A new multi-level autonomous control method for distribution network oriented to master-slave coordination

Abstract

The application provides a novel power distribution network multilevel autonomous control method facing main coordination, and belongs to the technical field of optimal operation control of power distribution networks. The method comprises the following steps: acquiring the multilevel network topology and operation data of the power distribution network; establishing an internal autonomous equivalent model at the transformer area layer to obtain the time sequence power interaction results between the transformer area and the feeder layer; constructing a comprehensive partition index based on the electrical distance, active power balance degree and reactive power balance degree at the feeder layer, and performing autonomous area division by using a genetic algorithm; establishing a main grid-multifeeder area collaborative optimization scheduling model; iteratively updating the area boundary variables and the Lagrange multiplier by using a distributed solving algorithm based on the alternating direction multiplier method to obtain the dynamic power interaction and resource output results of each level; and triggering weight updating and area rolling reconstruction according to the change of the operation state. The application can take into account the structural rationality and operation functionality of the area division, and improve the economy, safety and coordinated control efficiency of the power distribution network in the high proportion distributed resource access scenario.

Description

Technical Field

[0001] This invention relates to the field of distribution network optimization operation control technology, and in particular to a novel multi-level autonomous control method for distribution networks oriented towards main-distribution coordination. Background Technology

[0002] With the large-scale integration of flexible resources such as distributed photovoltaics, energy storage, electric vehicles, flexible loads, and reactive power compensation devices into the distribution network, traditional centralized operation and control methods face problems such as voltage exceeding limits, increased network losses, power flow reversal, and a sharp increase in computational scale. To balance the safe, economical, and flexible operation of the distribution network, a more efficient collaborative control mechanism needs to be established among different levels, including the main grid, feeders, and distribution substations. Existing regional division methods mostly focus on voltage control or single-service scenarios, lacking a unified characterization of the impact of feeder integration on the dynamic characteristics of source and load. Furthermore, existing control methods are mostly concentrated on substation-level control, feeder-level control, or a dual-layer control system (feeder-substation level), with few solutions simultaneously considering multi-layer collaborative optimization across the main grid, feeder, and substation levels.

[0003] In scenarios with a high proportion of distributed resources, the continued use of centralized solutions often leads to a significant increase in computational and communication burdens, making it difficult to meet the real-time operation requirements of large-scale new distribution networks. Furthermore, existing zoning indicators primarily focus on network structure characteristics, neglecting the local active and reactive power balancing capabilities within a region, resulting in low adjustment efficiency of the zoning results in actual operation. Therefore, there is an urgent need to propose a novel multi-level autonomous control method for distribution networks that can simultaneously consider equivalent modeling of transformer substations, functional zoning at the feeder level, and coordinated distributed solution for main and distribution systems, in order to improve the rationality of zoning results and control execution efficiency. Summary of the Invention

[0004] To address the aforementioned technical problems in related technologies, this invention proposes a novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, which can overcome the above-mentioned shortcomings of existing technologies.

[0005] To achieve the above-mentioned technical objectives, the technical solution of the present invention is implemented as follows: A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination; This novel multi-level autonomous control method for distribution networks, oriented towards master-distribution coordination, includes the following steps: Acquire network topology data, node and branch parameters, main grid boundary parameters, distributed photovoltaic parameters, energy storage parameters, load forecast data, source-load forecast data, and node operating status data of the distribution network; An internal autonomous equivalent model is established at the transformer substation level. The distributed photovoltaic and energy storage systems within the transformer substation are optimized with the goal of minimizing the daily operating cost of the substation. The time-series active and reactive power interaction results between the transformer substation level and the feeder level are obtained. Electrical distance index, active power balance index, reactive power balance index, and comprehensive modularity zoning index are constructed in the feeder layer, and autonomous regions are divided in the feeder layer according to the comprehensive modularity zoning index. Based on the results of feeder autonomous region division, a main grid-multi-feeder region collaborative optimization scheduling model is established to obtain the objective function and operational constraints of each feeder autonomous region. A distributed solution algorithm based on the alternating direction multiplier method is used to iteratively solve the cooperative optimization scheduling model, update the inter-regional boundary interaction variables and Lagrange multipliers, and obtain the dynamic power interaction results of the main network layer, feeder layer and transformer area layer, as well as the distributed resource output results. Based on changes in the operating status of the distribution network, the weights of various items in the comprehensive modularity zoning index are adaptively adjusted to achieve a new type of multi-level autonomous control of the distribution network oriented towards main and distribution coordination.

[0006] Furthermore, the autonomous equivalent model within the transformer substation layer includes power exchange constraints between the feeder layer and the transformer substation layer, the objective function of the transformer substation layer, and operational constraints of the transformer substation layer; wherein, the objective function of the transformer substation layer includes curtailment penalty cost, distributed energy storage operation and maintenance cost, and power interaction penalty cost between the transformer substation layer and the feeder layer; the operational constraints of the transformer substation layer include power balance constraints, energy storage charging and discharging constraints, and distributed photovoltaic output constraints.

[0007] Furthermore, the autonomous region division of the feeder layer includes: The modularity function is used to characterize the structural characteristics of strong coupling within a region and weak coupling between regions. The electrical distance between nodes is calculated based on the node voltage-reactive power sensitivity, and the edge weights in the weighted network are determined according to the electrical distance, wherein the electrical distance is expressed as the electrical distance. The calculation formula is: ; In the formula, , , respectively, represent the ratio of the degree of impact on the voltage of nodes i and j when the reactive power of node β changes, and N is the total number of nodes in the distribution network; Construct active power balance index and reactive power balance index to reflect the region's local supply and demand matching capability; A comprehensive modularity partitioning index is constructed based on the modularity function, active power balance index, and reactive power balance index. With the goal of maximizing the comprehensive modularity partitioning index, a genetic algorithm is used to solve the region partitioning problem, and the autonomous region partitioning result of the feeder layer is obtained.

[0008] Furthermore, the active power balance index is described based on typical time-varying scenarios of the network and is expressed as the ratio of the region's net active power to the maximum value of the region's active power demand; the reactive power balance index is expressed as the ratio of the region's net reactive power to the region's reactive power load demand value.

[0009] Furthermore, the main grid-multi-feeder regional collaborative optimization scheduling model aims to minimize the daily operating cost of each feeder autonomous region. The daily operating cost includes the cost of new energy abandonment penalty, energy storage operation and maintenance cost, main grid power purchase cost, and grid loss cost. The operating constraints include DistFlow power flow constraints, power balance constraints, energy storage state of charge constraints, node voltage constraints, photovoltaic power output constraints, and upper and lower limits of main grid-feeder interactive power constraints.

[0010] Furthermore, the distributed solution algorithm based on the alternating direction multiplier method includes: A distributed optimization sub-model with boundary consistency constraints is established for each feeder autonomous region; Construct the Lagrange augmented function corresponding to the sub-objectives of each region; The decision variables within each sub-region are updated in parallel to obtain the region boundary coupling variables; Update the region boundary reference value based on the boundary coupling variables; Update the Lagrange multipliers based on the updated region boundary reference values; The algorithm is judged to meet the convergence condition based on the original residual and the dual residual. If the convergence condition is met, the distributed coordinated optimization result is output.

[0011] Furthermore, a penalty parameter is introduced into the Lagrange augmented function, and the convergence accuracy of the original residual and the dual residual is preset to 10. -4 .

[0012] Furthermore, the changes in operating status include changes in distributed photovoltaic output, changes in energy storage state of charge, load fluctuations, voltage deviations, and changes in grid-feeder exchange power.

[0013] Furthermore, the weights in the comprehensive modularity partitioning index are adaptively adjusted according to the current operating status, and the current autonomous region division result is judged to meet the structural and functional requirements in combination with the preset update threshold.

[0014] Furthermore, when it is determined that the current autonomous region division result does not meet the structural and functional requirements, the steps of feeder layer autonomous region division and main grid-multi-feeder region collaborative optimization scheduling are re-executed to realize the dynamic reconstruction and rolling optimization control of multi-level autonomous regions.

[0015] The beneficial effects of this invention are as follows: By establishing an internal autonomous equivalent model at the transformer substation level, the dimensionality of optimization variables at the feeder level is effectively reduced; by constructing a comprehensive zoning index that integrates electrical distance, active power, and reactive power balance, the autonomous regions at the feeder level can maintain strong structural coupling while improving the local load-source balancing capability within the region; by establishing a main grid-multi-feeder region collaborative optimization scheduling model and using the alternating direction multiplier method for distributed solution, efficient decoupling and rapid convergence of boundary variables between different levels are achieved; furthermore, by adaptively adjusting zoning weights according to changes in operating status and triggering regional rolling reconfiguration, the operational economy, voltage safety, and multi-level coordinated control efficiency of the distribution network under high-proportion distributed resource access scenarios are significantly improved. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram illustrating the power exchange relationship between the feeder layer and the substation layer. Figure 2 This is a schematic diagram illustrating the change of the comprehensive regional division index with the number of iterations. Figure 3 A schematic diagram of the improved IEEE 33-node system architecture and controllable resource configuration; Figure 4 This is a schematic diagram of the power balance results in the transformer substation area. Figure 5 A schematic diagram showing the voltage comparison of 33 nodes in the distribution network before and after optimization; Figure 6 A diagram illustrating the comparison of network loss before and after optimization; Figure 7 A schematic diagram showing the optimization results for photovoltaic power plants and energy storage power plants; Figure 8 This is a schematic diagram showing the change curves of the objective function in each region of the feeder layer during the distributed coordination process. Figure 9 This is a schematic diagram of the convergence process of the original residual and the dual residual. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0019] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0020] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] like Figures 1 to 9 As shown in the embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination includes the following steps: Acquire network topology data, node and branch parameters, main grid boundary parameters, distributed photovoltaic parameters, energy storage parameters, load forecast data, source-load forecast data, and node operating status data of the distribution network; An internal autonomous equivalent model is established at the transformer substation level. The distributed photovoltaic and energy storage systems within the transformer substation are optimized with the goal of minimizing the daily operating cost of the substation. The time-series active and reactive power interaction results between the transformer substation level and the feeder level are obtained. Electrical distance index, active power balance index, reactive power balance index, and comprehensive modularity zoning index are constructed in the feeder layer, and autonomous regions are divided in the feeder layer according to the comprehensive modularity zoning index. Based on the results of feeder autonomous region division, a main grid-multi-feeder region collaborative optimization scheduling model is established to obtain the objective function and operational constraints of each feeder autonomous region. A distributed solution algorithm based on the alternating direction multiplier method is used to iteratively solve the cooperative optimization scheduling model, update the inter-regional boundary interaction variables and Lagrange multipliers, and obtain the dynamic power interaction results of the main network layer, feeder layer and transformer area layer, as well as the distributed resource output results. Based on changes in the operating status of the distribution network, the weights of various items in the comprehensive modularity zoning index are adaptively adjusted to achieve a new type of multi-level autonomous control of the distribution network oriented towards main and distribution coordination.

[0022] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards main-distribution coordination is described. In a specific implementation, the internal autonomous equivalent model of the transformer substation layer includes power exchange constraints between the feeder layer and the transformer substation layer, an objective function for the transformer substation layer, and operational constraints for the transformer substation layer. The objective function for the transformer substation layer includes curtailment penalty costs, distributed energy storage operation and maintenance costs, and power interaction penalty costs between the transformer substation and the feeder layer. The operational constraints for the transformer substation layer include power balance constraints, energy storage charging and discharging constraints, and distributed photovoltaic output constraints.

[0023] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination is described. In a specific implementation, the division of the autonomous region at the feeder layer includes: The modularity function is used to characterize the structural characteristics of strong coupling within a region and weak coupling between regions. The electrical distance between nodes is calculated based on the node voltage-reactive power sensitivity, and the edge weights in the weighted network are determined according to the electrical distance, wherein the electrical distance is expressed as the electrical distance. The calculation formula is: ; In the formula, , , respectively, represent the ratio of the degree of impact on the voltage of nodes i and j when the reactive power of node β changes, and N is the total number of nodes in the distribution network; Construct active power balance index and reactive power balance index to reflect the region's local supply and demand matching capability; A comprehensive modularity partitioning index is constructed based on the modularity function, active power balance index, and reactive power balance index. With the goal of maximizing the comprehensive modularity partitioning index, a genetic algorithm is used to solve the region partitioning problem, and the autonomous region partitioning result of the feeder layer is obtained.

[0024] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards primary and secondary coordination is provided. In a specific implementation, the active power balance index is described based on typical time-varying scenarios of the network and is expressed as the ratio of the region's net active power to the region's maximum active power demand; the reactive power balance index is expressed as the ratio of the region's net reactive power to the region's reactive power load demand.

[0025] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards main grid-feeder coordination is described. In a specific implementation, the main grid-multi-feeder regional collaborative optimization scheduling model aims to minimize the daily operating cost of each feeder autonomous region. The daily operating cost includes the cost of new energy abandonment penalty, energy storage operation and maintenance cost, main grid power purchase cost, and grid loss cost. The operating constraints include DistFlow power flow constraints, power balance constraints, energy storage state of charge constraints, node voltage constraints, photovoltaic power output constraints, and upper and lower limits of main grid-feeder interactive power constraints.

[0026] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination is described. In a specific embodiment, the distributed solution algorithm based on the alternating direction multiplier method includes: A distributed optimization sub-model with boundary consistency constraints is established for each feeder autonomous region; Construct the Lagrange augmented function corresponding to the sub-objectives of each region; The decision variables within each sub-region are updated in parallel to obtain the region boundary coupling variables; Update the region boundary reference value based on the boundary coupling variables; Update the Lagrange multipliers based on the updated region boundary reference values; The algorithm is judged to meet the convergence condition based on the original residual and the dual residual. If the convergence condition is met, the distributed coordinated optimization result is output.

[0027] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination is provided. In a specific implementation, a penalty parameter is introduced into the Lagrange augmented function, and the convergence accuracy of the original residual and the dual residual is preset to 10. -4 .

[0028] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards main grid-feeder coordination is described. In a specific implementation, the changes in operating status include changes in distributed photovoltaic power output, changes in energy storage state of charge, load fluctuations, voltage deviations, and changes in power exchanged between the main grid and feeders.

[0029] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards main-distribution coordination is described. In a specific implementation, the weights in the comprehensive modularity partitioning index are adaptively adjusted according to the current operating status, and a preset update threshold is used to determine whether the current autonomous region division result meets the structural and functional requirements.

[0030] According to an embodiment of the present invention, a novel multi-level autonomous control method for distribution networks oriented towards main grid-multi-feeder coordination is described. In a specific implementation, when it is determined that the current autonomous region division result does not meet the structural and functional requirements, the steps of feeder layer autonomous region division and main grid-multi-feeder region collaborative optimization scheduling are re-executed to realize dynamic reconstruction and rolling optimization control of multi-level autonomous regions.

[0031] To facilitate understanding of the above technical solutions of the present invention, the following detailed description of the above technical solutions of the present invention is provided through specific embodiments and working principles.

[0032] Example 1 S1. Obtain network topology data, node and branch parameters, source-load prediction data, and node operating status data for the main grid layer, feeder layer, and transformer area layer. The source-load prediction data includes at least distributed photovoltaic power output prediction, load prediction, and main grid boundary price information.

[0033] The acquired data is standardized, time-series scenario-based, and equipment parameter-verified to form an input dataset for subsequent autonomous modeling of transformer substations, feeder layer partitioning, and main-distributor collaborative optimization.

[0034] S2. At the coupling node between the transformer substation layer and the feeder layer, establish active power and reactive power exchange constraints between the feeder layer and the transformer substation layer.

[0035] Equation (1) In the formula, , These represent the active and reactive power switched on the feeder layer side, respectively. , These represent the active and reactive power exchanged at the transformer substation level, respectively, and M represents the total number of transformer substations.

[0036] Construct an objective function at the transformer substation level to minimize the daily operating cost F of the substation; the objective function at the transformer substation level includes at least the cost of curtailment penalty. Distributed energy storage operation and maintenance costs and the power interaction penalty cost between the transformer area and the feeder layer .

[0037] Equation (2) Equation (3) Equation (4) Equation (5) In the formula: This is the cost coefficient for the penalty of abandoning light; Let t be the curtailed power of the l-th photovoltaic cell within the distribution area at time t; T is the time length of a typical scenario. This refers to the number of distributed photovoltaic systems within the transformer substation area. and These represent the charging and discharging power of the s-th distributed energy storage unit at time t, respectively. The operation and maintenance cost per unit power of energy storage; The quantity of distributed energy storage within the transformer substation area; The power interaction penalty cost coefficient between the transformer area and the feeder side of the upstream power grid at time t; This represents the actual value of the interaction power between the station layer and the feeder layer at time t.

[0038] Establish power balance constraints, energy storage charging and discharging constraints, and distributed photovoltaic power output constraints at the transformer substation level.

[0039] Equation (6) In the formula: The actual active power output of the distributed photovoltaic system within the transformer area at time t; Let t be the total load demand within the transformer area at time t.

[0040] Equation (7) In the formula: The energy state of stored energy s at time t; , Δt represents the initial energy state and the final energy state of energy storage s, respectively; Δt is the energy storage charging and discharging duration. , These are the charging and discharging efficiencies of energy storage s, respectively. and These are the minimum and maximum values ​​of the state of charge, respectively; , Let t represent the charging and discharging states of the stored energy s at time t. Within a time period, the stored energy has only one state, either charging or discharging, represented by 0 and 1. , These are the maximum values ​​of the charging power and discharging power of the energy storage s, respectively.

[0041] Equation (8) In the formula: Let t be the maximum output of the distributed photovoltaic system within the transformer substation.

[0042] By autonomously optimizing the resources within the transformer substations at different time periods, the equivalent time-series active and reactive power transferred from the substations to the feeder layer are obtained; the power exchange relationship between the substation layer and the feeder layer can be found in [reference needed]. Figure 1 .

[0043] S3. In the feeder layer, construct the modularity function, voltage-reactive power sensitivity and their corresponding node electrical distance models, and determine the edge weights of the weighted network based on the node electrical distance.

[0044] Equation (9) Equation (10) In the formula: χ is the modularity function; Ω is the set of nodes; Let be the weight of the edge connecting nodes i and j; This is the sum of the weights of the edges connected to node i; δ(i,j) is the sum of the weights of the edges connected to node j; σ is the sum of the weights of the network edges; when i and j are in the same partition, δ(i,j) is 1, otherwise it is 0.

[0045] Electrical distance. Voltage-reactive power sensitivity is used. Calculating the electrical distance, which reflects the degree of electrical coupling between network nodes, can be done by first calculating the voltage sensitivity of each node using the Newton-Raphson method, and then calculating the electrical distance between the nodes. Equation (11) In the formula: The magnitude change of node voltage i; Let J be the reactive power change at node j; The percentage change in the voltage at node i is the ratio of the change in reactive power at node j to the change in reactive power at node j. , The values ​​represent the effects of reactive power changes at node j on the voltage amplitudes at nodes i and j, respectively. The larger, The smaller the value, the greater the impact of the reactive power change at node j on the voltage amplitude at node i.

[0046] In practice, the electrical coupling between two nodes depends not only on the two nodes themselves, but also on the power injected by other nodes in the network and the node locations. If there are N nodes in the distribution network, then under the combined influence of other nodes, the electrical distance between nodes i and j... for Equation (12) In the formula: , These represent the ratio changes in the degree of impact on the voltage at nodes i and j when the reactive power at node β changes.

[0047] Considering the system's electrical distance index to determine the edge weights, that is: Equation (13) Further develop reactive power balance index, active power balance index, and comprehensive modularity zoning index.

[0048] Reactive power balance index The ability to balance reactive power locally is specifically expressed as Equation (14) Equation (15) In the formula: This represents the number of feeder layer regions. Let ν be the net reactive power of region ν at time t; The maximum reactive power provided by region ν at time t; Let ν be the reactive power load demand value of region ν at time t.

[0049] Active power balance index It is an indicator that describes typical time-varying network scenarios, specifically expressed as: Equation (16) Equation (17) In the formula: Let be the net active power of the region at time t; The maximum active power provided to region ν at time t is the active power demand of region ν at time t.

[0050] The comprehensive modularity partitioning index γ is expressed as Equation (18) Equation (19) In the formula: , , These are the weighting coefficients for different indicators.

[0051] Due to the time-varying nature of power generation and load output, the regional division results will also change in real time. The time of day with the highest photovoltaic penetration rate, 13:00, is selected for regional division, at which time the photovoltaic penetration rate is 120%. The weights corresponding to the three proposed indicators can be selected according to the scheduling needs of the dispatch center. In this example, the weights of the modularity indicator, active power balance indicator, and reactive power balance indicator are 0.6, 0.2, and 0.2, respectively.

[0052] A genetic algorithm was used for region partitioning. The maximum number of iterations was set to 500, the population size to 40, the crossover probability to 0.3–0.6, and the mutation probability to (0.1, 0.5). To ensure convergence, the two best individuals after each iteration were not subjected to crossover or mutation. The resulting curve shows the change in the comprehensive region partitioning index with the number of iterations. As can be seen from the figure, the comprehensive partitioning index reached its maximum value of 0.869 after approximately 172 iterations. For details on the change in the comprehensive region partitioning index with the number of iterations, please refer to [link to relevant documentation]. Figure 2 .

[0053] S4. For each feeder autonomous region, establish a collaborative optimization scheduling model with the goal of minimizing daily operating costs.

[0054] Taking into account both the safety and economy of power system operation, the daily operating cost within each region is considered. If the objective is to minimize the output of adjustable resources in each region, and the decision variables are the output values ​​of adjustable resources in each region, then the objective function for the ν-th feeder region is: Equation (20) In the formula: The cost of abandoning new energy sources in the region; The cost of energy storage operation and maintenance for the region; The cost of purchasing electricity from the main grid in the region; The network loss cost for region ν.

[0055] The penalty cost for abandoning new energy is Equation (21) In the formula: The number of photovoltaic power stations within region ν; Cost per unit of electricity generation loss for photovoltaic power plants; Let be the predicted active power output of the photovoltaic power station at node j at time t; Let t represent the actual active power output of the photovoltaic power station at node j at time t. Let be the curtailed power of the photovoltaic power station at node j at time t.

[0056] The operating cost of an energy storage power station is Equation (22) In the formula: The number of internal energy storage power stations in region ν; The operating cost per unit of energy storage power station; and These represent the charging power and discharging power of the energy storage power station at node j at time t, respectively.

[0057] The cost of purchasing electricity from the main grid is Equation (23) In the formula: The active power purchased by region ν from the upper-level main grid at time t; Let t be the real-time electricity price of the power grid at time t.

[0058] Network loss cost is Equation (24) In the formula: The number of nodes within region ν; Let ν be the total network loss in region ν at time t; Cost per unit of network loss; Let be the resistance of the line between nodes i and j; Let be the current flowing between nodes i and j; , These represent the active power and reactive power flowing from upstream node i to node j, respectively. Let be the voltage at node i.

[0059] The collaborative optimization scheduling model further includes DistFlow power flow constraints, second-order cone relaxation constraints, power balance constraints, energy storage state of charge constraints, node voltage constraints, photovoltaic output constraints, and upper and lower limits of grid-feeder interaction power constraints.

[0060] Distflow current constraints are Equation (25) In the formula: i∈ζ(j) and i∈ψ(j) are the sets of first (last) nodes of the branch to which node i belongs, with node j as the last (first) node; Xij is the line reactance between nodes i and j; Pij, and Qij,t are the active and reactive power flowing from upstream node i to node j at time t; Pjk,t and Qjk,t are the active and reactive power transferred between nodes j and k at time t; Pj,t and Qj,t are the net active and reactive power injected into node j at time t; Iij,t is the current in branch ij at time t; and Ui,t and Uj,t are the upstream and downstream node voltages at time t.

[0061] Equation (25) is a nonlinear constraint, therefore, according to the second-order cone relaxation principle, it needs to be converted into an inequality constraint, i.e. Equation (26) Power balance constraint is Equation (27) In the formula: , These are the active power and reactive power of the load at node j at time t, respectively. For node j at time t, the power it purchases from the main grid. Let be the reactive power output of the photovoltaic system at node j at time t.

[0062] The state of charge constraint of the energy storage power station is Equation (28) In the formula: Let represent the energy state of the energy storage power station at node j at time t. , These are the charging and discharging efficiencies of the energy storage station at node j, respectively. and These represent the initial energy state and the end-of-cycle energy state of the energy storage power station at node j, respectively. and These represent the maximum charging and discharging power of the energy storage station at node j, respectively. and These are the minimum and maximum values ​​of the charge state at node j, respectively; , , , represent the charging and discharging states of the energy storage power station at node j at time t. Within a time period, the energy storage power station only has charging or discharging states, represented by 0 and 1 respectively.

[0063] Node voltage constraints are Equation (29) In the formula: and These are the minimum and maximum limits for the square of the voltage at node i, respectively; This refers to the voltage amplitude at the substation's outlet. This is the system's rated voltage amplitude.

[0064] Photovoltaic power plant output constraint Equation (30) In the formula: , These represent the maximum active and reactive power outputs of the photovoltaic power station at node j, respectively.

[0065] The upper and lower limits of the main grid-feeder interaction power are respectively Equation (31) In the formula: Let ν be the interaction power between the feeder region and the main grid at time t; and These are the minimum and maximum values ​​of active power exchanged between the feeder region ν and the main grid, respectively. and These represent the minimum and maximum values ​​of reactive power exchanged between the feeder region ν and the main grid, respectively.

[0066] S5. Establish a distributed optimization sub-model with boundary consistency constraints for each partitioned region.

[0067] To achieve friendly interaction between regions and improve the low computational efficiency caused by centralized control, a distributed inter-region solution algorithm based on ADMM is proposed. For partitioning region 'a', the distributed optimization model based on the partition coordination principle can be written as follows: Equation (32) In the formula: Let a be the objective function for region a; For the equality constraints of region a; Let be the decision variables for region a; There are inequality constraints on region a; there exists an inequality constraint between region a and its neighboring region b. , , Boundary equality constraints are used as convergence conditions for distributed coordinated optimization between regions.

[0068] The steps of the interval distributed coordination optimization algorithm based on ADMM are as follows.

[0069] Step 1: Initialization. Set the initial values ​​of global variables based on the measured data of the distribution network, and set the initial values ​​of the Lagrange multipliers and voltage compensation parameters of all area boundary data to 0.

[0070] Step 2: Establish the Lagrange augmented function corresponding to each regional sub-objective. Taking region a as an example, the adjacent region is region b. There is a boundary equality constraint relationship between region a and region b. Then the Lagrange augmented function corresponding to region a and region b is... , It is represented as.

[0071] Equation (33) Equation (34) Equation (35) In the formula: k is the number of iterations; ρ is the penalty parameter; For control variables within region b; , Let A and B be the autonomous objective functions within regions a and b, respectively. , These are the interaction variables between regions a and b, obtained from the kth iteration optimization calculation within regions a and b, respectively. , These are the fixed reference values ​​for the (k+1)th iteration of regions a and b, respectively. , These are the Lagrange multipliers for regions a and b corresponding to the k-th iteration, respectively.

[0072] Step 3: Update the decision variables within each sub-region. Starting from k=1 iteration, in the (k+1)th iteration, the Lagrange augmented function within each region is solved in parallel to obtain the decision variables within the region at the corresponding iteration step size. Simultaneously, the coupling variables at the region boundaries can also be obtained. and ,Right now Equation (36) Step 4: Update the boundary variables between each region based on the boundary values ​​obtained in Step 3. and The average value of the coupled branch state can be calculated using equation (36) and used as a reference value for the next iteration.

[0073] Step 5: Update the Lagrange multipliers for each region, i.e. Equation (37) In the formula, , These are the Lagrange multipliers for regions a and b corresponding to the (k+1)th iteration, respectively.

[0074] Step 6: Determine if the algorithm has converged. The convergence criterion is the original residual. and dual residual Whether it tends to 0, the residual is defined as Equation (38) In the formula, δ represents the convergence precision. If the convergence condition is met, the algorithm terminates; otherwise, it returns to step 2.

[0075] S6. Monitor operational status information such as changes in distributed photovoltaic output, changes in energy storage state of charge, load fluctuations, voltage deviations, and changes in power exchanged between the main grid and feeders.

[0076] Based on the current operating status, the weights in the comprehensive modularity partitioning index are adaptively adjusted, and the current autonomous region division results are judged in combination with the preset update threshold to determine whether they still meet the structural and functional requirements.

[0077] If the judgment result does not meet the requirements, repeat steps S3 to S5 to complete the dynamic reconstruction and rolling optimization control of the autonomous region.

[0078] In one specific embodiment, the method of the present invention is verified using an improved IEEE 33-bus distribution system. A schematic diagram of the system structure and controllable resource configuration can be found in [reference needed]. Figure 3 .

[0079] Power flow calculations used per-unit values, with the IEEE 33-bus system reference power set at 100 MV·A and the voltage reference value at 12.66 kV. According to the proposed indicators and algorithm, the distribution network was divided into three regions, as shown in Table 1. The modularity function in the comprehensive indicator function was 0.9504, meeting the regional structural requirements and ensuring strong coupling within regions and weak coupling between regions. The reactive power balance index was 0.8786, indicating good reactive power balance in each region; the active power balance index was 0.6151, also indicating good active power balance in each region; the optimal number of regions was three, which was reasonable; each region included both energy storage and photovoltaics, enabling photovoltaic-storage complementarity and meeting functional requirements. The location and capacity parameters of adjustable resources in each region are shown in Table 2.

[0080] Table 1. Regional Division Results

[0081] Table 2. Controllable resource allocation in the distribution network

[0082] The optimized scheduling results for the three transformer substations connected to the feeder layer are as follows: Figure 4 As shown. By Figure 4 It can be seen that during the periods of 01:00-04:00 and 16:00-18:00, the load within the transformer substation is at its lowest, electricity prices are low, and energy storage is charging to store electrical energy. During the periods of 11:00-15:00 and 19:00-21:00, residential load is at its peak, and the output of distributed photovoltaic power fluctuates significantly, so energy storage discharges to ensure power balance within the substation. Based on the power interaction between the transformer substation and the feeder layer, it can be observed that roughly between 12:00-15:00, because distributed photovoltaic power is near full capacity and cannot be absorbed by the substation, the substation feeds power back to the feeder layer, and the substation as a whole is in a power supply state. During other periods, in order to ensure the lowest overall operating cost for the substation, it is necessary to purchase electricity from the feeder layer when electricity prices are low to ensure the economic efficiency of the substation's operation. At this time, the substation is equivalent to a load connected to the feeder layer.

[0083] The system utilizes an inter-regional distributed coordinated control method to optimize the control of adjustable resources within a sub-region, yielding voltage values ​​before and after optimization, such as... Figure 5 As shown, network loss is compared to, for example Figure 6 As shown.

[0084] Depend on Figure 5 As can be seen, before optimization, the overall voltage of the system was too high due to the influence of photovoltaic output, with the highest voltage amplitude located at node 17, reaching 1.0638 at 12:00. After optimization, the voltage was within a safe range, and the grid loss decreased to varying degrees at different times. Figure 6As can be seen, the optimized network loss was reduced by 88.654 kW compared to the unoptimized version, which improved the economic efficiency of system operation and proved the effectiveness of the distributed coordination control strategy. Figure 7 The diagram shows the active power output of photovoltaic power stations and the output of energy storage power stations in each area of ​​the feeder layer after the autonomous optimization within the feeder layer area. During the peak load period from 18:00 to 21:00, the energy storage power stations release electricity to supply themselves and other loads in the system. During the daytime, the photovoltaic power stations are basically operating at full capacity, and the charging of the energy storage power stations promotes the system's absorption of photovoltaic power.

[0085] With an initial value of ρ of 0.5 and a convergence precision δ of 10⁻⁴, the changes in the objective functions F1, F2, and F3 of the three regions during the distributed coordination optimization process are as follows: Figure 8 As shown. By Figure 8 It can be seen that in the distributed coordination optimization process, each region will continuously adjust the active and reactive power output of adjustable resources within the region, eventually converging to the global optimal solution.

[0086] The ADMM algorithm is an iterative process that requires regions to cooperate and exchange boundary information. Therefore, the faster the convergence speed, the fewer times information needs to be exchanged at the region boundaries, and the smaller the communication burden on the entire system. Figure 9 As shown, the 33-node system basically reached convergence after 38 iterations. Since the objective function and constraints of the sub-region itself are linear programming problems, the solution is relatively fast, which is conducive to the rapid convergence of multiple regions.

[0087] To verify the accuracy and effectiveness of distributed coordinated control between regions, the control effect of the improved IEEE 33-node example was compared with that of the centralized algorithm control, as shown in Tables 3 and 4. The comparison time was chosen to be 12:00, at which time power backfeeding and voltage over-limit are relatively severe.

[0088] Table 3. Comparison of output between centralized control and distributed control

[0089] Table 4. Cost Comparison of Centralized and Distributed Algorithms

[0090] Tables 3 and 4 show that the results of distributed coordinated control and centralized optimization control are very close overall. The total time of the centralized algorithm is 36.908s, while the total time of the distributed coordinated control algorithm is 34.225s. This proves that the distributed coordinated control based on the ADMM algorithm can achieve rapid convergence of optimization control within each sub-region and regional boundary data through continuous optimization within the region, power interaction between regions, and alternating updates of virtual balance nodes.

[0091] Simulation results show that the method of the present invention can reduce system network loss while ensuring voltage safety constraints, and achieve efficient coordinated control of multi-level and multi-regional resources.

[0092] In summary, by utilizing the technical solutions described above, the dimensionality of optimization variables at the feeder layer is effectively reduced by establishing an internal autonomous equivalent model at the transformer substation level. By constructing a comprehensive zoning index that integrates electrical distance, active power, and reactive power balance, the autonomous regions at the feeder layer maintain strong structural coupling while enhancing local load balancing capabilities within the region. Furthermore, by establishing a main grid-multi-feeder region collaborative optimization scheduling model and employing the alternating direction multiplier method for distributed solution, efficient decoupling and rapid convergence of boundary variables between different levels are achieved. Finally, by adaptively adjusting zoning weights based on changes in operating status and triggering regional rolling reconfiguration, the operational economy, voltage safety, and multi-level coordinated control efficiency of the distribution network under high-proportion distributed resource access scenarios are significantly improved.

[0093] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, characterized in that, Includes the following steps: Acquire network topology data, node and branch parameters, main grid boundary parameters, distributed photovoltaic parameters, energy storage parameters, load forecast data, source-load forecast data, and node operating status data of the distribution network; An internal autonomous equivalent model is established at the transformer substation level. The distributed photovoltaic and energy storage systems within the transformer substation are optimized with the goal of minimizing the daily operating cost of the substation. The time-series active and reactive power interaction results between the transformer substation level and the feeder level are obtained. Electrical distance index, active power balance index, reactive power balance index, and comprehensive modularity zoning index are constructed in the feeder layer, and autonomous regions are divided in the feeder layer according to the comprehensive modularity zoning index. Based on the division of feeder autonomous regions, a collaborative optimization scheduling model of main grid-multi-feeder regions is established to obtain the objective function and operational constraints of each feeder autonomous region. A distributed solution algorithm based on the alternating direction multiplier method is used to iteratively solve the cooperative optimization scheduling model, update the inter-regional boundary interaction variables and Lagrange multipliers, and obtain the dynamic power interaction results of the main network layer, feeder layer and transformer area layer, as well as the distributed resource output results. Based on changes in the operating status of the distribution network, the weights of various items in the comprehensive modularity zoning index are adaptively adjusted to achieve a new type of multi-level autonomous control of the distribution network oriented towards main and distribution coordination.

2. The novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination as described in claim 1, characterized in that, The autonomous equivalent model within the transformer substation layer includes power exchange constraints between the feeder layer and the transformer substation layer, the objective function of the transformer substation layer, and operational constraints of the transformer substation layer. The objective function of the transformer substation layer includes curtailment penalty costs, distributed energy storage operation and maintenance costs, and power interaction penalty costs between the transformer substation layer and the feeder layer. The operational constraints of the transformer substation layer include power balance constraints, energy storage charging and discharging constraints, and distributed photovoltaic output constraints.

3. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 1, is characterized in that... The autonomous region division of the feeder layer includes: The modularity function is used to characterize the structural characteristics of strong coupling within a region and weak coupling between regions. The electrical distance between nodes is calculated based on the node voltage-reactive power sensitivity, and the edge weights in the weighted network are determined according to the electrical distance, wherein the electrical distance is expressed as the electrical distance. The calculation formula is: ； In the formula, , , respectively, represent the ratio of the degree of impact on the voltage of nodes i and j when the reactive power of node β changes, and N is the total number of nodes in the distribution network; Construct active power balance index and reactive power balance index to reflect the region's local supply and demand matching capability; A comprehensive modularity partitioning index is constructed based on the modularity function, active power balance index, and reactive power balance index. With the goal of maximizing the comprehensive modularity partitioning index, a genetic algorithm is used to solve the region partitioning problem, and the autonomous region partitioning result of the feeder layer is obtained.

4. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 3, is characterized in that... The active power balance index is described based on typical time-varying network scenarios and is expressed as the ratio of the region's net active power to the region's maximum active power demand; the reactive power balance index is expressed as the ratio of the region's net reactive power to the region's reactive power load demand.

5. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 1, is characterized in that... The main grid-multi-feeder regional collaborative optimization scheduling model aims to minimize the daily operating cost of each feeder autonomous region. The daily operating cost includes the cost of new energy abandonment penalty, energy storage operation and maintenance cost, main grid power purchase cost, and grid loss cost. The operating constraints include DistFlow power flow constraints, power balance constraints, energy storage state of charge constraints, node voltage constraints, photovoltaic power output constraints, and upper and lower limits of main grid-feeder interactive power constraints.

6. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 1, is characterized in that... The distributed solution algorithm based on the alternating direction multiplier method includes: A distributed optimization sub-model with boundary consistency constraints is established for each feeder autonomous region; Construct the Lagrange augmented function corresponding to the sub-objectives of each region; The decision variables within each sub-region are updated in parallel to obtain the region boundary coupling variables; Update the region boundary reference value based on the boundary coupling variables; Update the Lagrange multipliers based on the updated region boundary reference values; The algorithm is judged to meet the convergence condition based on the original residual and the dual residual. If the convergence condition is met, the distributed coordinated optimization result is output.

7. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 6, is characterized in that... A penalty parameter is introduced into the Lagrange augmented function, and the convergence accuracy of the original residual and the dual residual is preset to 10. -4 .

8. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 1, is characterized in that... The changes in operating status include changes in distributed photovoltaic output, changes in the state of charge of energy storage, load fluctuations, voltage deviations, and changes in the power exchange between the main grid and the feeder.

9. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 8, is characterized in that, The weights in the comprehensive modularity partitioning index are adaptively adjusted based on the current operating status, and the current autonomous region division results are judged to meet the structural and functional requirements in combination with the preset update threshold.

10. A novel multi-level autonomous control method for distribution networks oriented towards master-distribution coordination, as described in claim 9, is characterized in that... When it is determined that the current autonomous region division result does not meet the structural and functional requirements, the steps of feeder layer autonomous region division and main grid-multi-feeder area collaborative optimization scheduling are re-executed to realize the dynamic reconstruction and rolling optimization control of multi-level autonomous regions.

Images