A virtual power plant and power grid collaborative bidding decision method based on an equivalent regulation model

Abstract

The present application relates to the technical fields of power system operation control and power market transaction, and discloses a virtual power plant and power grid collaborative bidding decision method based on an equivalent regulation model, comprising generating an equivalent active-reactive feasible region with cost ladder attributes; reducing the dimension of the virtual power plant as a whole to an equivalent virtual energy storage model subject to power and electricity double constraints, and aggregating to generate a dynamic equivalent energy storage capacity constraint; constructing a double-layer master-slave game collaborative bidding model; equivalently converting the double-layer master-slave game collaborative bidding model into a single-layer mixed integer linear programming model and solving it to obtain the optimal bid amount and bid price curve of the virtual power plant and the collaborative scheduling instructions of the internal resources. The present application realizes accurate decoupling and dimension reduction of the underlying complex physical constraints, greatly reduces the complexity of power grid clearing calculation while retaining the time-varying energy boundary and cost characteristics, and effectively supports the inter-period arbitrage and system safety regulation of the virtual power plant.

Description

Technical Field

[0001] This invention relates to the fields of power system operation control and power market trading technology, and more specifically, to a virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model. Background Technology

[0002] Against the backdrop of accelerated construction of new power systems and the deepening of electricity market mechanisms, virtual power plants, as key carriers for aggregating massive heterogeneous distributed energy resources, are playing an increasingly important role in alleviating grid supply and demand imbalances and improving system operational flexibility. Virtual power plants aggregate dispersed energy storage, micro gas turbines, distributed photovoltaics, and flexible loads through advanced communication technologies, forming controllable virtual power generation and consumption entities. These entities participate in the optimized dispatching of the external power grid and spot market competition as independent stakeholders. However, in the process of virtual power plants being implemented in actual engineering projects, achieving efficient and secure interaction between internal heterogeneous resources and the external main grid level still faces significant challenges.

[0003] Currently, research on virtual power plants' participation in grid collaborative dispatch and bidding strategies faces two major technical bottlenecks: First, existing dispatch systems often simply equate virtual power plants to conventional generating units lacking energy constraints, ignoring the strong dual coupling constraints of "power-electricity" caused by energy storage and flexible loads. This makes it easy for grid-issued bidding clearing commands to violate underlying physical boundaries during actual execution, leading to serious safety risks. Second, if the constraints and cost tiers of massive underlying heterogeneous resources are fully integrated into the centralized clearing model of the grid spot market, it will not only greatly increase the computational dimension of the main grid model, making the solution difficult, but also fail to reflect the economic game demand of virtual power plants as independent entities making strategic bids. Therefore, it is urgent to explore a cross-level collaborative control and bidding mechanism that balances physical operational safety, clearing computation efficiency, and market game characteristics.

[0004] In the prior art, patent CN121097781A discloses a low-carbon dispatching method for mobile energy storage systems considering distribution network security constraints. This invention discloses a method for spatiotemporally dispatching mobile energy storage clusters using a two-layer multi-agent deep reinforcement learning decision-making framework. Although this method monitors power network security through AC optimal power flow analysis, it has significant limitations in practical engineering applications: First, this method relies on deep reinforcement learning for decision-making, and reinforcement learning algorithms typically require massive amounts of training samples. Furthermore, its decision-making process lacks rigorous mathematical optimality guarantees and environmental universality, making it difficult to guarantee the stability of the strategy when facing complex electricity spot market environments. Second, this invention focuses on the spatiotemporal movement decisions of individual mobile energy storage systems and does not propose a dimensionality reduction mechanism that can refine massive heterogeneous resources into a standardized equivalent control model. This results in its inability to effectively alleviate the computational pressure on the main grid dispatch center during large-scale spot market clearing and also makes it difficult to accurately characterize the "power-energy" dual-coupling physical boundary constraints generated by resource aggregation.

[0005] For example, patent CN118801437A discloses a method and device for joint optimization of shared energy storage and virtual power plants in the spot market. This invention achieves multi-party interest coordination and joint optimization by constructing a multi-entity hybrid game model among shared energy storage, virtual power plants, and producers and sellers. However, this method has the following technical bottlenecks when participating in main grid bidding: On the one hand, this invention mainly focuses on the interest game relationship between various entities and still tends to the modeling idea of ​​"full access". That is, the operational constraints of each micro-entity need to be explicitly handled in the clearing process, which causes the dimension of the main grid clearing model to increase geometrically with the expansion of the scale of access resources, seriously affecting the solution efficiency of the spot market; on the other hand, this equivalent model ignores the cross-voltage level physical characteristics of virtual power plants in the process of mapping from the access-side distribution network to the main grid hub point, and fails to consider the impact of leakage reactance and loss of the main transformer on the boundary of regulation capacity, which leads to the risk of distortion of the support boundary reported to the grid dispatch center, making it difficult to achieve accurate inter-period arbitrage under the premise of ensuring the underlying physical security.

[0006] In summary, existing technologies either rely on full-scale positive game models where computational complexity increases exponentially with resource scale, or employ macroscopic equivalent methods that ignore physical losses across voltage levels and the time-series coupling characteristics of equipment power and electricity. These approaches fail to achieve a deep integration of underlying physical constraints, computational efficiency, and macroscopic bidding strategies in the complex market environment of large-scale distributed resource access. Therefore, there is an urgent need for a virtual power plant equivalent control and collaborative bidding decision-making method capable of reducing and extracting massive underlying microscopic constraints and inversely retrieving the real adjustment boundaries on the main grid side through physical mapping. Summary of the Invention

[0007] To address the problems in related technologies, this invention proposes a virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model, in order to overcome the aforementioned technical problems existing in the existing related technologies.

[0008] This method integrates underlying physical constraint mapping with upper-level market game mechanism. By implementing a cost-layering strategy and a two-stage geometric projection method, and constructing an equivalent virtual energy storage model subject to dual constraints of power and energy, it achieves dimensionality reduction and extraction of the physical characteristics of massive heterogeneous resources at the underlying level on the main grid side. Combined with a two-layer master-slave game collaborative clearing strategy, it enables safe and refined quantitative assessment of the real regulation boundary and bidding potential of virtual power plants at multiple time scales. This effectively avoids the risks of physical overruns and computational failures caused by ignoring energy constraints or excessively high model dimensions in cross-level collaborative regulation.

[0009] Therefore, the specific technical solution adopted by the present invention is as follows:

[0010] A virtual power plant and grid collaborative bidding decision-making method based on an equivalent control model includes:

[0011] By using a cost stratification strategy and a two-stage geometric projection method, the feasible domain of the distribution network inside the virtual power plant is mapped to the main grid side, generating an equivalent active-reactive feasible domain with cost ladder attributes. The two stages include the feasible domain characterization stage of the distribution network inside the virtual power plant and the cross-voltage level projection mapping stage.

[0012] Based on the equivalent active-reactive feasible domain, the virtual power plant is reduced in dimension to be equivalent to a virtual energy storage model subject to dual constraints of power and energy, and dynamic equivalent energy storage capacity constraints are generated. The dynamic equivalent energy storage capacity constraints include equivalent charging and discharging power constraints and time-varying energy upper and lower limit constraints.

[0013] To address the two-way clearing mechanism of the electricity spot market, the dynamic equivalent energy storage capacity constraint is introduced as the physical boundary condition for virtual power plants to participate in market bidding. This leads to the construction of a two-layer master-slave game collaborative bidding model with the maximization of the comprehensive net profit of virtual power plants as the upper layer and the minimization of the total electricity purchase and operation cost of the power grid system as the lower layer.

[0014] The two-layer master-slave game collaborative bidding model is equivalently transformed into a single-layer mixed integer linear programming model and solved to obtain the optimal quantity and price quotation curve of the virtual power plant and the collaborative scheduling instructions for internal resources.

[0015] The beneficial effects of this invention are as follows:

[0016] 1) This invention adopts a cost stratification strategy and a two-stage geometric projection method to characterize the equivalent active-reactive feasible region. Taking into account the losses of hub transformers, it safely maps the micro-physical boundary of the underlying distribution network to the high-voltage main grid side, effectively avoiding the physical limit risk caused by ignoring network losses in cross-voltage level dispatch.

[0017] 2) This invention addresses the challenges of dimensional explosion and temporal coupling caused by massive distributed resource access. It reduces the overall dimensionality of the virtual power plant to an equivalent virtual energy storage model constrained by both power and energy. This significantly reduces the computational complexity of the main grid dispatch center while accurately preserving the time-varying energy boundaries and regulation cost characteristics of the underlying resources, thus avoiding the overestimation of actual output due to neglecting energy dissipation.

[0018] 3) This invention constructs a two-layer master-slave game-based collaborative bidding mechanism for the electricity spot market, and combines the strong duality theorem and the Big M method to rigorously transform it into a single-layer mixed integer linear programming model that is easy to solve. Under the premise of meeting the requirements of fast calculation and global optimization in the spot market, it fully leverages the strategic game capabilities of virtual power plants as independent interest entities, and uses the marginal electricity price signal of nodes to guide flexible resources to conduct precise intertemporal arbitrage. This drives virtual power plants to spontaneously complete the peak shifting and valley filling of grid load while ensuring physical security, significantly improving the flexibility of the new power system operation and the economy of market transactions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram illustrating the principle of a virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to an embodiment of the present invention.

[0021] Figure 2 This is one of the schematic diagrams of the 18hVPP cost stratification feasible region generated by projecting a virtual power plant onto the main grid side according to an embodiment of the present invention;

[0022] Figure 3 This is the second schematic diagram of the 18hVPP cost stratification feasible domain generated by projecting a virtual power plant onto the main grid side according to an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram of the equivalent VES capacity feasible region generated by projecting a virtual power plant onto the main grid side according to an embodiment of the present invention;

[0024] Figure 5 This is one of the feasible domain characterization results of the virtual energy storage (VES) model according to an embodiment of the present invention;

[0025] Figure 6This is the second feasible domain characterization result of the virtual energy storage (VES) model according to the embodiments of the present invention;

[0026] Figure 7 This is a schematic diagram of the dynamic equivalent capacity feasible region of the virtual energy storage (VES) model according to an embodiment of the present invention;

[0027] Figure 8 This is a diagram illustrating the clearing power results and the daily power conservation of the power grid during peak shaving and valley filling in the spot market environment according to an embodiment of the present invention.

[0028] Figure 9 This is a diagram illustrating the clearing power results and the effect of adjusting power distribution during load peak shifting in a spot market environment, according to an embodiment of the present invention.

[0029] Figure 10 This is a comparison chart of the sensitivity analysis of different resource endowment characteristics on the collaborative bidding results and dynamic physical boundaries according to embodiments of the present invention;

[0030] Figure 11 This is a strategy pricing diagram for Virtual Power Plant 1 (VPP1) according to an embodiment of the present invention;

[0031] Figure 12 This is a dynamic state-of-charge diagram of virtual power plant 1 according to an embodiment of the present invention;

[0032] Figure 13 This is a diagram showing the internal resource response of virtual power plant 1 at 18:00 according to an embodiment of the present invention;

[0033] Figure 14 This is a strategy pricing diagram for Virtual Power Plant 2 (VPP2) according to an embodiment of the present invention;

[0034] Figure 15 This is a dynamic state-of-charge diagram of the virtual power plant 2 according to an embodiment of the present invention;

[0035] Figure 16 This is a diagram showing the internal resource response of virtual power plant 2 at 18:00 according to an embodiment of the present invention;

[0036] Figure 17 This is a strategy pricing diagram for Virtual Power Plant 3 (VPP3) according to an embodiment of the present invention;

[0037] Figure 18 This is a dynamic state-of-charge diagram of the virtual power plant 3 according to an embodiment of the present invention;

[0038] Figure 19 This is a diagram showing the internal resource response of virtual power plant 3 at 18:00 according to an embodiment of the present invention. Detailed Implementation

[0039] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0040] According to an embodiment of the present invention, a virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model is provided.

[0041] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to an embodiment of the present invention, a virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model is provided, comprising:

[0042] Step S1: Using a cost stratification strategy and a two-stage geometric projection method, the feasible region of the distribution network inside the virtual power plant is mapped to the main grid side to generate an equivalent active-reactive feasible region with cost ladder attributes. The two stages include the feasible region characterization stage of the distribution network inside the virtual power plant and the cross-voltage level projection mapping stage.

[0043] In this embodiment, the feasible region characterization of the VPP (Virtual Power Plant) based on the main transformer projection is used. The PQ feasible region (i.e., the equivalent active-reactive feasible region) of the VPP represents the maximum equivalent active and reactive power regulation range that it can present at the PCC (i.e., the point of common coupling) without violating the various physical security constraints of the internal distribution network. Since the regulation costs of various distributed resources differ in the real market, this invention adopts a cost stratification strategy and a two-stage geometric projection method to accurately characterize the equivalent PQ feasible region of the VPP for the 220kV hub transmission network under different bidding tiers.

[0044] Specifically, by utilizing a cost stratification strategy and a two-stage geometric projection method, the feasible region of the internal distribution network of the virtual power plant is mapped to the main grid side, generating an equivalent active-reactive feasible region with cost tier attributes, including:

[0045] 1) The first stage is the characterization of the feasible region of the 35kV internal distribution network (this first stage of feasible region characterization is carried out in the space of the 35kV internal distribution network). This stage is based on the linearized branch power flow equation and uses the maximization of the projection of the virtual power plant grid connection point power polygon in each search direction as the objective function to construct a boundary search model. Under the premise of satisfying the system node voltage safety threshold, the branch apparent capacity thermal stability limit and the physical constraints of various internal distributed heterogeneous resources, the boundary search model is iteratively solved using the vertex search algorithm (that is, multiple search direction vectors are preset and uniformly distributed on the active-reactive plane. Under each search direction, the boundary vertex with the largest projection in that direction is obtained by solving the boundary search model. By traversing all search directions, the initial active and reactive power boundary vertex set of the internal distribution network side is obtained.

[0046] Specifically, the objective function expression for the boundary search model is:

[0047] ;

[0048] In the formula, , These represent the equivalent active and reactive power at the PCC on the 35kV side, respectively. , These are the weighting coefficients of active power and reactive power in the projection direction during the k-th iteration search;

[0049] The expression for the constraints of the boundary search model is:

[0050] ;

[0051] ;

[0052] ;

[0053] ;

[0054] ;

[0055] ;

[0056] ;

[0057] ;

[0058] In the formula, P ih,t Q ih,t P represents the active power and reactive power flowing through branch ih during time period t. ji,t Q ji,tLet be the active power and reactive power flowing through the upstream branch ji during time period t, respectively, and let h be the downstream child node of node i. Let be the set of child nodes of node i, j be the upstream node of node i, and π(i) be the set of upstream nodes of node i. , For the equivalent injected active and reactive power of node i, U h,t U is the per-unit voltage value of node h during time period t. i,t Let U be the per-unit voltage value of node i during time period t. min U max These are the lower and upper safety limits for node voltages, respectively, R ih X ih These are the branch resistance and reactance, respectively. N is the number of sides when performing polygon linearization using a circular constraint. The apparent capacity thermal stability limit of the branch or grid connection point. Based on rigid load, For the distributed resource set of access node i, Contributing to the foundation of distributed resources, s r Let r be the maximum apparent power capacity of the distributed resource during time period t. For the index of the pricing ladder, Let r be the actual active power adjustment of the distributed resource r during time period t. , These refer to the upward and downward adjustment of resources, respectively. , These are the maximum upward and maximum downward adjustment power of the resource, respectively. For reactive power support, This is the reactive power regulation margin coefficient. Let r be the maximum apparent power capacity of the distributed resource r during time period t.

[0059] 2) The second stage is cross-voltage level projection mapping. Due to the large leakage reactance and copper loss of the 220kV / 35kV hub transformer, significant active and reactive power margins will be consumed under heavy load conditions. In order to accurately evaluate the actual support capacity of the VPP to the upstream transmission network, this invention adopts the geometric projection mapping method to project the set of vertices of the 35kV side polygon obtained in the first stage onto the 220kV coordinate system.

[0060] Specifically, this stage takes into account the consumption of regulation margin by the equivalent resistance and leakage reactance loss of the hub transformer under heavy load conditions. Using geometric projection mapping technology, the initial boundary vertex set obtained in the first stage is projected from the internal distribution network side to the high-voltage main grid side coordinate system to obtain the active and reactive power boundary vertex set of the high-voltage main grid side. Based on the active and reactive power boundary vertex set of the high-voltage main grid side, combined with the cost layering strategy and traversing different bidding ladders, the virtual power plant equivalent dynamic active-reactive feasible region with cost ladder attributes is accurately characterized and generated. The maximum projection interval of the boundary of the virtual power plant equivalent dynamic active-reactive feasible region on the active power coordinate axis is extracted, and the maximum projection interval is used as the charging and discharging power extreme value of the subsequent virtual energy storage model in the corresponding time period.

[0061] Specifically, the calculation process for the consumption of regulation margin by equivalent resistance and leakage reactance loss is as follows: based on the equivalent resistance and leakage reactance of the hub transformer, the active power loss and reactive power loss of the grid connection point during the transmission process across voltage levels are calculated respectively, and the loss is deducted from the active power and reactive power at the initial boundary vertex.

[0062] Specifically, the expression for projecting the initial set of active and reactive power boundary vertices from the internal distribution network side to the high-voltage main grid side coordinate system is as follows:

[0063] ;

[0064] ;

[0065] In the formula, , These are the equivalent active and reactive power at the 220kV side PCC, respectively, which are the VPP true power boundaries ultimately fed back to the grid dispatching layer. T X T These are the equivalent resistance and leakage reactance of a 220kV transformer, respectively.

[0066] Specifically, the implementation method for generating the equivalent dynamic active-reactive feasible region of a virtual power plant with cost ladder attributes is as follows: First, the virtual power plant is segmented according to the unit adjustment cost of each heterogeneous distributed resource within the virtual power plant to construct a multi-level bidding ladder; second, under each bidding ladder, the corresponding resource output combination is traversed, and the boundary search model and cross-voltage level projection mapping are repeatedly executed to obtain the sub-feasible region under the corresponding cost segment; finally, the sub-feasible regions under each cost segment are superimposed and aggregated to generate the equivalent dynamic active-reactive feasible region of a virtual power plant with cost ladder attributes.

[0067] This invention obtains the feasible region of VPP dynamic PQ with cost thermal ladder attributes by traversing all time periods of the day and continuously improving differentiated pricing. The maximum projection interval of this feasible region on the active power axis constitutes the extreme value of charging and discharging power of the subsequent VES (Virtual Energy Storage) model in time period t.

[0068] Step S2: Based on the equivalent active-reactive feasible domain, the virtual power plant is reduced in dimension to be equivalent to a virtual energy storage model subject to dual constraints of power and energy, and dynamic equivalent energy storage capacity constraints are generated. The dynamic equivalent energy storage capacity constraints include equivalent charging and discharging power constraints and time-varying energy upper and lower limit constraints.

[0069] Specifically, based on the equivalent active-reactive feasible region, the virtual power plant is dimensionality-reduced and equivalent to a virtual energy storage model subject to dual constraints of power and energy, and the dynamic equivalent energy storage capacity constraints are aggregated and generated, including:

[0070] 1) VPP equivalent energy storage capacity constraint aggregation: Based on the equivalent active-reactive feasible region characterized by the main transformer projection mapping, the absolute active power boundary of the virtual power plant on each spatial section is determined. The power extreme value in a single time period cannot accurately reflect the time coupling and energy conservation characteristics of the actual energy storage system and flexible load inside the VPP. Since the physical ramping constraints of each distributed resource have been internalized in the underlying response model, this invention reduces the overall dimension of the VPP to an equivalent VES model subject to dual constraints of power and energy at the grid collaborative scheduling layer.

[0071] The specific structure and control mechanism of the virtual energy storage model are as follows: it consists of equivalent charge and discharge power constraints and time-varying energy boundary constraints. Specifically, the equivalent charge and discharge power constraint limits the actual equivalent regulation power of the virtual power plant within any scheduling period, ensuring it remains strictly between its maximum equivalent discharge power and maximum equivalent charging power. The time-varying energy boundary constraint limits the current equivalent energy state of the virtual energy storage, requiring it to always be between the dynamically aggregated upper and lower energy limits. Furthermore, the current equivalent energy state is iteratively calculated by subtracting the product of the equivalent regulation power and the scheduling time step from the equivalent energy state of the previous period.

[0072] 2) In this virtual energy storage model constrained by both power and energy, the decrease in the overall grid-connected power of the VPP 220kV side is defined as the equivalent discharge power of the VES; the increase in grid-connected power is defined as the equivalent charging power of the VES. Simultaneously, the cumulative adjustment of the VPP from the reference operating point is defined as the equivalent energy state of the VES. Based on the equivalent charging and discharging power and the equivalent energy state, the feasible domain of the virtual energy storage is determined.

[0073] Specifically, the expression for the feasible domain of virtual energy storage is:

[0074] ;

[0075] ;

[0076] In the formula, This represents the feasible solution set of the equivalent regulation power and state of charge for virtual energy storage during time period t. For each variable in the state set to satisfy the physical constraints on its right-hand side, Let be the equivalent regulating power of the virtual power plant during time period t. This indicates that VES is discharging. E indicates charging. VPP,t This represents the equivalent energy level of the virtual energy storage at the end of time period t. For virtual energy storage feasible domain, The active power of the virtual power plant at the 220kV side during time period t is the baseline operating power on the 220kV side. , These are the upper and lower limits of the absolute active power boundary, respectively. , These are the upper limit (i.e., the maximum equivalent discharge power) and the lower limit (i.e., the maximum equivalent charging power) of the mapped virtual energy storage, respectively. , These are the upper and lower limits of the time-varying energy boundary for virtual energy storage. For scheduling time steps, E VPP,0 and E VPP,t The energy balance is defined within the complete scheduling cycle to ensure the trans-day sustainability of VPP regulation.

[0077] 3) Because the VPP aggregates diverse heterogeneous resources, its overall energy boundary is not a constant. To accurately define the upper and lower limits of the time-varying energy feasible region in the above model... This invention constructs a polyhedral vertex mapping model with the goal of maximizing the extreme value of equivalent energy. Boundary optimization is performed along the directions of maximizing and minimizing energy, and the physical energy constraints of each heterogeneous distributed resource within the VPP are mapped and aggregated upwards. The upper and lower limits of the time-varying energy feasible domain are solved to obtain the dynamic equivalent energy storage capacity constraint.

[0078] Specifically, the objective function expression for the polyhedron vertex mapping model is:

[0079] ;

[0080] In the formula, vector Let t be the feasible boundary of the time-varying energy of the virtual power plant during time period t, Tr be the transpose of the matrix, [*] be the matrix, and v be the direction vector used to search for the energy boundary. These are used to search for the upper and lower bounds of the electrical quantity, respectively.

[0081] Step S3: For the two-way clearing mechanism of the electricity spot market, the dynamic equivalent energy storage capacity constraint is introduced as the physical boundary condition for virtual power plants to participate in market bidding. A two-layer master-slave game collaborative bidding model is constructed with the maximization of the comprehensive net profit of virtual power plants as the upper layer and the minimization of the total electricity purchase and operation cost of the power grid system as the lower layer.

[0082] Specifically, the two-layer master-slave game collaborative bidding model is composed of an upper-layer virtual power plant bidding model and a lower-layer power grid clearing model.

[0083] 1) Upper-level virtual power plant pricing model (i.e., VPP pricing model):

[0084] As an independent stakeholder, the VPP (Virtual Private Enterprise) externally functions as a single VES (Virtual Elastic Entity) participating in market bidding, and internally coordinates and allocates massive amounts of heterogeneous flexible resources. The upper-level model uses maximizing the VPP's overall intraday net profit as its objective function, mathematically expressed as:

[0085] ;

[0086] In the formula, F vpp N represents the intraday net profit of the virtual power plant. vpp Where T is the number of virtual power plants in the system, and T is the scheduling period. Let n(v) be the marginal electricity price of the node where the virtual power plant is located, and n(v) be the node number of the power grid to which the virtual power plant v is connected. , These represent the equivalent power output (selling electricity) and the equivalent power output (buying electricity) of the virtual power plant, respectively, N. int This refers to the number of tiers for flexible resources within the virtual power plant. , These represent the actual call cost during the discharge phase and the actual call cost during the charging phase of the i-th resource segment, respectively. , These are the upward and downward power boundaries for the internal k-th stage resources, respectively.

[0087] Constraints:

[0088] Strategic pricing constraints are implemented to prevent unreasonable buy-low-sell-high price inversions and to limit the unlimited amplification of market momentum. VPP-submitted buy and sell quotes must meet the following monotonicity and market limit constraints:

[0089] ;

[0090] In the formula, , Let represent the lowered and higher bid prices submitted by the virtual power plant during time period t, respectively; v is the set of indices for virtual power plants within the system; and C... cap The upper limit of the price set for the spot market;

[0091] Internal resource output constraints:

[0092] ;

[0093] ;

[0094] In the formula, , These are the maximum upward and downward power limits for the internal k-th tier resources, respectively.

[0095] The equivalent polymer physical mapping constraint stipulates that the equivalent power of the VPP in external bids must be strictly equal to the physical sum of the response power of each internal microscopic flexible resource segment at 220kV.

[0096] ;

[0097] ;

[0098] In the formula, E vpp,v,t The intraday dynamic equivalent energy state of the virtual power plant during time period t;

[0099] Dynamic equivalent charge state constraints:

[0100] ;

[0101] ;

[0102] In the formula, , These are the upper and lower limits of the dynamic energy boundary affected by users' electricity consumption behavior, respectively. This is the scheduling time step.

[0103] 2) Lower-level power grid clearing model

[0104] The lower-level model is the spot market centralized clearing model of the power grid dispatch center. Based on the principles of overall network security and economy, the power grid dispatch center aims to minimize the total system electricity purchase and operation cost, expressed as:

[0105] ;

[0106] In the formula, F iso N represents the total power purchase and operating cost of the system. g N bus These represent the total number of conventional units and the total number of nodes, respectively. gP represents the marginal cost of power generation by the generating unit. dg,g,t For the real-time absolute output of conventional units, , These are the discharge power and charging power of the grid-side independent energy storage system, respectively. , These are the cost coefficients for discharge power and charging power of grid-side independent energy storage systems, respectively. , These are the virtual unloaded variable and the power curtailment relaxation variable of the node, respectively, C slack The high cost of punishment is set;

[0107] Constraints:

[0108] Node power balance constraints:

[0109] ;

[0110] In the formula, Let g be the original, pure injected power after removing scheduling resources from node n, and g be the index of a conventional thermal power unit. B is the collection of various equipment resources (including generating units, energy storage, virtual power plants, etc.) of access node n. n,l , These represent the nodal admittance and voltage phase angle, respectively. The colon ":" at the end of the formula indicates the Lagrange dual variable associated with this constraint, and the Lagrange dual multiplier corresponding to this equality constraint. This refers to the marginal electricity price at the clearing point of the spot market;

[0111] Line transmission capacity constraints:

[0112] ;

[0113] In the formula, H l,n It is the line power transfer distribution factor (PTDF). This represents the maximum thermal stability limit of line l. These are the Lagrange multipliers corresponding to the lower and upper limits of the line transmission capacity constraints, respectively. The colon ":" at the end of the formula indicates the Lagrange dual variable associated with the constraint. For the power grid dispatch center, it can only receive the maximum external characteristic boundary declared by the VPP. and ;

[0114] Clearing boundary constraints for conventional thermal power units and virtual power plants:

[0115] ;

[0116] ;

[0117] ;

[0118] In the formula, These are the Lagrange multipliers corresponding to the lower and upper limits of the output of conventional thermal power units, respectively. These are the Lagrange multipliers corresponding to the lower and upper limits of the charging power reduction constraints of the virtual power plant, respectively; for the power grid dispatch center, it can only receive the maximum external characteristic boundary reported by the VPP. and ;

[0119] Operational constraints of grid-side independent energy storage systems:

[0120] ;

[0121] ;

[0122] ;

[0123] In the formula, E ess,t This represents the equivalent energy state of a grid-side independent energy storage system during time period t. These represent the lower and upper limits of the energy capacity of grid-side independent energy storage systems, respectively. ess,t The Lagrange multipliers corresponding to the constraints of the independent energy storage power evolution equation. These are the Lagrange multipliers corresponding to the lower and upper bound constraints of the independent energy storage capacity boundary, respectively. , These refer to the charging efficiency and discharging efficiency of an independent energy storage system, respectively.

[0124] System equilibrium node phase angle and slack variable nonnegativity constraint:

[0125] ;

[0126] ;

[0127] In the formula, The voltage phase angle at the system's slack node. The Lagrange multiplier is the one corresponding to the phase angle equality constraint of the balancing node.

[0128] Step S4: Transform the two-layer master-slave game collaborative bidding model into a single-layer mixed integer linear programming model and solve it to obtain the optimal quantity and price quotation curve of the virtual power plant and the collaborative scheduling instructions of internal resources.

[0129] The process involves transforming the two-layer master-slave game-theoretic collaborative bidding model into a single-layer mixed-integer linear programming model and solving it. This yields the optimal quantity and price quotation curve for the virtual power plant and the collaborative scheduling instructions for internal resources, including:

[0130] 1) The stationarity conditions and complementary relaxation conditions of the lower-level power grid clearing model are extracted using the Lagrange multiplier method, and the stationarity conditions and complementary relaxation conditions are introduced as additional constraints into the upper-level virtual power plant bidding model.

[0131] Specifically, the Lagrangian function used to construct the lower-level power grid clearing model. Its mathematical expression is as follows:

[0132]

[0133] In the formula, This refers to the original variable set of each operating parameter in the lower-level clearing model. To transform the node power balance equality constraint function into a zero-valued form, The Lagrange multiplier corresponding to this equation constraint (its economic meaning is the marginal electricity price at the clearing point of the spot market). The system operates on inequality constraint functions (covering the aforementioned line transmission capacity constraints, clearing boundary constraints, and independent energy storage operation constraints) in the form of less than or equal to zero for the j-th system. The Lagrange multipliers (dual variables) corresponding to the inequality constraints, and satisfying >0.

[0134] Based on the above Lagrange function By analyzing the original variables at the lower level Taking the partial derivative and setting it to zero yields the stationarity condition; simultaneously, based on the fact that the product of the Lagrange multiplier and the corresponding inequality constraint is zero (i.e., ... The complementary relaxation condition is derived by setting the stability condition to 0. Finally, the above stability condition and complementary relaxation condition are introduced as additional constraints into the upper-level virtual power plant pricing model.

[0135] 2) Based on the strong duality theorem, the nonlinear revenue term in the objective function of the upper-level virtual power plant pricing model is equivalently transformed into a linear combination containing system parameters and dual multipliers by using dual objective transformation.

[0136] Specifically, based on the strong duality theorem, the value of the objective function of the lower-level original problem is set to be equal to the value of the objective function of its dual problem. The nonlinear revenue term in the objective function of the upper-level virtual power plant pricing model, which is composed of the product of the node marginal electricity price and the clearing power, is equivalently transformed into a linear combination containing system parameters and dual multipliers by using the dual objective transformation. The upper-level objective function is then updated with this linear combination.

[0137] 3) The Big M method is used to linearize the nonlinear complementary relaxation conditions, and the two-layer nonconvex model is equivalently transformed into a single-layer mixed integer linear programming model for global optimization.

[0138] Specifically, the Big M method is adopted, which introduces a binary auxiliary variable and a sufficiently large positive real number M to linearize the nonlinear complementary relaxation conditions and convert them into equivalent linear inequality constraints. Then, combined with the updated upper-level linear objective function, the two-layer nonconvex model is equivalently transformed into a single-layer mixed-integer linear programming model for global optimization.

[0139] 4) Based on the global optimization results, output the optimal quantity and price quotation curve of the virtual power plant and the coordinated scheduling instructions of internal resources.

[0140] Furthermore, in order to verify the technical effects of the above-mentioned technical solution of the present invention, the following comparative experiments are conducted for demonstration:

[0141] A 5-node power grid system was used for simulation verification. This system comprises 5 nodes and 6 branches, integrating 4 conventional thermal power units (denoted as DG1, DG2, DG3, and DG4), 1 grid-side independent energy storage system (ESS), and three virtual power plants (denoted as VPP1, VPP2, and VPP3, representing three independent virtual power plant entities aggregating different underlying distributed resources) connected via 220-35kV transformers. The key parameter settings for the 5-node system and the inverse optimization model are shown in Table 1 below.

[0142] Table 1 Key parameter settings for the 5-node system and inverse optimization model

[0143]

[0144] First, a two-stage geometric projection method is used to search for and reduce the dimensionality of the massive heterogeneous distributed resources within the virtual power plant at the underlying physical boundary. Figures 2 to 4As shown, taking a typical time period (e.g., 18:00) as an example, as the cost of underlying resource allocation increases, the equivalent active power regulation boundary of the virtual power plant on the main grid side exhibits a layer-by-layer outward expansion trend, with its active power operation range eventually expanding to cover a wide range from 54.5MW to 76MW. Simultaneously, due to the physical constraints of equivalent leakage reactance and resistive losses under heavy load on the hub transformer, its reactive power regulation range exhibits a significant nonlinear contraction characteristic at active power extremes (e.g., close to 76MW). This dynamic boundary with cost-thermal attributes serves as the input boundary for subsequent equivalent control models and game-theoretic bidding, accurately covering everything from basic response conditions under low-price incentives to deep regulation conditions approaching voltage and branch thermal stability limits under high-price incentives. Based on this equivalent input boundary, it is further substituted into the constructed two-layer master-slave game-theoretic collaborative bidding model. By equivalently transforming this two-layer non-convex model into a single-layer mixed-integer linear programming model for global optimization, the final output is the system-wide collaborative clearing result under the spot market environment and the optimal bidding strategy for each virtual power plant. This serves as the input boundary for subsequent equivalent control models and game-theoretic bidding, covering operating conditions from basic response under low-price incentives to deep regulation conditions approaching voltage and branch thermal stability limits under high-price incentives. Based on this equivalent input boundary, it is further substituted into the constructed two-layer master-slave game-theoretic collaborative bidding model. By equivalently transforming this two-layer non-convex model into a single-layer mixed-integer linear programming model for global optimization, the final output is the system-wide collaborative clearing result under the spot market environment and the optimal bidding strategy for each virtual power plant.

[0145] Figures 5-7 The evolution of the dynamic energy envelope and equivalent energy of the virtual energy storage model with the scheduling time sequence is presented. Combined with... Figure 5 and Figure 6 It can be seen that, within a specific time period, the boundaries between the equivalent upward and downward active power adjustments of the virtual power plant are precisely defined within the ranges of approximately 66.5~70.5MW and 71~78MW, respectively, intuitively reflecting the tiered extreme values ​​under different dispatch costs. For example... Figure 7As shown in the feasible region of the equivalent capacity, the dimensionality reduction equivalence exhibits extremely high physical fit during the global mapping process, transforming the complex timing constraints of the underlying equipment into intuitive upper and lower bounds of the power capacity. Under the continuous constraint of polyhedral vertex aggregation, the maximum available equivalent power capacity of the virtual power plant exhibits significant time-varying characteristics: its upper bound of equivalent capacity (Emax) gradually climbs during the afternoon period (approximately 13:00-16:00) to a peak of approximately 60 MWh, while the lower bound of capacity (Emin) drops to a valley of approximately -40 MWh during the midday period (approximately 12:00-13:00). The dimensionality reduction equivalence exhibits extremely high physical fit during the global mapping process, transforming the complex timing constraints of the underlying equipment into intuitive upper and lower bounds of the power capacity. Under the continuous constraint of polyhedral vertex aggregation, the maximum available equivalent power capacity of the virtual power plant exhibits significant time-varying characteristics, robustly converging within the physical extremes of internal rigid loads and the energy conservation of heterogeneous resources, avoiding the misjudgment of long-term regulation potential by traditional equivalent models.

[0146] Figures 8-9 It demonstrates the spatial relationship between two-layer collaborative scheduling instructions and operational trajectories under the two-way clearing mechanism in the spot market. Figure 8 It is known that during the evening peak hours (approximately 17:00-19:00), actual net demand exceeds the previous forecast, resulting in a real-time upward adjustment (extra generation) demand of up to approximately 50MW; while during the nighttime off-peak hours (approximately 22:00-24:00), actual net demand drops sharply, leading to a real-time downward adjustment (reduction) rigid demand of approximately -150MW. Figure 9 As shown in the power allocation diagram, this drastic total system regulation demand is jointly borne by the conventional thermal power units (DG1-DG4), the energy storage system (ESS), and the three virtual power plants (VPP1, VPP2, VPP3). Figure 9 As shown in the power allocation diagram, the total system regulation demand is jointly borne by the conventional thermal power units (DG1-DG4), the energy storage system (ESS), and the three virtual power plants (VPP1, VPP2, VPP3). Figures 8-9 It can be seen that the clearing power exhibits a mirror structure in its temporal distribution, closely following the fluctuations in the marginal electricity price at each node. This directly reflects the absolute guidance of market price signals on the charging and discharging behavior of virtual power plants. For example, during the low-price period in the early morning, virtual power plants increase their grid-connected power; while during the high-price period in the evening, their grid-connected power is significantly reduced, demonstrating strong inter-period arbitrage characteristics. This confirms that under the master-slave game mechanism, virtual power plants can spontaneously complete peak-shifting and valley-filling of grid load while ensuring the underlying physical balance.

[0147] Figure 10 This study presents a comparative sensitivity analysis of the impact of different resource endowment characteristics on collaborative bidding results and dynamic physical boundaries. Combined with... Figure 10Specific data reveals significant differences in the market bidding returns of virtual power plants under different resource configurations: VPP1, with its high proportion of energy storage and ample flexible adjustment capabilities, achieved a daily comprehensive return of RMB 37,797.77, giving it an absolute advantage in inter-period arbitrage; while VPP2 and VPP3, primarily constrained by energy depletion and narrow adjustment boundaries, saw their returns drastically compressed, amounting to only RMB 10,377.96 and RMB 12,599 respectively. With changes in the internal flexible resource allocation ratio, the active support capabilities of virtual power plants exhibit distinct differences. When entering a high-proportion energy storage scenario, the system's maximum available output and inter-period arbitrage space show a non-linear surge, with a significant expansion of the envelope area; while the response capability of load-biased virtual power plants is strictly constrained by energy depletion. This effectively quantifies the impact of the underlying heterogeneous resource structure on the overall market game-playing ability.

[0148] Figures 11-19 This displays the strategic bidding, dynamic equivalent state of charge (SOC), and internal resource response of each virtual power plant (denoted as VPP1, VPP2, and VPP3) under different resource endowments. Figures 11-19 Specific curves and data analysis show that at the market clearing level ( Figure 11 , 14 (17) Each virtual power plant can effectively track the nodal marginal price (LMP) signal. For example, during the peak period from 17:00 to 18:00 when the price surges to approximately 3000 yuan / MW, the discharge price is adjusted synchronously in a stepwise manner, reflecting the significant strategic pricing characteristics under the master-slave game mechanism; at the power tracking level ( Figure 12 , 15 The actual SOC trajectory (solid line) obtained from the aggregation of physical constraints strictly falls within the dynamic energy upper and lower bound envelope generated by the aggregation of physical constraints. For example, the actual SOC of VPP2 drops to an extreme value of approximately -40MWh around 12:00, operating close to the lower capacity limit without any physical overruns, further verifying the precise control capability of the dimensionality reduction equivalent model over the physical safety boundary. In addition, by analyzing the internal resource allocation during the typical high-price period at 18:00 (when the marginal electricity price at the node is as high as 3000 yuan / MW), Figure 13 , 16 (19) It can be seen that the virtual power plant can perform orderly scheduling based on the differentiated cost tiers of micro-equipment. For example, VPP1 prioritizes the activation of tiered resources with actual execution volumes of 0.3MW and 3.0MW according to cost from low to high. VPP2 and VPP3 also accurately call up low-cost resource segments with execution volumes of 1.8MW and 0.6MW respectively, thereby achieving precise connection between macro-bidding strategy and underlying physical response.

[0149] This invention improves the scientific rigor and security of virtual power plant (VPS) control decisions in electricity spot market trading and complex grid interaction scenarios. The proposed method for collaborative bidding between VPS and the power grid, based on an equivalent control model, overcomes the engineering drawbacks of dimensional explosion faced by traditional full-access models. It identifies and accurately quantifies the dynamic constraints of cross-voltage level network losses and instantaneous energy states on supporting power during long-term scheduling, generating a safe scheduling boundary and bidding clearing strategy that combines underlying physical accuracy with upper-level game theory economics.

[0150] According to another embodiment of the present invention, a virtual power plant and power grid collaborative bidding decision-making system based on an equivalent control model is also provided, comprising:

[0151] The feasible region characterization module is used to map the feasible region of the distribution network inside the virtual power plant to the main grid side using a cost stratification strategy and a two-stage geometric projection method, generating an equivalent active-reactive feasible region with cost ladder attributes. The two stages include the feasible region characterization stage of the distribution network inside the virtual power plant and the cross-voltage level projection mapping stage.

[0152] The dimensionality reduction and equivalence module is used to reduce the overall virtual power plant to a virtual energy storage model subject to dual constraints of power and energy based on the equivalent active-reactive feasible domain, and aggregate and generate dynamic equivalent energy storage capacity constraints, which include equivalent charging and discharging power constraints and time-varying energy upper and lower limit constraints.

[0153] A two-layer master-slave game collaborative bidding model construction module is used to address the two-way clearing mechanism of the electricity spot market. It introduces the dynamic equivalent energy storage capacity constraint as the physical boundary condition for virtual power plants to participate in market bidding into the upper-layer model, and constructs a two-layer master-slave game collaborative bidding model with the maximization of the comprehensive net profit of virtual power plants as the upper layer and the minimization of the total electricity purchase and operation cost of the power grid system as the lower layer.

[0154] The model solving module is used to convert the two-layer master-slave game collaborative bidding model into a single-layer mixed integer linear programming model and solve it to obtain the optimal quantity and price quotation curve of the virtual power plant and the collaborative scheduling instructions of internal resources.

[0155] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0156] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0157] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0158] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0159] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model, characterized in that, include: By using a cost stratification strategy and a two-stage geometric projection method, the feasible domain of the distribution network inside the virtual power plant is mapped to the main grid side, generating an equivalent active-reactive feasible domain with cost ladder attributes. The two stages include the feasible domain characterization stage of the distribution network inside the virtual power plant and the cross-voltage level projection mapping stage. Based on the equivalent active-reactive feasible domain, the virtual power plant is reduced in dimension to be equivalent to a virtual energy storage model subject to dual constraints of power and energy, and dynamic equivalent energy storage capacity constraints are generated. The dynamic equivalent energy storage capacity constraints include equivalent charging and discharging power constraints and time-varying energy upper and lower limit constraints. To address the two-way clearing mechanism of the electricity spot market, the dynamic equivalent energy storage capacity constraint is introduced as the physical boundary condition for virtual power plants to participate in market bidding. This leads to the construction of a two-layer master-slave game collaborative bidding model with the maximization of the comprehensive net profit of virtual power plants as the upper layer and the minimization of the total electricity purchase and operation cost of the power grid system as the lower layer. The two-layer master-slave game collaborative bidding model is equivalently transformed into a single-layer mixed integer linear programming model and solved to obtain the optimal quantity and price quotation curve of the virtual power plant and the collaborative scheduling instructions for internal resources.

2. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 1, characterized in that, The method of using a cost stratification strategy and a two-stage geometric projection method to map the feasible region of the internal distribution network of the virtual power plant to the main grid side, generating an equivalent active-reactive feasible region with cost tier attributes, includes: Based on the linear branch power flow equation, a boundary search model is constructed with the objective function of maximizing the projection of the virtual power plant grid connection point power polygon in each search direction. The vertex search algorithm is used to iteratively solve the model to obtain the initial active and reactive power boundary vertex set of the internal distribution network side. Based on the consumption of regulation margin by the equivalent resistance and leakage reactance loss of the hub transformer under heavy load conditions, and combined with geometric projection mapping technology, the initial active and reactive power boundary vertex set is projected from the internal distribution network side to the high-voltage main grid side coordinate system to obtain the active and reactive power boundary vertex set on the high-voltage main grid side. Based on the set of active and reactive power boundary vertices on the high-voltage main grid side, combined with a cost tiering strategy and traversing different bidding tiers, an equivalent dynamic active-reactive feasible domain of a virtual power plant with cost tier attributes is generated. Extract the maximum projection interval of the boundary of the virtual power plant's equivalent dynamic active-reactive feasible domain on the active power coordinate axis, and use the maximum projection interval as the extreme value of the charging and discharging power of the virtual energy storage model in the corresponding time period.

3. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 2, characterized in that, The objective function expression of the boundary search model is: ； The expression for the constraints of the boundary search model is: ； ； ； ； ； ； ； ； In the formula, , These are the equivalent active and reactive power at the 35kV side common coupling point, respectively. , Let P be the weighting coefficients of active power and reactive power in the projection direction during the k-th iteration search. ih,t Q ih,t P represents the active power and reactive power flowing through branch ih during time period t. ji,t Q ji,t Let be the active power and reactive power flowing through the upstream branch ji during time period t, respectively, and let h be the downstream child node of node i. Let be the set of child nodes of node i, j be the upstream node of node i, and π(i) be the set of upstream nodes of node i. , For the equivalent injected active and reactive power of node i, U h,t U is the per-unit voltage value of node h during time period t. i,t Let U be the per-unit voltage value of node i during time period t. min U max These are the lower and upper safety limits for node voltages, respectively, R ih X ih These are the branch resistance and reactance, respectively. N is the number of sides when performing polygon linearization using a circular constraint. The apparent capacity thermal stability limit of the branch or grid connection point. Based on rigid load, For the distributed resource set of access node i, Contributing to the foundation of distributed resources, s r Let r be the maximum apparent power capacity of the distributed resource during time period t. For the index of the pricing ladder, Let r be the actual active power adjustment of the distributed resource r during time period t. , These refer to the upward and downward adjustment of resources, respectively. , These are the maximum upward and maximum downward adjustment power of the resource, respectively. For reactive power support, This is the reactive power regulation margin coefficient. Let r be the maximum apparent power capacity of the distributed resource r during time period t.

4. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 3, characterized in that, The expression for projecting the initial set of active and reactive power boundary vertices from the internal distribution network side to the high-voltage main grid side coordinate system is as follows: ； ； In the formula, , R represents the equivalent active power and reactive power at the 220kV side point of common coupling, respectively. T X T These are the equivalent resistance and leakage reactance of a 220kV transformer, respectively.

5. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 1, characterized in that, The method based on the equivalent active-reactive feasible region reduces the overall virtual power plant to a virtual energy storage model subject to dual constraints of power and energy, and aggregates and generates dynamic equivalent energy storage capacity constraints, including: Based on the equivalent active-reactive feasible region, the absolute active power boundary of the virtual power plant on each spatial section is determined. Based on the absolute active power boundary on each spatial section, the virtual power plant is reduced in dimensionality and equivalent to a virtual energy storage model subject to dual constraints of power and energy at the grid collaborative dispatch layer. In the virtual energy storage model, the equivalent charging and discharging power of the virtual energy storage is determined based on the change in the overall power output of the virtual power plant on the high-voltage main grid side, and the cumulative adjustment amount of the virtual power plant deviating from the benchmark operating point is set as the equivalent energy state of the virtual energy storage. Based on the equivalent charging and discharging power and the equivalent energy state, the feasible domain of virtual energy storage is determined. Using a polyhedral vertex mapping model with the goal of maximizing the extreme value of equivalent energy, boundary optimization is performed along the directions of maximizing and minimizing energy, respectively. The physical energy constraints of each heterogeneous distributed resource within the virtual power plant are mapped and aggregated upwards to obtain the upper and lower limits of the time-varying energy feasible domain, thus obtaining the dynamic equivalent energy storage capacity constraint.

6. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 5, characterized in that, The equivalent charge and discharge power of virtual energy storage is determined based on the change in the overall off-grid power of the virtual power plant on the high-voltage main grid side, including: The reduction in the overall off-grid power of the virtual power plant on the high-voltage main grid side is set as the equivalent discharge power of the virtual energy storage. The increase in the overall grid-connected power of the virtual power plant on the high-voltage main grid side is set as the equivalent charging power of the virtual energy storage.

7. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 5, characterized in that, The expression for the feasible domain of the virtual energy storage is: ； ； In the formula, This represents the feasible solution set of the equivalent regulation power and state of charge for virtual energy storage during time period t. For each variable in the state set to satisfy the physical constraints on its right-hand side, E represents the equivalent regulating power of the virtual power plant during time period t. VPP,t This represents the equivalent energy level of the virtual energy storage at the end of time period t. For virtual energy storage feasible domain, The active power of the virtual power plant at the 220kV side during time period t is the baseline operating power on the 220kV side. , These are the upper and lower limits of the absolute active power boundary, respectively. , These are the upper and lower limits of the maximum charging and discharging power of the mapped virtual energy storage, respectively. , These are the upper and lower limits of the time-varying energy boundary for virtual energy storage. This is the scheduling time step.

8. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 7, characterized in that, The objective function expression for the polyhedron vertex mapping model is: ； In the formula, Let v be the objective function of the polyhedron vertex mapping model, and v be the direction vector used to search for the charge boundary.

9. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 1, characterized in that, The two-layer master-slave game collaborative bidding model is composed of an upper-layer virtual power plant bidding model and a lower-layer power grid clearing model. The upper-level virtual power plant bidding model uses maximizing the intraday comprehensive net profit of the virtual power plant as its objective function, and the expression of the objective function is as follows: ； The constraints include strategic pricing constraints, internal resource output constraints, equivalent polymer physical mapping constraints, and dynamic equivalent state of charge constraints, expressed as follows: ； ； ； ； ； ； ； In the formula, F vpp N represents the intraday net profit of the virtual power plant. vpp Where T is the number of virtual power plants in the system, and T is the scheduling period. Let n(v) be the marginal electricity price of the node where the virtual power plant is located, and n(v) be the node number of the power grid to which the virtual power plant v is connected. , These represent the equivalent discharge power and charging power of the virtual power plant, respectively. int This refers to the number of tiers for flexible resources within the virtual power plant. , These represent the actual adjusted cost of resource usage for the i-th segment and the reduced cost of resource usage, respectively. , These are the upward and downward power boundaries for the internal k-th stage resources, respectively. , Let represent the lowered and higher bid prices submitted by the virtual power plant during time period t, respectively; v is the set of indices for virtual power plants within the system; and C... cap The upper limit of the price set by the spot market. , These are the maximum upward and downward power limits for the internal k-th tier resources, respectively, E vpp,v,t This represents the intraday dynamic equivalent energy state of the virtual power plant during time period t. , These are the upper and lower limits of the dynamic energy boundary affected by users' electricity consumption behavior, respectively. The scheduling time step; The lower-level grid clearing model aims to minimize the total system electricity purchase and operating cost. The objective function expression is as follows: ； The constraints include node power balance constraints, line transmission capacity constraints, clearing boundary constraints for conventional thermal power units and virtual power plants, operation constraints for grid-side independent energy storage systems, and non-negativity constraints for system balance node phase angles and relaxation variables. The expressions are as follows: ； ； ； ； ； ； ； ； ； ； In the formula, F iso N represents the total power purchase and operating cost of the system. g N bus These represent the total number of conventional units and the total number of nodes, respectively. g P represents the marginal cost of power generation by the generating unit. dg,g,t For the real-time absolute output of conventional units, , These are the discharge power and charging power of the grid-side independent energy storage system, respectively. , These are the cost coefficients for discharge power and charging power of grid-side independent energy storage systems, respectively. , These are the virtual unloaded variable and the power curtailment relaxation variable of the node, respectively, C slack The high cost of the penalty is set. Let g be the original, pure injected power after removing scheduling resources from node n, and g be the index of a conventional thermal power unit. B is the set of various device resources of access node n. n,l , These are the node admittance and voltage phase angle, respectively. The marginal electricity price at which the spot market clears out, H l,n The line power transfer distribution factor, This represents the maximum thermal stability limit of line l. These are the Lagrange multipliers corresponding to the lower and upper limits of the line transmission capacity constraints, respectively. These are the Lagrange multipliers corresponding to the lower and upper limits of the output of conventional thermal power units, respectively. These are the Lagrange multipliers E corresponding to the lower and upper limits of the charging power reduction constraints for virtual power plants. ess,t This represents the equivalent energy state of a grid-side independent energy storage system during time period t. These represent the lower and upper limits of the energy capacity of grid-side independent energy storage systems, respectively. ess,t The Lagrange multipliers corresponding to the constraints of the independent energy storage power evolution equation. These are the Lagrange multipliers corresponding to the lower and upper bound constraints of the independent energy storage capacity boundary, respectively. , These refer to the charging efficiency and discharging efficiency of an independent energy storage system, respectively. The voltage phase angle at the system's slack node. The Lagrange multiplier is the one corresponding to the phase angle equality constraint of the balancing node.

10. The virtual power plant and power grid collaborative bidding decision-making method based on an equivalent control model according to claim 1, characterized in that, The process of converting the two-layer master-slave game collaborative bidding model into a single-layer mixed-integer linear programming model and solving it to obtain the optimal quantity and price quotation curve of the virtual power plant and the collaborative scheduling instructions for internal resources includes: The stationarity conditions and complementary relaxation conditions of the lower-level power grid clearing model are extracted using the Lagrange multiplier method, and these conditions are introduced as additional constraints into the upper-level virtual power plant bidding model. Based on the strong duality theorem, the nonlinear revenue term in the objective function of the upper virtual power plant bidding model is equivalently transformed into a linear combination containing system parameters and dual multipliers by using dual objective transformation, and the objective function of the upper virtual power plant bidding model is updated by using the linear combination. The Big M method is used to linearize the nonlinear complementary relaxation conditions, and combined with the objective function of the updated upper-level virtual power plant bidding model, the two-level non-convex model is equivalently transformed into a single-level mixed integer linear programming model for global optimization. Based on the global optimization results, the optimal quantity and price quotation curve of the virtual power plant and the coordinated scheduling instructions for internal resources are output.

Images