A node baseline-guided power system main and microgrid collaborative scheduling method

Abstract

The application discloses a kind of node baseline guide power system main distribution microgrid collaborative scheduling method, it is related to the field of collaborative scheduling, this method faces three-level cooperation of main network, distribution network, microgrid, input is the unadjustable net load time series data of distribution network / microgrid area and each node flexible resource parameter, output is the load baseline LCDL of each node and response quality evaluation index, overall process is along the hierarchical structure of " distribution microgrid area aggregation layer→node and microgrid layer" development, first generate the shape target facing main-distribution boundary in area aggregation layer, then it is safely landed to each node and microgrid.This application uses the mirror symmetry closure mechanism of source and load about uniform line, under the balance and conservation constraint, the analytical expression of CDL is uniquely determined, without introducing preference objective function or iterative solver, realizes the paradigm conversion of adjusting responsibility from "statistical estimator" to "transmissible shape object".

Description

Technical Field

[0001] This invention relates to a collaborative scheduling method, specifically a power system master-distributor-micro collaborative scheduling method guided by node guidelines. Background Technology

[0002] When distribution networks and microgrids are in operation, the net power fluctuations of source loads within the area are absorbed free of charge by the upper-level main grid at the point of common coupling (PCC), leaving the "balancing responsibility" of the distribution / microgrid side to the main grid in a state of long-term limb. The key to coordinated scheduling of main-distribution-microgrids is to explicitly define this balancing responsibility across the main-distribution-microgrid boundaries as a calculable, transferable, and implementable object, so that the temporal shape target at the distribution / microgrid regional aggregation level and the safety implementation requirements at the spatial level of each node (microgrid) are unified under the same framework. This application defines this responsibility object as a "locational directrix load" for main-distribution-microgrid coordination, and expresses it in the form of load directrix (CDL) at the distribution / microgrid regional level and implements it in the form of node load directrix (LCDL) at the node and microgrid levels. This is the starting point of this application.

[0003] Regarding the definition, allocation, and evaluation of demand-side adjustment responsibilities under the main-distribution-micro network synergy, existing literature mainly focuses on the following dimensions, but none of them have formed a node alignment method that connects the three levels of main-distribution-micro network and can be both closed-loop calculated and securely implemented at the node (microgrid) level.

[0004] Regarding demand response baseline and price signaling methods, O'Connell et al. [1] systematically reviewed demand response baseline methods, pointing out that existing baseline methods mainly serve post-event settlement and response quantity verification. Essentially, they are statistical estimates that do not carry time-series shape information of regulation responsibility and cannot be directly used as day-ahead scheduling targets for resources under the coordination of main distribution micro-regions. Siano [2] provided a panoramic review of demand response projects, pointing out that although price mechanisms such as time-of-use pricing (TOU), peak pricing (CPP), and real-time pricing (RTP) can guide load shifting, their signals are scalar prices rather than time-series shapes. They cannot transmit complete instructions on "when to assume how much responsibility" to resources at each node in the distribution micro-region, and are even more difficult to coordinate with the power flow security of the distribution network.

[0005] Regarding the load guideline and nodal load guideline methods, in response to the above-mentioned shortcomings of the baseline method, reference [3] proposed the concept of load guideline (customer directrix load, CDL), which uses a system-friendly normalized load curve as the guiding target. However, it is uniformly published for the entire system and does not consider network topology and spatial security, making it difficult to directly guide the differentiated adjustment of each node within the distribution network. Reference [4] further proposed nodal customer directrix load, which publishes guidelines for different nodes of the transmission network to guide the overall participation of its downstream distribution network in demand response. However, there are significant differences between reference [4] and this application: First, in terms of application scenarios and hierarchical structure, reference [4] is geared towards the transmission network level and publishes a guideline for each transmission node in a unified calculation manner, and guides the distribution network under the node as a whole black box, without going deep into the distribution network or connecting to the microgrid and node level; this application is geared towards the coordination of main distribution microgrids, and adopts a hierarchical calculation structure of regional CDL and node LCDL, and implements the guideline from the distribution microregion aggregation layer to each node and microgrid step by step. Second, in terms of power flow model and constraints, reference [4] uses DC power flow and power transmission distribution factor (PTDF) modeling, only considering line transmission capacity constraints, while clearly leaving node voltage constraints, transformer capacity constraints, etc. to subsequent research; this application adopts the linearized branch power flow (LinDistFlow) model applicable to radial distribution networks, embedding node voltage amplitude constraints and branch active power capacity constraints into the guideline allocation at the same time, and node voltage exceeding the limit is one of the most critical issues in the safe operation of distribution networks and microgrids. Thirdly, while reference [4] takes demand response as the application background, this application takes the node guideline as the adjustment responsibility object of the main-partition-micro collaboration, and explicitly describes its transmission and implementation mechanism across the main-partition-micro boundary.

[0006] Regarding resource scheduling under power flow constraints in distribution networks, Baran and Wu[5] proposed a branch power flow (DistFlow) model for radial distribution networks, and its linearized version, LinDistFlow, has become a standard tool for optimal scheduling of distribution networks. Farivar and Low[6] further provided the accuracy conditions for convex relaxation of the branch power flow model. However, existing optimal scheduling methods for distribution networks based on LinDistFlow usually take economic / physical quantities such as minimizing total cost or maximizing new energy consumption as optimization objectives, and do not take "aligning the scheduling shape of each node with the regional guideline" as the optimization direction, nor do they reveal the structural law that "when network constraints are bound in the main distribution micro-coordination, the guidelines of each node will inevitably differentiate".

[0007] In terms of virtual power plants and flexible resource aggregation, existing virtual power plant (VPP) literature mainly focuses on the adjustable capacity / adjustable range modeling after resource aggregation, such as methods based on Minkowski and polyhedral approximation [7]. The output is a geometric description of the feasible region of aggregated power and power inverse allocation for the aggregate, rather than the time-series scheduling shape of each node and each time period within the distribution micro-region. Such methods usually do not embed distribution network power flow constraints, making it difficult to ensure that the aggregation and allocation results simultaneously meet the branch capacity and node voltage safety requirements, and also cannot directly serve the node baseline landing under the coordination of main distribution micro-region.

[0008] In terms of demand response assessment and evaluation, existing market rules and academic research mostly use ex-post indicators to evaluate the effectiveness of demand response, such as achievement rate (actual response volume / reported volume) and number of overruns. Existing literature lacks a systematic quantitative analysis framework for main grid-distribution-micro grid coordination regarding the quantitative mapping relationship between resource response level and the tracking error of distribution micro-regions to the main grid boundary (PCC), as well as the differentiated impacts (spatial heterogeneity) of insufficient resource response in different spatial locations on distribution network security (especially node voltage overruns).

[0009] Regarding the contribution of energy storage to the feasibility of alignment, the role of energy storage in distribution network demand response has been extensively studied. However, existing research mainly focuses on the economic dispatch value of energy storage (peak shaving and valley filling benefits) or its role as a control strategy for frequency regulation resources. The structurally necessary role of energy storage in "making the nodal alignment allocation problem under main-distribution-micro-network coordination from infeasible to feasible"—that is, the mechanism by which energy storage smooths the net load of nodes through intraday charging and discharging, thereby "buying" feasible alignment allocation space for each node that meets voltage and capacity constraints—has not yet been theoretically revealed in existing literature.

[0010] Defects and shortcomings of existing technology: 1) The regulation responsibility lacks a formal definition and closed-form calculation method for the coordination of primary, secondary, and micro-regions. Existing demand response baseline methods are essentially ex-post statistics, and price signals are scalar rather than time-series shapes. Neither can transmit a complete time-series instruction of "when to assume how much responsibility" to the resources of each node in the secondary and micro-regions. No method has been reported that defines the regulation responsibility across the primary-secondary-micro-region boundary as a normalized time-series shape object with total amount conservation and gives a closed-form analytical expression that does not depend on the optimization solver [1, 2].

[0011] 2) Existing load profile and node load profile methods are not geared towards the hierarchical coordination of main distribution and microgrids, do not delve into the internal distribution network and do not take into account voltage constraints. The unified load profile in reference [3] is published for the entire system without distinguishing between network topology and spatial security; the node load profile in reference [4] is geared towards the transmission network level, uses DC power flow and PTDF modeling, only takes into account line capacity constraints and leaves node voltage constraints for later, and treats the distribution network connected to the node as a black box to guide the whole, without connecting to the microgrid and node level, and without taking into account the node voltage amplitude constraints which are the most critical in the operation of the distribution network [3, 4].

[0012] 3) There is a lack of a structured connection mechanism that takes into account the power flow constraints of the distribution network between the regional balance target and the spatial security of the nodes. After solving the regional level scheduling target, the existing methods usually directly allocate the power flow to each node proportionally or according to the economic optimality, without embedding the power flow constraints of the distribution network (branch active power capacity limit and node voltage amplitude limit) into the allocation process. This may result in the allocation result violating network security constraints when the main distribution micro-coordination is implemented [5, 6].

[0013] 4) The differentiation of node guidelines lacks theoretical explanation and quantitative characterization. When the distribution network constraints are bound, the guidelines of each node will inevitably deviate from the unified guideline of the region. However, existing methods have not revealed the generation mechanism, quantitative structure (the double conservation law of net zero per node in the time dimension and cancellation per time period in the spatial dimension) and degradation conditions (differentiation disappears when the network is relaxed and the node guideline degenerates into the regional guideline) [5, 6].

[0014] 5) Quantitative assessment of the impact of response level on distribution network security is lacking. Existing demand response assessment methods lack quantitative mapping relationships from resource response level to system-level indicators such as PCC boundary tracking error, power liability leakage rate, and number of node voltage overruns in the distribution micro-region, making it difficult to support response quality assessment under the coordination of main distribution micro-regions [1, 2].

[0015] 6) The structural role of energy storage in the feasibility of quasi-line deployment has not been identified. In existing distribution network demand response studies, energy storage is usually regarded as an independent economic dispatch object or frequency regulation resource, and the structural necessity of energy storage in meeting voltage and capacity constraints has not been revealed from the perspective of "feasibility conditions of node quasi-line allocation problem under the coordination of main distribution and micro-distribution".[7]

[0016] 7) Insufficient carrying capacity and lack of quantitative output. When the distribution network constraints are too tight, making node alignment allocation infeasible, existing methods usually only output an "infeasible" judgment, without providing location information of binding constraints (over-limit branches or over-limit node voltages) and quantitative values ​​of responsibility residuals, which is not conducive to guiding network enhancement and planning under the coordination of main distribution and micro-distribution [5, 6].

[0017] References: [1] O'Connell N, Pinson P, Madsen H, et al. Benefits and challenges of electrical demand response: a critical review[J]. Renewable and Sustainable Energy Reviews, 2014, 39: 686-699. [2] Siano P. Demand response and smart grids—a survey[J]. Renewableand Sustainable Energy Reviews, 2014, 30: 461-478. [3] Fan Shuai, Jia Kunqi, Wang Fen, et al. Large-scale demand response based on load profile [J]. Automation of Electric Power Systems, 2020, 44(15): 19-27. [4] Meng Yan, Xiao Jucheng, Hong Juhua, et al. Nodal load guideline considering demand response uncertainty: concept and model [J]. Automation of Electric Power Systems, 2023, 47(13): 28-39. [5] Baran ME, Wu F F. Network reconfiguration in distributionsystems for loss reduction and load balancing[J]. IEEE Trans. Power Delivery, 1989, 4(2): 1401-1407. [6] Farivar M, Low S H. Branch flow model: relaxations andconvexification[J]. IEEE Trans. Power Systems, 2013, 28(3): 2554-2564. [7] Müller FL, Szabó J, Sundström O, et al. Aggregation and disaggregation of energetic flexibility from distributed energy resources[J]. IEEE Trans. Smart Grid, 2019, 10(2): 1205-1214. Summary of the Invention

[0018] The purpose of this invention is to provide a power system primary-distribution-micro-coordinated scheduling method guided by node guidelines, so as to solve the problems mentioned in the background art.

[0019] To achieve the above objectives, the present invention provides the following technical solution: A method for coordinated primary, secondary, and micro-level dispatching of a power system guided by node guidelines includes the following steps: Step S1: Oriented towards main-distribution coordination, in the distribution micro-region aggregation layer, with the regional non-adjustable net load time series data as input, establish a source-load mirror symmetric closed model, and output the injection shape of the main-distribution boundary common connection point PCC and the regional load guideline CDL as the regional time series responsibility target; Step S2: For distribution-micro collaboration, at the node and microgrid layers, with the regional CDL as input, establish a convex quadratic programming allocation model under the LinDistFlow power flow constraint, taking into account the node voltage amplitude constraint and the branch active power capacity constraint, and output the node-by-node adjustable scheduling scheme and the node load guideline LCDL. Step S3: Using the node load guideline LCDL output in step S2 as input, perform two-segment decomposition on it, verify the dual conservation laws of time and space dimensions and network security constraints, and output the verification conclusion of the node guideline differentiation. Step S4: Based on the actual operation of the microgrid, and taking the actual response data of each node LCDL and resources as input, establish a response quality assessment model based on the response degree parameter β, output system-level indicators such as PCC boundary tracking error, power responsibility leakage rate and voltage over-limit, and determine whether the network carrying capacity is sufficient.

[0020] As a further embodiment of the present invention, let the set of non-root nodes of the distribution network be N={1,...,N}, where node i also represents the user aggregate attached to that node, i.e., flexible resource i, and the resource number is the same as the node number. Node 0 is the root node, i.e., the common connection point PCC.

[0021] As a further embodiment of the present invention, step S1 is executed in the regional aggregation layer of the main distribution / microgrid collaboration. The inputs are the time-series data of the non-adjustable net load of the distribution network / microgrid region, the total power of the regional adjustable resources, and the adjustable power of the PCC at the main-distribution boundary. The PCC is regarded as an external adjustable source parallel to the adjustable resources in the region. The two jointly bear the fluctuation of the non-adjustable net load. Based on this, a mirror-symmetric closed model of the source load about the uniform line 1 / T is established. The regional closure relationship is that the adjustment capabilities of the PCC and the adjustable resources in the region are balanced and non-adjustable. The mirror symmetry closure condition of the aforementioned mirror symmetry closure model requires that the sum of the PCC boundary shape and the adjustable shape of the region be always equal to 2 / T, as shown in equation (1): (1) Solving the univariate equation (1) time-by-time yields a closed-form analytical expression that uniquely determines the injection shape at the boundary between the regional load guideline CDL and PCC, as shown in equations (2) and (3): (2) (3) In the formula, φ(t) is CDL, which is the regional load guideline, satisfying Σφ(t)=1; Shape is injected into the PCC boundary. The region's adjustable target r(t) is determined by the CDL and the total adjustable charge of the region, as shown in equation (4): (4) In the formula, the physical meaning of CDL is: when the net load of the region is lower than the average, φ(t) increases, indicating that more resources are used to absorb photovoltaics; when the net load is higher than the average, φ(t) decreases, indicating that resources are used to avoid peak loads. This step outputs the regional load guideline CDL, the PCC boundary injection shape and the regional adjustable target r(t), which are used as the regional time-series responsibility targets allocated by the node in step S2.

[0022] As a further aspect of the present invention, step S2 is executed at the node and microgrid layer. The inputs are the regional load baseline CDL and the regional adjustable target r(t), the distribution network topology and branch resistance and reactance parameters, the operating boundary and power budget of each node resource, output from step S1. With the goal of each node replicating the regional baseline, a convex quadratic programming allocation model under the LinDistFlow distribution network power flow constraints, taking into account node voltage amplitude constraints and branch active power capacity constraints, is established. This safely translates the regional time-series responsibility target into a node-by-node adjustable scheduling scheme. The convex quadratic programming allocation model consists of three parts: the LinDistFlow network model, constraints, and objective function, as detailed below: LinDistFlow network model The current flow recursion is shown in equations (5) and (6): (5) (6) The voltage recursion is shown in equation (7): (7) Constraints The active power limit constraint C1 for the branch is shown in equation (8): (8) The node voltage amplitude limit constraint C2 is shown in equation (9): (9) The resource operation boundary constraint C3 is shown in equation (10): (10) The energy budget conservation constraint C4 is shown in equation (11): (11) The region alignment constraint C5 is shown in equation (12): (12) Constraint C4 ensures that the daily power consumption of each resource remains constant, and constraint C5 ensures that the total power consumption of each node aligns with the regional target on a time-by-time basis.

[0023] objective function Based on the principle that each node adopts the CDL, the weighted deviation of the adjustable scheduling of each node from the CDL is minimized, as shown in Equation (13): (13) In the formula, >0 represents the weight coefficient of each node, ensuring that the objective function is strictly convex and the solution is unique. The whole problem constitutes a convex quadratic programming (QP) problem with a unique optimal solution. Solve the model and output the optimal adjustable schedule u for each node. (i,t) and the normalized node load guideline LCDL are used as the objects of safety verification in step S3.

[0024] As a further embodiment of the present invention, step S3 takes the node load guideline LCDL output in step S2 as input, performs structural decomposition and safety verification on it, and decomposes the node load guideline into two parts: regional guideline component and network correction component, as shown in equation (14): (14) The network correction component automatically satisfies two conservation laws. The time conservation law is shown in equation (15): (15) The law of conservation of space is shown in equation (16): (16) Degeneration Theorem: When network constraints are not bound, the network correction component is strictly zero, and the node load guideline LCDL degenerates into the regional load guideline CDL. This step verifies the two-segment decomposition structure of the network correction component, the dual conservation laws in the time and space dimensions, and the satisfaction of the node voltage amplitude constraints and branch active power capacity constraints, and outputs the verification conclusion of the node guideline differentiation.

[0025] As a further aspect of the present invention, step S4 evaluates the tracking quality of resources to the node load baseline LCDL during actual microgrid operation, with the input being the node-by-node load baseline LCDL verified in step S3 and the optimal adjustable scheduling u for each node. (i,t) and the actual response data of the resources are used to introduce the response degree parameter β to describe the degree of actual tracking of the resources to LCDL. A quantitative mapping evaluation model from the response degree to system-level indicators is established to output response quality indicators such as PCC boundary tracking error, power responsibility leakage rate and number of node voltage over-limit times, and the network carrying capacity is determined accordingly.

[0026] As a further aspect of the present invention, the response level parameter β∈[0,1] is defined to describe the degree to which the resource actually tracks the LCDL, and the actual adjustable active power of resource i under the response level β(i) is shown in equation (17): (17) The PCC boundary tracking error RMSE is shown in equation (18): (18) The electrical leakage rate is shown in equation (19): (19) This step scans the response level parameter β to obtain the monotonic variation of the three indicators with β, and quantifies the impact of spatial location on response quality by comparing homogeneous and heterogeneous resource spatial distribution scenarios.

[0027] Compared with the prior art, the beneficial effects of the present invention are: 1. A formal definition and closed-form calculation method (CDL) for regulating responsibility are proposed.

[0028] This application defines the regulation responsibility of the distribution network / microgrid area as a normalized time-series shape object—the load guideline CDL. It adopts a mirror-symmetric closed-loop mechanism of source load about the uniform line, and uniquely determines the analytical expression of the CDL under the constraints of balance and conservation. It does not require the introduction of a preference objective function or an iterative solver, and realizes the paradigm shift of regulation responsibility from "statistical estimate" to "transferable shape object".

[0029] 2. Establish a structured linkage mechanism (LCDL allocation) from regional balance objectives to node spatial security.

[0030] Based on the CDL, this application constructs a convex quadratic programming model under the LinDistFlow power flow constraint of the distribution network. With the goal of "each node adopting the CDL", it solves the network feasible allocation scheme that is closest to the CDL, realizing a two-stage progressive allocation of "time sequence first and space later".

[0031] 3. Reveal the generation mechanism and conservation structure of node shape differentiation.

[0032] This application characterizes the nodal directrix LCDL as two parts: CDL components and network correction components. It proves that the dual conservation law of net zero per node in the time dimension and net cancellation per time period in the spatial dimension of the network correction components is automatically guaranteed by the energy budget constraint and the region alignment constraint. At the same time, it proves the degradation theorem: when the network constraints are not bound, the LCDL degenerates into CDL.

[0033] 4. Establish a quantitative mapping framework from response level to system-level indicators.

[0034] This application introduces a response level parameter β to establish a quantitative mapping from β to three indicators: PCC boundary tracking error, power liability leakage rate, and voltage over-limit number. It also reveals the differential impact of spatial location on response quality through comparison of homogeneous / heterogeneous scenarios. Attached Figure Description

[0035] Figure 1 This is a diagram of the power system main-distribution-micro collaborative dispatching system architecture guided by node guidelines in an embodiment of the present invention.

[0036] Figure 2 This is a diagram showing the shape of the day-return curve and the region aggregation active power diagram in an embodiment of the present invention.

[0037] Figure 3 This is a heat map of the non-adjustable net active power d(i,t) of each node in an embodiment of the present invention.

[0038] Figure 4 The CDL logic diagram in this embodiment of the invention is a mirror symmetry diagram of φ(t) and M_PCC(t) and a regional power allocation conservation verification diagram.

[0039] Figure 5 This is a comparison diagram of LCDL and CDL for nodes 18 and 16 in an embodiment of the present invention.

[0040] Figure 6 This is a comparison of the main feeder voltage amplitude in an embodiment of the present invention: without coordination (left) and with coordination (right).

[0041] Figure 7 This is a β-scan of the response level in an embodiment of the present invention: a monotonically degrading graph of three indicators with β.

[0042] Figure 8 This is a comparison chart of the actual PCC injection and CDL target under different β values ​​in embodiments of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] In this embodiment of the invention, the method of this application is oriented towards three-level coordination of the main grid, distribution grid, and microgrid (hereinafter referred to as main-distribution-microgrid coordination). The inputs are the non-adjustable net load time-series data of the distribution network / microgrid area and the flexible resource parameters of each node. The outputs are the load baseline (LCDL) and response quality evaluation indicators for each node. The overall process unfolds along a hierarchical structure of "distribution-microgrid regional aggregation layer → node and microgrid layer". First, a shape target facing the main-distribution boundary (PCC) is generated at the regional aggregation layer. Then, it is safely implemented to each node and microgrid. Steps S1 to S4 are executed sequentially. Each step uses the output of the previous step as input, establishes the corresponding model, and solves the output result.

[0045] Step S1: Oriented towards main-distribution coordination, in the distribution micro-region aggregation layer, with the time series data of the region's non-adjustable net load as input, establish a source-load mirror symmetric closed model, and output the injection shape of the main-distribution boundary common connection point (PCC) and the regional load guideline CDL as the regional time series responsibility target; Step S2: For distribution-micro collaboration, at the node and microgrid layers, with the regional CDL as input, establish a convex quadratic programming allocation model under the LinDistFlow power flow constraint, taking into account the node voltage amplitude constraint and the branch active power capacity constraint, and output the node-by-node adjustable scheduling scheme and the node load guideline LCDL. Step S3: Using the node load guideline LCDL output in step S2 as input, perform two-segment decomposition on it, verify the dual conservation laws of time and space dimensions and network security constraints, and output the verification conclusion of the node guideline differentiation. Step S4: Based on the actual operation of the microgrid, and taking the actual response data of each node LCDL and resources as input, establish a response quality assessment model based on the response degree parameter β, output system-level indicators such as PCC boundary tracking error, power responsibility leakage rate and voltage over-limit, and determine whether the network carrying capacity is sufficient.

[0046] Let the set of non-root nodes in the distribution network be N = {1,...,N}, where node i also represents the user aggregate (flexible resource i) connected to that node, and the resource number is the same as the node number. Node 0 is the root node, i.e., the point of common connection (PCC).

[0047] Step S1: CDL modeling of the regional load guideline for primary and secondary load coordination This step is performed at the regional aggregation layer of the main-distribution-microgrid coordination. The inputs are the time-series data of the non-adjustable net load of the distribution network / microgrid region, the total amount of adjustable resources in the region, and the adjustable amount of the common connection point (PCC) at the main-distribution boundary. The PCC is regarded as an external adjustable source parallel to the adjustable resources in the region. The two share the non-adjustable net load fluctuations. Based on this, a mirror-symmetric closed-loop model of source and load about the uniform line 1 / T is established. The regional closure relationship is that the adjustable (PCC and resources) balances the non-adjustable.

[0048] The mirror symmetry closure condition of this model requires that the sum of the PCC boundary shape and the adjustable shape of the region be always equal to 2 / T, as shown in equation (1): (1) Solving the univariate equation (1) time-by-time uniquely determines the closed analytical expression of the injection shape at the boundary between the regional load guideline CDL and PCC, without the need to introduce a preference objective function or an iterative solver, as shown in equations (2) and (3): (2) (3) In the formula, φ(t) is CDL, which is the regional load guideline, satisfying Σφ(t)=1; Shape is injected into the PCC boundary. The region's adjustable target r(t) is determined by the CDL and the total adjustable charge of the region, as shown in equation (4): (4) In the formula, the physical meaning of CDL is: when the area's net load is lower than the average value. Increased load indicates that more resources are being used to absorb photovoltaic power; when the net load is higher than the average. The reduction indicates resource peak avoidance. This step outputs the regional load guideline CDL, PCC boundary injection shape, and regional adjustable target r(t), which serve as the regional time-series responsibility targets allocated by the node in step S2.

[0049] Step S2: Modeling the Node Load Guidelines (LCDL) for Micro-coordination This step is executed at the node and microgrid layers. The inputs are the regional load baseline (CDL) and regional adjustable target r(t) output from step S1, the distribution network topology and branch resistance and reactance parameters, and the operating boundaries and energy budgets of each node's resources. With the goal of each node replicating the regional baseline, a convex quadratic programming allocation model under the LinDistFlow distribution network power flow constraints, taking into account node voltage amplitude constraints and branch active power capacity constraints, is established. This model securely translates the regional time-series responsibility target into a node-by-node adjustable scheduling scheme. The model consists of three parts: the network model, constraints, and the objective function.

[0050] LinDistFlow network model The current flow recursion (from leaf to root convergence) is shown in equations (5) and (6): (5) (6) The voltage recursion (propagation from root to leaf) is shown in equation (7): (7) Constraints Constraint C1 (branch active power limit), as shown in equation (8): (8) Constraint C2 (node ​​voltage amplitude limit), as shown in equation (9): (9) Constraint C3 (resource operation boundary) is shown in equation (10): (10) Constraint C4 (energy budget conservation) is shown in equation (11): (11) Constraint C5 (Region Alignment (Tracing CDL)), as shown in Equation (12): (12) Constraint C4 ensures that the daily power consumption of each resource remains constant (the root of time conservation), and constraint C5 ensures that the total power consumption of each node aligns with the regional target on a time-by-time basis (the root of spatial conservation).

[0051] objective function Based on the principle that each node adopts the CDL, the weighted deviation of the adjustable scheduling of each node from the CDL is minimized, as shown in Equation (13): (13) In the formula, >0 represents the weight coefficient of each node, ensuring that the objective function is strictly convex and the solution is unique. The overall problem constitutes a convex quadratic programming (QP) problem with a unique optimal solution. Solving this model outputs the optimal adjustable schedule u for each node. (i,t) and the normalized node load guideline LCDL are used as the objects of safety verification in step S3.

[0052] Step S3: Decomposition and safety check of nodal line differentiation This step takes the nodal load guideline LCDL output from step S2 as input and performs structural decomposition and safety verification on it. The nodal load guideline is decomposed into two parts: the regional guideline component and the network correction component, as shown in equation (14): (14) The network correction component automatically satisfies two conservation laws. The time conservation law is shown in equation (15): (15) The law of conservation of space is shown in equation (16): (16) Degeneration Theorem (LCDL→CDL): When network constraints are not bound, the network correction component is strictly zero, and the node load profile LCDL degenerates into the regional load profile CDL. This step verifies the two-segment decomposition structure of the network correction component, the dual conservation laws in the time and space dimensions, and the satisfaction of node voltage magnitude constraints and branch active power capacity constraints, outputting the verification conclusion of the node profile differentiation.

[0053] Step S4: Response quality assessment of actual microgrid operation This step evaluates the tracking quality of resources to the node load baseline (LCDL) during actual microgrid operation. The inputs are the node-by-node load baseline (LCDL) verified in step S3 and the optimal adjustable scheduler u for each node. (i,t) and the actual response data of the resources. The response degree parameter β is introduced to describe the degree to which the resources actually track the LCDL. A quantitative mapping evaluation model from the response degree to system-level indicators is established. The response quality indicators such as PCC boundary tracking error, power responsibility leakage rate and number of node voltage over-limit times are output, and the network carrying capacity is determined accordingly.

[0054] Response degree parameters and indicator system The response level parameter β∈[0,1] is defined to describe the degree to which the resource actually tracks the LCDL. The actual adjustable active power of resource i under the response level β(i) is shown in equation (17): (17) The PCC boundary tracking error RMSE is shown in equation (18): (18) The electrical leakage rate is shown in equation (19): (19) Response level analysis This step scans the response level parameter β to obtain the monotonic variation of the three indicators with β, and quantifies the impact of spatial location on response quality by comparing homogeneous and heterogeneous resource spatial distribution scenarios.

[0055] System Implementation Architecture The method described in this application can be deployed and implemented in a distribution network / microgrid energy management system. Its system architecture, oriented towards main-distribution-microgrid collaboration, is as follows: Figure 1As shown, the self-configured micro-region aggregation layer is deployed to each node and the micro-network layer, and includes the following five modules: (1) Data acquisition module: Real-time acquisition of non-adjustable active and reactive power data of each node through SCADA system and power distribution automation terminal.

[0056] (2) CDL calculation module: calculate the boundary shapes of CDL and PCC in closed form according to equations (2)-(4), without iterative solution.

[0057] (3) LCDL allocation module: Construct a convex QP model under LinDistFlow constraints (Equations (5)-(13)), and call the QP solver to output LCDL. If it is not feasible, report insufficient bearing capacity.

[0058] (4) Response evaluation module: Calculate the three evaluation indicators under the response degree β according to formulas (17)-(19).

[0059] (5) Decision support module: The comprehensive calculation results provide dispatchers with suggestions for demand response strategy formulation and network planning.

[0060] Case Analysis Example system settings The test system used an IEEE 33-node radial distribution network. The base voltage was 12.66 kV, and the base capacity was 10 MVA. DR resources (total adjustable capacity 12.0 MWh) were installed at nodes 12, 16, 18, 24, 30, and 32; energy storage was installed at nodes 18 and 32; and distributed photovoltaic (PV) systems (total installed capacity 2.0 MW) were installed at nodes 18, 22, 25, and 32. The time resolution was 96 points (15 minutes). The shape of the regional return-to-date curve and the distribution of aggregated active power and non-adjustable net active power at each node are shown below. Figure 2 and Figure 3 As shown.

[0061] Results of main and auxiliary coordination and regional load profile Calculate the regional load guideline CDL and PCC boundary injection shape according to step S1. They are mirror-symmetric about the uniform line, and the regional power distribution satisfies the conservation check, as shown below. Figure 4 As shown.

[0062] Micro-coordination and node alignment verification results Following step S2, solve for the load guideline LCDL of each node under network constraints. Following step S3, verify the decomposed structure and network security. The LCDL of nodes bound to network constraints deviates from the region CDL, while the LCDL of unbound nodes coincides with the region CDL. Figure 5 As shown. After coordination, the main feeder voltage amplitude returns to the safe range, as... Figure 6 As shown.

[0063] Microgrid response quality assessment results Assess the response quality of the microgrid in actual operation according to step S4. Perform a homogeneous scan on the response level β. The PCC boundary tracking error, energy liability leakage rate, and number of node voltage overruns monotonically degrade as β decreases. Figure 7 As shown, the impact of insufficient resource response at different spatial locations on boundary tracking and node voltage varies in heterogeneous scenarios. A comparison of actual PCC injection and regional CDL targets under different β values ​​is presented. Figure 8 As shown.

[0064] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0065] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for coordinated primary, secondary, and micro-level dispatching of a power system guided by node guidelines, characterized in that, Includes the following steps: Step S1: Oriented towards main-distribution coordination, in the distribution micro-region aggregation layer, with the regional non-adjustable net load time series data as input, establish a source-load mirror symmetric closed model, and output the injection shape of the main-distribution boundary common connection point PCC and the regional load guideline CDL as the regional time series responsibility target; Step S2: For distribution-micro collaboration, at the node and microgrid layers, with the regional CDL as input, establish a convex quadratic programming allocation model under the LinDistFlow power flow constraint, taking into account the node voltage amplitude constraint and the branch active power capacity constraint, and output the node-by-node adjustable scheduling scheme and the node load guideline LCDL. Step S3: Using the node load guideline LCDL output in step S2 as input, perform two-segment decomposition on it, verify the dual conservation laws of time and space dimensions and network security constraints, and output the verification conclusion of the node guideline differentiation. Step S4: Based on the actual operation of the microgrid, and taking the actual response data of each node LCDL and resources as input, establish a response quality assessment model based on the response degree parameter β, output system-level indicators such as PCC boundary tracking error, power responsibility leakage rate and voltage over-limit, and determine whether the network carrying capacity is sufficient.

2. The power system primary-distribution-micro-coordinated dispatching method guided by node guidelines according to claim 1, characterized in that, Let the set of non-root nodes of the distribution network be N={1,...,N}, where node i also represents the user aggregate connected to that node, i.e., flexible resource i, and the resource number is the same as the node number. Node 0 is the root node, i.e., the point of common connection PCC.

3. The power system primary-distribution-micro-coordinated dispatching method guided by node guidelines according to claim 2, characterized in that, Step S1 is executed in the regional aggregation layer of the main distribution / microgrid collaboration. The inputs are the time series data of the non-adjustable net load of the distribution network / microgrid region, the total power of the regional adjustable resources, and the adjustable power of the PCC at the main-distribution boundary. The PCC is regarded as an external adjustable source parallel to the adjustable resources in the region. The two share the responsibility for the fluctuation of the non-adjustable net load. Based on this, a mirror-symmetric closed model of the source and load about the uniform line 1 / T is established. The regional closure relationship is that the adjustment capabilities of the PCC and the adjustable resources in the region are balanced and non-adjustable. The mirror symmetry closure condition of the aforementioned mirror symmetry closure model requires that the sum of the PCC boundary shape and the adjustable shape of the region be always equal to 2 / T, as shown in equation (1): （1） Solving the univariate equation (1) time-by-time yields a closed-form analytical expression that uniquely determines the injection shape at the boundary between the regional load guideline CDL and PCC, as shown in equations (2) and (3): （2） （3） In the formula, φ(t) is CDL, which is the regional load guideline, satisfying Σφ(t)=1; Shape is injected into the PCC boundary. The region's adjustable target r(t) is determined by the CDL and the total adjustable charge of the region, as shown in equation (4): （4） In the formula, the physical meaning of CDL is: when the net load of the region is lower than the average, φ(t) increases, indicating that more resources are used to absorb photovoltaics; when the net load is higher than the average, φ(t) decreases, indicating that resources are used to avoid peak loads. This step outputs the regional load guideline CDL, the PCC boundary injection shape and the regional adjustable target r(t), which are used as the regional time-series responsibility targets allocated by the node in step S2.

4. The power system primary-distribution-micro-coordinated dispatching method guided by node guidelines according to claim 3, characterized in that, Step S2 is executed at the node and microgrid layers. The inputs are the regional load baseline (CDL) and regional adjustable target r(t) output from step S1, the distribution network topology and branch resistance and reactance parameters, the operating boundaries and energy budgets of each node's resources, and the goal of each node replicating the regional baseline. A convex quadratic programming allocation model under LinDistFlow distribution network power flow constraints, taking into account node voltage amplitude constraints and branch active power capacity constraints, is established. This model securely translates the regional time-series responsibility target into a node-by-node adjustable scheduling scheme. The convex quadratic programming allocation model consists of three parts: the LinDistFlow network model, constraints, and the objective function, as detailed below: LinDistFlow network model The current flow recursion is shown in equations (5) and (6): （5） （6） The voltage recursion is shown in equation (7): （7） Constraints The active power limit constraint C1 for the branch is shown in equation (8): （8） The node voltage amplitude limit constraint C2 is shown in equation (9): （9） The resource operation boundary constraint C3 is shown in equation (10): （10） The energy budget conservation constraint C4 is shown in equation (11): （11） The region alignment constraint C5 is shown in equation (12): （12） Constraint C4 ensures that the daily power consumption of each resource remains constant, and constraint C5 ensures that the total power consumption of each node aligns with the regional target on a time-by-time basis. objective function Based on the principle that each node adopts the CDL, the weighted deviation of the adjustable scheduling of each node from the CDL is minimized, as shown in Equation (13): （13） In the formula, >0 represents the weight coefficient of each node, ensuring that the objective function is strictly convex and the solution is unique. The whole problem constitutes a convex quadratic programming (QP) problem with a unique optimal solution. Solve the model and output the optimal adjustable schedule u for each node. (i,t) and the normalized node load guideline LCDL are used as the objects of safety verification in step S3.

5. The power system primary-distribution-micro-coordinated dispatching method guided by node guidelines according to claim 4, characterized in that, Step S3 takes the node load guideline LCDL output in step S2 as input, performs structural decomposition and safety verification on it, and decomposes the node load guideline into two parts: regional guideline component and network correction component, as shown in equation (14): （14） The network correction component automatically satisfies two conservation laws. The time conservation law is shown in equation (15): （15） The law of conservation of space is shown in equation (16): （16） Degeneration Theorem: When network constraints are not bound, the network correction component is strictly zero, and the node load guideline LCDL degenerates into the regional load guideline CDL. This step verifies the two-segment decomposition structure of the network correction component, the dual conservation laws in the time and space dimensions, and the satisfaction of the node voltage amplitude constraints and branch active power capacity constraints, and outputs the verification conclusion of the node guideline differentiation.

6. The power system primary-distribution-micro-coordinated dispatching method guided by node guidelines according to claim 5, characterized in that, Step S4 evaluates the tracking quality of resources to the node load baseline LCDL during actual microgrid operation. The inputs are the node-by-node load baseline LCDL verified in step S3 and the optimal adjustable scheduling u for each node. (i,t) and the actual response data of the resources are used to introduce the response degree parameter β to describe the degree of actual tracking of the resources to LCDL. A quantitative mapping evaluation model from the response degree to system-level indicators is established to output response quality indicators such as PCC boundary tracking error, power responsibility leakage rate and number of node voltage over-limit times, and the network carrying capacity is determined accordingly.

7. The power system primary-distribution-micro-coordinated dispatching method guided by node guidelines according to claim 6, characterized in that, The response level parameter β∈[0,1] is defined to describe the degree to which the resource actually tracks the LCDL. The actual adjustable active power of resource i under the response level β(i) is shown in equation (17): （17） The PCC boundary tracking error RMSE is shown in equation (18): （18） The electrical leakage rate is shown in equation (19): （19） This step scans the response level parameter β to obtain the monotonic variation of the three indicators with β, and quantifies the impact of spatial location on response quality by comparing homogeneous and heterogeneous resource spatial distribution scenarios.

Images