A method and system for risk perception of power transmission and distribution coordination considering uncertainty

Abstract

The application discloses a kind of considering uncertainty transmission and distribution collaborative risk perception clearing method and system.System data acquisition module constructs photovoltaic moment information fuzzy set of power transmission system layer under transmission and distribution collaborative architecture;Distribution network aggregation module constructs the bottom scheduling model of power distribution system layer, and the power aggregation step curve is output after pre-clearing;Joint clearing module establishes the joint clearing model of power transmission system layer based on deterministic solvable constraint, power aggregation step curve and photovoltaic moment information fuzzy set are input into model and are solved jointly optimized, and boundary node scheduling instruction is obtained;Instruction issuing module issues boundary node scheduling instruction to bottom scheduling model, and after processing, bottom equipment scheduling instruction is output to virtual power plant, and transmission and distribution collaborative risk perception clearing is realized.The application resists extreme uncertainty risk by distributed robust opportunity constraint, and unlocks bottom flexibility in combination with transmission and distribution collaborative architecture, improves the risk resistance ability of coping with out-of-distribution prediction error, and is efficient and strong in physical executability.

Description

Technical Field

[0001] This invention relates to a power system clearing method, which relates to the field of power system operation and dispatching technology, and specifically to a transmission and distribution coordination risk perception clearing method and system that takes into account uncertainties. Background Technology

[0002] Existing power clearing and dispatch methods mainly suffer from the following two problems:

[0003] First, the decoupled operation of transmission and distribution networks makes it difficult to effectively utilize flexibility to address global risks. Traditional dispatching models typically decouple decision-making between the transmission and distribution networks. This model ignores the physical constraints of node voltage and line capacity within the distribution network and fails to accurately determine the true available flexibility boundaries of underlying distributed resources. This results in the transmission network often being unable to effectively utilize flexible resources on the distribution side when facing significant system-level power deficit risks; or it may blindly issue dispatching instructions, thereby triggering over-limit risks on the distribution network side.

[0004] Second, there is a lack of risk perception mechanisms to cope with the extreme uncertainties of large-scale new energy sources. With a high proportion of new energy sources connected to the grid, the prediction errors of power output from photovoltaic and other sources often fail to conform to a precise known probability distribution.

[0005] In summary, existing methods for handling uncertainty (such as traditional stochastic programming or standard robust optimization) have significant drawbacks: when faced with conventional prediction errors, traditional methods often lead to extremely high operating costs in pursuit of absolute safety; while when faced with out-of-distribution (OOD) prediction biases with unknown probabilities (such as a sudden drop in photovoltaic output caused by extreme weather), they often lack sufficient backup support, cannot guarantee the physical safety of the system, and are prone to causing serious consequences such as large-scale load shedding. Summary of the Invention

[0006] To address the problems existing in the background technology, this invention provides a risk perception and clearing method for transmission and distribution coordination that takes into account uncertainty. This invention is a risk perception and coordination clearing method capable of penetrating information barriers in the transmission and distribution network and effectively addressing extreme uncertainties. By precisely aggregating flexible resources on the distribution side, this invention provides physically feasible backup support for the transmission side to resist extreme off-probability risks, thereby achieving optimal allocation of global energy and reserves while ensuring extremely high grid reliability.

[0007] The technical solution adopted in this invention is:

[0008] The present invention provides a risk perception and clearing method for transmission and distribution coordination that takes into account uncertainty, comprising:

[0009] Step 1) Establish a transmission and distribution coordination architecture that includes the transmission system layer, the distribution system layer, and the virtual power plant (VPP) layer. Obtain historical photovoltaic operation data of the transmission system layer, and then construct a fuzzy set of photovoltaic moment information for the transmission system layer.

[0010] Step 2) Construct the underlying scheduling model of the power distribution system layer, input the resource regulation parameters reported by the virtual power plant into the underlying scheduling model for pre-clearing, and output the power aggregation step curve after processing.

[0011] Step 3) Establish a joint clearing model for transmission and distribution coordinated energy and reserves based on deterministic solvable constraints at the transmission system layer. Input the power aggregation step curve and the fuzzy set of photovoltaic moment information into the joint clearing model for joint optimization and solution. After solving, the boundary node scheduling instructions are obtained and then sent to the underlying scheduling model for processing. The underlying equipment scheduling instructions are then output to the virtual power plant, thereby realizing transmission and distribution coordinated risk perception and clearing.

[0012] In step 1), the transmission and distribution coordination architecture is a three-layer physical architecture. This architecture includes three layers from top to bottom: the transmission system layer, the distribution system layer, and the virtual power plant (VPP) layer. The transmission system layer includes z transmission network nodes, where a transmission network nodes are connected to traditional generator sets or large-capacity centralized photovoltaic systems composed of several photovoltaic units, b transmission network nodes are connected to only one of their respective distribution system layers, and c transmission network nodes are connected to both their respective distribution system layers and the transmission network energy storage system, z = a + b + c. The distribution system layer includes one distribution network root node and several distribution network nodes. The distribution network root node is connected to the transmission network nodes of the transmission system layer. Each transmission network node is connected to its own rigid load, photovoltaic unit, distribution network-side energy storage, or virtual power plant layer. The photovoltaic unit, distribution network-side energy storage, and virtual power plant layer serve as heterogeneous bottom-layer flexible resources. The virtual power plant layer includes distributed resources and a virtual power plant dispatch center. Distributed resources include distributed generator sets and flexible loads. The virtual power plant layer is connected to the transmission network nodes through the virtual power plant dispatch center.

[0013] In step 1), the historical photovoltaic operation data at the transmission system layer includes the actual and predicted historical active power output of centralized photovoltaic systems, and the fuzzy set of photovoltaic moment information at the transmission system layer includes the random variable of relative prediction error. probability distribution .

[0014] relative prediction error random variable probability distribution During the period The mean of the relative prediction error between the actual and predicted values ​​of active power output of centralized photovoltaic systems. and standard deviation Input the energy of the transmission and distribution coordination and the reserve joint clearing model.

[0015] In step 2), the underlying scheduling model of the power distribution system layer includes the overall net efficiency objective function, as follows:

[0016]

[0017] in, This represents the total net efficiency of the power distribution system layer within the scheduling cycle of the transmission and distribution coordination architecture. Represents the set of time periods within a scheduling period; Represents a set of steps with predefined segments; Indicates time period No. The electricity input cost of flexible loads for distributed resources in the virtual power plant layer; Indicates time period No. The power generation output cost of distributed generator sets in the distributed resources of the virtual power plant layer; Indicates time period The net operating and reserve costs of energy storage on the distribution network side of the distribution system layer; Indicates time period The power input cost of rigid loads in the power distribution system layer; Indicates time period The cost of network losses at the power distribution system layer; Indicates time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer; Indicates time period Net active power injected at the boundary tie line connecting the distribution system layer and the transmission system layer; Indicates the scheduling time step.

[0018] Resource adjustment parameters include time period No. The power input cost of flexible loads in the virtual power plant layer And the power generation output cost of distributed generator sets of distributed resources .

[0019] Boundary node scheduling instructions are time periods Net active power injected at the boundary tie line connecting the distribution system layer and the transmission system layer and time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer .

[0020] The power aggregation step curve includes the power distribution system layer. During the period The first submission to the power transmission system layer Section power generation output cost parameters and power input cost parameters And the first equal-width discretization on the power output side Power capacity width of single-segment equal-width stepped section The first equal-width discretization of the power input side Power capacity width of single-segment equal-width stepped section .

[0021] The underlying scheduling model of the power distribution system layer is based on AC power flow operation constraints, and it performs scheduling for each time period within the scheduling cycle. Both are valid.

[0022] The underlying device scheduling instructions include time periods in the AC power flow operation constraints. Distributed to distribution network nodes Active power output command of distributed generator set at the location Power consumption instructions for flexible loads and energy storage charging power commands and energy storage discharge power command .

[0023] In step 3), the specific model for the joint clearing of transmission and distribution coordinated energy and reserves at the transmission system layer is as follows:

[0024]

[0025] in, Represents the set of time periods within a scheduling period; Indicates time period The total net efficiency of the coupling between each power distribution system layer; Indicates time period The input cost of resilient loads at the transmission system layer; and They represent time periods respectively. The input cost and preset discharge cost of the input power when the energy storage system of the power grid connected to the power transmission system layer is charged; Indicates time period The cost of secondary power generation output of traditional generator sets; This represents the marginal cost coefficient of centralized photovoltaic power generation. Indicates time period The day-ahead clearing power of centralized photovoltaic power; This represents the frequency regulation reserve cost of the transmission and distribution coordination architecture; Indicates the scheduling time step.

[0026] Time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer These are the dual variables of the power balance constraint at the boundary nodes.

[0027] In step 3), the deterministically solvable constraints are as follows:

[0028]

[0029]

[0030] in, For the transmission and distribution coordination architecture in time period Up-modulation reserve capacity; For time period Solar power output cleared day-ahead; For time period The current forecast for the active power output of centralized photovoltaic power generation is as follows; and They represent time periods respectively. The mean and standard deviation of the relative prediction error between the actual and predicted values ​​of the active power output of centralized photovoltaic power generation; This refers to the safety margin factor. This is the preset risk tolerance level.

[0031] The present invention provides a transmission and distribution coordination risk perception and clearing system that takes into account uncertainties, comprising:

[0032] The data acquisition module is used to establish a transmission and distribution coordination architecture that includes the transmission system layer, the distribution system layer, and the virtual power plant layer, to obtain historical photovoltaic operation data of the transmission system layer, and then to construct a fuzzy set of photovoltaic moment information for the transmission system layer.

[0033] The distribution network aggregation module is used to construct the underlying scheduling model of the distribution system layer. It inputs resource adjustment parameters into the underlying scheduling model for pre-clearing and outputs a power aggregation step curve after processing.

[0034] The joint clearing module is used to establish a joint clearing model for transmission and distribution coordinated energy and reserves based on deterministic solvable constraints at the transmission system layer. The power aggregation step curve and the fuzzy set of photovoltaic moment information are input into the joint clearing model for joint optimization and solution to obtain the boundary node scheduling instructions.

[0035] The instruction issuance module is used to issue boundary node scheduling instructions to the underlying scheduling model. After processing, the underlying scheduling model outputs underlying equipment scheduling instructions to the virtual power plant, thereby realizing the risk perception and clearing of transmission and distribution coordination.

[0036] The electronic device of the present invention includes: a memory and a processor coupled to each other, wherein the memory stores program data, and the processor invokes the program data to execute the method described above.

[0037] The present invention provides a computer-readable storage medium having program data stored thereon, which, when executed by a processor, implements the method described above.

[0038] The beneficial effects of this invention are:

[0039] 1) Ensure the physical deliverability of underlying resources and improve scalability: Under the premise of strictly meeting the physical constraints of AC power flow, the distribution system performs deterministic step curve aggregation and equivalence of underlying distributed resources. This effectively avoids the direct uploading of underlying user privacy data, reduces the communication burden, and ensures the physical execution of flexible resources reported to the transmission system. It also reduces underlying safety hazards such as voltage overruns caused by macro-dispatch commands on the distribution network side.

[0040] 2) Balancing system safety and efficiency: The introduction of distributed bar opportunity constraints addresses the extreme uncertainty of large-capacity photovoltaic output on the main grid side. Without assuming that the prediction error follows a precise probability distribution, it can quantify and reserve system backup capacity in extreme scenarios such as out-of-distribution deviations. This effectively overcomes the shortcomings of traditional deterministic methods, such as the high risk of load shedding in extreme scenarios, and the low system efficiency caused by the overly conservative nature of traditional robust optimization.

[0041] 3) Ensuring model solution efficiency and engineering timeliness: The one-sided Cantelli-Chebyshev inequality is adopted to reconstruct the originally difficult-to-solve Bruker chance constraint into a deterministic solvable optimization problem, which improves computational efficiency, meets the computational timeliness requirements of actual day-ahead clearing, and has strong engineering application value.

[0042] 4) Achieve global coordinated optimization of transmission and distribution: By establishing a multi-level closed-loop interaction mechanism of "transmission system - distribution system - virtual power plant", the effective mapping of macro cost parameter signals and micro scheduling instructions is realized, ensuring the safety of joint operation of transmission and distribution networks. Attached Figure Description

[0043] Figure 1 A flowchart of a risk perception and clearing method for transmission and distribution coordination that takes into account uncertainties, provided in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of the three-layer physical architecture for transmission and distribution coordination provided in an embodiment of the present invention;

[0045] Figure 3 This is a logic diagram for generating the equivalent aggregation ladder curve of a single-sided distributed resource provided in an embodiment of the present invention.

[0046] Figure 4 This is a comparison chart of post-event average efficiency and expected unsupplied power under different clearing schemes provided in embodiments of the present invention.

[0047] Figure 5 A hardware structure block diagram of an electronic device for implementing the methods provided in the embodiments of the present invention. Detailed Implementation

[0048] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] like Figure 1 and Figure 2 As shown, the method of this invention mainly relies on a multi-level physical network architecture for implementation, aiming to solve the problem that traditional deterministic clearing models, when encountering out-of-distribution (OOD) prediction biases under high-proportion distributed resource access, lead to a surge in system load shedding risk and severe losses. The uncertainty-based transmission and distribution coordination risk perception clearing method of this invention is as follows:

[0050] First, a transmission and distribution coordination architecture is established, comprising a transmission system layer, a distribution system layer, and a virtual power plant (VPP) layer. Historical photovoltaic (PV) operation data for the transmission system layer is obtained, and then a fuzzy set of PV moment information for the transmission system layer is constructed. The historical PV operation data for the transmission system layer includes the actual and predicted historical active power output of centralized PV systems, such as the historical actual PV output over the past 30 days and the current-day predicted data sequence. The fuzzy set of PV moment information for the transmission system layer includes a random variable representing the relative prediction error. probability distribution The details are as follows:

[0051]

[0052] in, Indicates time period fuzzy set of photovoltaic moment information; Indicates time period The dimensionless relative prediction error random variable of active power output of centralized photovoltaic power generation; Represents a random variable of relative prediction error The actual but unknown probability distribution; It means that it is defined on the support set of random variables. The set of all possible probability distributions on the random variable support set. Several relative prediction error random variables taking values ​​within a preset error boundary range Composition, each relative prediction error random variable Each has its own probability distribution for taking values, and the preset error boundary range is obtained based on the statistical results of historical photovoltaic operation data; Represents probability distribution The mathematical expectation operator; and They represent time periods respectively. The actual and predicted values ​​of active power output from centralized photovoltaic systems are both based on pre-acquired input data, which are independently calculated using conventional numerical weather prediction or time series forecasting algorithms. and They represent time periods respectively. The mean and standard deviation of the relative prediction error between the actual and predicted values ​​of the active power output of centralized photovoltaic systems.

[0053] relative prediction error random variable probability distribution During the period The mean of the relative prediction error between the actual and predicted values ​​of active power output of centralized photovoltaic systems. and standard deviation Input the energy of the transmission and distribution coordination and the reserve joint clearing model.

[0054] The transmission and distribution coordination architecture is a three-layer physical architecture. From top to bottom, it comprises three layers: the transmission system layer, the distribution system layer, and the virtual power plant (VPP) layer. The transmission system layer includes z transmission network nodes. 'a' high-voltage transmission network nodes connect to traditional generating units or large-capacity centralized photovoltaic systems composed of several photovoltaic units. Traditional generating units refer to non-new energy generating units such as thermal power. 'b' transmission network nodes connect only to their respective distribution system layers. 'c' transmission network nodes connect to both their respective distribution system layers and the large-capacity energy storage system of the transmission network, where z = a + b + c. The distribution system layer includes one distribution network root node and several distribution network nodes. The distribution network root node connects to the transmission network nodes of the transmission system layer. Each transmission network node connects to its own rigid load, photovoltaic, distribution network-side energy storage, or virtual power plant layer. The photovoltaic, distribution network-side energy storage, and virtual power plant layers, as heterogeneous underlying flexible resources, are physically subject to strict constraints imposed by second-order cone AC power flow. The virtual power plant layer includes distributed resources and a virtual power plant dispatch center. Distributed resources include distributed generator sets and flexible loads. Specifically, the distributed resources within the virtual power plant layer (VPP) include, but are not limited to, rooftop distributed photovoltaic, gas turbines, and flexible loads composed of electric vehicles (EVs) and air conditioning systems. The virtual power plant layer connects to the transmission network nodes through the virtual power plant dispatch center.

[0055] The transmission system layer is responsible for the overall security scheduling of the transmission network, performing risk perception and joint clearing based on Distributed Robust Chance-Constraint (DRCC). The distribution system layer, as an intermediate layer, manages the active distribution network within its region, performing pre-clearing and aggregation of flexible resources at nodes. The Virtual Power Plant (VPP) layer is responsible for aggregating its massive distributed energy sources (DERs) and performing internal power balancing and scheduling. This hierarchical architecture ensures effective access to massive distributed resources while avoiding the dimensionality explosion problem at the transmission system level. To verify the effectiveness, advancement, and engineering applicability of the transmission-distribution coordinated risk perception and clearing method proposed in this embodiment, this invention was tested based on specific simulation examples. The simulation environment and parameter settings are as follows:

[0056] 1. Network topology and hardware environment:

[0057] This embodiment constructs a three-layer transmission and distribution coordination architecture: transmission system layer, distribution system layer, and virtual power plant (VPP). The transmission system layer adopts an IEEE 9-node system, coupled with three independent active distribution networks. Each distribution network uses an IEEE 33-node system; specifically, the three distribution networks are connected to nodes 5, 7, and 9 of the transmission system, respectively. Throughout the simulation scheduling process, the distribution network side employs second-order cone programming (SOCP) with precise relaxation techniques for solving the problem, and strictly constrains the voltage amplitude of all distribution network nodes to remain within a safe operating range of 0.95 pu to 1.05 pu to ensure the physical deliverability of the underlying flexible resources.

[0058] 2. Equipment parameter configuration for the power transmission system layer:

[0059] Traditional thermal power units are configured at nodes 1, 2, and 3 of the transmission system, with their power generation cost function coefficients (quadratic, linear, and constant terms) set as follows: Unit 1 (0.11, 5, 150), Unit 2 (0.085, 1.2, 600), and Unit 3 (0.1225, 1, 335). An additional large-scale centralized photovoltaic power plant with a peak capacity of 150 MW is connected at node 1. To provide system-level flexibility and backup support, a grid-side battery energy storage system (BESS) with a rated power / capacity of 70 MW / 280 MWh is configured at node 3, with its State of Charge (SOC) safety boundary set to 0%–100% and its charge / discharge efficiency set to 0.90.

[0060] 3. Distribution System Layer Equipment Parameter Configuration: Each distribution system is equipped with 8 virtual power plant (VPP) units and 5 distributed energy storage units. The specific parameter settings are as follows:

[0061] Virtual power plant (VPP) configurations: The VPPs located at nodes 8, 11, and 26 are configured with distributed photovoltaics, each with a rated installed capacity of 12 MW; the VPPs located at nodes 3, 4, 9, 19, and 29 are configured with micro gas turbines (GT), each with a rated installed capacity of 4 MW.

[0062] Distributed energy storage configuration: The battery energy storage system (BESS) located at nodes 3 and 29 has a rated capacity of 2.4 MWh, a rated charge / discharge power limit of 1.0 MW, and a minimum energy boundary of 0.9 MWh; the battery energy storage system (BESS) located at nodes 8, 11, and 26 has a rated capacity of 3.2 MWh, a rated charge / discharge power limit of 1.333 MW, and a minimum energy boundary of 1.2 MWh. The charge / discharge efficiency of all distributed energy storage systems is uniformly set to 0.90 (dimensionless).

[0063] In addition, during the pre-clearing phase of the distribution system, the predicted time-of-use electricity price is used as a reference variable for the electricity cost parameter at the boundary node, with its value range set between 25.39 (1 / MWh) and 54.35 (1 / MWh) to guide the participation of underlying flexible resources in the response.

[0064] 4. Uncertainty modeling and clearing model parameters:

[0065] To construct the moment information fuzzy set in the DRCC (Divided Broken Bar Opportunity Constraint) model, this embodiment extracts 200 historical samples of photovoltaic prediction errors from historical operating data for statistical analysis of empirical mean and standard deviation. The risk tolerance in the model... A value of 0.05 corresponds to a system reliability level of 95%. Based on the one-sided Cantelli-Chebyshev inequality, the corresponding safety margin coefficient can be further obtained. .

[0066] 5. Out-of-sample (OOS) testing and extreme distribution OOD stress testing solutions:

[0067] To verify the risk resistance capability of the proposed transmission and distribution coordination risk perception and clearing method in actual operation, this embodiment generates... =2000 out-of-sample scenes independent of the training set An assessment is conducted. If the operational reserve procured by the system is insufficient to cover the real-time power imbalance for any given hour, it is recorded as a system error. This embodiment introduces Expected Unserved Energy (EUE, in MWh / day) as a core indicator for quantifying the severity of system errors. ,as follows:

[0068]

[0069]

[0070] in, Representing a scene Down The actual real-time photovoltaic power deficit generated by the time system; for The photovoltaic power output cleared out day by day; for Moment Scene Real-time photovoltaic active power output; for The system's reserved backup capacity; This represents the time interval step size.

[0071] Meanwhile, to test the system's resilience under extreme conditions, this embodiment designed a distributed out-of-distribution (OOD) stress test. By injecting a significant distribution offset into the test data, it simulates an extreme scenario of a sudden drop in photovoltaic output. Actual photovoltaic output under stress test. as follows:

[0072]

[0073] in, express The raw prediction error sampled from the test dataset at each step; for The current photovoltaic (PV) power output is predicted in real time. The physical meaning of this operation is: while keeping the historical error fluctuation structure unchanged, the expected level of the predicted PV power output is systematically lowered by 5% to construct a more stringent off-grid test scenario.

[0074] 6. Hardware and software operating environment and computing efficiency:

[0075] The algorithm program in this embodiment was developed based on the MATLAB platform, modeled using YALMIP, and solved using the Gurobi optimization solver. The computer hardware configuration used for testing was an Intel Core i9-13980HX processor with 32 GB of RAM. Thanks to the relaxation and reconstruction techniques employed in this invention, the average solution time of the joint clearing model of the split-Bruker opportunity constraint at the transmission system layer with high computational load is only 8.6 seconds, indicating that this invention can meet the computational timeliness requirements of actual day-ahead scheduling.

[0076] Based on the day-ahead forecast data of flexible resources within the jurisdiction (such as flexible loads like electric vehicles and air conditioners, as well as power generation resources like distributed photovoltaics and gas turbines), the distribution system layer constructs a low-level scheduling model under the AC power flow operation constraints of the second-order cone programming (SOCP) form. This model constrains the physical operating boundary of the distribution network. The low-level scheduling model of the distribution system layer includes a total net efficiency objective function and a power aggregation ladder curve generation model. The specific total net efficiency objective function is as follows:

[0077]

[0078] in, This represents the total net efficiency of the power distribution system layer within the scheduling cycle of the transmission and distribution coordination architecture. Represents the set of time periods within a scheduling period; Represents a set of steps with predefined segments; Indicates time period No. The input cost of electricity consumption for flexible loads of distributed resources in the virtual power plant layer is the input cost, which is the utility. The cost can be specifically measured in terms of electricity consumption. Indicates time period No. The power generation output cost of distributed generator sets in the distributed resources of the virtual power plant layer; Indicates time period The net operating and reserve costs of energy storage on the distribution network side of the distribution system layer; Indicates time period The power input cost of rigid loads in the power distribution system layer; Indicates time period The cost of network loss at the power distribution system layer; Indicates time period The reference variable for the marginal energy cost parameter of the boundary node in the distribution network node of the distribution system layer is used as a parameter scan during the pre-clearing stage; Indicates time period The net active power injected at the boundary tie line connecting the distribution system layer to the transmission system layer. The boundary tie line refers to the boundary line connecting the boundary node to the outside. Indicates the scheduling time step.

[0079] After the overall net efficiency objective function is constructed, the power distribution system layer further performs equivalent aggregation and discretization of internal heterogeneous resources. For example... Figure 3 The diagram illustrates the logic for generating an equivalent aggregation ladder curve for single-sided distributed generation resources, as provided in this embodiment of the invention. Single-sided resource aggregation is used as an example. The Virtual Power Plant (VPP) first treats its internal resource blocks as independent rectangular capacity blocks, with the rectangle height representing a preset cost parameter. The width represents the adjustable active power of the resource block. Here, resource block 1, resource block 2, and resource block 3 represent different distributed generation resource units that participate in the sorting during aggregation processing. , , These represent the preset cost parameters for the corresponding resource blocks. , , These represent the adjustable active power of the corresponding resource blocks. Figure 3 This only illustrates the single-sided aggregation logic of the power generation side resources.

[0080] Subsequently, different sorting logics were used to aggregate the two types of resources: for flexible load resources, aggregation was performed according to preset cost parameters. The adjustable active power of each resource block is progressively spliced ​​and accumulated in ascending order to construct a monotonically increasing equivalent demand curve; for distributed generation resources, the preset cost parameters are used... The adjustable active power of each resource block is spliced ​​and accumulated in descending order to construct a monotonically decreasing equivalent supply curve.

[0081] Specifically, for any of the above curves, the equivalent aggregated first... Cumulative regulating power corresponding to each step All can be uniformly represented as:

[0082]

[0083] in, To set cost parameters The sorted number Adjustable active power of each resource block.

[0084] Furthermore, the equivalent supply curve and equivalent demand curve obtained from the initial aggregation are discretized with equal width to generate a power aggregation step curve generation model, as follows:

[0085]

[0086]

[0087]

[0088]

[0089] in, and These represent the power distribution system layer. During the period The first submission to the power transmission system layer The power generation output cost parameters and the power consumption input cost parameters of the segment; and These represent the power distribution system layer. During the period The first equal-width discretization of the power output side and the power input side The power capacity width of a single equal-width stepped section; and These represent the generation mapping function and the consumption mapping function, respectively, used at the power distribution system level. During the period Cost parameters are extracted at the midpoint of the cumulative power capacity of each equal-width step segment on both the power output and input sides, i.e., power generation output cost parameters are extracted. and power input cost parameters , Indicates an index for equal-width ladder segments. , This represents the total number of equal-width stepped segments in the equal-width discretization. and These represent the power distribution system layer. During the period After initial aggregation, the power output side and power input side... The remaining total available continuous active power capacity of the segment.

[0090] Using the power aggregation ladder curve generation model constructed above, the distribution system layer submits an equivalent, continuously adjustable, and physically executable virtual supply and demand resource declaration power aggregation ladder curve to the transmission system layer.

[0091] The resource regulation parameters reported by the virtual power plant are input into the underlying scheduling model for pre-clearing, and the processed parameters output a power aggregation step curve. Resource regulation parameters include time periods. No. The power input cost of flexible loads in the virtual power plant layer And the power generation output cost of distributed generator sets of distributed resources Boundary node scheduling instructions are for time periods. Net active power injected at the boundary tie line connecting the distribution system layer and the transmission system layer and time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer The power aggregation step curve includes the power distribution system layer. During the period The first submission to the power transmission system layer Section power generation output cost parameters and power input cost parameters And the first equal-width discretization on the power output side Power capacity width of single-segment equal-width stepped section and the first power input side Power capacity width of a single equal-width stepped segment discretized by equal-width segment .

[0092] Within the safe power flow boundary, the available adjustable capacity of heterogeneous underlying flexible resources is extracted. Specifically, for distributed resources, an equivalent supply curve is formed by accumulating from low to high according to preset power cost parameters. For flexible loads and energy storage charging resources, an equivalent demand curve is formed by accumulating from high to low according to preset input power cost parameters. The continuous available capacity obtained from the initial aggregation is discretized into equal-width stepped segments. By extracting representative cost parameters at the midpoint of each stepped segment, a power aggregation stepped curve submitted to the transmission system layer is generated.

[0093] The underlying scheduling model of the power distribution system layer is based on AC power flow operation constraints, and it applies to each time period within the scheduling cycle. All are true, as follows:

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100] in, Indicates the connection with the distribution network node The set of connected downstream distribution network nodes. and Distribution network nodes The virtual power plant, as a resource, is mounted inside the distribution network node, connecting the downstream distribution network node and the upstream transmission network node. and They represent time periods respectively. From distribution network nodes Flow to distribution network nodes The tributary currents are both beneficial and ineffective; and They represent time periods respectively. From transmission network nodes Flow to distribution network nodes The tributary currents are both beneficial and ineffective; and They represent time periods respectively. Distribution network nodes Net active and reactive power injection; and Representing the transmission network nodes and distribution network nodes The resistance and reactance of the branches between them; Indicates time period Transmission network nodes and distribution network nodes The square of the current amplitude in the branch between; and They represent time periods respectively. Distributed to distribution network nodes The active power output command of the distributed generator sets and the power consumption command of the flexible loads are given. and They represent time periods respectively. Distributed to distribution network nodes Energy storage charging power command and energy storage discharging power command at the location; Indicates time period Distribution network nodes The fixed load power, when the distribution network node When no device of the corresponding type is connected, its corresponding power value is zero; and They represent time periods respectively. Transmission network nodes and distribution network nodes The square of the voltage amplitude; and These represent the lower and upper limits of the voltage amplitude at a distribution network node, respectively. Represents a power transmission network node and distribution network nodes The upper limit of the current amplitude of the branches between them.

[0101] The underlying device scheduling instructions include time periods in the AC power flow operation constraints. Distributed to distribution network nodes Active power output command of distributed generator set at the location Power consumption instructions for flexible loads and energy storage charging power commands and energy storage discharge power command .

[0102] Then, a joint clearing model of transmission and distribution coordinated energy and reserves based on deterministic solvable constraints is established at the transmission system layer, as follows:

[0103]

[0104]

[0105]

[0106]

[0107]

[0108] in, Represents the set of time periods within a scheduling period; Indicates time period The total net efficiency of the coupling between each power distribution system layer; Indicates time period The input cost of resilient loads at the power transmission system layer, including large industrial loads and data centers connected to the power transmission network; and They represent time periods respectively. The input cost and preset discharge cost of the input power when the energy storage system of the power grid connected to the power transmission system layer is charged; Indicates time period The cost of secondary power generation output of traditional generator sets; This represents the marginal cost coefficient of centralized photovoltaic power generation. Indicates time period Day-ahead clearing power of centralized photovoltaic power, This represents the marginal operating cost of centralized photovoltaic power generation. This represents the frequency regulation reserve cost of the transmission and distribution coordination architecture; Indicates the scheduling time step; and These represent the power distribution system layer. During the period The first submission to the power transmission system layer The power generation output cost parameters and the power consumption input cost parameters of the segment; and These represent the power distribution system layer. During the period After initial aggregation, the power output side and power input side... Total remaining available continuous active power capacity and These represent the power distribution system layer. During the period The first equal-width discretization of the power output side and the power input side The power capacity width of a single equal-width stepped section; Indicates time period The net active power injected at the boundary tie line connecting the distribution system layer and the transmission system layer.

[0109] To address the stochastic nature of centralized photovoltaic (PV) power transmission, and to ensure sufficient reserve capacity even under extreme probability distributions, a distributed bar chance constraint (DRCC) is constructed as follows. A reserve capacity optimization model based on the DRCC is established to address the stochastic nature of centralized PV power transmission. The specific constraints are as follows:

[0110]

[0111] in, The infimum operator represents the lower bound of the probability among all possible probability distributions contained in the moment information fuzzy set.

[0112] Physically, this constraint guarantees that even under the worst probability distribution within the fuzzy set, the reserve capacity is still sufficient to compensate for the photovoltaic power output shortage caused by prediction errors. Then, using the one-sided Cantelli-Chebyshev inequality, the complex bibliophilic constraint is reconstructed into a deterministic solvable constraint, as follows:

[0113]

[0114]

[0115] in, For the transmission and distribution coordination architecture in time period Up-modulation reserve capacity; For time period Solar power output cleared day-ahead; For time period The current forecast for the active power output of centralized photovoltaic power generation is as follows; and They represent time periods respectively. The mean and standard deviation of the relative prediction error between the actual and predicted values ​​of the active power output of centralized photovoltaic power generation; This is the safety margin coefficient derived from risk tolerance; This is the preset risk tolerance level.

[0116] The power aggregation step curve and the fuzzy set of photovoltaic moment information are input into the joint optimization solution of the transmission and distribution coordinated energy and reserve clearing model. The boundary node scheduling instructions are obtained by the solver directly solving the problem. For boundary node power balance constraints, time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer As the dual variable of the power balance constraint of the boundary node, it also serves as the boundary node scheduling instruction. The boundary node scheduling instruction is sent to the underlying scheduling model for processing and then output to the virtual power plant, thereby realizing the risk perception and clearing of transmission and distribution coordination.

[0117] To further illustrate the data flow process of this invention in actual operation, the following uses a typical midday photovoltaic high-output period as an example to explain the single-period execution process. The final simulation verification conclusions and technical effect analysis are as follows:

[0118] Based on the above parameter configurations, under a stress test scenario with an extreme prediction bias of -5% out-of-distribution (OOD) in photovoltaic output, the post-event average efficiency (EUE) of different clearing schemes is compared with the expected unsupplied power (EUE) as follows: Figure 4 As shown.

[0119] To further quantify and verify the technical advantages of the transmission and distribution coordinated risk perception and clearing architecture proposed in this invention, Table 1 presents a comparison of the core operating indicators of four different clearing schemes when facing an extreme deviation of -5% in photovoltaic forecasts.

[0120] Table 1. Comparison of core indicators of different clearing schemes under extreme photovoltaic forecasting bias.

[0121] The system physical error probability is defined as the percentage of scenarios in which load shedding or voltage exceeding limits occurs at least once during the entire daily scheduling cycle.

[0122] from Figure 4 The following conclusions can be drawn intuitively from Table 1:

[0123] 1) Traditional deterministic clearing schemes have high system risks. Because this method does not reserve sufficient backup capacity for extreme prediction errors, it performs poorly when facing an extreme off-distribution OOD deviation of -5%, with the expected unsupplied electricity consumption (EUE) rising to 72.07 MWh / day, the system physical error probability reaching 100%, and resulting in a post-event average efficiency drop to -227.67 × 10⁻⁶. 4 This indicates that the system faces significant operational risks and performance losses under extreme scenarios.

[0124] 2) Traditional stochastic programming schemes offer some improvements, but their ability to withstand extreme risks remains limited. Compared to traditional deterministic clearing, its ex-post average efficiency turns positive, and the expected unsupplied electricity decreases to 3.61 MWh / day. However, the system's physical error probability remains high at 54.70%, indicating that it still has significant risk exposure when facing unknown extreme OOD deviations.

[0125] 3) The independent clearing scheme of the transmission system without distribution system coordination is limited. When the transmission and distribution coordination mechanism is turned off, the transmission system cannot fully utilize the flexible resources within the distribution system and can only rely on its own resources to reserve a higher margin to cope with the potential photovoltaic shortage risk, resulting in both the expected day-ahead efficiency and the post-event average efficiency being lower than the method of this invention.

[0126] 4) The transmission and distribution coordinated risk perception and clearing method proposed in this invention achieves a better trade-off between reliability and economy. Under the stress test of -5% OOD photovoltaic prediction deviation, the solution of this invention controls the expected unsupplied electricity to 0.88 MWh / day, reduces the system physical error probability to 5.70%, and achieves an average efficiency of 20.28 × 10⁻⁶ after the event. 4 Furthermore, the post-hoc performance in the worst-case 5% scenario is -37.06 × 10⁻⁶. 4 All of these methods outperform independent clearing schemes for transmission systems without distribution system coordination, and are also significantly better than traditional stochastic programming and traditional deterministic clearing schemes. This shows that the present invention can effectively reduce physical operation risks and improve system resilience in extremely unknown scenarios.

[0127] In summary, this invention, by combining the pluripotent rod chance constraint with a multi-level transmission and distribution collaborative architecture, effectively enhances the physical resilience of high-proportion renewable energy power systems to extreme distribution out-of-prediction errors while taking into account system economy. It achieves a better trade-off between reliability and economy and has good engineering practicality and application prospects.

[0128] This invention also designs a transmission and distribution coordination risk perception and clearing system that takes uncertainty into account, including a data acquisition module, a distribution network aggregation module, a joint clearing module, and an instruction issuance module. The data acquisition module is used to establish a transmission and distribution coordination architecture including a transmission system layer, a distribution system layer, and a virtual power plant layer, and to acquire historical photovoltaic operation data of the transmission system layer, thereby constructing a fuzzy set of photovoltaic moment information for the transmission system layer. The distribution network aggregation module is used to construct a bottom-level scheduling model for the distribution system layer, input resource adjustment parameters into the bottom-level scheduling model for pre-clearing, and output a power aggregation step curve after processing. The joint clearing module is used to establish a transmission and distribution coordination energy and reserve joint clearing model based on deterministic solvable constraints for the transmission system layer, input the power aggregation step curve and the fuzzy set of photovoltaic moment information into the transmission and distribution coordination energy and reserve joint clearing model for joint optimization and solution, and obtain boundary node scheduling instructions. The instruction issuance module is used to issue boundary node scheduling instructions to the bottom-level scheduling model, and the bottom-level scheduling model processes and outputs bottom-level equipment scheduling instructions to the virtual power plant, thereby realizing transmission and distribution coordination risk perception and clearing.

[0129] Corresponding to the aforementioned embodiments of the transmission and distribution coordination risk perception and clearing method considering uncertainty, this invention also provides embodiments of an electronic device for executing the method. The electronic device includes a memory and one or more processors. The memory stores executable code, and when the processor executes the executable code, it implements the clearing method described in the above embodiments. The electronic device for transmission and distribution coordination risk perception and clearing considering uncertainty provided by this invention can be applied to terminals or servers with data processing capabilities, such as dispatch center servers or cluster computing nodes. Embodiments of this invention can be implemented through software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it can be formed by the processor of a related device reading and running computer program instructions in non-volatile memory. From a hardware perspective, such as... Figure 5 As shown, in addition to the processor, internal bus, memory, network interface, and non-volatile memory, electronic devices may typically include other hardware as needed for actual functionality, which will not be elaborated further. For the device embodiments, since they basically correspond to the method embodiments, the relevant details can be found in the description of the method embodiments.

[0130] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the present invention according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0131] This invention also provides a computer-readable storage medium storing a program thereon, which, when executed by a processor, implements the methods described in the above embodiments. The computer-readable storage medium can be an internal storage unit of the relevant device in any of the foregoing embodiments, such as a hard disk or memory; or it can be an external storage device of the relevant device, such as a plug-in hard disk, smart memory card, SD card, or flash memory card. Further, the computer-readable storage medium may include both internal storage units and external storage devices of the relevant device. The computer-readable storage medium is used to store computer programs and other programs and data required by the relevant device, and can also be used to temporarily store data that has been output or will be output.

[0132] The above embodiments are used to explain and illustrate the present invention, but not to limit the present invention. Any modifications and changes made to the present invention within the spirit and scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A risk perception and clearing method for transmission and distribution coordination that takes into account uncertainty, characterized in that, include: Step 1) Establish a transmission and distribution coordination architecture that includes the transmission system layer, the distribution system layer, and the virtual power plant layer; obtain historical photovoltaic operation data of the transmission system layer; and then construct a fuzzy set of photovoltaic moment information for the transmission system layer. Step 2) Construct the underlying scheduling model of the power distribution system layer, input the resource regulation parameters reported by the virtual power plant into the underlying scheduling model for pre-clearing, and output the power aggregation step curve after processing. Step 3) Establish a joint clearing model for transmission and distribution coordinated energy and reserves based on deterministic solvable constraints at the transmission system layer. Input the power aggregation step curve and the fuzzy set of photovoltaic moment information into the joint clearing model for joint optimization and solution. After solving, the boundary node scheduling instructions are obtained and then sent to the underlying scheduling model for processing. The underlying equipment scheduling instructions are then output to the virtual power plant, thereby realizing transmission and distribution coordinated risk perception and clearing.

2. The transmission and distribution coordination risk perception and clearing method considering uncertainty according to claim 1, characterized in that: In step 1), the power transmission system layer includes z power transmission network nodes, of which a power transmission network nodes are connected to traditional generator sets or centralized photovoltaic systems, b power transmission network nodes are connected to only one of their respective distribution system layers, and c power transmission network nodes are connected to both their respective distribution system layers and the power transmission network energy storage system, z = a + b + c. The distribution system layer includes a distribution network root node and several distribution network nodes. The distribution network root node connects to the transmission network nodes of the transmission system layer. Each transmission network node connects to its own rigid load, photovoltaic, distribution network-side energy storage, or virtual power plant layer. The photovoltaic, distribution network-side energy storage, and virtual power plant layers serve as the underlying flexible resources. The virtual power plant layer includes distributed resources and a virtual power plant dispatch center. Distributed resources include distributed generator sets and flexible loads. The virtual power plant layer connects to the transmission network nodes through the virtual power plant dispatch center.

3. The transmission and distribution coordination risk perception and clearing method taking into account uncertainty as described in claim 2, characterized in that: In step 1), the historical photovoltaic operation data at the transmission system layer includes the actual and predicted historical active power output of centralized photovoltaic systems, and the fuzzy set of photovoltaic moment information at the transmission system layer includes the random variable of relative prediction error. probability distribution ; relative prediction error random variable probability distribution During the period The mean of the relative prediction error between the actual and predicted values ​​of active power output of centralized photovoltaic systems. and standard deviation Input the energy of the transmission and distribution coordination and the reserve joint clearing model.

4. The transmission and distribution coordination risk perception and clearing method considering uncertainty according to claim 2, characterized in that: In step 2), the underlying scheduling model of the power distribution system layer includes the overall net efficiency objective function, as follows: in, This represents the total net efficiency of the power distribution system layer within the scheduling cycle of the transmission and distribution coordination architecture. Represents the set of time periods within a scheduling period; Represents a set of steps with predefined segments; Indicates time period No. The electricity input cost of flexible loads for distributed resources in the virtual power plant layer; Indicates time period No. The power generation output cost of distributed generator sets in the distributed resources of the virtual power plant layer; Indicates time period The net operating and reserve costs of energy storage on the distribution network side of the distribution system layer; Indicates time period The power input cost of rigid loads in the power distribution system layer; Indicates time period The cost of network losses at the power distribution system layer; Indicates time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer; Indicates time period Net active power injected at the boundary tie line connecting the distribution system layer and the transmission system layer; Indicates the scheduling time step; Resource adjustment parameters include time period No. The power input cost of flexible loads in the virtual power plant layer And the power generation output cost of distributed generator sets of distributed resources ; Boundary node scheduling instructions are time periods Net active power injected at the boundary tie line connecting the distribution system layer and the transmission system layer and time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer ; The power aggregation step curve includes the power distribution system layer. During the period The first submission to the power transmission system layer Section power generation output cost parameters and power input cost parameters And the first equal-width discretization on the power output side Power capacity width of single-segment equal-width stepped section The first equal-width discretization of the power input side Power capacity width of single-segment equal-width stepped section .

5. The transmission and distribution coordination risk perception and clearing method considering uncertainty according to claim 4, characterized in that: The underlying scheduling model of the power distribution system layer is based on AC power flow operation constraints; The underlying device scheduling instructions include time periods in the AC power flow operation constraints. Distributed to distribution network nodes Active power output command of distributed generator set at the location Power consumption instructions for flexible loads and energy storage charging power commands and energy storage discharge power command .

6. The transmission and distribution coordination risk perception and clearing method taking into account uncertainty according to claim 2, characterized in that: In step 3), the specific model for the joint clearing of transmission and distribution coordinated energy and reserves at the transmission system layer is as follows: in, Represents the set of time periods within a scheduling period; Indicates time period The total net efficiency of the coupling between each power distribution system layer; Indicates time period The input cost of resilient loads at the transmission system layer; and They represent time periods respectively. The input cost and preset discharge cost of the input power when the energy storage system of the power grid connected to the power transmission system layer is charged; Indicates time period The cost of secondary power generation output of traditional generator sets; This represents the marginal cost coefficient of centralized photovoltaic power generation. Indicates time period The day-ahead clearing power of centralized photovoltaic power; This represents the frequency regulation reserve cost of the transmission and distribution coordination architecture; Indicates the scheduling time step; Time period Marginal energy cost parameters of boundary nodes in the distribution network nodes of the distribution system layer These are the dual variables of the power balance constraint at the boundary nodes.

7. The transmission and distribution coordination risk perception and clearing method taking into account uncertainty according to claim 2, characterized in that: In step 3), the deterministically solvable constraints are as follows: in, For the transmission and distribution coordination architecture in time period Up-modulation reserve capacity; For time period Solar power output cleared day-ahead; For time period The current forecast for the active power output of centralized photovoltaic power generation is as follows; and They represent time periods respectively. The mean and standard deviation of the relative prediction error between the actual and predicted values ​​of the active power output of centralized photovoltaic power generation; This refers to the safety margin factor. This is the preset risk tolerance level.

8. A transmission and distribution coordination risk perception and clearing system that takes into account uncertainty, characterized in that, include: The data acquisition module is used to establish a transmission and distribution coordination architecture that includes the transmission system layer, the distribution system layer, and the virtual power plant layer, to obtain historical photovoltaic operation data of the transmission system layer, and then to construct a fuzzy set of photovoltaic moment information of the transmission system layer; The distribution network aggregation module is used to construct the underlying scheduling model of the distribution system layer. It inputs resource adjustment parameters into the underlying scheduling model for pre-clearing and outputs a power aggregation step curve after processing. The joint clearing module is used to establish a joint clearing model of transmission and distribution coordinated energy and reserves based on deterministic solvable constraints at the transmission system layer. The power aggregation step curve and photovoltaic moment information fuzzy set are input into the joint clearing model of transmission and distribution coordinated energy and reserves for joint optimization and solution to obtain the boundary node scheduling instructions. The instruction issuance module is used to issue boundary node scheduling instructions to the underlying scheduling model. After processing, the underlying scheduling model outputs underlying equipment scheduling instructions to the virtual power plant, thereby realizing the risk perception and clearing of transmission and distribution coordination.

9. An electronic device, characterized in that, include: A memory and a processor are coupled to each other, wherein the memory stores program data, and the processor invokes the program data to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium storing program data thereon, characterized in that, When the program data is executed by the processor, the method as described in any one of claims 1-7 is implemented.

Images