A power distribution network operation optimization method and system based on partition electricity price

Abstract

This invention relates to the field of distribution network operation optimization technology, and provides a method and system for distribution network operation optimization based on regional pricing. The method includes: dividing the target distribution network into regions based on the line topology and day-ahead forecast data; processing the day-ahead forecast data and time-of-use pricing data according to the region division results to obtain regional day-ahead forecast data and regional time-of-use pricing data; performing operation status analysis based on a regional pricing optimization model to obtain a regional pricing adjustment strategy based on the region division results, regional day-ahead forecast data, and regional time-of-use pricing data; and optimizing and regulating the load of the target distribution network according to the regional pricing adjustment strategy. This invention enables a load demand regulation optimization mechanism that adjusts regional pricing based on the distribution network operation status, improving distribution network dispatch efficiency, ensuring efficient access and local consumption of renewable energy, and thus enhancing the stability of power grid operation.

Description

Technical Field

[0001] This invention relates to the field of distribution network operation optimization technology, and in particular to a distribution network operation optimization method and system based on regional electricity pricing. Background Technology

[0002] Distribution networks face the technical challenge of insufficient local absorption of renewable energy. Although some studies have found that electricity price dispatch strategies can guide changes in user load demand, these strategies only extend to the price setting level and lack an extension from price strategies to actual grid operation control. They fail to achieve price-based distribution network power flow optimization, resulting in price strategies being unable to solve practical distribution network operation problems. Furthermore, fixed-price strategies cannot flexibly respond to dynamic changes in electricity demand and supply, and do not fully consider the impact of renewable energy output fluctuations on grid operation. They are fundamentally unable to meet the practical application requirements of adaptive and efficient adjustments to address electricity demand fluctuations, thereby improving distribution network dispatch efficiency and the local absorption capacity of renewable energy. Therefore, there is an urgent need for a distribution network operation optimization method that can intelligently sense the grid operating status and adjust regional electricity prices to guide user load demand response, thereby improving distribution network dispatch efficiency and enhancing the grid's ability to cope with large-scale renewable energy integration. Summary of the Invention

[0003] The purpose of this invention is to provide a distribution network operation optimization method based on regional electricity pricing, and a load demand regulation optimization mechanism based on adjusting regional electricity prices according to the distribution network operation status. By reasonably optimizing regional electricity prices, the invention aims to motivate users to participate in the local consumption of renewable energy, improve the dispatch efficiency of the distribution network, solve problems such as insufficient local consumption of renewable energy, and ensure the efficient access and local consumption of renewable energy.

[0004] To achieve the above objectives, it is necessary to provide a method and system for optimizing the operation of distribution networks based on regional electricity pricing.

[0005] In a first aspect, embodiments of the present invention provide a method for optimizing the operation of a distribution network based on regional electricity pricing, the method comprising: Obtain the line topology, day-ahead forecast data, and time-of-use electricity price data of the target distribution network; the day-ahead forecast data includes load forecast data and photovoltaic output forecast data; Based on the line topology and the day-ahead forecast data, the regions are divided to obtain the region division results; the region division results include multiple control regions. Based on the regional division results, the day-ahead forecast data and the time-of-use electricity price data are processed to obtain the corresponding regional day-ahead forecast data and regional time-of-use electricity price data. Based on the regional division results, the day-ahead forecast data of the regional divisions, and the time-of-use electricity price data of the regional divisions, an operational status analysis is performed based on a pre-built regional electricity price optimization model to obtain a regional electricity price adjustment strategy; the regional electricity price optimization model is trained and constructed with the goal of maximizing the value of cross-regional power transmission and minimizing the cost of curtailment of solar power. Based on the aforementioned regional electricity price adjustment strategy, load optimization and control are performed on the target distribution network.

[0006] Further, the step of dividing the region based on the line topology and the day-ahead forecast data to obtain the region division result includes: With maximizing region partitioning performance as the optimization objective, a region partitioning objective function is constructed; the region partitioning performance is obtained by weighted summation based on preset region partitioning performance indicators. Based on the line topology and the day-ahead forecast data, the objective function for the region division is solved using the improved Grey Wolf algorithm to obtain the region division result.

[0007] Furthermore, the preset region division performance indicators include region modularity index, region active power balance index, and region reactive power balance index. The regional modularity index is constructed based on regional electrical distance analysis of the target distribution network. The regional active power balance index is constructed based on the matching relationship analysis of regional photovoltaic active power output and load active power demand of the target distribution network. The regional reactive power balance index is constructed based on the matching relationship analysis of regional photovoltaic reactive power output and load reactive power demand in the target distribution network.

[0008] Furthermore, the steps for constructing the regional modularity index include: Based on the line topology, the electrical distance between each node pair in the target distribution network is obtained, and the electrical edge weight of the corresponding node pair is obtained based on each electrical distance. Based on the electrical edge weights of all the node pairs, the sum of the distribution network edge weights is obtained, and the electrical edge weights of all edges connected to the same node are accumulated to obtain the weighted degree of the corresponding node. The expected electrical connection weight of the corresponding node pair is obtained by summing the weighted degree of the two nodes in each node pair and the edge weight of the distribution network. The deviations between the electrical edge weights and the corresponding expected electrical connection weights of each node pair are summarized and analyzed to obtain the regional modularity index.

[0009] Furthermore, the regional electricity price optimization model is based on reinforcement learning and is obtained by performing grid dispatch interaction simulation training on the target distribution network.

[0010] Furthermore, the state space and action space of the zonal electricity price optimization model are defined based on the real-time operating characteristics of the target distribution network and the electricity price adjustment strategy, respectively; the reward function of the zonal electricity price optimization model is constructed based on the principle of maximizing the difference between the value of cross-regional power transmission and the cost of curtailment of solar power. The real-time operating characteristics include day-ahead zone load data, day-ahead zone photovoltaic output data, zone time-of-use electricity price data, dispatch time, and regional division results; The electricity price adjustment strategy includes the amount of electricity price adjustment at any given time in each control zone.

[0011] Furthermore, the step of obtaining the execution reward of the electricity price adjustment strategy during the training of the regional electricity price optimization model includes: Obtain the photovoltaic output of each of the control zones at the time after the implementation of the electricity price adjustment strategy; Based on the pre-adjustment load active power, the pre-adjustment electricity price, and the time-based electricity price adjustment amount in the electricity price adjustment strategy for each control zone, the corresponding time-based response load active power is obtained based on the preset electricity price load response model. Based on the active power of the load and the photovoltaic output at the moment of all the control areas, the preset power balance equations are solved to obtain the transmission power at the sending end and the transmission power at the receiving end of each regional tie line; the preset power balance equations include inter-regional power balance equations and intra-regional power balance equations. Based on the time-sending and time-receiving transmission power of each regional tie line, the value of inter-regional tie line transmission power is summarized and analyzed to obtain the corresponding time-based cross-regional transmission value. The grid photovoltaic output absorption deviation is summarized and analyzed based on the real-time photovoltaic output and real-time response load active power of each control area to obtain the real-time curtailment cost; The execution reward for the corresponding electricity price adjustment strategy is obtained based on the cross-regional power transmission value at the specified time and the curtailment cost of the specified time.

[0012] Secondly, embodiments of the present invention provide a distribution network operation optimization system based on regional electricity pricing, the system comprising: The data acquisition module is used to acquire the line topology, day-ahead forecast data, and time-of-use electricity price data of the target distribution network; the day-ahead forecast data includes load forecast data and photovoltaic output forecast data. The region division module is used to divide regions according to the line topology and the day-ahead forecast data to obtain the region division result; the region division result includes multiple control regions. The data processing module is used to process the day-ahead forecast data and the time-of-use electricity price data according to the regional division results, so as to obtain the corresponding regional day-ahead forecast data and regional time-of-use electricity price data. The electricity price analysis module is used to perform operational status analysis based on the regional division results, the day-ahead forecast data of the regions, and the time-of-use electricity price data of the regions, and to obtain regional electricity price adjustment strategies based on a pre-built regional electricity price optimization model; the regional electricity price optimization model is trained and constructed with the goal of maximizing the value of cross-regional power transmission and minimizing the cost of curtailment of solar power. The load control module is used to perform load optimization control on the target distribution network according to the zonal electricity price adjustment strategy.

[0013] Furthermore, the region division module is also used for: With maximizing region partitioning performance as the optimization objective, a region partitioning objective function is constructed; the region partitioning performance is obtained by weighted summation based on preset region partitioning performance indicators. Based on the line topology and the day-ahead forecast data, the objective function for the region division is solved using the improved Grey Wolf algorithm to obtain the region division result.

[0014] Furthermore, the regional electricity price optimization model is based on reinforcement learning and is obtained by performing grid dispatch interaction simulation training on the target distribution network.

[0015] This invention provides a method and system for optimizing the operation of a distribution network based on regional electricity pricing. The method acquires the line topology of the target distribution network, day-ahead forecast data including load forecast data and photovoltaic output forecast data, and time-of-use electricity price data. Based on the line topology and day-ahead forecast data, the network is divided into regions, resulting in a network region division with multiple control areas. The day-ahead forecast data and time-of-use electricity price data are then processed according to the region division results to obtain regional day-ahead forecast data and regional time-of-use electricity price data. Based on the regional division results, regional day-ahead forecast data, and regional time-of-use electricity price data, an operational status analysis is performed using a pre-trained regional electricity pricing optimization model aimed at maximizing the value of inter-regional power transmission and minimizing the cost of curtailment of solar power. This yields a regional electricity pricing adjustment strategy, and a technical solution for optimizing and controlling the load of the target distribution network according to the regional electricity pricing adjustment strategy. Compared with existing technologies, this distribution network operation optimization method based on regional electricity pricing has a load demand regulation optimization mechanism based on the distribution network operation status to adjust regional electricity prices. By reasonably optimizing regional electricity prices, it mobilizes the enthusiasm of users to participate in the local consumption of renewable energy, improves the dispatch efficiency of the distribution network, solves the problem of insufficient local consumption of renewable energy, ensures the efficient access and local consumption of renewable energy, and thus improves the stability and economy of power grid operation. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the distribution network operation optimization method based on regional electricity pricing in an embodiment of the present invention. Figure 2 This is a schematic diagram of the line topology of the target distribution network in an embodiment of the present invention; Figure 3 This is a schematic diagram of the regional division results of the target power distribution network in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the operational optimization effect of the target distribution network in an embodiment of the present invention; Figure 5 This is a schematic diagram of the operation optimization effect of the control area 1 of the target distribution network in an embodiment of the present invention; Figure 6 This is a schematic diagram of the distribution network operation optimization system based on regional electricity pricing in an embodiment of the present invention; The attached figures are labeled as follows: 100. Data Acquisition Module; 200. Regional Division Module; 300. Data Processing Module; 400. Electricity Price Analysis Module; 500. Load Control Module. Detailed Implementation

[0017] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of this invention and are used to illustrate the invention, but are not intended to limit the scope of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0018] In one embodiment, such as Figure 1 As shown, a method for optimizing the operation of a distribution network based on regional electricity pricing is provided, including the following steps: S11. Obtain the line topology, day-ahead forecast data, and time-of-use pricing data of the target distribution network; whereby, the target distribution network can be understood as the distribution network that needs to undergo operational optimization, and the corresponding line topology can be understood as the transmission line network topology of the target distribution network, which varies depending on the actual distribution network and is not specifically limited here; the day-ahead forecast data can be understood as the time-series data of load distribution and photovoltaic output distribution of the target distribution network obtained based on existing distribution network load and photovoltaic output forecasting technology throughout the entire dispatch cycle (e.g., dispatching once per hour, with a 24-hour day as a complete dispatch cycle), that is, including load forecast data and photovoltaic output forecast data, and Load forecast data includes the active and reactive power forecast data of each load node at each dispatch time, and photovoltaic output forecast data includes the output forecast data of each distributed photovoltaic node at each dispatch time. The corresponding time-of-use electricity price data can be understood as the electricity price data of the target distribution network at each dispatch time corresponding to the day-ahead forecast data. Before the implementation of the zonal electricity price adjustment, the electricity price data of all load areas in the target distribution network at the same dispatch time are the same. After the implementation of the zonal electricity price adjustment, the electricity price data of different load areas in the target distribution network at the same dispatch time may be the same or different, depending on the actual control situation of the target distribution network.

[0019] S12. Based on the line topology and the day-ahead forecast data, regional division is performed to obtain regional division results. The regional division results include multiple control areas, and each control area can be understood as a sub-region of the distribution network that can be independently controlled, divided based on the network structure characteristics (node ​​electrical coupling strength) and electrical characteristics (renewable energy output and load matching) of the target distribution network. To ensure the scientific validity and feasibility of the regional division results, this embodiment preferably uses an improved gray wolf algorithm to perform multi-region division of the target distribution network based on the load distribution, spatiotemporal characteristics of photovoltaic power generation output, and line topology. Specifically, the step of dividing the region based on the line topology and the day-ahead forecast data to obtain regional division results includes: With maximizing the performance of regional division as the optimization objective, a regional division objective function is constructed. The regional division performance is obtained by weighted summation based on preset regional division performance indicators. The preset regional division performance indicators can be understood as multi-dimensional indicators designed to ensure that the obtained control area meets the regional division requirements of tight electrical coupling and high power balance within the area, and weak electrical correlation between areas. In this embodiment, the preset regional division performance indicators preferably include regional modularity indicators, regional active power balance indicators, and regional reactive power balance indicators.

[0020] In this embodiment, the regional modularity index can be understood as an index that measures the electrical tightness within a control area and the electrical separation between control areas, and is constructed based on regional electrical distance analysis of the target distribution network. To ensure the scientific and rational construction of the regional modularity index, it is preferable to construct the regional modularity index using an electrical weighted modularity mechanism that uses the electrical distance obtained from the line topology analysis of the target distribution network as the definition basis for the coupling weight between nodes in the distribution network. Specifically, the construction steps of the regional modularity index include: Based on the line topology, the electrical distances between each pair of nodes in the target distribution network are obtained, and the electrical edge weights of the corresponding pair of nodes are obtained based on each electrical distance. The electrical distances between node pairs can be obtained using existing power system electrical distance analysis techniques, such as calculations based on node reactive power and voltage sensitivity, and are not specifically limited here. In practical applications, after obtaining the electrical distances of all pairs of nodes in the target distribution network, the maximum electrical distance is first obtained, and then the electrical edge weights of each pair of nodes are calculated based on the following formula: In the formula, For nodes With nodes Formed node pairs Electrical edge weights; The maximum electrical distance within the target distribution network; For node pairs The corresponding electrical distance.

[0021] Based on the electrical edge weights of all node pairs, the sum of distribution network edge weights is obtained. Then, the electrical edge weights of all edges connected to the same node are summed to obtain the weighted degree of the corresponding node. The sum of distribution network edge weights can be understood as half the sum of the electrical edge weights of all node pairs, and the weighted degree of the corresponding node can be understood as the sum of the electrical edge weights of all edges connected to a given node, as specifically expressed below: In the formula, This is the sum of the edge weights of the distribution network; and They are nodes and nodes The weighting degree; and Each is a node pair and node pairs Electrical edge weights; The total number of nodes in the target distribution network.

[0022] Based on the weighted degree of the two nodes in each node pair and the sum of the distribution network edge weights, the expected electrical connection weight of the corresponding node pair is obtained. The expected electrical connection weight can be understood as the expected electrical connection strength between the two nodes in a random connection network, determined by the weighted degree of each node (overall node connection strength) and the sum of the distribution network edge weights (connection density) of the entire distribution network, and can be expressed as: In the formula, For node pairs The expected electrical connection weight.

[0023] The deviations between the electrical edge weights and the corresponding expected electrical connection weights of each node pair are summarized and analyzed to obtain the regional modularity index; wherein, the regional modularity index can be expressed as: In the formula, This is a regional modularity indicator; To represent nodes With nodes Whether a variable is in the same control region, a value of 0 indicates a node. With nodes Not in the same control region, a value of 1 indicates a node With nodes Within the same regulatory zone.

[0024] In this embodiment, the regional active power balance index can be understood as an index used to characterize whether the active power provided by distributed photovoltaics in all control areas within the entire target distribution network can meet the active power demand of the load in the corresponding area. Based on the matching relationship analysis of regional photovoltaic active power output and load active power demand in the target distribution network, it can be expressed as: In the formula, in, Set of control regions The first in One region; For the regulation area Active power balance; For the regulation area Regional photovoltaic active power output refers to the sum of active power provided by all distributed photovoltaic systems within the region. For the regulation area The active power demand of the load is the sum of the active power demand of all regions. This represents the total number of control areas. This represents the number of scheduling cycles. The regional active power balance index for the target distribution network.

[0025] In this embodiment, the regional reactive power balance index can be understood as an index used to characterize whether the reactive power provided by distributed photovoltaics in all control areas of the entire target distribution network can meet the reactive power demand of the load in the corresponding area. Based on the matching relationship analysis of regional photovoltaic reactive power output and load reactive power demand in the target distribution network, it can be expressed as: In the formula, in, For the regulation area The reactive power balance; For the regulation area The reactive power output of photovoltaic systems in the inner region is the sum of the reactive power provided by all distributed photovoltaic systems in the region. For the regulation area The reactive power demand of the load is the sum of the reactive power demand of all units in the area. The reactive power balance index for the target distribution network area.

[0026] By weighted summing the regional modularity index, regional active power balance index, and regional reactive power balance index obtained based on the above methods and steps, the regional partitioning performance can be obtained, and the objective function for regional partitioning can be determined as follows: In the formula, , and These are the weighting coefficients for the regional modularity index, the regional active power balance index, and the regional reactive power balance index, respectively. These can be set based on actual application requirements, as long as they meet the following criteria: .

[0027] Based on the line topology and the day-ahead forecast data, the objective function for region partitioning is solved using the improved Grey Wolf algorithm to obtain the region partitioning result. The process of solving the objective function for region partitioning using the improved Grey Wolf algorithm can refer to existing technologies for solving optimization problems using the improved Grey Wolf algorithm, and will not be described in detail here.

[0028] S13. Based on the regional division results, process the day-ahead forecast data and the time-of-use electricity price data to obtain corresponding zoned day-ahead forecast data and zoned time-of-use electricity price data; wherein, the zoned day-ahead forecast data includes zoned day-ahead load forecast data and zoned day-ahead photovoltaic output forecast data; the zoned day-ahead load forecast data includes day-ahead load forecast data of multiple control areas, which can be obtained by summarizing the active power forecast data of all load nodes in the target distribution network's load forecast data at each scheduling time according to the control area; the zoned day-ahead photovoltaic output forecast data includes day-ahead photovoltaic output forecast data of multiple control areas, which can be obtained by summarizing the output forecast data of all photovoltaic nodes in the target distribution network's photovoltaic output forecast data at each scheduling time according to the control area.

[0029] In this embodiment, the time-of-use electricity price data includes time-of-use electricity price data for multiple control areas. As mentioned above, before the first implementation of the zonal electricity price adjustment, the time-of-use electricity price data for each control area is equivalent to the actual time-of-use electricity price data of the target distribution network. After the implementation of the zonal electricity price adjustment, the time-of-use electricity price data for each control area can be directly adopted from the regional time-of-use electricity price data currently being implemented in each control area within the target distribution network, which will not be described in detail here.

[0030] S14. Based on the regional division results, the day-ahead forecast data of the regional divisions, and the time-of-use electricity price data of the regional divisions, an operational status analysis is performed based on a pre-constructed regional electricity price optimization model to obtain a regional electricity price adjustment strategy. The regional electricity price optimization model can be understood as an intelligent optimization model capable of analyzing relevant operational data of each control area within the target distribution network to determine the regional time-of-use electricity price applicable to load control of each control area. To ensure that regional electricity price adjustments can guide user load demand response to improve distribution network dispatch efficiency and effectively enhance the grid's ability to cope with large-scale renewable energy access, this embodiment preferably trains and constructs the regional electricity price optimization model with the goal of maximizing the value of cross-regional power transmission and minimizing the cost of curtailment.

[0031] Meanwhile, in order to better perceive the operational changes of the distribution network and to adaptively and efficiently adjust the electricity price dispatching strategies of each control area based on the fluctuations in electricity demand and renewable energy output of the target distribution network, this embodiment preferably uses reinforcement learning to train the target distribution network through grid dispatching interactive simulation to obtain a regional electricity price optimization model. That is, the regional electricity price optimization model in this embodiment is a reinforcement learning agent that makes decisions on the current operating state of the target distribution network based on the input regional division results, regional day-ahead forecast data, and regional time-of-use electricity price data to output the electricity price adjustment amount of each control area. This reinforcement learning agent learns the optimal strategy by repeatedly exploring the Markov game process (MGP).

[0032] In this embodiment, the state space and action space of the zonal electricity price optimization model are defined based on the real-time operating characteristics and electricity price adjustment strategy of the target distribution network, respectively. The state space, used to characterize the operating features of the target distribution network, is constructed using multi-dimensional real-time operating characteristic parameters, providing a learning foundation for the reinforcement learning agent to perceive the environment. This includes day-ahead zonal load data, day-ahead zonal photovoltaic output data, zonal time-of-use electricity price data, scheduling time, and regional division results, collectively forming a high-dimensional feature space that reflects the real-time operating status of the distribution network. The reinforcement learning agent performs learning decisions on behalf of the target distribution network, observing the current environment at each scheduling time to obtain the state variables at that moment. State variables used for analysis to determine the next action to be executed and for scheduling. It can be represented as: In the formula, In the formula, It is a vector composed of load data of each control area corresponding to the day-ahead load forecast data of the partition; For the regulation area During scheduling The sum of the predicted active power of the load; A vector composed of photovoltaic output data for each control region corresponding to the pre-districted photovoltaic output forecast data; For the regulation area During scheduling The sum of the photovoltaic output forecast data; For scheduling time Time-of-use electricity pricing; To represent nodes Is it classified as the first The variable in each control region, when the value is 1, indicates a node. Classified to the A control region, where a value of 0 indicates a node. Not classified as the first One regulatory zone; The total number of nodes in the target distribution network. K The total number of control zones in the target distribution network; This is the membership matrix corresponding to the region division results, used to characterize the membership relationships between each node and each control zone within the target distribution network. It should be noted that in practical applications, the membership matrix corresponding to the region division results can be flattened into a one-dimensional vector according to a preset order, and... and The state variables are concatenated and then input into the reinforcement learning agent for state analysis.

[0033] The action space in this embodiment can be understood as the electricity price adjustment actions of all control areas obtained through decision analysis based on the state variables and agent policies at the current scheduling moment and through linear mapping. It is an electricity price adjustment strategy that includes the time-based electricity price adjustment amounts for each control area, and can be expressed as: In the formula, For the regulation area During scheduling The amount of electricity price adjustment; For scheduling time Electricity price adjustment strategy.

[0034] In this embodiment, the reward function of the regional electricity price optimization model can be understood as an evaluation function that optimizes and updates the output actions of the reinforcement learning agent. It is the core evaluation index of the agent's decision-making effect and is used to guide strategy optimization through quantitative feedback signals. In order to improve the economic efficiency of the target distribution network while ensuring its operational stability, this embodiment preferably analyzes the comprehensive operational benefits of the target distribution network based on the difference between the value of cross-regional power transmission and the cost of curtailment, and constructs the reward function of the reinforcement learning agent based on the principle of maximizing the difference between the value of cross-regional power transmission and the cost of curtailment.

[0035] In the process of obtaining the regional electricity price optimization model through grid dispatch interactive simulation training of the target distribution network, the simulation environment of the target distribution network can be constructed using existing simulators. Within the simulation environment, the action instructions (electricity price adjustment strategies) output by the reinforcement learning agent are applied to calculate the cross-regional transmission value (the sum of the transmission power values ​​of the inter-regional tie lines) and the curtailment cost (the sum of the costs of photovoltaic power not being locally absorbed) based on the relevant operating state quantities of the target distribution network, thereby obtaining the corresponding action execution reward. Specifically, the steps for obtaining the execution reward of the electricity price adjustment strategy during the training of the regional electricity price optimization model include: The photovoltaic output of each control area after the implementation of the electricity price adjustment strategy is obtained; wherein, the photovoltaic output of each control area after the implementation of the electricity price adjustment strategy at the scheduling time can be obtained by summing and calculating the output data of all distributed photovoltaics in the target distribution network at the scheduling time obtained from the simulation operation according to the scheduling area, that is, the photovoltaic output of the corresponding control area is obtained by summing the distributed photovoltaic output data belonging to the same control area at the current scheduling time.

[0036] Based on the pre-adjustment active power of the load in each control zone, the pre-adjustment electricity price, and the time-based electricity price adjustment amount in the aforementioned electricity price adjustment strategy, the corresponding time-based response load active power is obtained based on the preset electricity price load response model. The preset electricity price load response model can be understood as a model constructed based on the analysis of the influence relationship between the regional electricity price change rate and the load change rate, used to quantify the load response to regional electricity price changes, and can be expressed as: In the formula, in, For scheduling time The composition of the pre-adjustment active power of the load in each control zone dimensional load vector; For A diagonal matrix created from the vector elements; For scheduling time The composition of electricity prices before adjustment in each control zone 3D electricity price vector; For scheduling time in the electricity price adjustment strategy The composition of the electricity price adjustment amount at different times in each control zone 3D electricity price change vector; The load response electricity price change coefficient matrix is ​​used to characterize the sensitivity of load in each control area to changes in electricity price. Its parameters can be identified offline based on historical operating data and can be updated periodically based on the collected data to adapt to changes in user electricity consumption behavior (user load). For scheduling time Lower control area Load response control area The coefficient of change in electricity price; For scheduling time Lower control area The change in load, For scheduling time Lower control area The original load; For scheduling time Lower control area The change in electricity price; For scheduling time Lower control area The original electricity price; For scheduling time The vector composed of the active power of the load at the corresponding time after the implementation of the electricity price adjustment strategy in each control area.

[0037] In practical applications, the active power of the load before adjustment, the electricity price before adjustment, and the time-based electricity price adjustment amount in the electricity price adjustment strategy of all control areas are substituted into the above-mentioned preset electricity price load response model for calculation, so as to obtain the required time-based response load active power of each control area. The specific calculation process will not be detailed here.

[0038] Based on the moment-response load active power and moment-time photovoltaic output of all the aforementioned control areas, the preset power balance equations are solved to obtain the moment-time sending-end transmission power and moment-time receiving-end transmission power of each regional tie line. The preset power balance equations can be understood as a set of power constraint equations for the target distribution network, constructed by treating each control area as a node in a regional equivalent network and the regional tie line between any two adjacent control areas as an edge in the regional equivalent network, based on the regional equivalent network. This set of equations includes inter-regional power balance equations and intra-regional power balance equations. It should be noted that for a regional tie line, the two control areas it connects are one receiving end and one sending end. For the regulation area and The connecting lines between them control the area. For regional connection lines The sending area, the control area For regional connection lines The receiving end region.

[0039] The inter-regional power balance equation in this embodiment can be understood as the balance constraint that the sending-end power and receiving-end power of the tie line between two arbitrarily connected control regions at each scheduling time must satisfy, which can be expressed as: In the formula, and The control time is respectively Lower area connection line The transmitting power at the sending end and the transmitting power at the receiving end; For power transmission loss of the regional tie line; For the collection of connecting lines; This is the set of scheduling times.

[0040] The regional power balance equation in this embodiment can be understood as the balance constraint of power within the region under arbitrary control at each scheduling time, and can be expressed as: In the formula, For the time of regulation Lower control area Photovoltaic power output within the time frame; For the time of regulation Lower control area The instantaneous response of the load active power; For all from other control areas Input to control area The sum of inter-regional power transmission capacity; For all areas under regulation Send to other control areas The sum of inter-regional power transmission capacity.

[0041] In the above set of power constraint equations for the target distribution network, for any area tie line In terms of its corresponding sending-end transmission power and receiving end transmission power All are unknowns to be determined; the scheduling time will be... Solving the system of equations simultaneously for all inter-regional power balance equations and intra-regional power balance equations yields the following result: The system of linear equations with unknowns can be solved using the regional equivalent network to obtain the sending-end and receiving-end transmission power of all regional tie lines at the current scheduling time. It should be noted that the specific solution can be found by referring to the existing solution process for linear equations, which will not be detailed here.

[0042] Based on the time-sending and time-receiving transmission power of each regional tie line, the value of inter-regional tie line transmission power is summarized and analyzed to obtain the corresponding time-based cross-regional transmission value. In the formula, The value of cross-regional power transmission at time t under scheduling time; and These are the regional tie lines at scheduling time t. Sending area and receiving region Electricity price after adjustment; For regional connection lines The unit transmission cost.

[0043] Based on the real-time photovoltaic output and real-time load active power of each control area, a summary analysis of the photovoltaic output absorption deviation of the power grid is performed to obtain the real-time curtailment cost. The real-time curtailment cost can be understood as the curtailment loss caused by the total real-time photovoltaic absorption deviation of the entire target distribution network, obtained by analyzing the actual real-time photovoltaic absorption of each control area based on the principle of local photovoltaic power absorption. The specific calculation process for the real-time curtailment cost includes: 1) The actual local photovoltaic power consumption is obtained by taking the minimum value of the instantaneous photovoltaic output and the instantaneous load active power in each control zone, and is expressed as: In the formula, For the regulation area During scheduling Actual local photovoltaic power consumption: When a certain control area is under dispatch... When the photovoltaic output at a given moment does not exceed the active power of the response load at that moment, it is considered that all the photovoltaic output in the control area has been absorbed by the local load. When the photovoltaic output exceeds the active power of the load response at a given moment, it is assumed that the entire load of the region is supplied by local photovoltaic power. The excess is transmitted to other control areas or the upper-level power grid through the corresponding regional interconnection line, and is not actually absorbed in the region. .

[0044] 2) The actual local photovoltaic power absorbed by the distribution network is obtained by summing up the actual local photovoltaic power absorbed by all control areas: In the formula, For scheduling time The downstream distribution network actually absorbs photovoltaic power locally; To regulate the total number of regions; The moment-to-moment photovoltaic output of the distribution network is obtained by summing the moment-to-moment photovoltaic output of all control areas. .

[0045] 3) The cost of curtailment at any given time is calculated based on the deviation between the photovoltaic output of the distribution network at that time and the actual local consumption of photovoltaic power by the distribution network: In the formula, The penalty coefficient for photovoltaic power not being consumed locally in the target distribution network; For scheduling time Cost of abandoning light at the next moment.

[0046] Based on the value of inter-regional power transmission at the specified time and the cost of curtailment of solar power at the specified time, the execution reward for the corresponding electricity price adjustment strategy is obtained; wherein, the execution reward is calculated based on the following reward function expression: In the formula, To execute the scheduling time The incentives that can be obtained from implementing the electricity price adjustment strategy.

[0047] It should be noted that the regional electricity price optimization model in this embodiment uses the value function-based DQN (Deep QNetwork) training algorithm to train the reinforcement learning agent, including an evaluation Q-network and a target Q-network. The evaluation Q-network is used to output the action value of each candidate action given a state. Based on this, the regional electricity price adjustment action (regional electricity price adjustment strategy) is generated; the target Q network is used to construct the temporal difference (TD) target value to improve training stability, and its network parameters are obtained by periodically and synchronously updating the evaluation Q network parameters according to a preset step size; the specific construction methods of the evaluation Q network and the target Q network can refer to the existing DQN training algorithm, which will not be detailed here.

[0048] In practical applications, to make DQN suitable for regional electricity price adjustment decisions, the electricity price adjustment for each control region is pre-discretely divided into several tiers, forming a finite set of candidate actions. The evaluation Q-network assesses the value of the candidate actions and outputs the corresponding Q-value. During training, one scheduling cycle is considered one round, and at each scheduling time... Get the current state vector An ε-greedy strategy is used to select regional electricity price adjustment actions from the candidate action set. Subsequently, the regional electricity price adjustment action is substituted into the aforementioned preset electricity price load response model to obtain the active power of the load response at any given time. Then, the preset power balance equations are solved to obtain the regional tie-line power (the time-based transmission power at the sending end and the time-based transmission power at the receiving end of each regional tie-line). Based on this, the time-based value of cross-regional power transmission and the time-based cost of curtailment are calculated to obtain the immediate reward. and form interactive samples The data is stored in the experience replay pool. During the parameter update phase, a small batch of samples is randomly sampled from the experience replay pool. A supervision signal is constructed based on the TD target. The mean square error loss between the action value output by the evaluation Q network and the target value is calculated, and the evaluation Q network parameters are updated using the gradient descent optimization method. By introducing the experience replay and target Q network update mechanism, the correlation of samples can be effectively reduced and training oscillations can be suppressed, thereby improving convergence stability. When the preset number of training rounds is reached or the cumulative reward meets the convergence criterion, the trained regional electricity price optimization model is obtained, which is used to output the electricity price adjustment amount for each regulation area online.

[0049] The zonal electricity price optimization model constructed based on the above methods and steps can be used interactively with the actual target distribution network. By analyzing the state variables transmitted in real time from the target distribution network, the optimal electricity price adjustment amount for each control area can be obtained, thereby generating a zonal electricity price adjustment strategy that can ensure the efficient access and local consumption of renewable energy in the target distribution network.

[0050] S15. Based on the zonal electricity price adjustment strategy, perform load optimization and control on the target distribution network. After obtaining the zonal electricity price adjustment strategy, it can be directly applied to guide the formulation of electricity price strategies in various control areas within the target distribution network. This allows for the adjustment of electricity prices in different control areas to guide changes in load demand response to adapt to fluctuations in photovoltaic output in different areas, thereby improving the distribution network dispatch efficiency and the access capability of large-scale renewable energy.

[0051] To verify the effectiveness of the distribution network operation optimization method based on regional electricity pricing provided by this invention, this embodiment also... Figure 2 The target distribution network shown is subjected to operational optimization simulation. Figure 2 The target distribution network shown has 3 distributed photovoltaic (PV) systems (PV1 at node 15, PV2 at node 20, and PV3 at node 30) and 33 load nodes. Based on the obtained distribution network line topology and day-ahead forecast data, regional division optimization is performed to obtain the following solution. Figure 3 The diagram showing the regional division results is as follows; Figure 3 The target distribution network shown is divided into four control zones, and the boundaries of these zones fall in areas with relatively weak electrical coupling. This results in stronger electrical connections between nodes within each zone, and each control zone has a high power self-balancing capability, which reduces reliance on cross-regional power exchange. This provides clear control boundaries and implementation targets for subsequent load control based on regional electricity pricing.

[0052] Based on the regional division results, the operation status of the target distribution network is analyzed using a regional electricity price optimization model to determine the regional electricity price adjustment strategy. Based on the obtained regional electricity price adjustment strategy, load optimization and control are then implemented on the target distribution network to obtain... Figure 4 The diagram showing the optimization effect is as follows; Figure 4 As shown, during peak photovoltaic (PV) output periods, significant curtailment occurs due to insufficient local load and limited inter-regional transmission capacity before regulation. By adopting the zonal pricing strategy of this invention, load is guided to shift to periods and regions with high PV output through price adjustments, thereby improving local PV absorption capacity. This demonstrates that this method can achieve better operational optimization results by considering the combined objectives of inter-regional transmission value and curtailment costs. Furthermore, taking regulation region 1 as an example... Figure 5The study demonstrates the operational status of control area 1 before and after the implementation of the regional pricing strategy. After adopting the regional pricing strategy, the overall electricity price in control area 1 decreased during the daytime hours when photovoltaic output was high, thus guiding the load towards the midday photovoltaic peak, resulting in a significant increase in load around 12:00 compared to before the adjustment. This change helps improve the local photovoltaic absorption capacity of the area and reduces reliance on inter-regional power transmission. Simultaneously, during the evening peak around 19:00, the electricity price in control area 1 increased relative to before the adjustment, and the load response exhibited peak shaving, with a decrease in load compared to before the adjustment. This reduces peak-hour power supply pressure and alleviates the peak power exchange spikes on inter-regional interconnection lines, verifying the feasibility and effectiveness of the proposed regional pricing optimization strategy at the regional scale.

[0053] This invention provides a method for obtaining the line topology of a target distribution network, day-ahead forecast data including load forecast data and photovoltaic output forecast data, and time-of-use electricity price data. Based on the line topology and day-ahead forecast data, the network is divided into regions, resulting in a network region division with multiple control areas. The day-ahead forecast data and time-of-use electricity price data are then processed according to the region division results to obtain regional day-ahead forecast data and regional time-of-use electricity price data. Based on the region division results, regional day-ahead forecast data, and regional time-of-use electricity price data, an operational status analysis is performed using a pre-trained regional electricity price optimization model aimed at maximizing the value of inter-regional power transmission and minimizing the cost of curtailment. This yields a regional electricity price adjustment strategy and a technical solution for load optimization and control of the target distribution network based on the regional electricity price adjustment strategy. This load demand control optimization mechanism, based on the distribution network operational status adjustment of regional electricity prices, aims to motivate users to participate in the local consumption of renewable energy by reasonably optimizing regional electricity prices, thereby improving the dispatch efficiency of the distribution network, solving problems such as insufficient local consumption of renewable energy, ensuring efficient access and local consumption of renewable energy, and ultimately improving the stability and economy of the power grid operation.

[0054] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.

[0055] In one embodiment, such as Figure 6 As shown, a distribution network operation optimization system based on regional electricity pricing is provided, the system comprising: The data acquisition module 100 is used to acquire the line topology, day-ahead forecast data, and time-of-use electricity price data of the target distribution network; the day-ahead forecast data includes load forecast data and photovoltaic output forecast data. The region division module 200 is used to divide regions according to the line topology and the day-ahead forecast data to obtain region division results; the region division results include multiple control regions. The data processing module 300 is used to process the day-ahead forecast data and the time-of-use electricity price data according to the regional division results, so as to obtain the corresponding regional day-ahead forecast data and regional time-of-use electricity price data. The electricity price analysis module 400 is used to perform operational status analysis based on the regional division results, the day-ahead forecast data of the regional divisions, and the time-of-use electricity price data of the regional divisions, and to obtain regional electricity price adjustment strategies; the regional electricity price optimization model is trained and constructed with the goal of maximizing the value of cross-regional power transmission and minimizing the cost of curtailment of solar power. The load control module 500 is used to perform load optimization control on the target distribution network according to the zonal electricity price adjustment strategy.

[0056] In one embodiment, the region division module is further configured to: With maximizing region partitioning performance as the optimization objective, a region partitioning objective function is constructed; the region partitioning performance is obtained by weighted summation based on preset region partitioning performance indicators. Based on the line topology and the day-ahead forecast data, the objective function for the region division is solved using the improved Grey Wolf algorithm to obtain the region division result.

[0057] In one embodiment, the regional electricity price optimization model is obtained by performing grid dispatch interaction simulation training on the target distribution network based on reinforcement learning.

[0058] Specific limitations regarding the distribution network operation optimization system based on regional electricity pricing can be found in the limitations of the distribution network operation optimization method based on regional electricity pricing mentioned above. The corresponding technical effects are equivalent and will not be repeated here. Each module in the aforementioned distribution network operation optimization system based on regional electricity pricing can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0059] In summary, the present invention provides a distribution network operation optimization method and system based on regional electricity pricing. This method utilizes a load demand regulation optimization mechanism that adjusts regional electricity prices based on the distribution network's operating status. By rationally optimizing regional electricity prices, it motivates users to participate in the local consumption of renewable energy, improves the dispatch efficiency of the distribution network, solves problems such as insufficient local consumption of renewable energy, ensures efficient access and local consumption of renewable energy, and ultimately enhances the stability and economy of the power grid operation.

[0060] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0061] The above-described embodiments are merely preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. A power distribution network operation optimization method based on partitioned electricity price, characterized in that, The method includes: Obtain the line topology, day-ahead forecast data, and time-of-use electricity price data of the target distribution network; the day-ahead forecast data includes load forecast data and photovoltaic output forecast data; Based on the line topology and the day-ahead forecast data, the regions are divided to obtain the region division results; the region division results include multiple control regions. Based on the regional division results, the day-ahead forecast data and the time-of-use electricity price data are processed to obtain the corresponding regional day-ahead forecast data and regional time-of-use electricity price data. Based on the regional division results, the day-ahead forecast data of the regional divisions, and the time-of-use electricity price data of the regional divisions, an operational status analysis is performed based on a pre-built regional electricity price optimization model to obtain a regional electricity price adjustment strategy; the regional electricity price optimization model is trained and constructed with the goal of maximizing the value of cross-regional power transmission and minimizing the cost of curtailment of solar power. Based on the aforementioned regional electricity price adjustment strategy, load optimization and control are performed on the target distribution network. The zonal electricity price optimization model is based on reinforcement learning and is trained through grid dispatch interaction simulation on the target distribution network. The state space and action space of the zonal electricity price optimization model are derived from the real-time operating characteristics and electricity price adjustment strategy of the target distribution network, respectively. The reward function of the zonal electricity price optimization model is constructed based on maximizing the difference between the value of cross-regional power transmission and the cost of curtailment of solar power. The real-time operating characteristics include day-ahead zonal load data, day-ahead zonal photovoltaic output data, zonal time-of-use electricity price data, dispatch time, and regional division results. The electricity price adjustment strategy includes the time-based electricity price adjustment amount for each control zone. The steps for obtaining the execution reward of the electricity price adjustment strategy during the training process of the zonal electricity price optimization model include: Obtain the photovoltaic output of each of the control zones at the time after the implementation of the electricity price adjustment strategy; Based on the pre-adjustment load active power, the pre-adjustment electricity price, and the time-based electricity price adjustment amount in the electricity price adjustment strategy for each control zone, the corresponding time-based response load active power is obtained based on the preset electricity price load response model. Based on the active power of the load and the photovoltaic output at the moment of all the control areas, the preset power balance equations are solved to obtain the transmission power at the sending end and the transmission power at the receiving end of each regional tie line; the preset power balance equations include inter-regional power balance equations and intra-regional power balance equations. Based on the time-sending and time-receiving transmission power of each regional tie line, the value of inter-regional tie line transmission power is summarized and analyzed to obtain the corresponding time-based cross-regional transmission value. The grid photovoltaic output absorption deviation is summarized and analyzed based on the real-time photovoltaic output and real-time response load active power of each control area to obtain the real-time curtailment cost; The execution reward for the corresponding electricity price adjustment strategy is obtained based on the cross-regional power transmission value at the specified time and the curtailment cost of the specified time.

2. The distribution network operation optimization method based on regional electricity pricing as described in claim 1, characterized in that, The step of dividing the region based on the line topology and the day-ahead forecast data to obtain the region division result includes: With maximizing region partitioning performance as the optimization objective, a region partitioning objective function is constructed; the region partitioning performance is obtained by weighted summation based on preset region partitioning performance indicators. Based on the line topology and the day-ahead forecast data, the objective function for the region division is solved using the improved Grey Wolf algorithm to obtain the region division result.

3. The distribution network operation optimization method based on regional electricity pricing as described in claim 2, characterized in that, The preset region division performance indicators include region modularity index, region active power balance index, and region reactive power balance index. The regional modularity index is constructed based on regional electrical distance analysis of the target distribution network. The regional active power balance index is constructed based on the matching relationship analysis of regional photovoltaic active power output and load active power demand of the target distribution network. The regional reactive power balance index is constructed based on the matching relationship analysis of regional photovoltaic reactive power output and load reactive power demand in the target distribution network.

4. The distribution network operation optimization method based on regional electricity pricing as described in claim 3, characterized in that, The steps for constructing the regional modularity index include: Based on the line topology, the electrical distance between each node pair in the target distribution network is obtained, and the electrical edge weight of the corresponding node pair is obtained based on each electrical distance. Based on the electrical edge weights of all the node pairs, the sum of the distribution network edge weights is obtained, and the electrical edge weights of all edges connected to the same node are accumulated to obtain the weighted degree of the corresponding node. The expected electrical connection weight of the corresponding node pair is obtained by summing the weighted degree of the two nodes in each node pair and the edge weight of the distribution network. The deviations between the electrical edge weights and the corresponding expected electrical connection weights of each node pair are summarized and analyzed to obtain the regional modularity index.

5. A distribution network operation optimization system based on regional electricity pricing, characterized in that, The system, employing the distribution network operation optimization method based on regional electricity pricing as described in claim 1, comprises: The data acquisition module is used to acquire the line topology, day-ahead forecast data, and time-of-use electricity price data of the target distribution network; the day-ahead forecast data includes load forecast data and photovoltaic output forecast data. The region division module is used to divide regions according to the line topology and the day-ahead forecast data to obtain the region division result; the region division result includes multiple control regions. The data processing module is used to process the day-ahead forecast data and the time-of-use electricity price data according to the regional division results, so as to obtain the corresponding regional day-ahead forecast data and regional time-of-use electricity price data. The electricity price analysis module is used to perform operational status analysis based on the regional division results, the day-ahead forecast data of the regions, and the time-of-use electricity price data of the regions, and to obtain regional electricity price adjustment strategies based on a pre-built regional electricity price optimization model; the regional electricity price optimization model is trained and constructed with the goal of maximizing the value of cross-regional power transmission and minimizing the cost of curtailment of solar power. The load control module is used to perform load optimization control on the target distribution network according to the zonal electricity price adjustment strategy.

6. The distribution network operation optimization system based on regional electricity pricing as described in claim 5, characterized in that, The region division module is also used for: With maximizing region partitioning performance as the optimization objective, a region partitioning objective function is constructed; the region partitioning performance is obtained by weighted summation based on preset region partitioning performance indicators. Based on the line topology and the day-ahead forecast data, the objective function for the region division is solved using the improved Grey Wolf algorithm to obtain the region division result.

Images