Power distribution network optimization scheduling method based on swarm intelligence and source network load storage coordination

Abstract

The invention relates to the technical field of power distribution network optimization scheduling, and discloses a power distribution network optimization scheduling method based on swarm intelligence and source network load storage coordination, and the method comprises the steps: collecting the multi-dimensional operation data of a source network load storage side operation unit, and carrying out the time-space alignment based on the data transmission time delay; missing value completion is performed on the multi-dimensional operation data after space-time alignment by using a multi-scale interpolation mode based on transmission confidence; evaluating the complemented multi-dimensional operation data by using a swarm intelligence collaborative evaluation function to generate collaborative fitness of the intelligent agent; and constructing a power distribution network operation constraint, carrying out joint iterative optimization on the scheduling strategy of each agent by adopting a swarm intelligent cooperation mechanism, and outputting an agent scheduling strategy meeting the power distribution network operation constraint. According to the method, unified perception of the source network load storage state is realized through space-time alignment and data completion, and on the basis of swarm intelligent evaluation and collaborative optimization, source network load storage collaborative adaptive optimization scheduling is realized on the premise of satisfying the operation constraint of the power distribution network.

Description

Technical Field

[0001] This invention relates to the field of distribution network optimization scheduling, and in particular to a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, belonging to a large-scale power grid intelligent scheduling system. Background Technology

[0002] With the widespread integration of distributed power sources, large-scale flexible loads, and energy storage systems into distribution networks, the operation mode of distribution networks has gradually evolved from unidirectional power supply and centralized control to a collaborative operation mode of multiple entities coexisting and strongly coupled interaction, characterized by coordinated operation of power sources, grids, loads, and storage. Significant differences exist among various operating units in terms of power regulation capabilities, response speeds, and operational objectives, resulting in distribution network dispatching exhibiting characteristics of multiple objectives, multiple constraints, and strong uncertainty. Traditional dispatching methods centered on single operational indicators or centralized optimization are insufficient to characterize the collaborative behavior of multiple entities and their dynamic game relationships, necessitating the introduction of novel optimization dispatching mechanisms with distributed decision-making and group collaboration capabilities.

[0003] In practical applications, due to inconsistencies in sampling periods and communication delays among different devices and systems, traditional scheduling models that assume data synchronization and time consistency struggle to guarantee the real-time performance and feasibility of scheduling results. Specifically, the data sampling periods and upload delays of various subsystems, such as photovoltaic power generation, energy storage systems, and adjustable loads, differ significantly. When communication links are congested or nodes fail, some critical data may be delayed or even missing, leading to inconsistent input states for scheduling optimization and causing problems such as power allocation deviations, voltage exceeding limits, or system oscillations.

[0004] Existing research, such as patent CN118763704A, proposes an active distribution network scheduling method and medium based on Stackelberg game theory and multi-agent interests. This method takes maximizing the economic benefits for distribution network operators as the first objective function, and maximizing the daily operating revenue of charging stations and minimizing the electricity costs for power users as the second and third objective functions, respectively. It constructs a master-slave game structure by building an upper-level (leader) model and a lower-level (follower) model, and uses an improved particle swarm optimization algorithm to solve the game equilibrium under constraints, achieving active distribution network scheduling with optimal multi-agent interests. This type of method is representative in multi-agent interest modeling and provides an effective approach to active distribution network scheduling. However, it has high requirements for the quality and synchronization of operational data, does not fully consider the impact of communication latency and data gaps on the reliability of scheduling decisions, and the optimization process is based on global game equilibrium, resulting in high computational complexity. Its scalability and real-time performance are limited in large-scale distributed scenarios.

[0005] To address this issue, this invention proposes a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, which helps to build a smart distribution network with adaptive sensing and dynamic optimization capabilities, providing solid support for the safe, economical and stable operation of the distribution network. Summary of the Invention

[0006] This invention proposes a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination. Step S1 introduces a spatiotemporal alignment mechanism based on data transmission delay during the multi-dimensional operation data acquisition stage of source-grid-load-storage, solving the data time offset problem caused by differences in communication links and inconsistent sampling periods among different operating units, and providing a unified and comparable data foundation for subsequent swarm intelligence collaborative evaluation. Step S2 adopts a multi-scale interpolation method based on transmission confidence to adaptively fill in missing values ​​in the spatiotemporally aligned multi-dimensional operation data, avoiding the error diffusion introduced by traditional single-scale interpolation under complex operating conditions. Step S3 abstracts the source-side, grid-side, load-side, and storage-side operating units as intelligent agents and constructs a swarm intelligence collaborative evaluation function, mapping heterogeneous operation characteristics to a unified collaborative index vector, realizing quantifiable evaluation and comparison of different types of intelligent agents under the same collaborative framework. Step S4, while taking the maximization of the collaborative fitness of intelligent agents as the local optimization objective, introduces distribution network operation constraints such as source-grid-load-storage power balance, voltage safety, energy storage energy, and operating power, and performs joint iterative optimization through the swarm intelligence collaborative mechanism to avoid the local optima of a single intelligent agent from destroying the overall operational safety and stability of the system.

[0007] To achieve the above objectives, this invention provides a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, comprising the following steps: S1: Collect multi-dimensional operation data from the source-grid-load-storage side operation unit, and perform spatiotemporal alignment on the multi-dimensional operation data based on the data transmission delay to obtain spatiotemporally aligned multi-dimensional operation data; S2: Use a multi-scale interpolation method based on transmission confidence to fill in missing values ​​in the spatiotemporally aligned multidimensional running data to obtain the filled multidimensional running data. S3: The source-grid-load-storage side operation unit is used as an intelligent agent. The completed multi-dimensional operation data is evaluated using the swarm intelligence collaborative evaluation function to generate the collaborative fitness of the intelligent agent. S4: Maximizing the cooperative fitness of the agents is taken as the local optimization objective of the agents. The power distribution network operation constraints are constructed, and the scheduling strategies of each agent are jointly iteratively optimized using a swarm intelligence cooperative mechanism to output an agent scheduling strategy that satisfies the power distribution network operation constraints.

[0008] As a further improvement of the present invention: Furthermore, the multi-dimensional operational data collected in step S1 from the source-grid-load-storage side operation units includes: S11: Collect multi-dimensional operation data from the source-grid-load-storage side operation units respectively, and generate the collection timestamp of the multi-dimensional operation data. The source-grid-load-storage side operation units include source-side operation units, grid-side operation units, load-side operation units and storage-side operation units respectively. S12: The source-grid-load-storage side operation unit transmits the collected multi-dimensional operation data to the distribution network, and the distribution network records the timestamp of the received multi-dimensional operation data; S13: Calculate the difference between the receiving timestamp and the acquisition timestamp of the multidimensional running data, and use it as the data transmission delay of the multidimensional running data.

[0009] Furthermore, the spatiotemporal alignment of multidimensional operational data based on the data transmission delay also includes: S14: Generate the optimized scheduling cycle of the distribution network and construct a global time axis; S15: Based on the collection timestamp of the multidimensional running data, the multidimensional running data is allocated to the specified position of the global time axis to obtain multidimensional running data with a unified time axis; S16: Based on the multidimensional operation data of the unified time axis, construct the spatiotemporally aligned multidimensional operation data of the source-grid-load-storage side operation units. The spatiotemporally aligned multidimensional operation data is in matrix form. Each row of matrix elements describes the position coordinates of an operation unit, the multidimensional operation data at different time points in the global time axis, and the data transmission delay of the multidimensional operation data. The first column of the spatiotemporally aligned multidimensional operation data describes the position coordinates of all operation units. The other columns of matrix elements describe the multidimensional operation data of different operation units at the same time point and the data transmission delay of the multidimensional operation data.

[0010] Furthermore, in step S2, missing values ​​are filled in for the spatiotemporally aligned multidimensional running data using a multi-scale interpolation method, including: S21: Extract missing values ​​from the spatiotemporally aligned multidimensional running data; S22: Extract the rows before and after the missing values ​​in the matrix. Multidimensional operational data are used to form a short-term tensor sequence with missing values; S23: Perform temporal operations on the short-term tensor sequence using temporal convolution to obtain the temporal interpolation result of the missing values; Specifically, the formula for the temporal convolution operation of the short-term tensor sequence is: ; ; in, Indicates missing values short-term tensor sequences, In short-term tensor sequences Temporal convolution weights, This represents an exponential function with the natural constant as its base. Indicates the time decay coefficient. Represents multidimensional operational data The indicator function value, if multidimensional running data If it is a missing value, then =0, otherwise =1, Indicates missing values The timing interpolation results; S24: Select multiple running units of the same type as the running unit corresponding to the missing value and whose spatial distance is lower than the preset distance threshold as neighboring running units. Extract the multidimensional running data of the neighboring running units at the time point corresponding to the missing value as neighboring multidimensional running data. Use a spatial interpolation method based on spatial weight to perform weighted interpolation on the extracted neighboring multidimensional running data to obtain the spatial interpolation result of the missing value.

[0011] Furthermore, step S2, which calculates the transmission confidence of missing values ​​and performs missing value completion based on the temporal and spatial interpolation results of the missing values, also includes: S25: Obtain the average data transmission latency of the matrix row containing the missing values; S26: Convert the average data transmission delay into a transmission confidence level for missing values; Specifically, the conversion formula for the transmission confidence is: ; in, This represents the average data transmission latency of the matrix row containing the missing values. Represents the smoothing coefficient. Indicates standard data transmission latency. This represents an exponential function with the natural constant as its base. This indicates the average data transmission delay. Transmission confidence level of the conversion; S27: Using the transmission confidence as the imputation weight for the missing value, the temporal imputation result and the spatial imputation result of the missing value are weighted to obtain the missing value completion result.

[0012] Furthermore, in step S3, the source-grid-load-storage side operation unit is treated as an intelligent agent, and the completed multi-dimensional operation data is evaluated using a swarm intelligence collaborative evaluation function, including: S31: Calculate the mean value of the multidimensional running data of each row of matrix elements in the completed multidimensional running data, and use it as the running state of the agent corresponding to each row of matrix elements; S32: Convert the operating state of the intelligent agent into a collaborative index vector, wherein the collaborative index vector includes a power balance contribution index, a voltage support deviation index, an operating constraint occupancy index, and a collaborative adjustment potential index. S33: Receive the collaborative index vector of the agent using the swarm intelligence collaborative evaluation function, and generate the collaborative fitness of the agent.

[0013] Furthermore, the expression for the swarm intelligence collaborative evaluation function is: ; in, This represents the collaborative evaluation function of swarm intelligence. A vector representing the collaborative index of an agent. This represents the cooperative fitness of the agents. , The following are the collaborative index vectors respectively. The indicators include power balance contribution, voltage support deviation, operational constraint occupancy, and coordinated regulation potential. The weights of the power balance contribution index, voltage support deviation index, operating constraint occupancy index, and coordinated regulation potential index are represented in order, respectively. Indicates the desired power coordination level. This represents an exponential function with the natural constant as its base. This represents the voltage deviation penalty coefficient.

[0014] Furthermore, in step S4, maximizing the cooperative fitness of the agent is used as the local optimization objective of the agent, and power distribution network operation constraints are constructed, including: The power distribution network operation constraints include power balance constraints of source, grid, load and storage, operating voltage constraints, energy constraints of storage side and operating power constraints; The power balance constraint between the source, grid, load, and storage means that the total active power provided to the distribution network by the intelligent agent corresponding to the source-side operating unit and the intelligent agent corresponding to the storage-side operating unit should be balanced with the power demand of the intelligent agent corresponding to the load-side operating unit and the operating loss of the distribution network. That is, there is no power deficit or power redundancy in the power system as a whole, ensuring the energy conservation and stable supply and demand relationship of the distribution network in the current dispatch cycle. The operating voltage constraint means that the operating voltage of the agent corresponding to each operating unit is always within the allowable safe operating range; The energy constraint on the storage side means that the energy storage charge state of the intelligent agent corresponding to the storage side operation unit must be maintained between the preset minimum and maximum energy boundaries to prevent the intelligent agent corresponding to the storage side operation unit from overcharging, over-discharging or unsustainable operation. The operating power constraint means that the operating power of the agent corresponding to each operating unit is always within the allowable safe operating range.

[0015] Furthermore, step S4, which involves jointly iteratively optimizing the scheduling strategies of each agent using a swarm intelligence collaborative mechanism based on the distribution network operation constraints, also includes: S41: Initialize the scheduling strategy and scheduling parameters for each agent, where the scheduling strategy includes the agent's operating voltage and power, and the scheduling parameters are the iteration step size of the scheduling strategy during joint iterative optimization. S42: Calculate the cooperative fitness of the agents after adjustment according to the scheduling strategy, and calculate the maximum cooperative fitness of different types of agents after adjustment according to the scheduling strategy. The types of agents include source-side agents, network-side agents, load-side agents and storage-side agents, which correspond to source-side operation units, network-side operation units, load-side operation units and storage-side operation units respectively. Calculate the total cooperative fitness of all agents after adjustment according to the scheduling strategy, and calculate the change in the total cooperative fitness. If the change is less than the preset change threshold, terminate the iteration and perform distribution network optimization scheduling according to the agent's current scheduling strategy. S43: Calculate the group communication coefficient between the agent corresponding to the maximum cooperative fitness and agents of the same type, and use the group cooperative update method to iteratively optimize the scheduling parameters and scheduling strategy; S44: Based on the distribution network operation constraints, determine whether the iteratively optimized scheduling strategy meets the distribution network operation constraints. If it does not meet the constraints, map the iteratively optimized scheduling strategy to a feasible scheduling strategy that meets the distribution network operation constraints, and proceed to step S42; if it does meet the constraints, proceed directly to step S42.

[0016] Compared with existing technologies, this invention proposes a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, which has the following beneficial effects: First, this invention effectively improves the completeness and reliability of multi-dimensional operational data of source-grid-load-storage systems by dividing the missing value imputation process into two complementary stages: temporal-scale imputation and spatial-scale imputation. Specifically, in the temporal dimension, the temporal convolution operation based on short-term tensor sequences fully utilizes the operational correlation before and after the missing value, and uses exponentially decaying weights to characterize the influence intensity of operational status changes over time. Simultaneously, an indicator function is introduced to avoid interference from missing data on the imputation results, making the temporal imputation results more consistent with operational evolution patterns. In the spatial dimension, by selecting similar, spatially limited neighboring operational units, and combining spatial distance decay with nominal attribute similarity to construct spatial weights, a joint characterization of the operational environment and equipment characteristics is achieved, thereby obtaining more physically consistent spatial imputation results. This spatiotemporal dual-scale imputation mechanism reduces the sensitivity of a single imputation method to abnormal data and enhances the transferability of imputation results between different operational units.

[0017] Meanwhile, this invention introduces an agent-based collaborative optimization mechanism for multiple types of agents (source, grid, load, and storage) to unify the modeling and joint iteration of the local scheduling strategies of each agent with the overall operational goals of the distribution network, effectively improving the global consistency and adaptability of distribution network optimization scheduling. Specifically, by introducing collaborative fitness and group communication coefficients, this invention enables agents of the same type to achieve orderly collaboration while considering communication latency and operational differences, avoiding the slow convergence and scheduling conflicts caused by traditional centralized or isolated optimization. Furthermore, by combining multi-dimensional distribution network operational constraints with feasibility mapping of the iteration results, this ensures that the scheduling strategy achieves optimal performance under the conditions of meeting safety, voltage, energy, and power constraints, thereby significantly enhancing the stability, economy, and robustness of the distribution network in scenarios with a high proportion of distributed power sources and diverse load access. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, provided as an embodiment of the present invention.

[0019] Figure 2 This is a coordination relationship diagram of various types of operating units provided in an embodiment of the present invention.

[0020] Figure 3 This is a flowchart of an intelligent agent scheduling process provided in an embodiment of the present invention. Detailed Implementation

[0021] The realization of the objectives, functional characteristics, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0022] This invention provides a distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this invention: a server, a terminal, etc. In other words, the distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0023] Reference Figure 1 , Figure 2 as well as Figure 3 Embodiment 1 of the present invention is as follows: A distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, the method comprising: S1: Collect multi-dimensional operation data from the source-grid-load-storage side operation unit, and perform spatiotemporal alignment on the multi-dimensional operation data based on the data transmission delay to obtain spatiotemporally aligned multi-dimensional operation data.

[0024] Collect multi-dimensional operational data from source-grid-load-storage side operation units, including: S11: Collect multi-dimensional operation data from the source-grid-load-storage side operation units respectively, and generate the collection timestamp of the multi-dimensional operation data. The source-grid-load-storage side operation units include source-side operation units, grid-side operation units, load-side operation units and storage-side operation units respectively. It should be noted that, for example, Figure 2 The diagram shows the coordination relationship between various types of operating units. Solid lines represent power transmission and dashed lines represent signal transmission. Source-side operating units refer to various power supply equipment and their control systems connected to the distribution network, including distributed photovoltaic power generation units, wind power generation units, gas micro generator sets, and corresponding inverters and power generation control devices, which are used to provide power to the distribution network and are the energy supply source of the distribution network. The grid-side operation unit is based on the distribution network and includes distribution lines, transformers, switching equipment, and distribution automation terminals. It is used to collect power from the source side, distribute power to the load side, and provide physical channels and operational constraints for energy interaction between the source, load, and storage side operation units. Load-side operating units refer to various types of electrical loads connected to the distribution network and their management systems, including industrial loads, commercial loads, and residential loads, which consume electrical energy and generate electricity demand. The energy storage side operation unit refers to the electrochemical energy storage device and its energy management system configured on the distribution network side, including battery packs, power conversion devices and energy storage controllers, used to store and release electrical energy, realize energy time shift, smooth source-side fluctuations and support the stable operation of the distribution network; The source-side, grid-side, load-side, and storage-side operating units form a close coupling relationship through the distribution network. The source-side operating unit provides electrical energy, the grid-side operating unit undertakes the transmission and constraint of electrical energy, the load-side operating unit consumes electrical energy and regulates demand, and the storage-side operating unit absorbs or releases electrical energy at different operating stages. Specifically, the multi-dimensional operating data of the source-side operating unit includes the active power output, reactive power output, operating voltage, and available power capacity; the multi-dimensional operating data of the grid-side operating unit includes power, operating voltage, current, and power flow value; the multi-dimensional operating data of the load-side operating unit includes load power, operating voltage, and adjustable load capacity; and the multi-dimensional operating data of the energy storage-side operating unit includes the energy storage state of charge, charging and discharging power, operating voltage, and maximum stored energy. The active power output refers to the power actually output by the power source to the distribution network per unit time, which is consumed by the load or converted into other forms of energy. The reactive power output refers to the power provided or absorbed by the power source to the distribution network that does not directly perform work but is used to maintain electromagnetic energy exchange. As an embodiment of the present invention, multi-dimensional operating data of the source-side operating unit is collected in real time through the power generation unit controller, inverter and SCADA system; multi-dimensional operating data of the grid-side operating unit is obtained by the distribution automation terminal, PMU, smart switch and online monitoring device; multi-dimensional operating data of the load-side operating unit is obtained by the smart meter and electricity consumption information collection system; and multi-dimensional operating data of the energy storage-side operating unit is collected in real time by the energy storage management system EMS / BMS.

[0025] S12: The source-grid-load-storage side operation unit transmits the collected multi-dimensional operation data to the distribution network, and the distribution network records the timestamp of the received multi-dimensional operation data; S13: Calculate the difference between the receiving timestamp and the acquisition timestamp of the multidimensional running data, and use it as the data transmission delay of the multidimensional running data.

[0026] Optionally, the transmission delay of different operating units can be evaluated by active network measurement and used as the data transmission delay of the multi-dimensional operating data collected by the operating units. Specifically, the data management center in the distribution network sends measurement and detection messages to the operating units. After receiving the messages, the operating units immediately send back response messages. The end-to-end delay is calculated based on the sending and receiving timestamps and used as the transmission delay of the operating units.

[0027] Based on the data transmission delay, spatiotemporal alignment of multidimensional operational data also includes: S14: Generate the optimized scheduling cycle of the distribution network and construct a global time axis; As an embodiment of the present invention, an optimized scheduling cycle for the distribution network is set. The following constraints must be met: and ; in, These represent, in turn, the minimum time scales for significant changes in the active / reactive power output of the source-side operating unit, the minimum time scale for significant changes in the load of the load-side operating unit, and the minimum time scale for significant changes in the charging and discharging power of the storage-side operating unit. These represent the minimum data sampling periods for the source, grid, load, and storage side operating units, respectively. This indicates that the maximum value in the set is selected; it should be noted that both the time scale and the data sampling period are converted into timestamp format. The global timeline is represented as follows: ; in, This indicates the initial scheduling time point, which can be set. This is the end timestamp of the last optimized dispatch of the distribution network. This represents the b-th time point on the global timeline, and B represents the length of the global timeline. Let B be 50. Optionally, it can be based on the initial scheduling time point. The data transmission latency of each operating unit and the computational load of the distribution network generate the optimal scheduling cycle of the distribution network. : ; in, Indicates the initial scheduling time point Average data transmission latency of each operating unit Indicates the initial scheduling time point The number of running units in operation. This represents the computational load of the distribution network, which is set as the average time it takes for the distribution network to complete one optimized dispatch cycle. All represent control coefficients, set They are 0.4 and 0.5 respectively; S15: Based on the collection timestamp of the multidimensional running data, the multidimensional running data is allocated to the specified position of the global time axis to obtain multidimensional running data with a unified time axis; Specifically, the formula for generating the time point position of multidimensional runtime data on the global time axis is: ,and satisfy ; in, This indicates the position of the multidimensional runtime data at a specific point in time on the global time axis. This represents the timestamp of the collection of the multidimensional operational data. Indicates that extraction makes To reach the minimum value of k, where ; The specified time point is used as the location of the multidimensional runtime data on the global time axis; S16: Based on the multidimensional operation data of the unified time axis, construct the spatiotemporally aligned multidimensional operation data of the source-grid-load-storage side operation units. The spatiotemporally aligned multidimensional operation data is in matrix form. Each row of matrix elements describes the position coordinates of an operation unit, the multidimensional operation data at different time points in the global time axis, and the data transmission delay of the multidimensional operation data. The first column of the spatiotemporally aligned multidimensional operation data describes the position coordinates of all operation units. The other columns of matrix elements describe the multidimensional operation data of different operation units at the same time point and the data transmission delay of the multidimensional operation data.

[0028] It should be noted that this invention achieves standardized alignment of asynchronous operation data from multiple sources (source, grid, load, and storage) in the time dimension by introducing a unified construction mechanism for optimizing the scheduling cycle and the global time axis. Specifically, this invention improves scheduling timeliness and stability by simultaneously constraining the scheduling cycle to the physical change timescale of each operating unit and the minimum data sampling period, avoiding response lag caused by incomplete data due to an excessively short scheduling cycle or a response lag caused by an excessively long scheduling cycle. Simultaneously, an optional scheme can adaptively correct the scheduling cycle by combining the average data transmission latency of the operating unit and the distribution network's computational load, ensuring that the scheduling frequency matches communication conditions and computing power, effectively reducing the risk of scheduling failures caused by communication congestion or computational overload. Furthermore, by mapping the collection timestamps of multi-dimensional operation data based on a unified time axis, operation data from different sources and with inconsistent arrival times can be accurately assigned to the corresponding time points on the time axis, reducing the impact of time series drift on subsequent analysis and significantly improving the accuracy and robustness of distribution network collaborative scheduling.

[0029] S2: Use a multi-scale interpolation method based on transmission confidence to fill in missing values ​​in the spatiotemporally aligned multidimensional running data to obtain the filled multidimensional running data.

[0030] The missing values ​​in the spatiotemporally aligned multidimensional running data are filled using a multi-scale interpolation method, including: S21: Extract missing values ​​from the spatiotemporally aligned multidimensional running data; S22: Extract the rows before and after the missing values ​​in the matrix. Multidimensional operational data are used to form a short-term tensor sequence with missing values; S23: Perform temporal operations on the short-term tensor sequence using temporal convolution to obtain the temporal interpolation result of the missing values; Specifically, the formula for the temporal convolution operation of the short-term tensor sequence is: ; ; in, Indicates missing values short-term tensor sequences, In short-term tensor sequences Temporal convolution weights, This represents an exponential function with the natural constant as its base. Indicates the time decay coefficient. Represents multidimensional operational data The indicator function value, if multidimensional running data If it is a missing value, then =0, otherwise =1, Indicates missing values The timing interpolation results; S24: Select multiple running units of the same type as the running unit corresponding to the missing value and whose spatial distance is lower than the preset distance threshold as neighboring running units. Extract the multidimensional running data of the neighboring running units at the time point corresponding to the missing value as neighboring multidimensional running data. Use a spatial interpolation method based on spatial weight to perform weighted interpolation on the extracted neighboring multidimensional running data to obtain the spatial interpolation result of the missing value.

[0031] As one embodiment of the present invention, the types of operating units include source-side operating units, grid-side operating units, load-side operating units, and storage-side operating units. A preset distance threshold of 2 kilometers is set, and the spatial interpolation method based on spatial weights is as follows: ; ; in, Indicates missing values Spatial interpolation results Indicates missing values The neighboring multidimensional operational data set, , Represents the neighboring multidimensional running data set Any neighboring multidimensional running data in the data, Represents neighboring multidimensional operational data Spatial weights, Represents neighboring multidimensional operational data Corresponding operating unit and missing value The spatial distance between the corresponding operating units Indicates the spatial attenuation coefficient, set It is 0.4. Represents neighboring multidimensional operational data Basic operational attributes of the corresponding operational unit With missing values Basic operational attributes of the corresponding operational unit The similarity between them, where the basic operating attributes of the operating unit are the rated attributes of the operating unit, including rated capacity, rated voltage, rated power, etc.; Optionally, the similarity is calculated using cosine similarity.

[0032] Step S2, which calculates the transmission confidence of missing values ​​and performs missing value completion based on the temporal and spatial interpolation results, also includes: S25: Obtain the average data transmission latency of the matrix row containing the missing values; S26: Convert the average data transmission delay into a transmission confidence level for missing values; Specifically, the conversion formula for the transmission confidence is: ; in, This represents the average data transmission latency of the matrix row containing the missing values. Represents the smoothing coefficient. Indicates standard data transmission latency. This represents an exponential function with the natural constant as its base. This indicates the average data transmission delay. Transmission confidence level of the conversion; S27: Using the transmission confidence as the imputation weight for the missing value, the temporal imputation result and the spatial imputation result of the missing value are weighted to obtain the missing value completion result.

[0033] Specifically, the missing value Temporal interpolation results and spatial interpolation results The weighting formula is: , Indicates missing values The result of filling in missing values.

[0034] It should be noted that this invention introduces a transmission confidence calculation mechanism based on average data transmission delay during the missing value completion process, explicitly integrating communication reliability into the weight allocation of missing value imputation, thus achieving adaptive completion with data quality awareness. When the average data transmission delay is small and the data temporal continuity is strong, the weight of the temporal imputation result is increased; when the average data transmission delay is large and the data reliability of the unit itself is reduced, the participation of the spatial imputation result is enhanced, thereby avoiding the introduction of errors by a single imputation method under abnormal communication conditions. This effectively improves the stability and accuracy of multi-dimensional operation data completion of source-grid-load-storage, reduces the impact of communication uncertainty on subsequent group intelligent scheduling and collaborative optimization results, and enhances the robustness and engineering applicability of distribution network optimization scheduling.

[0035] S3: The source-grid-load-storage side operation unit is used as an intelligent agent. The completed multi-dimensional operation data is evaluated using a swarm intelligence collaborative evaluation function to generate the collaborative fitness of the intelligent agent.

[0036] In step S3, the source-grid-load-storage side operation unit is treated as an intelligent agent, and the completed multi-dimensional operation data is evaluated using a swarm intelligence collaborative evaluation function, including: S31: Calculate the mean value of the multidimensional running data of each row of matrix elements in the completed multidimensional running data, and use it as the running state of the agent corresponding to each row of matrix elements; S32: Convert the operating state of the intelligent agent into a collaborative index vector, wherein the collaborative index vector includes a power balance contribution index, a voltage support deviation index, an operating constraint occupancy index, and a collaborative adjustment potential index. As an embodiment of the present invention, the types of intelligent agents include source-side intelligent agents, grid-side intelligent agents, load-side intelligent agents and storage-side intelligent agents, which correspond to source-side operation units, grid-side operation units, load-side operation units and storage-side operation units respectively. The formula for transforming the collaborative index vector of the source-side agent is: ; in, In order, they are the power balance contribution index, voltage support deviation index, operational constraint occupancy index, and collaborative adjustment potential index of the source-side intelligent agent. This represents the average active power output of the source-side intelligent agent. This indicates the maximum allowable active power output of the source-side intelligent agent. This represents the average operating voltage of the source-side intelligent agent. Indicates the rated voltage of the source-side intelligent agent. This represents the average reactive power output of the source-side intelligent agent. This indicates the maximum allowable reactive power output of the source-side intelligent agent; The formula for transforming the collaborative index vector of the network-side intelligent agent is: ; in, In order, they are the power balance contribution index, voltage support deviation index, operational constraint occupancy index, and collaborative adjustment potential index of the network-side intelligent agent. This represents the average power of the network-side agents. This indicates the maximum allowable power for the network-side agent. Indicates the rated voltage of the network-side intelligent agent. This represents the average operating voltage of the network-side intelligent agent. This represents the average current of the network-side intelligent agent. This indicates the rated current of the network-side intelligent agent. This represents the mean power flow value of the network-side agents. This represents the maximum allowable power flow value for the network-side agent; The formula for transforming the collaborative index vector of the load-side agent is: ; in, In order, they are the power balance contribution index, voltage support deviation index, operational constraint occupancy index, and collaborative adjustment potential index of the load-side intelligent agent. This represents the average load power of the load-side agent. This indicates the maximum allowable load power for the load-side agent. Indicates the rated voltage of the load-side intelligent agent. This represents the average operating voltage of the load-side intelligent agent. This represents the average adjustable load capacity of the load-side agent. The formula for transforming the collaborative index vector of the storage-side intelligent agent is: ; in These are the power balance contribution index, voltage support deviation index, operational constraint occupancy index, and collaborative adjustment potential index of the energy storage-side intelligent agent. This represents the average charging and discharging power of the storage-side intelligent agent. This indicates the maximum allowable charging and discharging power for the energy storage agent. Indicates the rated voltage of the storage-side intelligent agent. This represents the average operating voltage of the storage-side intelligent agent. This represents the average state of charge of the energy storage agent on the energy storage side. Indicates the target state of charge of the energy storage agent (set to 0.9); S33: Receive the collaborative index vector of the agent using the swarm intelligence collaborative evaluation function, and generate the collaborative fitness of the agent.

[0037] It should be noted that this invention performs row-level statistical processing on the completed multidimensional operational data, compressing complex temporal and spatial operational information into a mean value that can characterize the overall characteristics of the operational unit, serving as the operational state of the agent, effectively reducing data dimensionality and computational complexity. Based on this, differentiated collaborative index vector conversion formulas are constructed for the different functional positioning of source-side, grid-side, load-side, and storage-side agents, uniformly mapping physical quantities such as power, voltage, current, power flow values, and state of charge into comparable collaborative indices. This maintains the physical meaning of various agents while achieving consistent expression of indices across multiple types of agents. Furthermore, by introducing operational constraint occupancy and collaborative adjustment potential indicators, this invention ensures that collaborative evaluation not only focuses on the compliance of the current operational state but also considers future adjustment capabilities and system flexibility.

[0038] The expression for the swarm intelligence collaborative evaluation function is: ; in, This represents the collaborative evaluation function of swarm intelligence. A vector representing the collaborative index of an agent. This represents the cooperative fitness of the agents. , The following are the collaborative index vectors respectively. The indicators include power balance contribution, voltage support deviation, operational constraint occupancy, and coordinated regulation potential. The weights of the power balance contribution index, voltage support deviation index, operating constraint occupancy index, and coordinated regulation potential index are represented in order, respectively. This indicates the desired level of power coordination (set to 0.7). This represents an exponential function with the natural constant as its base. This represents the voltage deviation penalty coefficient, set. It is 0.2.

[0039] As an embodiment of the present invention, based on the roles of the source-side intelligent agent, grid-side intelligent agent, load-side intelligent agent, and storage-side intelligent agent in the coordination of power generation, grid, load, and storage, different index weights are assigned to the source-side intelligent agent, grid-side intelligent agent, load-side intelligent agent, and storage-side intelligent agent. Specifically, the index weights of the source-side intelligent agent in the power balance contribution index, voltage support deviation index, operational constraint occupancy index, and collaborative adjustment potential index are 0.40, 0.30, 0.10, and 0.20, respectively, emphasizing the controllability and planning tracking capability of power output and suppressing voltage disturbances and output fluctuations. The index weights of the grid-side intelligent agent in the power balance contribution index, voltage support deviation index, operational constraint occupancy index, and collaborative adjustment potential index are 0.15, 0.40, 0.35, and 0.10, respectively, with the core objective being power... For load-side agents, the weights of power balance contribution, voltage support deviation, operational constraint occupancy, and collaborative adjustment potential are 0.30, 0.15, 0.15, and 0.40 respectively, emphasizing load response and peak shaving / valley filling capabilities. For storage-side agents, the weights of power balance contribution, voltage support deviation, operational constraint occupancy, and collaborative adjustment potential are 0.25, 0.25, 0.10, and 0.40 respectively. As the core of distribution network dispatch, increasing the weight of the collaborative adjustment potential index ensures that the collaborative fitness truly reflects the actual contribution of each agent in system collaborative dispatch, effectively avoiding dispatch bias caused by unified weight evaluation, and improving the accuracy and stability of multi-agent collaborative optimization.

[0040] It should be noted that the scheduling strategy of an agent with higher cooperative fitness has a stronger guiding effect on the global search direction, thereby improving the convergence speed and stability of cooperative optimization. Overall, higher cooperative fitness indicates that, under the premise of satisfying its own operational constraints, the corresponding agent's power response characteristics, operational stability, and regulation capabilities are more closely matched with the overall cooperative scheduling objectives of the distribution network, resulting in a greater comprehensive contribution to the safe, economical, and stable operation of the power system.

[0041] S4: Maximizing the cooperative fitness of the agents is taken as the local optimization objective of the agents. The power distribution network operation constraints are constructed, and the scheduling strategies of each agent are jointly iteratively optimized using a swarm intelligence cooperative mechanism to output an agent scheduling strategy that satisfies the power distribution network operation constraints.

[0042] Maximizing the cooperative fitness of the agent is used as the local optimization objective of the agent, and power distribution network operation constraints are constructed, including: The power distribution network operation constraints include power balance constraints of source, grid, load and storage, operating voltage constraints, energy constraints of storage side and operating power constraints; The power balance constraint between the source, grid, load, and storage means that the total active power provided to the distribution network by the intelligent agent corresponding to the source-side operating unit and the intelligent agent corresponding to the storage-side operating unit should be balanced with the power demand of the intelligent agent corresponding to the load-side operating unit and the operating loss of the distribution network. That is, there is no power deficit or power redundancy in the power system as a whole, ensuring the energy conservation and stable supply and demand relationship of the distribution network in the current dispatch cycle. The operating voltage constraint means that the operating voltage of the agent corresponding to each operating unit is always within the allowable safe operating range; The energy constraint on the storage side means that the energy storage charge state of the intelligent agent corresponding to the storage side operation unit must be maintained between the preset minimum and maximum energy boundaries to prevent the intelligent agent corresponding to the storage side operation unit from overcharging, over-discharging or unsustainable operation. The operating power constraint means that the operating power of the agent corresponding to each operating unit is always within the allowable safe operating range.

[0043] Step S4, based on the power distribution network operation constraints, employs a swarm intelligence collaborative mechanism to jointly iteratively optimize the scheduling strategies of each agent, and further includes: S41: Initialize the scheduling strategy and scheduling parameters for each agent, where the scheduling strategy includes the agent's operating voltage and power, and the scheduling parameters are the iteration step size of the scheduling strategy during joint iterative optimization. S42: Calculate the cooperative fitness of the agents after adjustment according to the scheduling strategy, and calculate the maximum cooperative fitness of different types of agents after adjustment according to the scheduling strategy. The types of agents include source-side agents, network-side agents, load-side agents and storage-side agents, which correspond to source-side operation units, network-side operation units, load-side operation units and storage-side operation units respectively. Calculate the total cooperative fitness of all agents after adjustment according to the scheduling strategy, and calculate the change in the total cooperative fitness. If the change is less than the preset change threshold, terminate the iteration and perform distribution network optimization scheduling according to the agent's current scheduling strategy. Specifically, the formula for calculating the change in the total cooperative fitness is as follows: ,in The sum of the collaborative fitness of all agents after the d-th joint iteration optimization and the (d-1)-th joint iteration optimization are respectively, and the preset change threshold is set to 10. For reference Figure 3 The diagram shows the intelligent agent scheduling flowchart. Based on the scheduling process in the intelligent agent scheduling flowchart, the power distribution network is optimized and scheduled. S43: Calculate the group communication coefficient between the agent corresponding to the maximum cooperative fitness and agents of the same type, and use the group cooperative update method to iteratively optimize the scheduling parameters and scheduling strategy; In one embodiment of the present invention, the data transmission delay between the agent corresponding to the maximum cooperative fitness and agents of the same type is obtained through an active network measurement method, and the data transmission delay is converted into a group communication coefficient. The iterative optimization formulas for the scheduling parameters and scheduling strategy are as follows: ; ; in, Represents a random number between 0 and 1. All represent learning factors, set The values ​​are 0.2 and 0.3 respectively. Denotes the scheduling parameters of the agent obtained from the d-th joint iteration optimization. This represents the scheduling strategy of the agent obtained in the d-th joint iteration optimization. Denotes the scheduling parameters of the agent obtained from the (d-1)th joint iteration optimization. This represents the scheduling strategy of the agent obtained through the (d-1)th joint iteration optimization. This represents the scheduling strategy corresponding to the agent with the highest cooperative fitness after the (d-1)th joint iteration optimization. This represents the scheduling policy corresponding to the agent with the highest cooperative fitness after the (d-1)th joint iteration optimization, which is of the same type as the agent in the current iteration optimization. This represents the group communication coefficient between the agent currently undergoing optimization and the agent of the same type that has the highest cooperative fitness after the (d-1)th joint iteration. This represents the data transmission delay between the agent currently undergoing iteration optimization and the agent of the same type that has the highest cooperative fitness after the (d-1)th joint iteration optimization. Indicates the delay penalty coefficient, set It is 0.3; S44: Based on the distribution network operation constraints, determine whether the iteratively optimized scheduling strategy meets the distribution network operation constraints. If it does not meet the constraints, map the iteratively optimized scheduling strategy to a feasible scheduling strategy that meets the distribution network operation constraints, and proceed to step S42; if it does meet the constraints, proceed directly to step S42.

[0044] It should be noted that the terms "comprising," "including," or any other variations thereof used herein are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0045] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0046] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination, characterized in that, The method includes: S1: Collect multi-dimensional operation data from the source-grid-load-storage side operation unit, and perform spatiotemporal alignment on the multi-dimensional operation data based on the data transmission delay to obtain spatiotemporally aligned multi-dimensional operation data; S2: Use a multi-scale interpolation method based on transmission confidence to fill in missing values ​​in the spatiotemporally aligned multidimensional running data to obtain the filled multidimensional running data. S3: The source-grid-load-storage side operation unit is used as an intelligent agent. The completed multi-dimensional operation data is evaluated using the swarm intelligence collaborative evaluation function to generate the collaborative fitness of the intelligent agent. S4: Maximizing the cooperative fitness of the agents is taken as the local optimization objective of the agents. The power distribution network operation constraints are constructed, and the scheduling strategies of each agent are jointly iteratively optimized using a swarm intelligence cooperative mechanism to output an agent scheduling strategy that satisfies the power distribution network operation constraints.

2. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 1, characterized in that, The multi-dimensional operation data collected in step S1 from the source-grid-load-storage side operation units includes: S11: Collect multi-dimensional operation data from the source-grid-load-storage side operation units respectively, and generate the collection timestamp of the multi-dimensional operation data. The source-grid-load-storage side operation units include source-side operation units, grid-side operation units, load-side operation units and storage-side operation units respectively. S12: The source-grid-load-storage side operation unit transmits the collected multi-dimensional operation data to the distribution network, and the distribution network records the timestamp of the received multi-dimensional operation data; S13: Calculate the difference between the receiving timestamp and the acquisition timestamp of the multidimensional running data, and use it as the data transmission delay of the multidimensional running data.

3. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 2, characterized in that, Based on the data transmission delay, spatiotemporal alignment of multidimensional operational data also includes: S14: Generate the optimized scheduling cycle of the distribution network and construct a global time axis; S15: Based on the collection timestamp of the multidimensional running data, the multidimensional running data is allocated to the specified position of the global time axis to obtain multidimensional running data with a unified time axis; S16: Based on the multidimensional operation data of the unified time axis, construct the spatiotemporally aligned multidimensional operation data of the source-grid-load-storage side operation units. The spatiotemporally aligned multidimensional operation data is in matrix form. Each row of matrix elements describes the position coordinates of an operation unit, the multidimensional operation data at different time points in the global time axis, and the data transmission delay of the multidimensional operation data. The first column of the spatiotemporally aligned multidimensional operation data describes the position coordinates of all operation units. The other columns of matrix elements describe the multidimensional operation data of different operation units at the same time point and the data transmission delay of the multidimensional operation data.

4. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 1, characterized in that, Step S2 involves using multi-scale interpolation to fill in missing values ​​in the spatiotemporally aligned multidimensional running data, including: S21: Extract missing values ​​from the spatiotemporally aligned multidimensional running data; S22: Extract the rows before and after the missing values ​​in the matrix. Multidimensional operational data are used to form a short-term tensor sequence with missing values; S23: Perform temporal operations on the short-term tensor sequence using temporal convolution to obtain the temporal interpolation result of the missing values; S24: Select multiple running units of the same type as the running unit corresponding to the missing value and whose spatial distance is lower than the preset distance threshold as neighboring running units. Extract the multidimensional running data of the neighboring running units at the time point corresponding to the missing value as neighboring multidimensional running data. Use a spatial interpolation method based on spatial weight to perform weighted interpolation on the extracted neighboring multidimensional running data to obtain the spatial interpolation result of the missing value.

5. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 4, characterized in that, Step S2, which calculates the transmission confidence of missing values ​​and performs missing value completion based on the temporal and spatial interpolation results, also includes: S25: Obtain the average data transmission latency of the matrix row containing the missing values; S26: Convert the average data transmission delay into a transmission confidence level for missing values; S27: Using the transmission confidence as the imputation weight for the missing value, the temporal imputation result and the spatial imputation result of the missing value are weighted to obtain the missing value completion result.

6. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 1, characterized in that, In step S3, the source-grid-load-storage side operation unit is treated as an intelligent agent, and the completed multi-dimensional operation data is evaluated using a swarm intelligence collaborative evaluation function, including: S31: Calculate the mean value of the multidimensional running data of each row of matrix elements in the completed multidimensional running data, and use it as the running state of the agent corresponding to each row of matrix elements; S32: Convert the operating state of the intelligent agent into a collaborative index vector, wherein the collaborative index vector includes a power balance contribution index, a voltage support deviation index, an operating constraint occupancy index, and a collaborative adjustment potential index. S33: Receive the collaborative index vector of the agent using the swarm intelligence collaborative evaluation function, and generate the collaborative fitness of the agent.

7. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 6, characterized in that, The expression for the swarm intelligence collaborative evaluation function is: ； in, This represents the collaborative evaluation function of swarm intelligence. A vector representing the collaborative index of an agent. This represents the cooperative fitness of the agents. , The following are the collaborative index vectors respectively. The indicators include power balance contribution, voltage support deviation, operational constraint occupancy, and coordinated regulation potential. The weights of the power balance contribution index, voltage support deviation index, operating constraint occupancy index, and coordinated regulation potential index are represented in order, respectively. Indicates the desired power coordination level. This represents an exponential function with the natural constant as its base. This represents the voltage deviation penalty coefficient.

8. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 1, characterized in that, In step S4, maximizing the cooperative fitness of the agent is used as the local optimization objective of the agent, and power distribution network operation constraints are constructed, including: The power distribution network operation constraints include power balance constraints of source, grid, load and storage, operating voltage constraints, energy constraints of storage side and operating power constraints; The source-grid-load-storage power balance constraint means that the total active power provided to the distribution network by the intelligent agent corresponding to the source-side operating unit and the intelligent agent corresponding to the storage-side operating unit is balanced with the power demand of the intelligent agent corresponding to the load-side operating unit and the operating loss of the distribution network. The operating voltage constraint means that the operating voltage of the agent corresponding to each operating unit is always within the allowable safe operating range; The energy constraint on the storage side means that the energy storage charge state of the intelligent agent corresponding to the storage side operation unit must be maintained between the preset minimum and maximum energy boundaries to prevent the intelligent agent corresponding to the storage side operation unit from overcharging, over-discharging or unsustainable operation. The operating power constraint means that the operating power of the agent corresponding to each operating unit is always within the allowable safe operating range.

9. The distribution network optimization scheduling method based on swarm intelligence and source-grid-load-storage coordination as described in claim 8, characterized in that, Step S4, based on the power distribution network operation constraints, employs a swarm intelligence collaborative mechanism to jointly iteratively optimize the scheduling strategies of each agent, and further includes: S41: Initialize the scheduling strategy and scheduling parameters for each agent, where the scheduling strategy includes the agent's operating voltage and power, and the scheduling parameters are the iteration step size of the scheduling strategy during joint iterative optimization. S42: Calculate the cooperative fitness of the agents after adjustment according to the scheduling strategy, and calculate the maximum cooperative fitness of different types of agents after adjustment according to the scheduling strategy. The types of agents include source-side agents, network-side agents, load-side agents and storage-side agents, which correspond to source-side operation units, network-side operation units, load-side operation units and storage-side operation units respectively. Calculate the total cooperative fitness of all agents after adjustment according to the scheduling strategy, and calculate the change in the total cooperative fitness. If the change is less than the preset change threshold, terminate the iteration and perform distribution network optimization scheduling according to the agent's current scheduling strategy. S43: Calculate the group communication coefficient between the agent corresponding to the maximum cooperative fitness and agents of the same type, and use the group cooperative update method to iteratively optimize the scheduling parameters and scheduling strategy; S44: Based on the distribution network operation constraints, determine whether the iteratively optimized scheduling strategy meets the distribution network operation constraints. If it does not meet the constraints, map the iteratively optimized scheduling strategy to a feasible scheduling strategy that meets the distribution network operation constraints, and proceed to step S42; if it does meet the constraints, proceed directly to step S42.

Images