Coordinated optimization control method and device for power distribution network and microgrid

Abstract

The application provides a collaborative optimization control method and device for a power distribution network and a microgrid, relates to the technical field of smart grids, and comprises the following steps: obtaining power micro monitoring data flow; performing multi-level fault prediction and tracing; based on load time-varying characteristics, performing multi-level power flow reconstruction scheduling on the power micro collaborative power grid according to a power micro fault prediction and tracing map, and constructing a power micro reconstruction scheduling space; introducing a reconstruction scheduling risk analysis model to perform reconstruction scheduling risk filtering; performing power supply quality fitting evaluation filtering based on load time-varying characteristics; performing double-branch guided difference breeding optimization on the power micro reconstruction scheduling space to obtain a power micro reconstruction scheduling optimization result. The application solves the technical problem that the power distribution network and the microgrid are difficult to effectively respond to faults and load fluctuations in the prior art, leading to insufficient stability of the power micro collaborative power grid, improves the intelligent scheduling and management capability of the power micro collaborative power grid, and thus improves the power supply reliability and stability.

Description

Technical Field

[0001] This application relates to the field of smart grid technology, specifically to a collaborative optimization control method and apparatus for distribution networks and microgrids. Background Technology

[0002] Currently, existing distribution networks and microgrids typically employ passive response and centralized optimization architectures for coordinated operation, which struggle to cope with the rapid and uncertain nature of faults and load fluctuations. Due to the lack of accurate fault prediction and proactive coordination mechanisms across multiple time scales, they often exhibit insufficient stability under disturbances, leading to power outages or voltage quality degradation, severely hindering the improvement of the resilience and reliability of distribution-microgrid coordinated systems. With the application of renewable energy in microgrids, grid load fluctuations have become more severe. Existing coordinated dispatch methods fail to adequately consider load changes and struggle to adjust the grid's operating status in real time during load fluctuations and grid topology changes, thus impacting grid stability and power quality.

[0003] In summary, existing technologies suffer from insufficient stability of the distribution network and microgrid collaborative grid due to the inability of the distribution network and microgrid to effectively cope with faults and load fluctuations. Summary of the Invention

[0004] The purpose of this application is to provide a collaborative optimization control method and apparatus for distribution networks and microgrids, in order to solve the technical problem in the prior art where the distribution network and microgrid are unable to effectively cope with faults and load fluctuations, resulting in insufficient stability of the distribution-microgrid collaborative network.

[0005] To achieve the above objectives, this application provides a collaborative optimization control method and apparatus for distribution networks and microgrids.

[0006] Firstly, this application provides a coordinated optimization control method for distribution networks and microgrids. This method is implemented using a coordinated optimization control device for distribution networks and microgrids. The method includes: real-time monitoring of the distribution-microgrid coordinated network to obtain a distribution-microgrid monitoring data stream, wherein the distribution-microgrid coordinated network includes a distribution network, a first microgrid group, and a second microgrid group; performing multi-level fault prediction and tracing of the distribution-microgrid coordinated network based on the distribution-microgrid monitoring data stream to establish a distribution-microgrid fault prediction and tracing map; and based on the time-varying characteristics of the load, ... Based on the distribution micro-fault prediction and tracing map, multi-level power flow reconfiguration scheduling is performed on the distribution micro-coordinated power grid to construct a distribution micro-reconfiguration scheduling space. A reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk in the distribution micro-reconfiguration scheduling space, establishing a first filtered reconfiguration scheduling branch. Based on the time-varying characteristics of the load, power quality fitting evaluation filtering is performed on the distribution micro-reconfiguration scheduling space to establish a second filtered reconfiguration scheduling branch. Based on the first filtered reconfiguration scheduling branch and the second filtered reconfiguration scheduling branch, dual-branch guided differential propagation optimization is performed on the distribution micro-reconfiguration scheduling space to obtain the distribution micro-reconfiguration scheduling optimization result.

[0007] Optionally, the first microgrid group includes multiple grid-connected microgrids, and the second microgrid group includes multiple islanded microgrids.

[0008] Optionally, the distribution network monitoring data stream is cleaned and classified to obtain a distribution network monitoring sequence, multiple first microgrid monitoring sequences, and multiple second microgrid monitoring sequences; based on a future time zone, fault prediction and tracing are performed on the distribution network according to the distribution network monitoring sequences to obtain a distribution network fault prediction and tracing path; based on the future time zone, fault prediction and tracing are performed on the first microgrid group according to the multiple first microgrid monitoring sequences to obtain a first microgrid fault prediction and tracing distribution; based on the future time zone, fault prediction and tracing are performed on the second microgrid group according to the multiple second microgrid monitoring sequences to obtain a second microgrid fault prediction and tracing distribution; and a distribution network fault prediction and tracing map is generated based on the distribution network fault prediction and tracing path, the first microgrid fault prediction and tracing distribution, and the second microgrid fault prediction and tracing distribution.

[0009] Optionally, load monitoring data of the distribution micro-network is obtained, and the load monitoring data is trend-predicted based on the future time zone to generate the time-varying load characteristics.

[0010] Optionally, based on the time-varying load characteristics, power flow reconfiguration scheduling is performed on the distribution network according to the distribution network fault prediction and tracing path to obtain a first reconfiguration scheduling decision domain; based on the time-varying load characteristics, power flow reconfiguration scheduling is performed on the first microgrid group according to the first microgrid fault prediction and tracing distribution to obtain a second reconfiguration scheduling decision domain; based on the time-varying load characteristics, power flow reconfiguration scheduling is performed on the second microgrid group according to the second microgrid fault prediction and tracing distribution to obtain a third reconfiguration scheduling decision domain; and a multi-level reconfiguration scheduling decision combination is performed based on the first reconfiguration scheduling decision domain, the second reconfiguration scheduling decision domain, and the third reconfiguration scheduling decision domain to generate the distribution-microgrid reconfiguration scheduling space.

[0011] Optionally, a distribution-micro coordinated model is established by performing twin modeling based on the distribution-micro coordinated power grid; the distribution-micro coordinated model is simulated and controlled according to each distribution-micro reconfiguration scheduling scheme in the distribution-micro reconfiguration scheduling space to obtain simulation data for each reconfiguration; the reconfiguration scheduling risk analysis model is activated, and the reconfiguration scheduling risk analysis indicators of the reconfiguration scheduling risk analysis model include electrical safety risk, power flow oscillation risk, and power supply loss risk; the simulation data for each reconfiguration are input into the reconfiguration scheduling risk analysis model to obtain multiple reconfiguration scheduling risk sequences; based on the multiple reconfiguration scheduling risk sequences, the distribution-micro reconfiguration scheduling space is selected according to the reconfiguration scheduling risk constraints to generate the first filtered reconfiguration scheduling branch.

[0012] Optionally, the a-th distribution micro-reconfiguration scheduling scheme is extracted from the distribution micro-reconfiguration scheduling space, where a is a positive integer; the power supply evaluation retrieval of the distribution micro-cooperative grid is performed using the load time-varying characteristics and the a-th distribution micro-reconfiguration scheduling scheme as the a-th power supply evaluation retrieval conditions to obtain the a-th historical power supply quality evaluation sequence; support evaluation is performed on each historical power supply quality evaluation coefficient in the a-th historical power supply quality evaluation sequence to obtain the a-th evaluation support sequence; credible attention allocation is performed on the a-th historical power supply quality evaluation sequence based on the a-th evaluation support sequence to obtain the a-th evaluation attention allocation sequence; weighted aggregation is performed on the a-th historical power supply quality evaluation sequence based on the a-th evaluation attention allocation sequence to generate the a-th power supply quality fitting evaluation coefficient; if the a-th power supply quality fitting evaluation coefficient is greater than or equal to the power supply quality fitting evaluation threshold, the a-th distribution micro-reconfiguration scheduling scheme is added to the second filtering reconfiguration scheduling branch.

[0013] Optionally, the intersection of the first filtering reconfiguration scheduling branch and the second filtering reconfiguration scheduling branch is analyzed to obtain a first optimization domain for reconfiguration scheduling; guided differential propagation optimization is performed on the micro-reconfiguration scheduling space based on the first filtering reconfiguration scheduling branch to obtain a second optimization domain for reconfiguration scheduling; guided differential propagation optimization is performed on the micro-reconfiguration scheduling space based on the second filtering reconfiguration scheduling branch to obtain a third optimization domain for reconfiguration scheduling; the first optimization domain, the second optimization domain, and the third optimization domain are merged to establish a fourth optimization domain for reconfiguration scheduling; the reconfiguration scheduling cost analysis condition is activated, which includes electrical safety risk weight, power flow oscillation risk weight, and power supply loss risk weight; the reconfiguration scheduling cost minimization optimization is performed on the fourth optimization domain based on the reconfiguration scheduling cost analysis condition to generate the micro-reconfiguration scheduling optimization result.

[0014] Optionally, the distribution micro-reconfiguration scheduling space is differentiated according to the first filtering reconfiguration scheduling branch to obtain a first scheduling difference feature distribution; the distribution micro-reconfiguration scheduling space is guided to perform differential mutation propagation according to the first scheduling difference feature distribution to obtain a first reconfiguration scheduling propagation group; the reconfiguration scheduling risk analysis model is introduced to filter the first reconfiguration scheduling propagation group for reconfiguration scheduling risk to establish a second reconfiguration scheduling propagation group; the second reconfiguration scheduling propagation group is filtered for power supply quality fitting evaluation based on the load time-varying characteristics to obtain the second optimization domain of the reconfiguration scheduling.

[0015] Secondly, this application also provides a coordinated optimization control device for distribution networks and microgrids, used to execute the coordinated optimization control method for distribution networks and microgrids as described in the first aspect, wherein the coordinated optimization control device for distribution networks and microgrids includes: a data monitoring module for real-time monitoring of the distribution-microgrid coordinated network and obtaining distribution-microgrid monitoring data streams, wherein the distribution-microgrid coordinated network includes a distribution network, a first microgrid group, and a second microgrid group; a multi-level fault prediction and tracing module for performing multi-level fault prediction and tracing of the distribution-microgrid coordinated network based on the distribution-microgrid monitoring data streams and establishing a distribution-microgrid fault prediction and tracing map; and a multi-level power flow reconfiguration scheduling module for reconfiguring and scheduling based on the time-varying characteristics of the load and the distribution-microgrid fault prediction data streams. The system employs a traceability mapping method to perform multi-level power flow reconfiguration scheduling on the distribution-micro collaborative power grid, constructing a distribution-micro reconfiguration scheduling space. A reconfiguration scheduling risk filtering module is used to introduce a reconfiguration scheduling risk analysis model to filter the distribution-micro reconfiguration scheduling space, establishing a first filtered reconfiguration scheduling branch. A power quality fitting evaluation filtering module is used to perform power quality fitting evaluation filtering on the distribution-micro reconfiguration scheduling space based on the load's time-varying characteristics, establishing a second filtered reconfiguration scheduling branch. A dual-branch guided differential propagation optimization module is used to perform dual-branch guided differential propagation optimization on the distribution-micro reconfiguration scheduling space based on the first and second filtered reconfiguration scheduling branches, obtaining the distribution-micro reconfiguration scheduling optimization result.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages: By real-time monitoring of the distribution-microgrid coordinated network, a distribution-microgrid monitoring data stream is obtained. The distribution-microgrid coordinated network includes a distribution network, a first microgrid group, and a second microgrid group. Based on the distribution-microgrid monitoring data stream, multi-level fault prediction and tracing are performed on the distribution-microgrid coordinated network to establish a distribution-microgrid fault prediction and tracing map. Based on the time-varying characteristics of the load, multi-level power flow reconfiguration scheduling is performed on the distribution-microgrid coordinated network according to the distribution-microgrid fault prediction and tracing map, constructing a distribution-microgrid reconfiguration scheduling space. A reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk in the distribution-microgrid reconfiguration scheduling space, establishing a first filtered reconfiguration scheduling branch. Based on the time-varying characteristics of the load, a power quality fitting evaluation filter is performed on the distribution-microgrid reconfiguration scheduling space, establishing a second filtered reconfiguration scheduling branch. Based on the first filtered reconfiguration scheduling branch and the second filtered reconfiguration scheduling branch, a dual-branch guided differential propagation optimization is performed on the distribution-microgrid reconfiguration scheduling space to obtain the distribution-microgrid reconfiguration scheduling optimization result. In other words, by monitoring the distribution-micro-grid in real time and acquiring the distribution-micro-grid monitoring data stream, adopting multi-level fault prediction and tracing technology, introducing load time-varying characteristics to support the distribution-micro-grid fault prediction and tracing map, constructing a distribution-micro-grid reconfiguration scheduling space, introducing a reconfiguration scheduling risk analysis model for reconfiguration scheduling risk filtering, performing power supply quality fitting evaluation filtering based on load time-varying characteristics, and adopting a dual-branch guided differential propagation optimization algorithm to optimize the scheduling decision space, the intelligent scheduling and management capabilities of the distribution-micro-grid are improved, thereby enhancing the reliability and stability of power supply.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the collaborative optimization control method for distribution networks and microgrids used in this application.

[0020] Figure 2 This is a schematic diagram of the structure of the collaborative optimization control device for distribution networks and microgrids used in this application.

[0021] Figure labeling: Data monitoring module 11, multi-level fault prediction and tracing module 12, multi-level power flow reconstruction and scheduling module 13, reconstruction and scheduling risk filtering module 14, power quality fitting evaluation and filtering module 15, dual-branch guided differential propagation optimization module 16. Detailed Implementation

[0022] This application provides a collaborative optimization control method and device for distribution networks and microgrids, solving the technical problem of insufficient stability in existing distribution-microgrid collaborative power grids due to the inability of distribution networks and microgrids to effectively cope with faults and load fluctuations. By real-time monitoring of the distribution-microgrid collaborative power grid and acquiring distribution-microgrid monitoring data streams, a multi-level fault prediction and tracing technology is adopted. Load time-varying characteristics are introduced to support the distribution-microgrid fault prediction and tracing map, constructing a distribution-microgrid reconfiguration scheduling space. A reconfiguration scheduling risk analysis model is introduced for reconfiguration scheduling risk filtering, power quality fitting evaluation filtering is performed based on load time-varying characteristics, and a dual-branch guided differential propagation optimization algorithm is used to optimize the scheduling decision space. This improves the intelligent scheduling and management capabilities of the distribution-microgrid collaborative power grid, thereby enhancing power supply reliability and stability.

[0023] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0024] Example 1, please refer to the appendix. Figure 1 This application provides a coordinated optimization control method for distribution networks and microgrids, wherein the coordinated optimization control method for distribution networks and microgrids is executed by a coordinated optimization control device for distribution networks and microgrids, and the coordinated optimization control method for distribution networks and microgrids specifically includes the following steps: Real-time monitoring of the distribution-microgrid collaborative network to obtain distribution-microgrid monitoring data streams, wherein the distribution-microgrid collaborative network includes a distribution network, a first microgrid group, and a second microgrid group.

[0025] Furthermore, this application also includes the following steps: the first microgrid group includes multiple grid-connected microgrids, and the second microgrid group includes multiple islanded microgrids.

[0026] Specifically, a distribution-microgrid coordinated system is a system of coordinated operation between a distribution network and microgrids. It typically consists of a regional distribution network and multiple microgrids within its jurisdiction, including the distribution network, a first microgrid group, and a second microgrid group. The distribution network is an intermediate network that receives electrical energy from the high-voltage transmission network and distributes it to user areas at different voltage levels. It usually has a radial structure and is a key link connecting the main grid and users. The first microgrid group includes multiple grid-connected microgrids, which are electrically connected to the distribution network through a point of common coupling. They can exchange power bidirectionally with the main grid and receive coordination and dispatch commands from the distribution network. The second microgrid group includes multiple islanded microgrids, which are a group of microgrids that are electrically disconnected from the regional distribution network and operate independently.

[0027] For example, suppose a region contains a 10kV distribution network with 5 feeders and 5 microgrids connected to it. The distribution network is a network model containing one 110kV / 10kV step-down substation with a main transformer capacity of 20MVA and 5 10kV feeders. Status identification confirms that 3 microgrids are in grid-connected state: Microgrid 1 is an industrial park with a maximum load of 2.1MW, connected to the grid via feeder F1; Microgrid 2 is a data center with a maximum load of 1.5MW, connected to the grid via feeder F2; Microgrid 3 is a residential community with 500kW photovoltaic power, connected to the grid via feeder F3, forming the first microgrid group. The other 2 microgrids are in islanded state: Microgrid 4 is a hospital equipped with a 1MW gas turbine and energy storage; due to scheduled line maintenance, it was disconnected from feeder F4 at 08:00 that day and entered islanded operation; Microgrid 5 is a remote rural area relying on photovoltaic power and energy storage; due to its remote location, it normally operates as an island, forming the second microgrid group.

[0028] By deploying high-precision sensors and communication modules at key nodes of the distribution network and grid connection points of various microgrids, panoramic operational data, including voltage, current, power, frequency, switch status, and temperature of key equipment, is collected synchronously at a high frequency of tens to hundreds of times per second. This data is then aggregated in real time to a central data processing platform via a high-speed communication network, forming a distribution-microgrid monitoring data stream. This data stream is a continuous stream of collected grid operation data, including information on voltage, current, load power, frequency, and equipment status. By clearly defining and delineating the distribution network and microgrid groups in different operating states, a standardized and modular description of the complex grid structure is achieved.

[0029] Based on the distribution micro-monitoring data stream, multi-level fault prediction and tracing are performed on the distribution micro-coordinated power grid to establish a distribution micro-fault prediction and tracing map.

[0030] Furthermore, this application also includes the following steps: cleaning and classifying the distribution network monitoring data stream to obtain a distribution network monitoring sequence, multiple first microgrid monitoring sequences, and multiple second microgrid monitoring sequences; based on a future time zone, performing fault prediction and tracing on the distribution network according to the distribution network monitoring sequences to obtain a distribution network fault prediction and tracing path; based on the future time zone, performing fault prediction and tracing on the first microgrid group according to the multiple first microgrid monitoring sequences to obtain a first microgrid fault prediction and tracing distribution; based on the future time zone, performing fault prediction and tracing on the second microgrid group according to the multiple second microgrid monitoring sequences to obtain a second microgrid fault prediction and tracing distribution; and generating the distribution network fault prediction and tracing map based on the distribution network fault prediction and tracing path, the first microgrid fault prediction and tracing distribution, and the second microgrid fault prediction and tracing distribution.

[0031] Specifically, the data stream from the distribution network and microgrid monitoring is preprocessed, including data cleaning and classification, removing obvious outliers, filling in temporary data gaps caused by communication interruptions, and categorizing and organizing the data according to its source. The data is then classified according to the grid unit to which it belongs, generating distribution network monitoring sequences, first microgrid monitoring sequences, and second microgrid monitoring sequences, corresponding to the distribution network, the first microgrid group, and the second microgrid group, respectively.

[0032] Using distribution network monitoring sequences, multiple first microgrid monitoring sequences, and multiple second microgrid monitoring sequences, prediction models are driven for different objects to perform fault prediction and tracing. Based on future time zone settings, time-series prediction models are used to predict and trace faults for each monitoring sequence. For the distribution network, a simulation model based on time-series prediction and power flow calculation is used to extrapolate potential faults such as equipment overload and voltage exceedances caused by load growth or distributed generation fluctuations in future time zones, and to simulate their cascading development process, generating distribution network fault prediction and tracing paths. For the first microgrid cluster, multiple first microgrid monitoring sequences are input into a trained fault probability prediction model, outputting the probability of each microgrid experiencing a fault in the future time zone, forming the first microgrid fault prediction and tracing distribution. Similarly, for the second microgrid cluster, multiple second microgrid monitoring sequences are input into a trained fault probability prediction model, outputting the probability of each microgrid experiencing a fault in the future time zone, forming the second microgrid fault prediction and tracing distribution.

[0033] Taking a distribution network as an example, the distribution network monitoring sequence is input into the prediction engine, and a time-series prediction algorithm is used to make ultra-short-term predictions of load and distributed power output in the future time zone. Based on the prediction data, the distribution network power flow calculation and static security analysis modules are driven to perform rolling simulations, scanning for risks of equipment overload or voltage exceeding limits in the distribution network. Once a risk point is identified, such as a line current exceeding its limit in 8 minutes, fault tracing analysis is immediately initiated. This simulates the action logic and delay of the protection devices related to the identified risk point, and deduces the expansion path of the fault, such as from a single line overload to bus voltage loss and load loss caused by protection tripping. Finally, the event sequence is linked together according to timeline and logical relationship, and integrated into a clear distribution network fault prediction and tracing path. The distribution network fault prediction and tracing path is a structured description of one or more fault development sequences from potential causes to consequences, not only indicating the fault but also showing its development path. The future time zone is a future time window for fault prediction, such as the next 15 minutes or 30 minutes. The fault prediction and tracing process for the first and second microgrid groups is similar to that for the distribution network.

[0034] By integrating the distribution network fault prediction and tracing paths, the first microgrid fault prediction and tracing distribution, and the second microgrid fault prediction and tracing distribution, a distribution-microgrid fault prediction and tracing map is formed. This map integrates the fault paths of the distribution network with the fault probability distribution of the microgrid group into a single, visualized graph, clearly showing the fault risk correlations across the entire network.

[0035] By implementing hierarchical fault prediction and global correlation tracing, we have achieved accurate insight and visualization of future risks in the distribution-microgrid collaborative network, clarifying the root cause, development path and scope of impact of faults, while covering distribution networks, grid-connected microgrids and islanded microgrids, ensuring the controllability of the entire network.

[0036] Based on the time-varying characteristics of the load, the distribution micro-fault prediction and tracing map is used to perform multi-level power flow reconfiguration scheduling on the distribution micro-coordinated power grid, and a distribution micro-reconfiguration scheduling space is constructed.

[0037] Furthermore, this application also includes the following steps: obtaining load monitoring data of the distribution micro-network, and performing trend prediction on the load monitoring data according to the future time zone to generate the load time-varying characteristics.

[0038] Specifically, load monitoring data for the distribution-micro-grid collaborative system is obtained from the distribution-micro-grid monitoring data stream. This data includes power consumption data for each load node, such as active power, reactive power, and current, and is then organized into a time series. The load monitoring data undergoes preprocessing, including cleaning and smoothing to eliminate noise. The load forecasting engine is activated, and by combining historical load data, real-time load data, and potential external factors, time series analysis is used to perform rolling forecasts of the load in future time zones at the minute or 15-minute level, generating a high-time-resolution load forecasting baseline curve. For example, an LSTM network is used to perform time series forecasting of the active power of each node, combined with historical load curves, to generate predicted values ​​for future loads. Through the load forecasting baseline curve, time-varying load characteristics are extracted, such as the load ramp-up rate, the expected peak load size and occurrence time, and the intensity of load fluctuations. These time-varying load characteristics are a quantitative description of the load's variation patterns in future time zones, including key feature parameters such as the load forecasting curve, rate of change, peak / valley values, and their occurrence times.

[0039] For example, the total load of a certain region at the current moment is 28.5MW. Historical data shows that the load in this region typically rises steadily during weekday afternoons due to the activation of commercial air conditioning, peaking around 16:00-17:00. Currently, it is a summer weekday, and the weather forecast indicates that the afternoon temperature will rise from 35℃ to 40℃. The load forecasting model, incorporating historical patterns and temperature effects, outputs a baseline load forecast curve for the next 4 hours. The load will rise steadily from the current 28.5MW: 29.1MW at 4:30, 30.2MW at 15:00, 31.8MW at 15:30, 32.5MW at 6:00, 32.0MW at 16:30, 30.5MW at 17:00, and 28.0MW at 17:30. The maximum load was 32.5MW, reaching its peak at 16:30. The maximum ramp rate was from 15:00 to 16:00, with the load increasing by 2.3MW / h. The load fluctuation was 1.52MW over the next 4 hours.

[0040] By accurately predicting short-term loads and extracting features, load fluctuation trends can be grasped in advance, reducing the impact of sudden load increases or decreases on overall stability. The time-varying characteristics of loads can be used for power flow reconfiguration and optimized scheduling of distribution networks and microgrids, thereby improving scheduling accuracy.

[0041] Furthermore, this application also includes the following steps: based on the time-varying load characteristics, performing power flow reconfiguration scheduling on the distribution network according to the distribution network fault prediction tracing path to obtain a first reconfiguration scheduling decision domain; based on the time-varying load characteristics, performing power flow reconfiguration scheduling on the first microgrid group according to the first microgrid fault prediction tracing distribution to obtain a second reconfiguration scheduling decision domain; based on the time-varying load characteristics, performing power flow reconfiguration scheduling on the second microgrid group according to the second microgrid fault prediction tracing distribution to obtain a third reconfiguration scheduling decision domain; and performing multi-level reconfiguration scheduling decision combination based on the first reconfiguration scheduling decision domain, the second reconfiguration scheduling decision domain, and the third reconfiguration scheduling decision domain to generate the distribution-microgrid reconfiguration scheduling space.

[0042] Specifically, based on the distribution network fault prediction tracing path and load time-varying characteristics, and with the goal of load transfer and overload elimination, a series of network reconfiguration schemes are generated through heuristic search, forming the first reconfiguration scheduling decision domain. Power flow reconfiguration scheduling is a control process that optimizes power flow distribution, eliminates equipment overruns, and restores out-of-power loads by changing the state of switches in the distribution network, adjusting the network topology, and reallocating the output of controllable resources such as generators and energy storage. Specifically, based on the load time-varying characteristics, the predicted load curves for each node in the future time zone are determined. Potential risk nodes are identified based on the distribution network fault prediction tracing path, and power flow reconfiguration calculations are performed on the distribution network in conjunction with grid operation constraints. Power flow reconfiguration scheduling optimizes and adjusts the power flow of the distribution network under the premise of considering grid constraints. By changing the state of switches in the distribution network, adjusting the network topology, and reallocating controllable resources, the grid can still operate safely and stably under future load fluctuations and potential faults. Through heuristic search, all feasible switching operation combinations after reconfiguration are traversed. Each combination is evaluated to determine whether it can safely transfer the load on the fault path to other healthy feeders under the future scenario described by the time-varying characteristics of the load, and to meet all safety constraints such as line capacity and node voltage. All feasible switching operation combinations that meet the requirements and their corresponding power flow results are summarized to form the first reconfiguration scheduling decision domain for the distribution network.

[0043] For example, suppose the fault prediction and tracing map shows that a certain fault path is due to an overcurrent in feeder F1, leading to cable overheating, protection tripping, and a load loss of 4.8MW. The load time-varying characteristics show that the current load of feeder F1 is 4.8MW, and it is predicted to rise to 5.0MW in 15 minutes. The adjacent feeder F2 currently has a load of 3.5MW, with a 1.5MW margin; feeder F3 currently has a load of 4.0MW, with a 1.0MW margin. The core risk is overload of feeder F1, and the goal is to transfer at least 0.2MW of load from feeder F1 within a 15-minute time window. Scheme 1: Close tie switch SW12 to connect F1 and F2. Simulation calculations show that 1.0MW of load at the end of feeder F1 can be transferred to F2. After the transfer, the current of F1 drops to a safe range, approximately 260A, while the current of F2 rises to 4.5MW, still within the safe limit of 5.0MW. Option 2 involves closing the tie switch SW13 to connect F1 and F3. Simulation calculations show that the 1.5MW load in the middle section of feeder F1 can be transferred to F3. After the transfer, the current in F1 drops to a safe range of approximately 250A, while the current in F3 rises to 5.5MW, exceeding its safety limit of 5.2MW, thus violating the constraint. Therefore, this option is eliminated. Option 3 involves simultaneously closing SW12 and the sectionalizing switch SW11 to segment F1. Simulation calculations show that the 1.8MW load downstream of F1 can be transferred to F2 and powered by the upstream section of F1, respectively, by 1.0MW and 0.8MW, with all lines operating safely. Therefore, the first reconfiguration scheduling decision domain obtained in the end contains two feasible schemes, namely Scheme 1 and Scheme 2. The operation command of Scheme 1 is to close the tie switch SW12, and the expected effect is to reduce the F1 load by 1.0MW to 3.8MW and eliminate the overload risk. The operation command of Scheme 2 is to close the tie switch SW12 and at the same time close the sectionalizing switch SW11, and the expected effect is to reduce the F1 load by 1.8MW to 3.0MW and completely eliminate the overload risk.

[0044] Similarly, for the first microgrid group, based on the time-varying characteristics of the load and the fault prediction and tracing distribution of the first microgrid, a series of internal scheduling schemes are generated with the aim of adjusting the exchange power between each microgrid and the distribution network and smoothing internal fluctuations. These schemes include operations such as energy storage charging and discharging plans and controllable distributed power generation output adjustments, forming the second reconfiguration scheduling decision domain for the first microgrid group. For the second microgrid group, based on the time-varying characteristics of the load and the fault prediction and tracing distribution of the second microgrid, and with the aim of ensuring its own continuous and stable operation, a series of autonomous scheduling schemes are generated, including operations such as diesel generator start-up and shutdown and load hierarchical control, forming the third reconfiguration scheduling decision domain.

[0045] The first, second, and third reconfiguration scheduling decision domains are combined into a complete global scheduling plan, forming the distribution microgrid reconfiguration scheduling space. This space includes multiple distribution microgrid reconfiguration scheduling schemes for the distribution network, the first microgrid group, and the second microgrid group. By generating corresponding distribution microgrid reconfiguration scheduling schemes for the distribution network, the first microgrid group, and the second microgrid group respectively, the distribution microgrid reconfiguration scheduling space is constructed, reducing the probability of potential overloads, line oscillations, and voltage anomalies.

[0046] A reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk of the micro-reconfiguration scheduling space and establish a first filtering reconfiguration scheduling branch.

[0047] Furthermore, this application also includes the following steps: performing twin modeling based on the distribution-micro coordinated power grid to establish a distribution-micro coordinated model; performing simulation control on the distribution-micro coordinated model according to each distribution-micro reconfiguration scheduling scheme in the distribution-micro reconfiguration scheduling space to obtain simulation data for each reconfiguration; activating the reconfiguration scheduling risk analysis model, wherein the reconfiguration scheduling risk analysis indicators of the reconfiguration scheduling risk analysis model include electrical safety risk, power flow oscillation risk, and power supply loss risk; inputting the simulation data for each reconfiguration into the reconfiguration scheduling risk analysis model to obtain multiple reconfiguration scheduling risk sequences; and selecting the distribution-micro reconfiguration scheduling space based on the multiple reconfiguration scheduling risk sequences and according to the reconfiguration scheduling risk constraints to generate the first filtered reconfiguration scheduling branch.

[0048] Specifically, based on the distribution-microgrid coordinated network, twin modeling is performed. This involves using digital technology to construct a virtual model in the information space that completely corresponds to the physical distribution-microgrid coordinated network, resulting in a distribution-microgrid coordinated model. This model can accurately simulate the dynamic behavior of the physical grid for simulation testing and analysis. Each distribution-microgrid reconfiguration scheduling scheme within the distribution-microgrid reconfiguration scheduling space is input into the distribution-microgrid coordinated model for simulation control, simulating the dynamic operation process in future time zones and recording detailed reconfiguration simulation data. Each reconfiguration simulation data includes simulation data corresponding to each scheme obtained after simulation operation, including node voltage, line load, energy storage SOC, and power loss.

[0049] A reconfiguration scheduling risk analysis model is activated to quantify the grid operation risk of each distribution / microgrid reconfiguration scheduling scheme. The reconfiguration scheduling risk analysis indicators of the model include electrical safety risk, power flow oscillation risk, and power supply loss risk. Data is obtained from historical distribution network and microgrid monitoring systems, including node voltage, line power, energy storage SOC, load curves, historical fault events, historical scheduling schemes, and corresponding system operation indicators. The data is preprocessed, including data cleaning and normalization, to obtain processed standard data. Feature extraction is performed from the preprocessed standard data, including node voltage, line power, energy storage SOC, microgrid status indicators, and potential fault prediction indicators, with corresponding risk indicators identified: electrical safety risk, power flow oscillation risk, and power supply loss risk. The dataset is divided into a 70% training set, a 15% validation set, and a 15% test set. Minimizing the mean squared error is used as the loss function, and hyperparameters are optimized with tree depths of 6–12, learning rates of 0.1–0.3, and the number of trees ranging from 200 to 500. Optimal parameters are selected through cross-validation on the validation set.

[0050] Electrical safety risk assessment primarily evaluates the degree of violation of steady-state safety criteria, such as equipment overload and voltage exceeding limits. Power flow oscillation risk assessment evaluates whether power and voltage exhibit a sustained or increasing oscillation trend after small disturbances or faults, reflecting dynamic stability. Power supply loss risk assessment evaluates the economic risks arising from increased power transmission losses after network reconfiguration. The reconfiguration scheduling risk analysis model does not learn all risk tasks simultaneously. Instead, it first focuses on the most destructive electrical safety risk. Once its predictive performance stabilizes, more refined power flow oscillation and power supply loss risk tasks are gradually introduced. This strategy of progressing from easy to difficult effectively avoids the model getting stuck in local optima during the early stages of training. In each training round, batches of sample data are input into the multi-task learning architecture. Forward propagation outputs predicted values ​​in each of the three risk branches, and then the weighted total loss between these predicted values ​​and the true risk labels is calculated. This loss function is composed of the binary cross-entropy loss of the three tasks with preset weights, such as 0.4 for electrical safety, 0.4 for power flow oscillation, and 0.2 for power supply loss. Using backpropagation and the Adam optimizer, model parameters are iteratively updated based on the loss gradient. Exponential learning rate decay and gradient pruning techniques are employed to ensure the stability and convergence of the training process. Throughout the process, a validation set is used to continuously monitor the model's generalization ability and trigger an early stopping mechanism to prevent overfitting. Training is terminated once the validation set loss stops decreasing for several consecutive rounds. On independent test sets, the model achieved excellent performance with an accuracy exceeding 92% and an AUC-ROC exceeding 0.95. Interpretive techniques such as SHAP were used to analyze its decision-making logic, ensuring that its predictions are not only accurate but also reliable.

[0051] The trained reconfiguration scheduling risk analysis model is integrated into the distribution micro-distribution collaborative scheduling system. The reconfiguration simulation data of each distribution micro-distribution reconfiguration scheduling scheme is input into the reconfiguration scheduling risk analysis model, and multiple reconfiguration scheduling risk sequences are output. Each reconfiguration scheduling risk sequence includes the electrical safety risk, power flow oscillation risk and power supply loss risk of its corresponding distribution micro-distribution reconfiguration scheduling scheme.

[0052] The reconfiguration scheduling risk sequence corresponding to each micro-distribution reconfiguration scheduling scheme is automatically compared with the preset reconfiguration scheduling risk constraints. Schemes that do not meet the reconfiguration scheduling risk constraints are eliminated, and those that meet the constraints on electrical safety risk, power flow oscillation risk, and power supply loss risk are retained and added to the first filtering reconfiguration scheduling branch. The reconfiguration scheduling risk constraints include electrical safety risk constraints, power flow oscillation risk constraints, and power supply loss risk constraints, such as electrical safety risk ≤ 0.7, power flow oscillation risk ≤ 0.7, and power supply loss risk ≤ 0.8. For example, the reconfiguration scheduling risk sequence of Scheme 1 is 0.15 for electrical safety risk, 0.08 for power flow oscillation risk, and 0.1 for power supply loss risk, all of which meet the constraints. The reconfiguration scheduling risk sequence of Scheme 2 is 0.12 for electrical safety risk, 0.75 for power flow oscillation risk, and 0.05 for power supply loss risk, but the power flow oscillation risk does not meet the constraint. The reconfiguration scheduling risk sequence of Scheme 3 is 0.82 for electrical safety risk, 0.05 for power flow oscillation risk, and 0.12 for power supply loss risk, but the electrical safety risk does not meet the constraint. The reconfiguration scheduling risk sequence of Scheme 4 is 0.78 for electrical safety risk, 0.72 for power flow oscillation risk, and 0.09 for power supply loss risk, but neither electrical safety nor power flow oscillation risk meets the constraints. Therefore, Scheme 1 is added to the first filtering reconfiguration scheduling branch. By combining digital twin simulation with machine learning-driven rapid risk analysis, the vast space of alternative schemes is tested, potentially unsafe schemes are screened in advance, and overall energy efficiency is optimized through reconfiguration scheduling risk filtering.

[0053] Based on the time-varying characteristics of the load, the power quality fitting evaluation and filtering of the distribution micro-reconfiguration scheduling space are performed to establish a second filtered reconfiguration scheduling branch.

[0054] Furthermore, this application also includes the following steps: extracting the a-th distribution micro-reconfiguration scheduling scheme according to the distribution micro-reconfiguration scheduling space, where a is a positive integer; using the load time-varying characteristics and the a-th distribution micro-reconfiguration scheduling scheme as the a-th power supply evaluation retrieval conditions, performing power supply quality evaluation retrieval on the distribution micro-cooperative grid to obtain the a-th historical power supply quality evaluation sequence; performing support evaluation on each historical power supply quality evaluation coefficient in the a-th historical power supply quality evaluation sequence to obtain the a-th evaluation support sequence; performing credible attention allocation on the a-th historical power supply quality evaluation sequence according to the a-th evaluation support sequence to obtain the a-th evaluation attention allocation sequence; performing weighted aggregation on the a-th historical power supply quality evaluation sequence according to the a-th evaluation attention allocation sequence to generate the a-th power supply quality fitting evaluation coefficient; if the a-th power supply quality fitting evaluation coefficient is greater than or equal to the power supply quality fitting evaluation threshold, adding the a-th distribution micro-reconfiguration scheduling scheme to the second filtering reconfiguration scheduling branch.

[0055] Specifically, the a-th micro-reconfiguration scheduling scheme is extracted from the micro-reconfiguration scheduling space, where a is a positive integer. Using the load time-varying characteristics and the a-th micro-reconfiguration scheduling scheme as evaluation retrieval conditions, historical power quality data similar to this scheme are queried. Typically, this involves retrieving multiple historical similar scenarios and their corresponding historical power quality evaluation coefficients from a large historical operational database using a similarity matching algorithm, resulting in the a-th historical power quality evaluation sequence. Each historical power quality evaluation coefficient in the a-th historical power quality evaluation sequence represents the power quality of a historical scheduling scheme. It is a comprehensive value, usually calculated based on indicators such as voltage deviation, frequency deviation, and harmonic distortion rate, ranging from [0,1]. A higher value indicates better power quality.

[0056] Support is evaluated for each historical power quality evaluation coefficient within the 'a'-th historical power quality evaluation sequence. This involves counting the frequency of each historical power quality evaluation coefficient within the 'a'-th historical power quality evaluation sequence to obtain the 'a'-th evaluation support sequence. Higher support indicates that the score value appears more frequently in history, thus increasing its statistical reliability. Based on the 'a'-th evaluation support sequence, a reliable attention allocation is performed on the 'a'-th historical power quality evaluation sequence, assigning weights to each power quality evaluation coefficient. Higher weights indicate a greater impact of the coefficient on the scheduling scheme; typically, more frequent scores are assigned higher attention weights.

[0057] Based on the 'a'-th evaluation attention allocation sequence, a weighted aggregation of the 'a'-th historical power quality evaluation sequence is performed. That is, the 'a'-th evaluation attention allocation sequence is used as the weight to perform a weighted average of the 'a'-th historical power quality evaluation sequences, yielding the 'a'-th power quality fitting evaluation coefficient. This coefficient represents the power quality level achievable after the 'a'-th distribution micro-reconfiguration scheduling scheme is executed. The 'a'-th power quality fitting evaluation coefficient is compared with a power quality fitting evaluation threshold. If the threshold is met (i.e., the 'a'-th power quality fitting evaluation coefficient is greater than or equal to the power quality fitting evaluation threshold), the 'a'-th distribution micro-reconfiguration scheduling scheme is added to the second filtering reconfiguration scheduling branch. For other schemes in the distribution micro-reconfiguration scheduling space, the above process is repeated, adding all distribution micro-reconfiguration scheduling schemes greater than or equal to the power quality fitting evaluation threshold to the second filtering reconfiguration scheduling branch. The power quality fitting evaluation threshold is a preset threshold value used to determine whether the power quality of the distribution micro-reconfiguration scheduling scheme reaches an acceptable level. For example, if the power quality score sequence of Scheme 1 is 0.88, 0.90, 0.87, 0.91, ..., and after support evaluation, the scores 0.88 appear 12 times, 0.90 appear 15 times, 0.87 appear 10 times, ..., the attention weights are calculated, and the scores 0.90 and 0.88 with high support receive high weights, which are 0.25 and 0.22 respectively. Then the power quality fitting evaluation coefficient of Scheme 1 is 0.89. If the power quality fitting evaluation threshold is 0.85, then Scheme 1 is greater than or equal to the power quality fitting evaluation threshold and is added to the second filtering reconfiguration scheduling branch.

[0058] By employing power quality evaluation and support evaluation mechanisms, dispatching schemes with good historical performance can be effectively selected, thereby improving the stability and reliability of the power grid. Through weighted aggregation of power quality data, the power quality of different dispatching schemes under various operating conditions can be accurately quantified, thus optimizing the operational efficiency of the power grid.

[0059] Based on the first filtering and reconfiguration scheduling branch and the second filtering and reconfiguration scheduling branch, the micro-reconfiguration scheduling space is subjected to dual-branch guided differential propagation optimization to obtain the micro-reconfiguration scheduling optimization result.

[0060] Furthermore, this application also includes the following steps: performing intersection analysis on the first filtering reconfiguration scheduling branch and the second filtering reconfiguration scheduling branch to obtain a first optimization domain for reconfiguration scheduling; performing guided differential propagation optimization on the micro-reconfiguration scheduling space according to the first filtering reconfiguration scheduling branch to obtain a second optimization domain for reconfiguration scheduling; performing guided differential propagation optimization on the micro-reconfiguration scheduling space according to the second filtering reconfiguration scheduling branch to obtain a third optimization domain for reconfiguration scheduling; merging the first optimization domain, the second optimization domain, and the third optimization domain for reconfiguration scheduling to establish a fourth optimization domain for reconfiguration scheduling; activating the reconfiguration scheduling cost analysis conditions, which include electrical safety risk weight, power flow oscillation risk weight, and power supply loss risk weight; performing reconfiguration scheduling cost minimization optimization on the fourth optimization domain for reconfiguration scheduling according to the reconfiguration scheduling cost analysis conditions to generate the micro-reconfiguration scheduling optimization result.

[0061] Furthermore, this application also includes the following steps: identifying differences in the distribution micro-reconfiguration scheduling space based on the first filtering reconfiguration scheduling branch to obtain a first scheduling difference feature distribution; guiding the distribution micro-reconfiguration scheduling space to perform differential mutation propagation based on the first scheduling difference feature distribution to obtain a first reconfiguration scheduling propagation group; introducing the reconfiguration scheduling risk analysis model to perform reconfiguration scheduling risk filtering on the first reconfiguration scheduling propagation group to establish a second reconfiguration scheduling propagation group; and performing power quality fitting evaluation filtering on the second reconfiguration scheduling propagation group based on the load time-varying characteristics to obtain the second optimization domain of the reconfiguration scheduling.

[0062] Specifically, the intersection of the first and second filtering reconstruction scheduling branches is analyzed to find the solutions that both filtering reconstruction scheduling branches have, i.e., the solutions that simultaneously meet the safety and quality requirements, thus obtaining the first optimization domain of the reconstruction scheduling.

[0063] The first filtering reconfiguration scheduling branch identifies differences in the micro-distribution reconfiguration scheduling space. This involves comparing the schemes in the first filtering reconfiguration scheduling branch with the original schemes in the micro-distribution reconfiguration scheduling space, analyzing the differences in characteristics such as the magnitude and direction of changes in node voltage, line power, and energy storage status, to generate a first scheduling difference feature distribution. This first distribution is derived by comparing the schemes in the safety branch with other schemes in the original reconfiguration scheduling space, extracting key feature difference patterns. Using this first distribution as the target, differential mutation propagation is performed on the micro-distribution reconfiguration scheduling space. This involves differentially mutating the schemes in the micro-distribution reconfiguration scheduling space to generate a new group of scheduling schemes, resulting in the first reconfiguration scheduling propagation group. This first group is a set of new schemes generated by mutating the schemes in the micro-distribution reconfiguration scheduling space towards the direction of the first filtering reconfiguration scheduling branch. All micro-distribution reconfiguration scheduling schemes in the first propagation scheduling group are simulated and controlled. The simulation data is input into a reconfiguration scheduling risk analysis model to obtain multiple reconfiguration scheduling risk sequences. Micro-distribution reconfiguration scheduling schemes that meet the reconfiguration scheduling risk constraints are then selected and added to the second reconfiguration scheduling propagation group. For each micro-reconfiguration scheduling scheme in the second reconfiguration scheduling breeding group, a power quality fitting evaluation filter is performed. That is, multiple historical similar scenarios and their corresponding multiple historical power quality evaluation coefficients are retrieved for each micro-reconfiguration scheduling scheme, and support evaluation is performed. By calculating the power quality fitting evaluation coefficient, all micro-reconfiguration scheduling schemes that meet the power quality fitting evaluation threshold are selected to obtain the second optimization domain of reconfiguration scheduling.

[0064] Similarly, based on the second filtering reconfiguration scheduling branch, the distribution micro-reconfiguration scheduling space is differentiated and mutated in the direction of the second filtering reconfiguration scheduling branch to generate a new group of scheduling schemes. The mutated schemes are then subjected to reconfiguration scheduling risk filtering and power quality fitting evaluation filtering to obtain the third optimization domain for reconfiguration scheduling.

[0065] The first optimization domain of the reconfiguration scheduling is a set of safe and high-quality solutions obtained through intersection analysis. The second optimization domain is a set of new solutions obtained by the safety branch (i.e., the first filtering reconfiguration scheduling branch), guided by propagation and then filtered. The third optimization domain is a set of new solutions obtained by the quality branch (i.e., the second filtering reconfiguration scheduling branch), guided by propagation and then filtered. Merging the first, second, and third optimization domains yields the fourth optimization domain. The reconfiguration scheduling cost analysis condition is activated, which includes the reconfiguration scheduling cost analysis function: Reconfiguration scheduling cost = Electrical safety risk weight. Electrical safety risk + power flow oscillation risk weighting Power flow oscillation risk + power supply loss risk weight Power supply loss risk. For example, the cost of rescheduling is 0.5. Electrical safety risk +0.3 Current oscillation risk +0.2 Risk of power supply loss.

[0066] Based on the analytical conditions for reconfiguration scheduling costs, the fourth optimization domain of reconfiguration scheduling is optimized to minimize the reconfiguration scheduling cost. This involves calculating the reconfiguration scheduling cost of all schemes in the fourth optimization domain, ranking them, and selecting the scheme with the minimum reconfiguration scheduling cost as the output for execution. For example, if scheme P1 has an electrical safety risk of 0.15, a power flow oscillation risk of 0.08, and a power supply loss risk of 0.10, then the reconfiguration scheduling cost is 0.5. 0.15 + 0.3 0.08 + 0.2 0.10 = 0.119; Scheme P2 has an electrical safety risk of 0.10, a power flow oscillation risk of 0.12, and a power supply loss risk of 0.15, therefore the reconfiguration scheduling cost = 0.5. 0.10 + 0.3 0.12 + 0.2 0.15 = 0.116; Scheme P3 has an electrical safety risk of 0.18, a power flow oscillation risk of 0.05, and a power supply loss risk of 0.08, therefore the reconfiguration scheduling cost = 0.5. 0.18 + 0.3 0.05 + 0.2 0.08 = 0.121. Comparing the reconfiguration scheduling cost, scheme P2 is selected as the optimal result of the micro-reconfiguration scheduling search. The optimal result of the micro-reconfiguration scheduling search is the final generated optimal scheduling scheme, which satisfies both safety constraints and power supply quality requirements, and minimizes the overall scheduling cost, and is used as the final operation instruction.

[0067] By employing intersection analysis and dual-branch guided differential propagation, and considering both electrical safety risks and power supply quality, multi-objective optimization is achieved. A large number of candidate solutions are generated through differential propagation and merging strategies, improving the coverage of the optimization space and avoiding local optima. Through weight setting and cost calculation, the overall scheduling cost is minimized, ensuring that the generated optimization results achieve the optimal balance expected by the user across multiple dimensions.

[0068] In summary, the collaborative optimization control method for distribution networks and microgrids provided in this application has the following technical effects: By real-time monitoring of the distribution-microgrid coordinated network, a distribution-microgrid monitoring data stream is obtained. The distribution-microgrid coordinated network includes a distribution network, a first microgrid group, and a second microgrid group. Based on the distribution-microgrid monitoring data stream, multi-level fault prediction and tracing are performed on the distribution-microgrid coordinated network to establish a distribution-microgrid fault prediction and tracing map. Based on the time-varying characteristics of the load, multi-level power flow reconfiguration scheduling is performed on the distribution-microgrid coordinated network according to the distribution-microgrid fault prediction and tracing map, constructing a distribution-microgrid reconfiguration scheduling space. A reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk in the distribution-microgrid reconfiguration scheduling space, establishing a first filtered reconfiguration scheduling branch. Based on the time-varying characteristics of the load, a power quality fitting evaluation filter is performed on the distribution-microgrid reconfiguration scheduling space, establishing a second filtered reconfiguration scheduling branch. Based on the first filtered reconfiguration scheduling branch and the second filtered reconfiguration scheduling branch, a dual-branch guided differential propagation optimization is performed on the distribution-microgrid reconfiguration scheduling space to obtain the distribution-microgrid reconfiguration scheduling optimization result. In other words, by monitoring the distribution-micro-grid in real time and acquiring the distribution-micro-grid monitoring data stream, adopting multi-level fault prediction and tracing technology, introducing load time-varying characteristics to support the distribution-micro-grid fault prediction and tracing map, constructing a distribution-micro-grid reconfiguration scheduling space, introducing a reconfiguration scheduling risk analysis model for reconfiguration scheduling risk filtering, performing power supply quality fitting evaluation filtering based on load time-varying characteristics, and adopting a dual-branch guided differential propagation optimization algorithm to optimize the scheduling decision space, the intelligent scheduling and management capabilities of the distribution-micro-grid are improved, thereby enhancing the reliability and stability of power supply.

[0069] Example 2: Based on the same inventive concept as the coordinated optimization control method for distribution networks and microgrids in Example 1, this application also provides a coordinated optimization control device for distribution networks and microgrids. Please refer to the appendix. Figure 2 The coordinated optimization control device for distribution networks and microgrids includes: Data monitoring module 11 is used to monitor the distribution-microgrid collaborative network in real time and obtain distribution-microgrid monitoring data streams. The distribution-microgrid collaborative network includes a distribution network, a first microgrid group, and a second microgrid group. Multi-level fault prediction and tracing module 12 is used to perform multi-level fault prediction and tracing of the distribution-microgrid collaborative network based on the distribution-microgrid monitoring data streams and to establish a distribution-microgrid fault prediction and tracing map. Multi-level power flow reconfiguration scheduling module 13 is used to perform multi-level power flow reconfiguration scheduling of the distribution-microgrid collaborative network based on the load time-varying characteristics and the distribution-microgrid fault prediction and tracing map, and to construct a distribution-microgrid reconfiguration scheduling space. Reconfiguration scheduling risk... The filter module 14 is used to introduce a reconfiguration scheduling risk analysis model to filter the reconfiguration scheduling risk of the distribution micro reconfiguration scheduling space and establish a first filtered reconfiguration scheduling branch; the power quality fitting evaluation filter module 15 is used to perform power quality fitting evaluation filtering on the distribution micro reconfiguration scheduling space based on the time-varying characteristics of the load and establish a second filtered reconfiguration scheduling branch; the dual-branch guided differential reproduction optimization module 16 is used to perform dual-branch guided differential reproduction optimization on the distribution micro reconfiguration scheduling space according to the first filtered reconfiguration scheduling branch and the second filtered reconfiguration scheduling branch to obtain the distribution micro reconfiguration scheduling optimization result.

[0070] Furthermore, the data monitoring module 11 in the coordinated optimization control device for distribution networks and microgrids is also used for: the first microgrid group includes multiple grid-connected microgrids, and the second microgrid group includes multiple islanded microgrids.

[0071] Furthermore, the multi-level fault prediction and tracing module 12 in the collaborative optimization control device for distribution networks and microgrids is also used for: cleaning and classifying the distribution and microgrid monitoring data streams to obtain distribution network monitoring sequences, multiple first microgrid monitoring sequences, and multiple second microgrid monitoring sequences; performing fault prediction and tracing on the distribution network based on the distribution network monitoring sequences according to the future time zone to obtain distribution network fault prediction and tracing paths; performing fault prediction and tracing on the first microgrid group based on the multiple first microgrid monitoring sequences according to the future time zone to obtain first microgrid fault prediction and tracing distributions; performing fault prediction and tracing on the second microgrid group based on the multiple second microgrid monitoring sequences according to the future time zone to obtain second microgrid fault prediction and tracing distributions; and generating the distribution network fault prediction and tracing map based on the distribution network fault prediction and tracing paths, the first microgrid fault prediction and tracing distributions, and the second microgrid fault prediction and tracing distributions.

[0072] Furthermore, the multi-level power flow reconfiguration scheduling module 13 in the collaborative optimization control device for distribution networks and microgrids is also used to: obtain load monitoring data of the distribution-microgrid collaborative network, and perform trend prediction on the load monitoring data according to the future time zone to generate the load time-varying characteristics.

[0073] Furthermore, the multi-level power flow reconfiguration scheduling module 13 in the collaborative optimization control device for distribution networks and microgrids is also used for: performing power flow reconfiguration scheduling on the distribution network based on the time-varying load characteristics and according to the distribution network fault prediction tracing path to obtain a first reconfiguration scheduling decision domain; performing power flow reconfiguration scheduling on the first microgrid group based on the time-varying load characteristics and according to the first microgrid fault prediction tracing distribution to obtain a second reconfiguration scheduling decision domain; performing power flow reconfiguration scheduling on the second microgrid group based on the time-varying load characteristics and according to the second microgrid fault prediction tracing distribution to obtain a third reconfiguration scheduling decision domain; and performing a multi-level reconfiguration scheduling decision combination based on the first reconfiguration scheduling decision domain, the second reconfiguration scheduling decision domain, and the third reconfiguration scheduling decision domain to generate the distribution-microgrid reconfiguration scheduling space.

[0074] Furthermore, the reconfiguration scheduling risk filtering module 14 in the collaborative optimization control device for distribution networks and microgrids is also used for: performing twin modeling based on the distribution-microgrid collaborative network to establish a distribution-microgrid collaborative model; performing simulation control on the distribution-microgrid collaborative model according to each distribution-microgrid reconfiguration scheduling scheme in the distribution-microgrid reconfiguration scheduling space to obtain each reconfiguration simulation data; activating the reconfiguration scheduling risk analysis model, wherein the reconfiguration scheduling risk analysis indicators of the reconfiguration scheduling risk analysis model include electrical safety risk, power flow oscillation risk, and power supply loss risk; inputting the reconfiguration simulation data into the reconfiguration scheduling risk analysis model to obtain multiple reconfiguration scheduling risk sequences; and, based on the multiple reconfiguration scheduling risk sequences, selecting the distribution-microgrid reconfiguration scheduling space according to the reconfiguration scheduling risk constraints to generate the first filtered reconfiguration scheduling branch.

[0075] Furthermore, the power quality fitting evaluation filtering module 15 in the collaborative optimization control device for distribution networks and microgrids is also used for: extracting the a-th distribution-microgrid reconfiguration scheduling scheme according to the distribution-microgrid reconfiguration scheduling space, where a is a positive integer; performing power quality evaluation retrieval on the distribution-microgrid collaborative network using the load time-varying characteristics and the a-th distribution-microgrid reconfiguration scheduling scheme as the a-th power quality evaluation retrieval conditions to obtain the a-th historical power quality evaluation sequence; performing support evaluation on each historical power quality evaluation coefficient in the a-th historical power quality evaluation sequence to obtain the a-th evaluation support sequence; performing credible attention allocation on the a-th historical power quality evaluation sequence according to the a-th evaluation support sequence to obtain the a-th evaluation attention allocation sequence; performing weighted aggregation on the a-th historical power quality evaluation sequence according to the a-th evaluation attention allocation sequence to generate the a-th power quality fitting evaluation coefficient; and adding the a-th distribution-microgrid reconfiguration scheduling scheme to the second filtering reconfiguration scheduling branch if the a-th power quality fitting evaluation coefficient is greater than or equal to the power quality fitting evaluation threshold.

[0076] Furthermore, the dual-branch guided differential propagation optimization module 16 in the collaborative optimization control device for distribution networks and microgrids is also used for: performing intersection analysis on the first filtering reconfiguration scheduling branch and the second filtering reconfiguration scheduling branch to obtain a first optimization domain for reconfiguration scheduling; performing guided differential propagation optimization on the distribution-microgrid reconfiguration scheduling space according to the first filtering reconfiguration scheduling branch to obtain a second optimization domain for reconfiguration scheduling; performing guided differential propagation optimization on the distribution-microgrid reconfiguration scheduling space according to the second filtering reconfiguration scheduling branch to obtain a third optimization domain for reconfiguration scheduling; merging the first optimization domain, the second optimization domain, and the third optimization domain to establish a fourth optimization domain for reconfiguration scheduling; activating the reconfiguration scheduling cost analysis conditions, which include electrical safety risk weight, power flow oscillation risk weight, and power supply loss risk weight; performing reconfiguration scheduling cost minimization optimization on the fourth optimization domain according to the reconfiguration scheduling cost analysis conditions to generate the optimization result for the distribution-microgrid reconfiguration scheduling.

[0077] Furthermore, the dual-branch guided differential propagation optimization module 16 in the collaborative optimization control device for distribution networks and microgrids is also used for: identifying differences in the distribution-microgrid reconfiguration scheduling space based on the first filtered reconfiguration scheduling branch to obtain a first scheduling difference feature distribution; guiding the distribution-microgrid reconfiguration scheduling space to perform differential mutation propagation based on the first scheduling difference feature distribution to obtain a first reconfiguration scheduling propagation group; introducing the reconfiguration scheduling risk analysis model to perform reconfiguration scheduling risk filtering on the first reconfiguration scheduling propagation group to establish a second reconfiguration scheduling propagation group; and performing power quality fitting evaluation filtering on the second reconfiguration scheduling propagation group based on the load time-varying characteristics to obtain the second optimization domain of the reconfiguration scheduling.

[0078] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The collaborative optimization control method and specific examples for distribution networks and microgrids in the foregoing Embodiment 1 are also applicable to the collaborative optimization control device for distribution networks and microgrids in this embodiment. Through the foregoing detailed description of the collaborative optimization control method for distribution networks and microgrids, those skilled in the art can clearly understand the collaborative optimization control device for distribution networks and microgrids in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0079] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0080] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A coordinated optimization control method for distribution networks and microgrids, characterized in that, include: Real-time monitoring of the distribution-microgrid collaborative network to obtain distribution-microgrid monitoring data streams, wherein the distribution-microgrid collaborative network includes a distribution network, a first microgrid group, and a second microgrid group; Based on the distribution micro-monitoring data stream, multi-level fault prediction and tracing are performed on the distribution micro-coordinated power grid to establish a distribution micro-fault prediction and tracing map. Based on the time-varying characteristics of the load, the distribution micro-fault prediction and tracing map is used to perform multi-level power flow reconfiguration scheduling on the distribution micro-coordinated power grid, and a distribution micro-reconfiguration scheduling space is constructed. A reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk of the micro-reconfiguration scheduling space and establish a first filtering reconfiguration scheduling branch. Based on the time-varying characteristics of the load, the power quality fitting evaluation and filtering of the distribution micro-reconfiguration scheduling space are performed to establish a second filtered reconfiguration scheduling branch. Based on the first filtering and reconfiguration scheduling branch and the second filtering and reconfiguration scheduling branch, the micro-reconfiguration scheduling space is subjected to dual-branch guided differential propagation optimization to obtain the micro-reconfiguration scheduling optimization result.

2. The coordinated optimization control method for distribution networks and microgrids as described in claim 1, characterized in that, Based on the distribution microgrid monitoring data stream, multi-level fault prediction and tracing are performed on the distribution microgrid collaborative power grid, and a distribution microgrid fault prediction and tracing map is established, including: The distribution network monitoring data stream is cleaned and classified to obtain a distribution network monitoring sequence, multiple first microgrid monitoring sequences, and multiple second microgrid monitoring sequences. Based on the future time zone, fault prediction and tracing of the distribution network are performed according to the distribution network monitoring sequence to obtain the distribution network fault prediction and tracing path. Based on the future time zone, fault prediction and tracing of the first microgrid group are performed according to the multiple first microgrid monitoring sequences to obtain the first microgrid fault prediction and tracing distribution. Based on the future time zone, fault prediction and tracing of the second microgrid group are performed according to the multiple second microgrid monitoring sequences to obtain the second microgrid fault prediction and tracing distribution; Based on the distribution network fault prediction and tracing path, the first microgrid fault prediction and tracing distribution, and the second microgrid fault prediction and tracing distribution, the distribution microgrid fault prediction and tracing map is generated.

3. The coordinated optimization control method for distribution networks and microgrids as described in claim 1, characterized in that, Based on the time-varying characteristics of the load, a multi-level power flow reconfiguration scheduling is performed on the distribution micro-fault prediction and tracing map of the distribution micro-fault coordinated power grid, constructing a distribution micro-fault reconfiguration scheduling space, including: Based on the time-varying characteristics of the load, the power flow reconfiguration scheduling of the distribution network is performed according to the distribution network fault prediction and tracing path to obtain the first reconfiguration scheduling decision domain. Based on the time-varying characteristics of the load, the first microgrid group is reconfigured and scheduled according to the fault prediction and tracing distribution of the first microgrid, and a second reconfiguration and scheduling decision domain is obtained. Based on the time-varying characteristics of the load, the second microgrid group is reconfigured and scheduled according to the fault prediction and tracing distribution of the second microgrid to obtain the third reconfiguration and scheduling decision domain. The micro-reconfiguration scheduling space is generated by combining multi-level reconfiguration scheduling decisions based on the first reconfiguration scheduling decision domain, the second reconfiguration scheduling decision domain, and the third reconfiguration scheduling decision domain.

4. The coordinated optimization control method for distribution networks and microgrids as described in claim 1, characterized in that, A reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk in the micro-reconfiguration scheduling space, establishing a first filtering reconfiguration scheduling branch, including: Based on the aforementioned distribution-micro-grid coordination, a twin model is constructed to establish a distribution-micro-grid coordination model; The micro-allocation collaborative model is simulated and controlled according to each micro-allocation reconfiguration scheduling scheme in the micro-allocation reconfiguration scheduling space to obtain simulation data for each reconfiguration. The reconfiguration scheduling risk analysis model is activated, and the reconfiguration scheduling risk analysis indicators of the reconfiguration scheduling risk analysis model include electrical safety risk, power flow oscillation risk, and power supply loss risk; The reconfiguration simulation data are input into the reconfiguration scheduling risk analysis model to obtain multiple reconfiguration scheduling risk sequences. Based on the multiple reconfiguration scheduling risk sequences, the micro-reconfiguration scheduling space is selected according to the reconfiguration scheduling risk constraints to generate the first filtered reconfiguration scheduling branch.

5. The coordinated optimization control method for distribution networks and microgrids as described in claim 1, characterized in that, Based on the time-varying characteristics of the load, the power quality fitting evaluation filter is performed on the distribution micro-reconfiguration scheduling space, and a second filtered reconfiguration scheduling branch is established, including: The a-th micro-reconfiguration scheduling scheme is extracted from the micro-reconfiguration scheduling space, where a is a positive integer; Using the load time-varying characteristics and the a-th distribution micro-reconfiguration scheduling scheme as the a-th power supply evaluation retrieval conditions, the power supply quality evaluation retrieval of the distribution micro-coordinated power grid is performed to obtain the a-th historical power supply quality evaluation sequence. Based on the support evaluation of each historical power supply quality evaluation coefficient in the a-th historical power supply quality evaluation sequence, the a-th evaluation support sequence is obtained. Based on the a-th evaluation support sequence, perform credible attention allocation on the a-th historical power quality evaluation sequence to obtain the a-th evaluation attention allocation sequence; Based on the a-th evaluation attention allocation sequence, the a-th historical power supply quality evaluation sequence is weighted and aggregated to generate the a-th power supply quality fitting evaluation coefficient; If the power quality fitting evaluation coefficient of the a-th power supply is greater than or equal to the power quality fitting evaluation threshold, the a-th micro-reconfiguration scheduling scheme is added to the second filtering reconfiguration scheduling branch.

6. The coordinated optimization control method for distribution networks and microgrids as described in claim 1, characterized in that, Based on the first and second filtering reconstruction scheduling branches, a dual-branch guided differential propagation optimization is performed on the micro-reconstruction scheduling space to obtain the micro-reconstruction scheduling optimization result, including: The first optimization domain of the reconstruction scheduling is obtained by performing an intersection analysis on the first filtering reconstruction scheduling branch and the second filtering reconstruction scheduling branch. Based on the first filtering reconstruction scheduling branch, the micro-reconstruction scheduling space is guided by differential propagation optimization to obtain the second optimization domain of reconstruction scheduling. Based on the second filtering reconstruction scheduling branch, the micro-reconstruction scheduling space is guided by differential propagation optimization to obtain the third optimization domain of reconstruction scheduling. Merge the first optimization domain, the second optimization domain, and the third optimization domain of the reconfigured scheduling to establish the fourth optimization domain of the reconfigured scheduling; Activate the reconfiguration scheduling cost analysis conditions, which include electrical safety risk weights, power flow oscillation risk weights, and power supply loss risk weights. Based on the reconfiguration scheduling cost analysis conditions, the fourth optimization domain of the reconfiguration scheduling is optimized to minimize the reconfiguration scheduling cost, and the micro-reconfiguration scheduling optimization result is generated.

7. The coordinated optimization control method for distribution networks and microgrids as described in claim 6, characterized in that, Based on the first filtering and reconfiguration scheduling branch, the micro-reconfiguration scheduling space is guided by differential propagation optimization to obtain the second optimization domain of the reconfiguration scheduling, including: Based on the first filtering and reconstructing scheduling branch, the micro-reconstruction scheduling space is differentiated to obtain the first scheduling difference feature distribution; Based on the first scheduling difference feature distribution, the micro-reconstruction scheduling space is guided to perform differential mutation reproduction to obtain the first reconstruction scheduling reproduction group; The reconfiguration scheduling risk analysis model is introduced to filter the reconfiguration scheduling risk of the first reconfiguration scheduling breeding group, and a second reconfiguration scheduling breeding group is established. Based on the time-varying characteristics of the load, the power quality fitting evaluation filter is applied to the second reconfiguration scheduling breeding group to obtain the second optimization domain of the reconfiguration scheduling.

8. The coordinated optimization control method for distribution networks and microgrids as described in claim 1, characterized in that, The first microgrid group includes multiple grid-connected microgrids, and the second microgrid group includes multiple islanded microgrids.

9. The coordinated optimization control method for distribution networks and microgrids as described in claim 2, characterized in that, The load monitoring data of the distribution micro-network is obtained, and the load monitoring data is trend predicted according to the future time zone to generate the load time-varying characteristics.

10. A coordinated optimization control device for distribution networks and microgrids, characterized in that, The step of implementing the coordinated optimization control method for distribution networks and microgrids according to any one of claims 1 to 9, wherein the coordinated optimization control device for distribution networks and microgrids comprises: The data monitoring module is used to monitor the distribution-microgrid collaborative network in real time and obtain the distribution-microgrid monitoring data stream. The distribution-microgrid collaborative network includes a distribution network, a first microgrid group, and a second microgrid group. The multi-level fault prediction and tracing module is used to perform multi-level fault prediction and tracing on the distribution micro-network based on the distribution micro-monitoring data stream, and to establish a distribution micro-fault prediction and tracing map. The multi-level power flow reconfiguration scheduling module is used to perform multi-level power flow reconfiguration scheduling on the distribution micro-collaborative power grid based on the load time-varying characteristics and the distribution micro-fault prediction and tracing map, and to construct the distribution micro-reconfiguration scheduling space. The reconfiguration scheduling risk filtering module is used to introduce a reconfiguration scheduling risk analysis model to perform reconfiguration scheduling risk filtering on the micro-reconfiguration scheduling space and establish a first filtering reconfiguration scheduling branch. The power quality fitting evaluation and filtering module is used to perform power quality fitting evaluation and filtering on the distribution micro-reconfiguration scheduling space based on the time-varying characteristics of the load, and to establish a second filtered reconfiguration scheduling branch. The dual-branch guided differential reproduction optimization module is used to perform dual-branch guided differential reproduction optimization on the micro-reconstruction scheduling space according to the first filtering reconstruction scheduling branch and the second filtering reconstruction scheduling branch, so as to obtain the micro-reconstruction scheduling optimization result.

Images