projects the first ordered load point set along the normal direction onto the plane containing the second ordered load point set using a coordinate transformation matrix, generating a third ordered load point set with a unified spatial reference. The data encapsulation operation compresses and stores the time-power coordinates of the third ordered load point set using a binary encoding format and transmits it to the visualization terminal via an encrypted channel.

[0117] Specifically, the load point set of the initial response curve sequence and the load point set of the pre-coupled response curve have spatial offsets in the time-power coordinate system. By setting a projection reference direction, a unified spatial reference is ensured for the mapping process of the two point sets. During the point set mapping process, an affine transformation algorithm is used to project each coordinate point of the first ordered load point set along the normal direction to the coordinate plane corresponding to the second ordered load point set, eliminating the spatial misalignment caused by curve fusion. In the generation of the third ordered load point set, the coordinate values ​​of each projected point are calculated using a weighted average algorithm, and the weighting coefficients are dynamically adjusted according to the time interval of the point sets. Discrete load data is represented by a piecewise linear interpolation method to convert the third ordered load point set into a time-continuous power change sequence, ensuring that the load curve displayed on the user interface is synchronized with the actual power grid response data. A cyclic redundancy check code is added during the data encapsulation process to prevent data packet loss or damage during transmission.

[0118] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0119] Perform point set acquisition operation: Acquire the first ordered load point set of the initial response curve sequence and the second ordered load point set of the pre-coupled response curve. Specifically, through the data acquisition module of the virtual power plant coordination system, 100 load points are collected from the initial response curve sequence and the pre-coupled response curve at a sampling interval of 1 second, forming the first ordered load point set and the second ordered load point set.

[0120] Setting the projection reference direction: The response normal of the first ordered load point set is set as the projection reference direction. Further, the vectors between adjacent points in the first ordered load point set are calculated, the average value of these vectors is taken as the response normal, and this average is normalized to serve as the projection reference direction.

[0121] Perform a point set mapping operation: Map the first ordered load point set to the second ordered load point set along the projection reference direction to generate a third ordered load point set as the representation point set of the target response curve. Thus, for each point in the first ordered load point set, find its projection point on the second ordered load point set along the projection reference direction to form a new third ordered load point set.

[0122] Perform data encapsulation operations: Generate discrete load data representation based on the third ordered load point set and transmit it to the user interface. Specifically, convert the third ordered load point set into a time-power corresponding data table, encapsulate it in JSON format, and transmit it to the user interface for display over the network.

[0123] This application further proposes that after generating discrete load data representation, it also includes: performing a morphology comparison operation, comparing the first load morphology of the response curve to be optimized with the second load morphology of the initial response curve sequence; when the morphology consistency condition is not met, triggering a curve correction operation, adjusting the target response curve according to the control rules until the load adaptation requirements are met; the control rules include the insertion of control points, the displacement of control points, or the correction of response vectors.

[0124] The morphology comparison operation calculates the similarity metric between the first and second load morphologies, using algorithms such as dynamic time warping or Euclidean distance, to determine whether they meet a preset morphology consistency threshold. The curve correction operation selects a specific adjustment method based on the control rules: the control point insertion operation adds a new control point at the curvature abrupt change position of the target response curve; the control point displacement operation translates the coordinates of an existing control point along the time or power axis; and the response vector correction operation adjusts the response direction or amplitude parameters of the load nodes.

[0125] Specifically, the morphology comparison operation first extracts the power-time distribution characteristics of the first load morphology and the sequence characteristics of the second load morphology, and generates a morphology difference index through a similarity measurement algorithm. When the difference index exceeds a preset threshold, a curve correction operation is triggered. For example, control points are inserted in sections where the rate of curvature change exceeds a critical value to increase the control accuracy of the curve in that section; the load nodes are adjusted to positions that match the distribution density of the initial response curve sequence through control point displacement operations; and the dynamic response direction of the load nodes is optimized through response vector correction operations to adapt to the trend of grid frequency fluctuations. The corrected target response curve is then subjected to morphology comparison again until the load morphology difference index is below the threshold and the load smoothness index meets the requirements, thereby ensuring that the optimized curve maintains the original load characteristics while adapting to the real-time state of the grid.

[0126] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0127] After generating discrete load data representations, a morphology comparison operation is performed. The first load morphology of the response curve to be optimized is compared with the second load morphology of the initial response curve sequence. If the morphology consistency condition is not met, a curve correction operation is triggered, adjusting the target response curve according to the control rules until the load adaptation requirements are met.

[0128] Specifically, morphological alignment is performed by calculating the morphological similarity between the response curve to be optimized and the initial response curve sequence. Morphological similarity can be quantified using characteristic parameters such as the slope and curvature of the curves. For example, multiple points can be sampled at equal intervals on two curves, the slope and curvature of these points can be calculated, and then the correlation coefficient between the two sets of data can be compared. If the correlation coefficient is lower than a preset threshold, the morphological consistency condition is considered not met.

[0129] Furthermore, when the morphological consistency condition is not met, a curve correction operation is triggered. The curve correction operation is performed according to predetermined control rules, including the insertion of control points, the displacement of control points, or the correction of response vectors.

[0130] The operation of inserting control points involves adding new control points to the target response curve. This increases the flexibility of the curve, allowing it to better match the shape of the initial response curve sequence. For example, new control points can be inserted in regions where the curve slope changes significantly, enabling finer adjustments to the curve shape.

[0131] Control point displacement is an operation that moves existing control points on the target response curve. This operation allows adjustment of the curve shape while keeping the number of control points constant. For example, a control point can be moved in a direction perpendicular to the curve to change the local curvature.

[0132] Response vector correction adjusts the overall shape of the target response curve. This operation is achieved by modifying the curve's response vector, thus altering the curve's overall trend. For example, the response vector can be rotated or scaled to make the overall shape of the target response curve more closely resemble the initial response curve sequence.

[0133] As a preferred implementation, the three control rules described above can be used in combination. First, a response vector correction operation is performed to adjust the overall trend of the curve. Then, a control point displacement operation is executed to fine-tune the local shape of the curve. Finally, if the shape consistency requirement is still not met, a control point insertion operation is performed for further refined adjustment.

[0134] These control operations are repeated until the load adaptation requirements are met. The determination of load adaptation requirements can be based on the morphological similarity index. When the morphological similarity reaches a preset threshold, the load adaptation requirements are considered to be met.

[0135] This application further proposes specific methods for achieving load adaptation requirements, including performing smoothness calculation operations, delay monitoring operations, condition judgment operations, and density adaptation operations. The smoothness calculation operation calculates the load smoothness index of the target response curve through a virtual power plant coordination system. This index is obtained based on the integral of the absolute value of the second derivative of load power with respect to time. The delay monitoring operation monitors grid response delay data in real time. The condition judgment operation determines whether to generate a load adaptation confirmation signal based on the comparison between the load smoothness index and a preset smoothness threshold, and the comparison between the response delay data and a preset delay upper limit. The density adaptation operation increases the distribution density of control points in sections where the load curvature change rate exceeds a critical value when both conditions are not simultaneously met, until the load smoothness index reaches the preset smoothness threshold.

[0136] The smoothness calculation operation quantifies the degree of curve fluctuation through integral operations. The second derivative reflects the curvature change, and the absolute value integral can eliminate the mutual cancellation of positive and negative fluctuations, accurately representing the overall smoothness. The delay monitoring operation uses a real-time data acquisition module to update delay data at a frequency of seconds. In the density adaptive operation, the distribution density of control points is positively correlated with the rate of curvature change, specifically achieved by dynamically inserting control points or adjusting the positions of existing control points. For example, when the rate of curvature change in a certain segment exceeds 0.5% / second, the spacing between control points is shortened from the default 30 seconds to 15 seconds.

[0137] Specifically, after the target response curve is generated, its load smoothness index is first calculated. If this index exceeds a preset threshold of 0.8 and the response delay is less than 200 milliseconds, the adaptation requirements are met. When either condition is not met, the system automatically analyzes the curvature change rate distribution of the load curve, identifies high-fluctuation segments with a curvature change rate exceeding 1.2% / second, and doubles the density of control points in these segments. New control points are inserted using a cubic spline interpolation algorithm to ensure the curve is continuous and differentiable. After three iterative adjustments, the system recalculates the smoothness index until it drops below 0.8. The entire process is completed automatically under the closed-loop control of the virtual power plant coordination system, without manual intervention, effectively balancing the requirements for curve smoothness and real-time response.

[0138] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0139] In a virtual power plant coordination system, load adaptation requirements are met through the following steps:

[0140] First, a smoothness calculation is performed. The load smoothness index of the target response curve is calculated using the virtual power plant coordination system. This index is obtained based on the integral of the absolute value of the second derivative of load power with respect to time. Specifically, the system uses a numerical integration method to discretely sample and accumulate the absolute values ​​of the second derivative of the target response curve over the time interval to obtain the smoothness index value.

[0141] Secondly, delay monitoring is performed. The system monitors grid response delay data in real time. By deploying delay detection devices at key nodes of the grid, the system collects the time difference between grid command issuance and load response, and transmits the data to the virtual power plant coordination system in real time.

[0142] Next, the system determines whether the load smoothness index exceeds a preset smoothing threshold and whether the response latency is below a preset latency upper limit. When both conditions are met, the system generates a load adaptation confirmation signal. This signal is used to notify the relevant modules that the load adjustment has achieved the expected results.

[0143] If the above two conditions are not simultaneously met, the system initiates density adaptive operation. Specifically, the system first calculates the rate of change of curvature of the target response curve and identifies segments where the rate of change of curvature exceeds a critical value. Then, the distribution density of control points is increased in these segments. The system adds new control points between the existing control points using an interpolation algorithm until the load smoothness index reaches a preset smoothing threshold.

[0144] This application further proposes a preset response time threshold setting including: performing frequency acquisition operation: continuously acquiring power grid frequency fluctuation data; performing offset calculation operation: calculating the absolute offset between the actual frequency and the reference frequency every minute; when the absolute offset exceeds the reference frequency deviation: activating the threshold scaling engine, multiplying the basic response time threshold by the product of the scaling factor and the absolute offset divided by the reference frequency deviation, and adding one to the result, dynamically updating the response time threshold; performing parameter synchronization operation: writing the updated response time threshold into the response constraint parameter library of all virtual load nodes.

[0145] The frequency acquisition operation collects frequency data in real time through sensors deployed at grid nodes, forming a timestamp-linked fluctuation dataset. The offset calculation operation uses a sliding window mechanism to calculate the absolute difference between the actual frequency and the reference frequency every minute, generating an offset sequence. The threshold scaling engine has a built-in scaling factor adjustment module that dynamically adjusts the scaling ratio of the basic response time threshold based on the magnitude of the offset deviation from the reference frequency. The parameter synchronization operation uses a distributed database update mechanism to ensure that all virtual load nodes complete constraint parameter synchronization within milliseconds after the threshold update.

[0146] Specifically, when grid frequency fluctuations exceed the reference frequency deviation, the threshold scaling engine is triggered. For example, if the reference frequency deviation is set to 0.2Hz, and the actual frequency offset in a certain minute is 0.3Hz, with a scaling factor set to 0.5, the updated response time threshold is calculated as: base threshold × (0.5 × (0.3 / 0.2) + 1). This calculation method ensures that the threshold increases linearly with the offset, while controlling the growth rate through the scaling factor. The updated threshold is written to the constraint library of the virtual load node in real time through parameter synchronization operations, ensuring that subsequent response time interval verification uses the latest threshold. This process, through a dynamic adjustment mechanism, ensures that the response time condition verification always matches the actual grid state, avoiding scheduling strategy failures caused by fixed thresholds.

[0147] As a preferred embodiment, the solution of this application is implemented as follows: Power grid frequency fluctuation data is continuously collected at a sampling rate of once per second using frequency sensors installed on the main power grid line. The collected data is transmitted to the frequency analysis module of the virtual power plant coordination system. The frequency analysis module calculates the absolute offset between the actual frequency of all sampling points and the reference frequency of 50Hz per minute. When this offset exceeds a preset reference frequency deviation of 0.2Hz, the threshold scaling engine is triggered. The scaling factor is set to 0.8, and the basic response time threshold is 30 seconds. The dynamically updated response time threshold is then calculated using the formula 30 × (0.8 × absolute offset / 0.2 + 1). The updated threshold is synchronously written into the response constraint parameter library of all virtual load nodes via an encrypted communication protocol. The parameter library is stored on the coordination node servers in each region using a distributed database architecture.

[0148] This application further proposes a preset response time threshold setting including: performing a frequency acquisition operation to continuously acquire grid frequency fluctuation data; performing an offset calculation operation to calculate the absolute offset between the actual frequency and the reference frequency per minute; when the absolute offset exceeds the reference frequency deviation, activating the threshold scaling engine, multiplying the basic response time threshold by the product of the scaling factor and the absolute offset divided by the reference frequency deviation, and adding one to the result, dynamically updating the response time threshold; and performing a parameter synchronization operation to write the updated response time threshold into the response constraint parameter library of all virtual load nodes.

[0149] The frequency acquisition operation uses sensors deployed at grid nodes to acquire frequency data at a sampling rate of at least twice per second. The data is then filtered to generate a frequency fluctuation curve. The offset calculation operation employs a sliding time window mechanism, extracting the average frequency data within a 60-second window as the actual frequency. This average is then compared to a preset reference frequency to calculate the absolute offset. The threshold scaling engine incorporates a linear scaling algorithm. When the detected offset exceeds 1.5 times the reference frequency deviation, a scaling factor is generated according to a formula. The scaling factor is composed of the ratio of the absolute offset to the reference frequency deviation plus a scaling coefficient, with a value ranging from 0.2 to 0.5. The parameter synchronization operation uses a distributed data broadcast protocol to synchronize the updated thresholds to the local storage units of all virtual load nodes within 50 milliseconds.

[0150] Specifically, when the grid frequency fluctuates, the frequency acquisition module captures frequency change data in real time, and generates a frequency fluctuation curve after filtering to eliminate noise interference. The offset calculation module extracts the average frequency every minute and compares it with the reference frequency. When the detected offset exceeds a preset deviation threshold, the threshold scaling engine is triggered. This engine dynamically adjusts the response time threshold based on the ratio of the current offset to the reference deviation. For example, when the reference frequency deviation is 0.5Hz and the detected offset is 0.8Hz, the scaling factor is 0.3, and the calculated scaling factor is (0.8 / 0.5)*0.3+1=1.48. The basic response time threshold of 10 seconds is updated to 14.8 seconds. The updated threshold is synchronized to all virtual load nodes through a high-speed communication network to ensure that the nodes use time constraint parameters that match the current grid state during subsequent scheduling. This dynamic adjustment mechanism enables the response time threshold to adapt to grid frequency fluctuations, avoiding premature or delayed responses caused by fixed thresholds, and improving the coordination efficiency between load scheduling and grid frequency.

[0151] As a preferred embodiment, the specific implementation of this application is as follows: In constructing the target response curve, the parameter set consisting of the response direction angle and amplitude is first initialized, with the response direction angle value limited to between 0 and 360 degrees. Subsequently, a stochastic gradient descent algorithm is used for iterative optimization. In each iteration, the mean square error between the current actual load curve and the target load curve is calculated as the objective function value. The learning rate is preset to a dynamic decay mode, with an initial value of 0.1 that decays to 90% of its original value every five iterations. During the parameter update phase, the optimized parameter set is corrected along the gradient descent direction of the objective function, and the grid response delay data is sampled in real time after each correction. When the change in the objective function value is less than 0.001 for three consecutive iterations, it is determined to be in a convergent state, and the optimization process is terminated. At this point, the final response direction angle and amplitude parameters are fixed and output.

[0152] This application further proposes user interface operations including: performing parameter input operations to receive load priority parameters input by the operator; performing weight conversion operations to convert the parameters into a weight sequence for each load node; performing dynamic visualization operations to display the distribution status of the third ordered load point set in the time-power coordinate system in real time, and superimposing a color-scale map of grid frequency fluctuation intensity; and activating a protection mechanism, triggering an audible and visual alarm signal, and suspending the execution of the current dispatching command when the load change per unit time exceeds the safety threshold.

[0153] The parameter input operation receives load priority parameters input by the operator through an interactive interface, and the parameters are converted into numerical data. The weight conversion operation maps the numerical parameters into a weight sequence for each load node based on a preset conversion algorithm. The weight sequence is used for subsequent load scheduling decisions. The dynamic visualization operation uses a graphics engine to dynamically render the third ordered load point set according to the time-power coordinate system, and simultaneously overlays a color gradient map of the power grid frequency fluctuation intensity. The color gradient of the color gradient map is positively correlated with the frequency fluctuation amplitude. The protection mechanism activation condition calculates the load power change rate per unit time in real time. When the change rate exceeds a preset safety threshold, it triggers the alarm signal generation module and the scheduling command interruption module.

[0154] Specifically, the load priority parameters input by the operator are converted into a weighted sequence, which is then used to dynamically adjust the display priority of each load node in the visualization interface. In the time-power coordinate system, the position of the third ordered load point set is determined by both the time axis and the power axis. The grid frequency fluctuation intensity color-coded map is displayed as a semi-transparent layer, allowing the operator to visually identify high-fluctuation areas. When the load power change rate monitoring module detects that the power change exceeds the safety threshold within a certain period, the alarm signal generation module immediately activates the audible and visual alarm device, while the dispatch command interruption module cuts off the currently executing dispatch command transmission channel. This process, through real-time monitoring and rapid response mechanisms, effectively prevents grid equipment overload caused by sudden load changes, ensuring system operational stability.

[0155] In a preferred embodiment, the solution of this application is implemented as follows: The user interface receives load priority parameters input by the operator through interactive controls, wherein the priority parameters include integer values ​​ranging from one to five, corresponding to five priority levels; the input priority parameters are converted into a normalized weight sequence, wherein the weight value of each virtual load node is generated by a linear mapping algorithm, and the mapping interval is from 0.1 to 1.0; the time power distribution map of the third ordered load point set is rendered in real time in the visualization engine, wherein the time axis uses minute-level scales, the power axis uses kilowatt units, and a grid frequency fluctuation intensity color scale map is superimposed, wherein the color mapping relationship of the color scale map is a gradient from blue to red, corresponding to a frequency offset of 0 Hz to 0.5 Hz; when the load power difference between adjacent time points is detected to exceed a preset safety threshold, the protection mechanism execution module is triggered, wherein the protection mechanism includes starting the red warning light on the operating console to flash, activating the buzzer alarm sound, and automatically cutting off the transmission channel of the current response curve replacement command.

[0156] Through the above technical solution, this application realizes the operator's real-time intervention capability on load priority, intuitively displays the correlation between load distribution and power grid frequency fluctuations through a dynamic visualization interface, and immediately blocks potentially risky operations when abnormal load fluctuations occur, effectively avoiding the problem of power grid equipment overload caused by load changes, and improving the safety control capability of the virtual power plant system under complex operating conditions.

[0157] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A charging load demand response method based on virtual power plant technology, characterized in that, include: Step S1: Perform load matching operation at the preset response scheduling layer through the virtual power plant coordination system: generate virtual load nodes corresponding to the charging load to be regulated, and synchronously generate the initial response curve sequence; Step S2: Rearrange the initial response curve sequence according to the time axis; The topology coupling engine performs coupling operations on the rearranged sequence and outputs the topology coupling results. Step S3: Verify whether the topology coupling result matches the preset response time conditions: During adaptation, a node replacement strategy is used to process the initial response curve sequence within the topological coupling result, and the first response curve to be optimized is output. When there is a mismatch, a second response curve to be optimized is constructed based on the virtual load nodes of adjacent sequences; The verification process includes: Detect the response time gaps between sequences in the topological coupling results; When the response time interval is less than the preset response time threshold: perform node replacement operation in the preset response scheduling layer to generate the replacement response curve, and implement node blanking processing to obtain the first response curve to be optimized. When the response time interval is not less than the preset response time threshold: a connection curve is inserted between the virtual load nodes of each two sequences to form a supplementary response curve, and node blanking processing is performed to obtain a second response curve to be optimized; the second response curve to be optimized is a smooth and continuous curve. Step S4: Perform optimization operations according to the processing mode adapted to the response curve to be optimized to generate the target response curve, and replace the charging load to be regulated with the target response curve; the processing mode includes aggregation mode and curve reconstruction mode.

2. The method according to claim 1, characterized in that, Step S1 includes: Analyze the charging load to be regulated to extract discrete load datasets; The virtual power plant coordination system performs data transformation operations at the preset response scheduling layer: creating virtual load nodes and initial response curve sequences based on discrete load datasets; Establish a first mapping relationship between virtual load nodes and charging load nodes to be regulated, and establish a second mapping relationship between the initial response curve sequence and the charging load to be regulated; Alternatively, a parameterized construction operation can be performed at the preset response scheduling layer through a virtual power plant coordination system: the initial response curve sequence and the corresponding virtual load nodes can be directly created based on the preset curve generation parameters.

3. The method according to claim 1, characterized in that, Step S4, which involves performing optimization operations according to the processing mode, includes: Identify the processing mode of the response curve to be optimized; When the processing mode is aggregation mode: perform a sequence merging operation to fuse all initial response curve sequences in the topological coupling result to form the target response curve; When the processing mode is curve reconstruction mode: Extract the first virtual load node and its response vector from the first initial response curve sequence in the topological coupling results; Extract the second virtual load node and its response vector from the last initial response curve sequence; Vector integration is performed through a virtual power plant coordination system to construct a target response curve based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector.

4. The method according to claim 3, characterized in that, The constructed target response curve includes: A transition curve generation operation is performed in the preset response scheduling layer: a transition response curve is output based on the first virtual load node, the first response vector, the second virtual load node, and the second response vector; Perform curve fusion operation: fuse the transition response curve with the initial response curve sequence to form a precoupled response curve; Perform underlying binding operations: Connect the initial response curve sequence with the pre-coupled response curve through the topology binding engine, and output the target response curve.

5. The method according to claim 4, characterized in that, After generating the target response curve, the following is also included: Perform point set acquisition operation: acquire the first ordered load point set of the initial response curve sequence and the second ordered load point set of the precoupled response curve; Setting the projection reference operation: Set the response normal of the first ordered load point set as the projection reference direction; Perform point set mapping operation: Map the first ordered load point set to the second ordered load point set along the projection reference direction to generate the third ordered load point set as the representation point set of the target response curve; Perform data encapsulation operations: Generate discrete load data representation based on the third ordered load point set and transmit it to the user interface.

6. The method according to claim 5, characterized in that, After generating the discrete load data representation, the following is also included: Perform a morphology comparison operation: compare the first load morphology of the response curve to be optimized with the second load morphology of the initial response curve sequence; When the morphological consistency condition is not met: trigger curve correction operation, adjust the target response curve according to the control rules until the load adaptation requirement is met; The control rules include: inserting control points, shifting control points, or correcting response vectors.

7. The method according to claim 6, characterized in that, The requirements for load adaptation include: Perform smoothness calculation: Calculate the load smoothness index of the target response curve through the virtual power plant coordination system. This index is obtained based on the integral of the absolute value of the second derivative of load power with respect to time. Perform delay monitoring operations: monitor power grid response delay data in real time; When the load smoothness index exceeds the preset smoothness threshold and the response delay is lower than the preset delay limit: generate a load adaptation confirmation signal; When both conditions are not met simultaneously: Initiate density adaptive operation, increase the distribution density of control points in the section where the rate of change of load curvature exceeds the critical value, until the load smoothness index reaches the preset smoothness threshold.

8. The method according to claim 1, characterized in that, The preset response time threshold settings include: Perform frequency acquisition operation: continuously acquire power grid frequency fluctuation data; Perform offset calculation: Calculate the absolute offset between the actual frequency and the reference frequency per minute; When the absolute offset exceeds the reference frequency deviation: activate the threshold scaling engine, multiply the base response time threshold by the product of the scaling factor and the absolute offset divided by the reference frequency deviation, and add one to the value to dynamically update the response time threshold. Perform parameter synchronization operation: Write the updated response time threshold into the response constraint parameter library of all virtual load nodes.

9. The method according to claim 3, characterized in that, Constructing the target response curve also includes: Perform parameter initialization: Set the optimized parameter set, including the response direction angle and amplitude; Perform iterative optimization operations: use the stochastic gradient descent algorithm to perform the operation cyclically. The sum of squares of the differences between the actual load and the target load is used as the objective function value; The optimization parameter set is updated by multiplying the learning rate by the negative direction of the objective function gradient; When the change in the objective function value is less than the convergence threshold for three consecutive iterations: lock the response vector and output the final optimized parameter set.

10. The method according to claim 5, characterized in that, User interface operations include: Perform parameter input operation: Receive load priority parameters input by the operator; Perform a weight transformation operation: convert the parameters into a weight sequence for each load node; Perform dynamic visualization operations: display the distribution status of the third ordered load point set in the time power coordinate system in real time, and overlay a color scale map of grid frequency fluctuation intensity; When the load change per unit time exceeds the safety threshold: activate the protection mechanism, trigger an audible and visual alarm signal, and suspend the execution of the current scheduling command.

Images