User-side power dispatch method based on power flow section constraint

Abstract

This invention discloses a user-side quantity and price bidding power dispatching method based on power flow section constraints, belonging to the field of power system operation and control technology. The method includes: obtaining the node section sensitivity matrix and the step-like quantity and price bidding curves of heterogeneous resources on the user side; using a high-dimensional topology manifold mapping algorithm to perform conformal aggregation of heterogeneous resources to obtain a standard bilateral resource dispatching model; using Clifford's algebra to construct a dispatching optimization model with the objective of maximizing total social welfare and including multiple vector-state power flow section constraints; solving the model to obtain the node marginal electricity price and preliminary dispatching plan, and implementing the dispatching. This invention addresses the spatial response topology distortion during the dimensionality reduction process of large-scale resource dispatching, and the difficulty in balancing clearing efficiency and global economic optimization under complex nonlinear security constraints of the power grid.

Description

Technical Field

[0001] This invention relates to the field of power system operation and control technology, and more specifically, to a user-side power dispatching method based on power flow section constraints and user-side quantity reporting and bidding. Background Technology

[0002] The new power system is characterized by a high proportion of new energy sources interacting with massive loads. The participation of heterogeneous resources on the user side (such as air conditioning clusters and distributed energy storage) in day-ahead and intraday market clearing has become an inevitable requirement for maintaining system power balance. In complex bilateral trading scenarios, the dispatch system must not only handle the massive economic allocation of resources with tiered quantity and price characteristics, but also strictly safeguard the safety bottom line of key power flow sections of the underlying physical power grid.

[0003] Current mainstream power dispatch architectures typically employ a hierarchical dimensionality reduction strategy. For massive numbers of nodes, traditional clustering algorithms based on Euclidean distance are often used to coarsely aggregate heterogeneous resources to alleviate the "curse of dimensionality." For complex power grid security constraints, DC optimal power flow (DC-OPF) is commonly used to meet the computational timeliness requirements of high-frequency clearing, or a first-order Taylor series expansion is used to locally linearize the nonlinear thermal limit boundary of AC apparent power.

[0004] However, the above-mentioned schemes have significant underlying technical limitations. On the one hand, traditional Euclidean spatial clustering is completely detached from the physical topology of the power grid, failing to accurately reflect the sensitivity and resistance differences of dispersed nodes to key transmission sections, and easily leading to topological distortion in the spatial response dimension. On the other hand, DC clearing forcibly removes the reactive power-voltage coupling relationship, while Taylor expansion suffers from uncontrollable high-order truncation errors at nonlinear physical boundaries. This lack of fidelity in the mathematical modeling process makes it easy for the scheduling results to trigger apparent power exceeding limits at transmission sections during actual execution, making it difficult to support a panoramic closed-loop scheduling that deeply integrates the hard physical constraints of the power grid with the economic value of bilateral market competition. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a user-side quantity reporting and bidding power dispatching method based on power flow section constraints. This method utilizes a high-dimensional topology manifold mapping algorithm to perform conformal dimensionality reduction and aggregation of massive heterogeneous resources, and combines Clifford's algebra to construct a social welfare maximization dispatching model with multiple vector state constraints. This addresses the issues of spatial response topology distortion during the dimensionality reduction process of large-scale resource dispatching, as well as the difficulty in balancing high-frequency clearing efficiency and global economic optimization due to complex nonlinear security constraints of the power grid.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The user-side power dispatching method based on power flow section constraints includes the following steps: obtaining the node section sensitivity matrix for the current dispatching period and the stepped power flow quotation curves for multiple user-side heterogeneous resources; based on the node section sensitivity matrix, using a high-dimensional topological manifold mapping algorithm to perform conformal aggregation of user-side heterogeneous resources to obtain a standard bilateral resource dispatching model; based on the standard bilateral resource dispatching model and the stepped power flow quotation curves, using Clifford's algebra to construct a dispatching optimization model with the objective function of maximizing total social welfare and including multiple vector state power flow section constraints; solving the dispatching optimization model to obtain the node marginal electricity price and preliminary power generation and consumption dispatching plan for the current dispatching period, and performing power dispatching on user-side heterogeneous resources based on the preliminary power generation and consumption dispatching plan.

[0008] In a preferred embodiment, the step of performing conformal aggregation of heterogeneous user-side resources based on the node cross-section sensitivity matrix and using a high-dimensional topological manifold mapping algorithm to obtain a standard bilateral resource scheduling model includes: extracting the power transfer distribution factor of user nodes corresponding to each heterogeneous user-side resource for multiple power flow cross-sections; and constructing a cross-section sensitivity tensor corresponding to each user node based on the power transfer distribution factor of the multiple power flow cross-sections.

[0009] In a preferred embodiment, the step of performing conformal aggregation of the multiple user-side heterogeneous resources based on the node cross-sectional sensitivity matrix using a high-dimensional topological manifold mapping algorithm to obtain a standard bilateral resource scheduling model further includes: mapping the cross-sectional sensitivity tensor corresponding to each user node to a non-Euclidean geometric manifold space to obtain the mapped coordinates of each user node in the non-Euclidean geometric manifold space; and calculating the curvature distance of different user nodes in the non-Euclidean geometric manifold space based on the mapped coordinates.

[0010] In a preferred embodiment, the step of performing conformal aggregation of the multiple user-side heterogeneous resources based on the node cross-section sensitivity matrix and using a high-dimensional topology manifold mapping algorithm to obtain a standard bilateral resource scheduling model further includes: if the curvature distance between multiple user nodes is less than or equal to a preset distance threshold, then the user-side heterogeneous resources corresponding to the multiple user nodes are aggregated into the same generalized energy storage node, so as to generate a standard bilateral resource scheduling model based on the multiple generalized energy storage nodes.

[0011] In a preferred embodiment, the parameters of the generalized energy storage node include an upper limit of equivalent power, a lower limit of equivalent power, and an equivalent energy capacity; the equivalent energy capacity is calculated based on the state of charge or operating temperature of the physical devices corresponding to the plurality of user nodes.

[0012] In a preferred embodiment, the step of constructing a scheduling optimization model with the objective function of maximizing total social welfare and including multiple vector state power flow section constraints using Clifford algebra includes: constructing a multiple vector space containing scalars, vectors, and double vectors under the Clifford algebra system; and in the multiple vector space, uniformly defining the active power, reactive power, and network topology parameters in the standard bilateral resource scheduling model as multiple vector variables.

[0013] In a preferred embodiment, the step of constructing a scheduling optimization model with the objective function of maximizing total social welfare and including multi-vector power flow section constraints using Clifford algebra further includes: extracting the apparent power thermal limit boundary based on the multi-vector variables; and linearizing the apparent power thermal limit boundary using algebraic rotation operations in the multi-vector space to generate the multi-vector power flow section constraints.

[0014] In a preferred embodiment, the objective function is constructed as follows: the difference between the total utility benefits on the user side and the total production costs of the generator set is calculated as the total social welfare; wherein the electricity consumption on the user side is limited by the declaration range of the tiered electricity quotation curve.

[0015] In a preferred embodiment, after solving the scheduling optimization model to obtain the node marginal electricity price and preliminary power generation and consumption scheduling plan for the current scheduling period, the method further includes: inputting the preliminary power generation and consumption scheduling plan as initial conditions into the reliability unit combination verification model for verification; if a physical limit violation is detected at the target power flow section, the physical limit of the target power flow section is locked, and the scheduling optimization model is re-solved until no physical limit violation is detected at the power flow section, and the target power dispatch instruction is obtained.

[0016] In a preferred embodiment, obtaining the node marginal price for the current scheduling period includes: when the scheduling optimization model detects that any of the multi-vector power flow section constraints have reached their limits, generating a shadow price corresponding to the power flow section constraint; and determining the node marginal price corresponding to each user-side heterogeneous resource based on the system-wide power balance marginal price and the shadow prices corresponding to each power flow section.

[0017] The technical effects and advantages of the user-side power dispatching method based on power flow section constraints in this invention are as follows:

[0018] This invention deeply integrates the underlying physical security boundary of the power grid with the economic demands of a bilateral market by acquiring the node section sensitivity matrix and the stepped reporting and pricing curves. It utilizes a high-dimensional topology manifold mapping algorithm to perform conformal aggregation of massive, dispersed heterogeneous user-side resources, significantly reducing the global scheduling computation dimensionality to overcome the "curse of dimensionality" while preserving the physical sensitivity response characteristics of underlying load nodes to key power flow sections of the power grid, thus solving the topology distortion problem caused by traditional clustering. Furthermore, it introduces Clifford's algebra to construct a scheduling optimization model with multiple vector-state power flow constraints, overcoming the pain point of difficulty in balancing computational efficiency with complex nonlinear power flow extreme constraints. Under the hard constraint of ensuring the thermal stability and security of the power grid, it maximizes total social welfare. Finally, by solving for accurate node marginal prices, it transforms rigid physical grid congestion into flexible economic price signals, efficiently and reliably guiding heterogeneous user-side resources to participate in system-level power dispatch. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the user-side power dispatching method based on power flow section constraints provided in an embodiment of the present invention.

[0020] Figure 2 This is a simulation diagram of a five-segment stepped quantity quotation curve provided in an embodiment of the present invention;

[0021] Figure 3 The simulation results of high-dimensional topological manifold dimensionality reduction and conformal aggregation provided in the embodiments of the present invention are shown in the figure.

[0022] Figure 4 The simulation comparison diagram of apparent power limit linearization provided for embodiments of the present invention;

[0023] Figure 5 The system load optimization and LMP evolution diagram provided for embodiments of the present invention. Detailed Implementation

[0024] 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1, Figure 1 The present invention provides a user-side power dispatching method based on power flow section constraints, comprising the following steps:

[0026] Step S1: Obtain the node section sensitivity matrix for the current scheduling period, as well as the stepped quantity and price quotation curves for multiple user-side heterogeneous resources.

[0027] In this embodiment, the power dispatch execution terminal first obtains multi-source heterogeneous parameters from the underlying data base. Specifically, for the security-constrained operating state of the transmission network, this method interacts with the secure isolation data interface of the power grid energy management system (EMS) to obtain the node section sensitivity matrix, which is pre-calculated based on the power flow of the basic operating sections. This matrix essentially reflects the linear transfer distribution effect of the small increment of active power injection at each physical node in the power grid on the active power flow of the monitored key transmission corridors or transmission sections. Assume that the current power grid topology contains... One injection node and For each key constrained power flow section, the obtained nodal section sensitivity matrix is... It is characterized by the following mathematical structure:

[0028] (1)

[0029] in, Indicates the first When a unit of active power is injected into a power grid physical node, at the th The incremental distribution of active power flow transfer caused by key power flow sections. This represents the total number of effectively controlled network nodes within the current scheduling period. This represents the total number of critical power flow sections subject to over-limit monitoring during the current scheduling cycle.

[0030] Furthermore, for massive, dispersed, heterogeneous user-side resources, such resources specifically include, but are not limited to, central air conditioning clusters in industrial and commercial buildings, distributed electrochemical energy storage power stations, and virtual power plants (VPPs) represented by load aggregators. This method connects with the business lines of the power trading platform to retrieve in real time the tiered quantity and price quotation curves submitted by each resource entity based on its own energy demand and physical boundaries. These tiered quantity and price quotation curves are mapped in the underlying data structure of the power trading platform as a set of discrete, increasing or decreasing segmented quantity-price pairs. Taking a large central air conditioning temperature-controlled load cluster represented by a virtual power plant as an example, its typical five-segment quantity and price quotation curve submitted to the trading platform can be abstracted as a mathematical set containing power adjustment ranges and corresponding marginal bid prices:

[0031] (2)

[0032] in, Indicates the first A complete set of quotation curves for heterogeneous resources on the user side during the current scheduling period; This indicates the physical segment number of the tiered pricing, with the highest segment number being 5. and They represent the first The user-side heterogeneous resources in the first The lower and upper limits of the absolute response power that can be provided within each price segment, affected by its state of charge or temperature comfort. Indicates the first The user-side heterogeneous resources in the first The marginal dispatching bid price (usually in yuan / megawatt-hour) corresponds to the response power range.

[0033] To more intuitively demonstrate the underlying market interaction data structure, such as Figure 2 The figure shown is a simulation diagram of a five-segment tiered pricing curve for a large central air conditioning cluster represented by a virtual power plant in this embodiment. From... Figure 2 As can be seen, with the increasing depth of load reduction (or provision of reverse support power) on the user side, the sacrifice to indoor temperature comfort gradually increases; therefore, its marginal declared price... It exhibits a significant nonlinear, monotonically increasing ladder-like characteristic. This structured form of quantity-price alignment provides a discrete but definite economic optimization boundary for the subsequent solution of the overall social welfare objective function.

[0034] By accurately acquiring and structurally modeling the aforementioned multi-source heterogeneous parameters, the long-standing problems of information silos and time scale mismatches between underlying physical power grid operation boundary data (such as EMS power flow information) and demand-side market-based transaction data (such as trading platform quantity and price information) are effectively solved. This lays a high-fidelity, multi-dimensional data foundation for guiding massive heterogeneous resources to participate in global economic optimal scheduling under the premise of strictly ensuring that the multiple power flow sections of the power grid do not exceed the limits.

[0035] Step S2: Based on the node cross-section sensitivity matrix, use the high-dimensional topology manifold mapping algorithm to perform conformal aggregation of heterogeneous resources on the user side to obtain a standard bilateral resource scheduling model.

[0036] Specifically, the above aggregation process first includes: extracting the power transfer distribution factor (PTDF) of user nodes corresponding to heterogeneous resources on each user side for multiple power flow sections; and constructing a section sensitivity tensor corresponding to each user node based on the PTDF of the multiple power flow sections. In the node section sensitivity matrix obtained in the aforementioned steps, its core element is the power transfer distribution factor (PTDF). The physical meaning of the power transfer distribution factor characterizes the linear proportionality coefficient that causes the change in active power flow at a specific monitored transmission section in the power grid when a user node injects a unit of active power under a specific network topology. For the first... Each user node extracts its data on the network. The power transfer distribution factors of each constrained section form a column vector. To fully reflect the spatial coupling characteristics and topological constraints between multiple cross-sections, the column vector is upgraded to a second-order cross-section sensitivity tensor through an outer product operation. Its mathematical expression is as follows:

[0037] (3)

[0038] In the formula, Indicates the first The cross-sectional sensitivity tensor corresponding to each user node. Indicates the first A column vector of power transfer distribution factors for each user node. This represents the tensor outer product operation. This represents the transpose of the column vector. This tensor matrix correlates the isolated cross-sectional influence, mapping the combined blockage effect of the node's power fluctuations across multiple cross-sections of the entire network.

[0039] Furthermore, to address the physical distortion problem caused by traditional Euclidean distance clustering in complex power grid topologies, the cross-sectional sensitivity tensor corresponding to each user node is mapped to a non-Euclidean geometric manifold space, obtaining the mapped coordinates of each user node in the non-Euclidean geometric manifold space; the curvature distance of different user nodes in the non-Euclidean geometric manifold space is then calculated. Traditional Euclidean distance metrics assume the parameter space is isotropic and only calculate absolute straight-line deviations. However, in actual power grids, electrical distances are constrained by network topology and transmission capacity limits, exhibiting highly anisotropic nonlinear characteristics. If two nodes are adjacent in the parameter space, but their power injection influence on a specific congested cross-section is opposite in direction or significantly different, their joint regulation resistance in the real physical environment is extremely high. Therefore, this embodiment introduces Riemannian manifold theory, mapping the high-dimensional cross-sectional sensitivity tensor to a non-Euclidean geometric manifold space through a locally preserving projection mapping function. The specific mapping transformation formula is as follows:

[0040] (4)

[0041] In the formula, Indicates the first The mapped coordinates of each user node on a non-Euclidean geometric manifold space; This represents a high-dimensional topological conformal mapping function used to preserve local electrical constraints in the tensor space during nonlinear dimensionality reduction. This is the cross-sectional sensitivity tensor constructed above.

[0042] After coordinate mapping, the curvature distance between two nodes is calculated using geodesic integration. This process directly assigns the physical congestion properties of the sensitivity tensor to the spatial measure. The specific mathematical formula is as follows:

[0043] (5)

[0044] In order for the curvature distance to reflect the underlying physical constraints, the Riemannian metric tensor in the manifold space is directly defined jointly by the section sensitivity tensors around the mapped coordinates:

[0045] (6)

[0046] In the formula, Indicates the first The and the first The curvature distance of each user node in the non-Euclidean manifold space; Indicates the connection node With nodes The manifold geodesic trajectory, parameters ; This is a Riemannian metric tensor defined on the geodesic trajectory; It is the identity matrix; This refers to the network congestion penalty weighting coefficient. Coordinates in manifold space The set of local neighborhood nodes; For neighboring nodes The coordinates; For nodes The cross-sectional sensitivity tensor. By introducing By reshaping the metric tensor, the topological regions that are more prone to cross-sectional over-limiting have greater spatial curvature, thus reflecting the congestion resistance of the power grid topology more realistically and rigorously.

[0047] According to the calculation results, if the curvature distance between multiple user nodes is less than or equal to a preset distance threshold, the heterogeneous resources on the user side corresponding to the multiple user nodes are aggregated into the same generalized energy storage node, so as to generate the standard bilateral resource scheduling model based on the multiple generalized energy storage nodes.

[0048] This method uses a manifold clustering algorithm to aggregate heterogeneous resources in non-Euclidean space that have close curvature distances (i.e., highly homogeneous impact on the power flow of the entire network section and without congestion blockage) into clusters. Since the resources within the same cluster have extremely high homogeneity in electrical topology, their joint output will not cause internal contradictions in the allocation of power flow at the section. Therefore, they can be regarded as a single aggregate node participating in bilateral scheduling in the power grid topology.

[0049] Figure 3 This is a simulation result of dimensionality reduction and conformal aggregation of a high-dimensional topology manifold generated based on a real provincial power grid topology. The scattered points in the figure represent the mapped coordinates of tens of thousands of dispersed heterogeneous load nodes in the non-Euclidean geometric manifold space of the entire network. Under traditional Euclidean distance, these nodes may exhibit a chaotic distribution (as shown by the gray projection at the bottom of the scatter plot); however, after introducing a Riemannian metric tensor that incorporates the congestion resistance of the power grid section, the nodes, which were originally highly coupled in terms of electrical constraints, spontaneously cluster in curvature space to form five high-density feature clusters with clear boundaries (identified by different colors and envelope surfaces in the figure). This intuitively demonstrates that the manifold clustering algorithm proposed in this invention can perfectly preserve the common physical congestion characteristics of nodes in relation to key transmission sections while significantly compressing the dimensionality of control variables.

[0050] In this embodiment, the parameters of the generalized energy storage node include an upper limit of equivalent power, a lower limit of equivalent power, and an equivalent energy capacity; the equivalent energy capacity is calculated based on the state of charge or operating temperature of the physical devices corresponding to the multiple user nodes. Specifically, for underlying heterogeneous resources with different physical characteristics, this invention employs parallel parameter equivalence calculation logic:

[0051] On the one hand, for distributed electrochemical energy storage devices, their equivalent energy capacity is directly calculated dynamically based on the real-time state of charge (SOC) of the physical cells. The charge-discharge time-coupled difference equation and equivalent energy model are as follows:

[0052] (7)

[0053] (8)

[0054] In the formula, and They represent the first The state of charge of an energy storage device in the current time period and the next time period; Its grid-connected active power (assuming discharge is positive and charging is negative). The scheduling time step; This refers to the rated physical capacity of the energy storage device. The overall conversion efficiency of charging and discharging for energy storage devices; This is the equivalent energy state obtained through conversion; The minimum threshold for safe state of charge to ensure battery life.

[0055] On the other hand, for temperature-controlled loads such as central air conditioning clusters, which do not possess the physical cells of traditional batteries, it is necessary to utilize the Equivalent Thermal Parameter (ETP) model to dynamically convert their building thermodynamic equations into energy capacity parameters equivalent to a battery. The thermodynamic time-coupled difference equation for a single air conditioning unit is as follows:

[0056] (9)

[0057] Based on the indoor temperature comfort range, the equivalent energy capacity parameter conversion model for air conditioner virtual energy storage is as follows:

[0058] (10)

[0059] In the formula, and They represent the first The indoor temperature set by the air conditioner in the current and next time periods; Outdoor ambient temperature; and These represent the building's equivalent thermal resistance and equivalent heat capacity, respectively. The scheduling time step; This refers to the energy efficiency ratio of the air conditioner. This refers to the active power of the air conditioner during operation. This is the equivalent energy state obtained through conversion; The lower limit of the comfort temperature set for users. The external characteristic boundary of the generalized energy storage node can be obtained by algebraically summing the equivalent parameters of all heterogeneous devices within the cluster.

[0060] Furthermore, after completing the aforementioned high-dimensional manifold aggregation and parameter reduction, the standard bilateral resource scheduling model is generated. Mathematically, this model is abstracted as a joint operating baseline state encompassing source-side generators, demand-side generalized energy storage nodes, and grid topology branches. Specifically, this model reconstructs the originally massively dispersed physical devices into a more dimensional construct. A generalized energy storage node. For any generalized energy storage node in the standard bilateral resource scheduling model. The node load it injects into the grid topology is manifested as active power. With reactive power Furthermore, its active power output is strictly limited by the upper and lower limits of the equivalent power obtained from the aforementioned aggregation and conversion. These node-level active and reactive power injections, through network topology parameters (such as branch admittance and correlation matrix), directly drive the active and reactive power distribution of branches at each power flow section of the entire network, thereby providing a standardized physical source-end boundary and parameter basis for subsequently introducing higher-order algebra to construct global section constraints.

[0061] Through the above steps, the deep integration of high-dimensional topological manifold dimensionality reduction and generalized energy storage conversion significantly reduces the dimensionality of scheduling decision-making solutions and avoids the "curse of dimensionality," while preserving the physical sensitivity response of the massive heterogeneous loads at the underlying level to critical grid sections without loss. This effectively solves the computational explosion problem that easily occurs when directly scheduling massive distributed resources and the physical aggregation distortion caused by traditional clustering, thus constructing a standardized and efficient dimensionality reduction model for subsequent global economic optimization.

[0062] Step S3: Based on the standard bilateral resource scheduling model and the tiered quantity and price quotation curve, a scheduling optimization model with the objective function of maximizing total social welfare and including multiple vector state power flow section constraints is constructed using Clifford's algebra.

[0063] It should be noted that traditional optimal power flow calculations often face irreconcilable technical contradictions when dealing with grid security constraints: nonlinear models based on AC power flow are extremely complex and time-consuming, making it difficult to meet the high-frequency dispatch requirements of day-ahead or intraday power markets; while traditional DC power flow models, in pursuit of ultimate efficiency, are oversimplified, neglecting reactive power transmission and apparent power thermal limit boundaries, which can easily trigger significant grid cascading exceedances in scenarios with high penetration of renewable energy. The method provided in this embodiment effectively resolves this contradiction by introducing higher-order algebraic geometry theory.

[0064] In this embodiment, the objective function is constructed as follows: the difference between the total utility benefits on the user side and the total production costs of the generator sets is calculated as the total social welfare; wherein, the electricity consumption on the user side is limited by the declaration range of the tiered electricity quotation curve. Specifically, guided by maximizing the economic efficiency of the entire system's resource allocation, the following objective function for total social welfare based on mixed-integer linear programming is established:

[0065] (11)

[0066] In the formula, This represents the total social welfare target value for the current scheduling period; This represents the total number of generalized energy storage nodes generated in the aforementioned standard bilateral resource scheduling model; Indicates the first The total number of tiered price quotations for each generalized energy storage node; and They represent the first The node at the th The bid price for each tier and the actual optimized active power consumption in the bid; Indicates the scheduling time resolution; This indicates the total number of generator sets on the grid side; Indicates the first The total production cost of each generator set is linearized piecewise; and They represent the first The generator set was in the first The marginal generation cost and actual active power output of each segment are considered. Strict requirements are placed on the solution process of this model. The declaration must be within the corresponding tiered quotation curve range.

[0067] Furthermore, to overcome the nonlinear barrier in AC power flow calculation, a multi-vector space containing scalars, vectors, and double vectors is constructed under the Clifford algebraic framework. In this multi-vector space, active power, reactive power, and network topology parameters from the standard bilateral resource scheduling model are uniformly defined as multi-vector variables. This embodiment constructs a two-dimensional Clifford space. Establish an orthogonal vector basis and and a double vector basis characterizing planar rotation properties. (satisfy and In the aforementioned standard bilateral resource scheduling model, the node power injected by the generalized energy storage node ( and After coupling through network topology parameters, in the first... The active power of the branch circuit is formed on the constrained cross section. With reactive power Within this algebraic space, it is uniformly projected as multiple vector variables. :

[0068] (12)

[0069] In the formula, Indicates the first Multiple vector power of a monitored power flow section; Mapped to active power scalar components along the real number axis; The mapping is a double vector component characterizing the reactive power exchange characteristics.

[0070] Furthermore, based on the multiple vector variables, the apparent power thermal limit boundary is extracted; the apparent power thermal limit boundary is linearized using algebraic rotation operations in the multiple vector space to generate the multiple vector state power flow section constraint. In the conventional physical domain, the apparent power thermal limit boundary manifests as a nonlinear circular inequality constraint (i.e., To avoid the high-order truncation error caused by traditional Taylor expansion, this embodiment utilizes the rotor operation in Clifford algebra to directly transform the nonlinear boundary into a polyhedral hyperplane envelope in the multivector space. Specifically, the process involves constructing an algebraic rotor. The multiple vector power is calculated by performing a Clifford inner product with the basis after rotation via a spinor two-sided operation, thereby extracting its scalar part to generate a system of purely linear inequalities:

[0071] (13)

[0072] Its algebraic expansion is physically equivalent to the following multivariate linear constraint family:

[0073] (14)

[0074] In the formula, The operator indicates that only the scalar portion of the result of multiple vector operations is extracted; Representing algebraic spinor The converse operator; These are the initial basis vectors; For the first Discrete rotation angles of each projected tangent plane, according to Uniformly distributed values; This represents the total number of discrete tangent planes. For the first The maximum apparent power thermal stability extreme value of each tidal flow section.

[0075] To further demonstrate the advantages of this method in avoiding nonlinear extreme value errors, such as Figure 4 The figure shows a comparison of simulations of linearized apparent power limits for multi-vector power flow sections. The solid red arcs in the figure represent the nonlinear apparent power thermal stability boundary in the real AC physical domain. The traditional Taylor first-order expansion method (the dashed blue line in the figure) produces severe geometric truncation errors when the reference operating point deviates, causing some over-limit regions to be misjudged as safe regions. However, after employing the Clifford algebraic spinor operation of this invention, when the number of polygon faces... When set to 12 or 24 (green polygon envelope in the figure), the generated multivariate pure linear hyperplane is not only 100% enclosed within the absolute safety domain (i.e., inscribed polygon property, ensuring physical absolute safety), but the area error rate with the real nonlinear boundary can be controlled within 1.5%.

[0076] This step introduces Clifford's algebra technique and utilizes spinor projection of multiple vector spaces to reduce the originally highly challenging nonlinear apparent power limit to a purely linear combination of hyperplane constraints. This core operation maintains the rapid clearing capability of the purely linear optimization model while also ensuring the rigid physical safety of nonlinear reactive power occupancy and the apparent power thermal limit. It addresses the fundamental algorithmic challenge of traditional DC clearing failing to guarantee grid thermal stability and safety.

[0077] Step S4: Solve the scheduling optimization model to obtain the node marginal electricity price and preliminary power generation and consumption scheduling plan for the current scheduling period, and perform power scheduling on heterogeneous resources on the user side based on the preliminary power generation and consumption scheduling plan.

[0078] In this embodiment, the power dispatch execution terminal efficiently solves the aforementioned optimization model by calling a commercial large-scale mathematical programming solver (such as CPLEX or Gurobi), thereby outputting preliminary power generation and consumption dispatch plans for both the generating unit side and the user side. To ensure the absolute reliability of the power grid dispatch instructions in the actual physical environment, the preliminary power generation and consumption dispatch plan is input as an initial condition into the reliability unit commitment verification model for verification. If a physical limit violation is detected at the target power flow section, the physical limit of the target power flow section is locked, and the dispatch optimization model is resolved until no physical limit violation is detected at the power flow section, thus obtaining the target power dispatch instruction. Specifically, the reliability unit commitment (RUC) verification model is a mature application module for power system safety verification. In the closed-loop architecture of this method, the system first encapsulates the preliminary plan in a structured manner and inputs initial boundary data, including the generating unit start-stop state matrix (Boolean discrete variable) and the power injection vector of the generating and consuming nodes (continuous variable), into the RUC model. After receiving the data, the RUC model uses these parameters as constants to call the network-wide high-order AC power flow (AC-PF) equations for cross-sectional security scanning.

[0079] During the verification output and iteration interaction phase, if no limit exceedances occur across the entire network, the RUC model outputs a verification pass flag and terminates the iteration. If a physical limit exceedance is detected at the target power flow section, a verification failure flag is output, and a safety correction mechanism is initiated to forcibly wake up the spinning standby units in the system. More importantly, the program extracts the limit exceedance power deviation and the corresponding node sensitivity coefficient for the limit exceedance section, automatically generating a safety secant, i.e., a safety constraint secant plane (BendersCut). Mathematically, this secant plane is represented by an augmented linear inequality constraint, the specific form of which is:

[0080] (15)

[0081] In the formula, For nodes For cross-sections exceeding the limit The transition distribution factor; This refers to the difference in node power variables that the master model is required to adjust in the next iteration; This represents the safety recovery margin for the cross section. The system uses this inequality constraint as a feedback data stream, directly superimposed onto the next-level scheduling optimization model, driving the main solver to perform a new round of recalculation within the tightened feasible region until all safety checks pass, and finally issuing the target power dispatch command.

[0082] At the pricing mechanism and market guidance level, when the scheduling optimization model detects that any of the multi-vector power flow section constraints has reached its limit, a shadow price corresponding to that power flow section constraint is generated. Based on the system-wide power balance marginal price and the shadow prices corresponding to each power flow section, the nodal marginal price corresponding to each user-side heterogeneous resource is determined. Mathematically, the shadow price corresponds to the Lagrange dual multiplier of the constraint condition during optimization, quantifying the marginal contribution rate of each unit relaxation of the section limit to total social welfare. The mathematical solution space formula for the nodal marginal price (LMP) is as follows:

[0083] (16)

[0084] In the formula, Indicates the first The marginal electricity price of a generalized energy storage node; This represents the Lagrange multiplier (i.e., the system benchmark electricity price) corresponding to the global power balance equation constraint of the system. The total number of critical sections where blockage occurred; Power transfer distribution factor; That is, the cross-section The shadow price generated when the physical limit is reached. The penalty term with the summation sign accurately reflects the high unit replacement cost (spatial congestion cost) caused by the physical transmission bottleneck of the power grid.

[0085] To verify the actual execution performance of the scheduling architecture of this invention, such as Figure 5 The figure shows a comparison of the system-level load curve before and after optimization using this scheduling optimization model, as well as the evolution trend of some nodal marginal prices (LMP). The simulation curve evolution shows that during the grid's evening peak hours from 18:00 to 21:00 daily, when the system predicts that a key power flow section is about to exceed its limit, the cut plane triggered by the RUC verification model reduces the nodal marginal price in that receiving-end area. A sharp surge occurred (high price peaks are shown in the bar chart). Driven by this economic price signal, the aggregated generalized energy storage nodes automatically implemented peak shaving response (the concave part of the green curve in the chart), shifting some air conditioning loads and energy storage charging and discharging cycles to the low-price period in the early morning. This caused the maximum cross-sectional power flow load rate during the evening peak to smoothly drop from the edge of 102.3% to a safe 94.1%.

[0086] This step, through the deep integration of a closed-loop verification and iteration mechanism and dual pricing theory, transforms the rigid physical congestion and limit-breaking constraints in power grid dispatch into flexible economic price signals that characterize the degree of supply and demand scarcity at different spatial locations. This mechanism, without relying on traditional administrative orders for power rationing, spontaneously incentivizes massive amounts of heterogeneous user-side resources to perform peak shaving, valley filling, and spatial transfer through differentiated nodal marginal electricity prices, thus establishing a two-way closed-loop feedback mechanism between power grid physical security and the economic game of the electricity market.

[0087] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0088] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0089] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0090] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0091] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0092] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A user-side power dispatching method based on power flow section constraints, characterized in that, include: Obtain the node cross-section sensitivity matrix for the current scheduling period, as well as the stepped quantity and price quotation curves for multiple user-side heterogeneous resources; Based on the node cross-section sensitivity matrix, a high-dimensional topological manifold mapping algorithm is used to perform conformal aggregation of heterogeneous resources on the user side to obtain a standard bilateral resource scheduling model. The conformal aggregation includes: extracting the power transfer distribution factors of user nodes corresponding to each user-side heterogeneous resource for multiple power flow sections; performing an outer product operation on the column vectors formed by the power transfer distribution factors to generate a second-order section sensitivity tensor, which serves as the section sensitivity tensor for each user node; mapping the section sensitivity tensor corresponding to each user node to a non-Euclidean geometric manifold space to obtain the mapped coordinates of each user node in the non-Euclidean geometric manifold space; calculating the curvature distance between different user nodes in the non-Euclidean geometric manifold space based on the mapped coordinates; if the curvature distance between multiple user nodes is less than or equal to a preset distance threshold, then aggregating the user-side heterogeneous resources corresponding to multiple user nodes into the same generalized energy storage node, and generating the standard bilateral resource scheduling model based on multiple generalized energy storage nodes, system generator parameters, and grid topology branch parameters. Based on the standard bilateral resource scheduling model and the step-by-step quantity and price quotation curve, a scheduling optimization model with the objective function of maximizing total social welfare and containing multiple vector state power flow section constraints is constructed using Clifford algebra. Solve the aforementioned scheduling optimization model to obtain the node marginal electricity price and preliminary power generation and consumption scheduling plan for the current scheduling period, and perform power scheduling on heterogeneous resources on the user side based on the preliminary power generation and consumption scheduling plan.

2. The user-side power dispatching method based on power flow section constraints according to claim 1, characterized in that, The parameters of the generalized energy storage node include the upper limit of equivalent power, the lower limit of equivalent power, and the equivalent energy capacity. The equivalent energy capacity is calculated based on the state of charge or operating temperature of the physical devices corresponding to the multiple user nodes.

3. The user-side power dispatching method based on power flow section constraints according to claim 1, characterized in that, The scheduling optimization model constructed using Clifford's algebra, with the objective function of maximizing total social welfare and incorporating multiple vector-state power flow section constraints, includes: Under the Clifford algebra system, construct a multivector space containing scalars, vectors, and double vectors; In the multi-vector space, the active power, reactive power, and network topology parameters in the standard bilateral resource scheduling model are uniformly defined as multi-vector variables.

4. The user-side power dispatching method based on power flow section constraints according to claim 3, characterized in that, The method of constructing a scheduling optimization model using Clifford's algebra with the objective function of maximizing total social welfare and including multiple vector-state power flow section constraints also includes: Based on the aforementioned multiple vector variables, the apparent power thermal limit boundary is extracted; The apparent power thermal limit boundary is linearized by using algebraic rotation operations in the multi-vector space to generate the multi-vector power flow section constraint.

5. The user-side power dispatching method based on power flow section constraints according to claim 1, characterized in that, The objective function is constructed as follows: The difference between the total utility benefits on the user side and the total production costs of the generator set is calculated as the total social welfare. The electricity consumption on the user side is limited by the reporting range of the tiered reporting and pricing curve.

6. The user-side power dispatching method based on power flow section constraints according to claim 1, characterized in that, After solving the scheduling optimization model to obtain the node marginal electricity price and preliminary power generation and consumption scheduling plan for the current scheduling period, the process further includes: The preliminary power generation and consumption dispatch plan is input as an initial condition into the reliability unit combination verification model for verification. If a physical limit violation is detected at the target power flow section, the physical limit of the target power flow section is locked, and the scheduling optimization model is re-solved until no physical limit violation is detected at the power flow section, and the target power dispatch instruction is obtained.

7. The user-side power dispatching method based on power flow section constraints according to claim 1, characterized in that, The process of obtaining the node marginal electricity price for the current scheduling period includes: When the scheduling optimization model detects that any of the multi-vector power flow section constraints has reached its limit, it generates a shadow price corresponding to that power flow section constraint. Based on the system-wide power balance marginal price and the shadow price corresponding to each power flow section, the node marginal price corresponding to each user-side heterogeneous resource is determined.

Images