Method for participating in real-time optimization scheduling of power distribution network by power conversion cabinet cluster and related equipment

Abstract

The invention relates to the technical field of power distribution networks, and discloses a power conversion cabinet cluster participation power distribution network real-time optimization scheduling method and related equipment, and the method comprises the steps: carrying out the modeling of the schedulable capability of power conversion cabinets based on a power conversion demand prediction curve, and carrying out the clustering and dynamic grouping of the power conversion cabinets according to the schedulable capability, obtaining power conversion cabinet cluster resources with differentiated scheduling potential; establishing an optimal scheduling model of the real-time peak regulation of the power distribution network in which the power conversion cabinets participate based on the power conversion cabinet cluster resources, and obtaining a scheduling constraint framework considering the operation safety of the power distribution network and the power conversion service quality; and based on the optimal scheduling model and the scheduling constraint framework, a soft actor-commentator algorithm is adopted to quickly generate the charging power of each power conversion cabinet, and the power conversion cabinet cluster participates in real-time optimal scheduling of the power distribution network. The method ensures that the voltage of the low-voltage node does not exceed the limit, realizes efficient support and stable control of the voltage of the power distribution network, and ensures that the voltage of the power distribution network is in a safe and stable range.

Description

Technical Field

[0001] This invention relates to the field of power distribution network technology, specifically to a method and related equipment for real-time optimization scheduling of power distribution networks involving a cluster of battery swapping cabinets. Background Technology

[0002] As the penetration rate of distributed renewable energy in distribution networks continues to increase, its inherent volatility and uncertainty have brought significant challenges to distribution network operation. This has not only led to severe wind and solar curtailment and wasted energy resources, but also exacerbated voltage fluctuations in the distribution network, placing considerable pressure on the safe, stable, and economical operation of the power grid. To effectively address these negative impacts, fully exploring and rationally utilizing various flexible resources to participate in the real-time optimization and dispatch of the distribution network has become an important technical approach in the current power system field to solve the grid integration challenges of distributed renewable energy and improve the operational performance of the distribution network.

[0003] In recent years, electric bicycles have rapidly gained popularity in daily commuting and commercial services due to their green, environmentally friendly, energy-efficient, and flexible characteristics, resulting in a continuous increase in their social ownership. Especially driven by the rapid development of the express delivery and food delivery industries, electric bicycle battery swapping services have emerged and expanded rapidly. As a key infrastructure for battery swapping services, the deployment scale of electric bicycle battery swapping stations has been expanding under the guidance of relevant industrial policies and the active promotion of companies such as China Tower Energy Corporation, forming a massive distributed network.

[0004] It is worth noting that during daily operation, battery swapping cabinets often have idle fully charged batteries and flexible charging capacity, giving them a natural advantage in flexible charging. At the same time, battery swapping cabinets also have the technical advantages of fast response speed and high control precision. Therefore, the massive distributed deployment of battery swapping cabinets has become a type of flexible scheduling resource with considerable potential, and has a good foundation for participating in the optimization and scheduling of the distribution network.

[0005] However, current technologies lack systematic exploration and development of the scheduling potential of battery swapping cabinets, particularly in their application to real-time optimization and scheduling of distribution networks. This has resulted in the underutilization of this valuable flexibility resource, hindering its positive role in improving distribution network performance. Therefore, constructing a refined scheduling model for battery swapping cabinets can significantly enhance the distribution network's ability to regulate flexibility resources, achieving efficient peak shaving and valley filling of distribution network loads. Furthermore, it can provide rapid power support to the distribution network without additional investment in distributed energy storage facilities, effectively improving voltage quality and ultimately enhancing the economic efficiency and security of distribution network operation. This has significant practical implications and application value. Summary of the Invention

[0006] The purpose of this invention is to provide a method and related equipment for the real-time optimization and scheduling of power distribution networks by a cluster of battery swapping cabinets, so as to solve the technical problem of how to ensure that the voltage of low-voltage nodes does not exceed the limit.

[0007] This invention is achieved through the following technical solution: In a first aspect, the present invention provides a method for a battery swapping cabinet cluster to participate in real-time optimized scheduling of a power distribution network, comprising: The schedulable capability of battery swapping cabinets is modeled based on the battery swapping demand forecast curve, and the battery swapping cabinets are clustered and dynamically grouped according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. Based on the battery swapping cabinet cluster resources, an optimized scheduling model for battery swapping cabinets to participate in real-time peak shaving of the distribution network is established, resulting in a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. Based on the aforementioned optimized scheduling model and scheduling constraint framework, the soft actor-commentator algorithm is used to quickly generate the charging power of each battery swapping cabinet, enabling the battery swapping cabinet cluster to participate in the real-time optimized scheduling of the power distribution network.

[0008] Preferably, modeling the schedulable capacity of battery swapping cabinets based on the battery swapping demand forecast curve, and dynamically grouping the battery swapping cabinets according to the schedulable capacity specifically includes: Modeling the dispatchable capacity of the battery swapping cabinet, the modeling process includes calculating the battery swapping demand capacity, defining the minimum backup power capacity, setting power and energy constraints, and calculating the valley filling capacity and peak shaving capacity in different scenarios. Clustering and dynamic grouping of distributed battery swapping cabinets are performed, including standardization of battery swapping prediction curves, clustering analysis based on K-means algorithm, extraction of swapping trigger time, scheduling capability scoring, and adaptive level grouping.

[0009] Furthermore, the formula for calculating the battery swapping capacity demand is as follows:

[0010] in, This corresponds to the battery swapping capacity requirement; ; The number of batteries required by the battery swapping cabinet user at any given time; The formula for defining the minimum reserve capacity is as follows:

[0011] in, For battery swapping cabinet i exist t Minimum backup power capacity at any given time; The formulas for setting the capacity and power constraints of the battery swapping cabinet are as follows:

[0012] in, For battery swapping cabinet i exist t The capacity of a given moment; For this battery swapping cabinet t The charging power at any given time; The charging coefficient of the battery swapping cabinet; The battery swapping cabinet's power capacity meets the following constraints:

[0013] in, The formula represents the total number of batteries in the battery swapping cabinet. It clarifies the upper limit of the battery swapping cabinet's charge capacity at any given time, as well as the lower limit of the charge capacity that does not affect the quality of the battery swapping service at any given time. The charging power of the battery swapping cabinet meets the following constraints:

[0014] In the formula, This formula defines the maximum charging power of the battery swapping cabinet. t The upper limit of charging power at any given time is limited by the amount of electricity that can be charged at that time. The maximum rated charging power limit is limited by the lower limit of the charging power, which is limited by the battery swapping capacity required by the user at that moment. Both valley filling capacity and peak shaving capacity are based on and The size relationship is calculated by dividing the range of values, including: when At that time, the power swapping cabinet was in t Peak capacity at any time The range is:

[0015] when At that time, the power swapping cabinet was in t Peak capacity at any time The range is:

[0016] The battery swapping cabinet can respond to peak shaving tasks by reducing or stopping the default charging power to empty batteries. At this time, the lower limit of the charging power of the battery swapping cabinet at time t becomes 0. when At time t, the peak-shaving capacity of the power swapping cabinet is... The range is:

[0017] when At time t, the peak-shaving capacity of the power swapping cabinet is... The range is: .

[0018] Furthermore, the standardization process for the battery swapping prediction curve specifically includes: Battery swapping prediction curve After normalization, the processing formula is as follows:

[0019] Similarity is measured using Euclidean distance and clustered using the K-means algorithm. The specific formula for Euclidean distance is as follows:

[0020] Using volatility indicators Extract the set of group switching trigger times by combining thresholds;

[0021] Scheduling capability is scored using a function, expressed as follows:

[0022] in, Indicates battery swapping cabinet i At any moment t The score of regulatory ability, ∈[0.5,1.5] represents the service guarantee weighting factor, used to measure the degree to which future demand compresses the current adjustable space, and is taken as an empirical value. =1; Based on the scoring results, the battery swapping cabinets are classified into three categories at each time point: Category A, Category B, and Category C. The conditions for class A are met. The Class A battery swapping cabinet currently has a large amount of adjustable power resources and the rigid demand for service in the next period is relatively small. It has sufficient charging adjustment space and can take priority in undertaking peak shaving tasks. The conditions for class B are met. The Class B battery swapping cabinet still has a certain degree of controllability during the current period; The conditions for class C are met. The C-type battery swapping cabinet currently has a small number of empty batteries, or the demand will be high in the next period. It is necessary to prioritize basic battery swapping services, and the ability to participate in regulation is limited. Among them, the scoring threshold and scoring threshold The dynamic classification boundary is as follows:

[0023] in, , These are the current regulatory capacity score sequences. The mean and standard deviation.

[0024] Furthermore, the process of establishing an optimized scheduling model for battery swapping cabinets to participate in real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources includes: The objective function of the optimized scheduling model for the participation of the battery swapping cabinet in real-time peak shaving of the distribution network is set as the average voltage deviation of each node in the distribution network. The constraints include the charging power constraints of the battery swapping cabinet, the power flow conservation constraints of the distribution network, and the safe operation constraints of the distribution network. The charging power constraint of the battery swapping cabinet meets the following conditions:

[0025] in, The formula represents the total number of batteries in the battery swapping cabinet. It clarifies the upper limit of the battery swapping cabinet's charge capacity at any given time, as well as the lower limit of the charge capacity that does not affect the quality of the battery swapping service at any given time.

[0026] in, This formula defines the maximum charging power of the battery swapping cabinet. t The upper limit of charging power at any given time is limited by the amount of electricity that can be charged at that time. The maximum rated charging power limit is limited by the lower limit of the charging power, which is limited by the battery swapping capacity required by the user at that moment. The power flow conservation constraints of the distribution network satisfy:

[0027] in, , For nodes Flow to Node Active and reactive power; , For nodes Load at the location; , For connecting nodes With nodes Power loss on the branch; , This is the impedance of the branch; For nodes Voltage; The aforementioned power distribution network security constraints satisfy:

[0028] in, For this branch road The maximum safe current it can carry; , These are the minimum and maximum voltages allowed by the distribution network, respectively.

[0029] Preferably, based on the optimized scheduling model and scheduling constraint framework, the soft actor-commentator algorithm is used to quickly generate the charging power of each battery swapping cabinet, and the specific process of real-time optimized scheduling of the battery swapping cabinet cluster participating in the distribution network is as follows: S1, obtain system operation and environmental status data such as battery swapping demand forecast, battery status, node voltage, and voltage sensitivity coefficient calculated by dendritic network; S2. Construct the state space and action space. The state space contains information about the remaining power of the battery swapping cabinet, and the action space contains the charging power adjustment amount for each battery swapping cabinet. And is subject to adjustable capacity constraints; S3, initialize the various networks and experience replay buffers of the soft actor-critic algorithm, and set hyperparameters such as the learning rate; S4, Define the reward function S5 executes the strategy and collects experience data such as status and actions, storing it in the replay buffer; S6, sampled data is used to perform gradient updates on the policy network and value network, and the target network parameters are softly updated; S7: If the maximum number of training steps is reached or the strategy converges, training stops, the optimal scheduling strategy is output and deployed to the battery swapping cabinet controller.

[0030] Furthermore, the reward function is as follows:

[0031] in, As a reward for peak shaving; Reward voltage usage and punish voltage exceeding limits; encourage voltage stability. As a service reward, penalties will be imposed for battery swapping service failures caused by regulatory adjustments; As an efficiency incentive, priority is given to using Class A / B battery swapping cabinets to improve control efficiency.

[0032] Secondly, the present invention provides a system for real-time optimization and scheduling of power distribution networks involving a cluster of battery swapping cabinets, comprising: The schedulable modeling and grouping module is used to model the schedulable capability of the battery swapping cabinet based on the battery swapping demand prediction curve, and to cluster and dynamically group the battery swapping cabinets according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. The optimized scheduling model module is used to establish an optimized scheduling model for the battery swapping cabinets to participate in the real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources, and to obtain a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. The SAC algorithm scheduling module, based on the optimized scheduling model and scheduling constraint framework, uses the soft actor-commentator algorithm to quickly generate the charging power of each battery swapping cabinet, and performs real-time optimized scheduling of the battery swapping cabinet cluster participating in the distribution network.

[0033] Thirdly, the present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for the real-time optimization scheduling of the power distribution network by the battery swapping cabinet cluster.

[0034] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for the real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets.

[0035] Compared with the prior art, the present invention has the following beneficial technical effects: This invention provides a method for real-time optimization and scheduling of power distribution networks by clustering battery swapping cabinets. By aggregating a large number of dispersed battery swapping cabinets to form large-scale schedulable resources, it fully taps the scheduling potential of idle fully charged batteries and rechargeable capacity in the battery swapping cabinets. This allows such infrastructure, which originally only served electric bicycle battery swapping, to become an important supplementary force for peak shaving in the power distribution network. In actual scheduling, it can accurately realize peak shaving and valley filling operations, effectively offsetting the operational pressure on the power distribution network caused by the inherent volatility and uncertainty of renewable energy, significantly improving the adaptability and adjustment flexibility of the power distribution network to various load fluctuations, reducing wind and solar curtailment, and improving the overall energy utilization efficiency.

[0036] Furthermore, this invention adopts a data-driven control mode, which does not rely on the complex and precise parameters of the distribution network. It can learn the voltage sensitivity relationship between nodes simply through smart meter data, breaking the limitation of traditional voltage regulation that depends on precise parameters, reducing the implementation threshold of the regulation scheme, simplifying the regulation process, and achieving efficient support and stable control of the distribution network voltage. This avoids voltage fluctuation problems caused by distributed energy fluctuations and ensures that the distribution network voltage is within a safe and stable range.

[0037] Furthermore, the day-ahead clustering and dynamic grouping strategy adopted in this invention can pre-store various control-related information in the local controller of each battery swapping cabinet. When the distribution network is actually in operation and being dispatched, the battery swapping cabinet does not need to frequently interact with the aggregator. It can quickly determine the feasibility of peak-shaving commands based on the locally stored information and respond in a timely manner. This mode not only significantly reduces the communication burden of the system and reduces the hardware and maintenance costs caused by frequent communication, but also improves the response speed of the battery swapping cabinet to dispatching commands, ensuring the efficient advancement of real-time optimized dispatching of the distribution network. It takes into account the safety, economy and stability of distribution network operation and battery swapping services, providing a low-cost and highly reliable new path for flexible dispatching of the distribution network. Attached Figure Description

[0038] Figure 1 This is a flowchart of the method for the battery swapping cabinet cluster to participate in the real-time optimization scheduling of the power distribution network in an embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of the battery swapping cabinet cluster participating in the power distribution network optimization scheduling in an embodiment of the present invention; Figure 3 This is a flowchart of the SAC algorithm solution in an embodiment of the present invention; Figure 4 This is a schematic diagram of the real-time optimization and scheduling system for the power distribution network involving a cluster of battery swapping cabinets, as described in this embodiment of the invention. In the diagram: 1. Scheduling modeling grouping module; 2. Optimized scheduling model module; 3. SAC algorithm scheduling module. Detailed Implementation

[0039] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0040] The purpose of this invention is to provide a method and related equipment for the real-time optimization and scheduling of power distribution networks by a cluster of battery swapping cabinets, so as to solve the technical problem of how to ensure that the voltage of low-voltage nodes does not exceed the limit.

[0041] The present invention will now be described in further detail with reference to the accompanying drawings: Example 1 See Figure 1 In one embodiment of the present invention, a method for a battery swapping cabinet cluster to participate in real-time optimized scheduling of a power distribution network is provided, comprising: Step 1: Model the schedulable capability of the battery swapping cabinets based on the battery swapping demand forecast curve, and cluster and dynamically group the battery swapping cabinets according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. Step 11: Model the dispatchability of the battery swapping cabinet; The daily scheduling of the battery swapping cabinets is divided into 24 equal-length time periods. t ∈ T ={1, 2, ..., 24}, the time interval between adjacent moments The number of batteries required by users at each battery swapping station at any given time. This can be represented as follows: The required capacity of the battery swapping cabinet for each user at any given time can be obtained:

[0042] In the formula: This corresponds to the battery swapping capacity requirement;

[0043] The fastest charging rate from an empty battery to a fully charged battery is 1C. Whenever an empty battery is swapped in, the battery swapping station charges it at a default rate of 0.5C. When scheduling battery swapping stations, it is essential to ensure that the quality of the battery swapping service is not affected. To guarantee that users' battery swapping needs are met at all times, a minimum backup power capacity for a single battery swapping station is proposed, i.e., the number of batteries that can be swapped at any given time is reserved in advance:

[0044] In the formula: For battery swapping cabinet i exist t Minimum backup power capacity at any given time.

[0045] During the operation and scheduling of the battery swapping cabinet, it is assumed that the required batteries for swapping at any given time are swapped out at the start of that time. The battery swapping cabinet must meet the following capacity and power constraints:

[0046] In the formula: For battery swapping cabinet i exist t The capacity of a given moment; For this battery swapping cabinet t The charging power at any given time; This is the charging coefficient of the battery swapping cabinet.

[0047]

[0048] In the formula: The formula represents the total number of batteries in the battery swapping cabinet. It clarifies the upper limit of the battery swapping cabinet's charge capacity at any given time, as well as the lower limit of the charge capacity that does not affect the quality of the battery swapping service at any given time.

[0049]

[0050] In the formula: This formula defines the maximum charging power of the battery swapping cabinet. t The upper limit of charging power at any given time is limited by the amount of electricity that can be charged at that time. The lower limit of the charging power is limited by the rated maximum charging power and the battery swapping capacity required by the user at that moment.

[0051] Battery swapping stations can respond to valley filling tasks by increasing charging power. At that time, the power swapping cabinet was in t Valley capacity at any given time The range is:

[0052] when At that time, the power swapping cabinet was in t Valley capacity at any given time The range is:

[0053] The battery swapping cabinet can respond to peak shaving tasks by reducing or stopping the default charging power to empty batteries. At this time, the battery swapping cabinet... t The lower limit of charging power at any given time becomes 0.

[0054] when At that time, the power swapping cabinet was in t Peak capacity at any time The range is:

[0055] when At that time, the power swapping cabinet was in t Peak capacity at any time The range is:

[0056] The battery swapping cabinet can respond to peak shaving tasks by reducing or stopping the default charging power to empty batteries. At this time, the lower limit of the charging power of the battery swapping cabinet at time t becomes 0. when At time t, the peak-shaving capacity of the power swapping cabinet is... The range is:

[0057] when At time t, the peak-shaving capacity of the power swapping cabinet is... The range is: .

[0058] Step 12: Cluster and dynamically group the large number of dispersed battery swapping cabinets; Let the total number of battery swapping cabinets be N, and each battery swapping cabinet have a battery swapping prediction curve for the current day. To eliminate the impact of differences in the total number of batteries swapped between different battery swapping cabinets, and to retain only the curve shape information, the original curve was first subjected to maximum-minimum normalization:

[0059] Constructing a normalized matrix Calculate the Euclidean distance between any two curves:

[0060] Based on this distance metric, the K-means algorithm is used to cluster the battery swapping curves to obtain cluster centers representing different battery swapping demand curve shapes. To further identify key time points affecting the pace of regulation, a volatility index is introduced:

[0061] Combined with set threshold Extract the set of group switching trigger moments from each type of typical curve:

[0062] Will By writing to the corresponding battery swapping cabinet, the timing of the battery swapping behavior is discretized, and the control group update is triggered only at certain key moments, reducing the frequent changes and transmission requirements of control information.

[0063] In actual operation, the control capability of the battery swapping cabinet is not only affected by the number of batteries swapped at any given moment. The impact of this needs to be considered, and the battery swapping needs in the next period should also be taken into account. The service guarantee constraints are as follows. Specifically, the more empty batteries are swapped in at the current moment, the more abundant the system's adjustable resources; conversely, the greater the battery swapping demand at the next moment, the more batteries need to be fully charged in advance, and the smaller the charging adjustment margin left for the system. Therefore, the following schedulable capacity evaluation function is constructed to characterize the battery swapping cabinet. i At any moment t Overall response potential:

[0064] In the formula: Indicates battery swapping cabinet i At any moment t The score of regulatory ability, ∈[0.5,1.5] represents the service guarantee weighting factor, used to measure the degree to which future demand compresses the current adjustable space, and is taken as an empirical value. =1.

[0065] This function comprehensively evaluates the difference between adjustable power resources and the service load that must be guaranteed. The higher the score, the greater the control margin and the stronger the ability to participate in peak shaving scheduling.

[0066] Based on the above capability scores, the battery swapping cabinets are divided into the following three level groups at any given time: Type A: Satisfied These battery swapping cabinets currently have a lot of adjustable power resources and the rigid demand for service in the next period is relatively small. They have sufficient charging adjustment space and can take priority in peak shaving tasks. Category B: Satisfied These types of battery swapping cabinets still offer some flexibility in terms of adjustment during the current period. Class C: Satisfied These types of battery swapping stations currently have a limited number of empty batteries, or the demand may be high in the next period. Therefore, priority should be given to ensuring basic battery swapping services, and the ability to participate in regulation is limited.

[0067] Among them, the scoring threshold , The settings can be flexibly configured according to the control strategy. To enhance group adaptability, this paper adopts the following dynamic classification boundary based on standard deviation:

[0068] In the formula: , These are the current regulatory capacity score sequences. The mean and standard deviation. This setting can adaptively adjust the capacity distribution characteristics under different daily operating environments, avoiding group imbalance caused by a fixed threshold under extreme load scenarios.

[0069] Ultimately, the capability level classification result will be generated and written to the local controller of each battery swapping cabinet along with the demand clustering information. During operation, there is no need to communicate frequently with the battery swapping cabinet aggregator. The feasibility of peak shaving commands can be quickly determined locally, achieving efficient control response with low communication volume.

[0070] Step 2: Based on the battery swapping cabinet cluster resources, establish an optimized scheduling model for battery swapping cabinets to participate in real-time peak shaving of the distribution network, and obtain a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services, such as... Figure 2 As shown; The specific process is as follows: Step 21, the objective function is the average voltage deviation of each node in the distribution network; Step 22, the constraints include the charging power constraint of the battery swapping cabinet, the power flow conservation constraint of the distribution network, the safe operation constraint of the distribution network, and the service quality constraint of the battery swapping cabinet.

[0071] (1) Charging power constraints of battery swapping cabinet

[0072] In the formula: The formula represents the total number of batteries in the battery swapping cabinet. It clarifies the upper limit of the battery swapping cabinet's charge capacity at any given time, as well as the lower limit of the charge capacity that does not affect the quality of the battery swapping service at any given time.

[0073]

[0074] In the formula: This formula defines the maximum charging power of the battery swapping cabinet. t The upper limit of charging power at any given time is limited by the amount of electricity that can be charged at that time. The lower limit of the charging power is limited by the rated maximum charging power and the battery swapping capacity required by the user at that moment.

[0075] (2) The power flow constraints of the distribution network are as follows:

[0076] In the formula: , For nodes Flow to Node Active and reactive power; , For nodes Load at the location; , For connecting nodes With nodes Power loss on the branch; , This is the impedance of the branch; For nodes Voltage.

[0077] (3) The aforementioned distribution network security constraints are:

[0078] In the formula: For this branch road The maximum safe current it can carry; , These are the minimum and maximum voltages allowed by the distribution network, respectively.

[0079] Step 3: Based on the optimized scheduling model and scheduling constraint framework, the soft actor-commentator algorithm is used to quickly generate the charging power of each battery swapping cabinet, and to carry out real-time optimized scheduling of the battery swapping cabinet cluster participating in the power distribution network.

[0080] according to Figure 3 As shown, the specific process is as follows: Step 31: Obtain system operation data and environmental status: Obtain the predicted battery swapping demand, actual battery swapping quantity, charging power, battery status, node voltage, control group information, and voltage sensitivity coefficient calculated by dendritic network (DN) for each battery swapping cabinet at the current moment.

[0081] Step 32: Construct the state space and action space: The state space includes: the remaining power of each battery swapping cabinet, the current charging power, the predicted value of battery swapping demand in the next time period, the voltage of the node, and the voltage sensitivity coefficient.

[0082] The operating range refers to the amount of charging power adjustment for each battery swapping cabinet. The range is constrained by its adjustable capacity:

[0083] Step 33: Use the Soft Actor-Critic (SAC) algorithm to initialize the policy network, value network, target network, and experience replay buffer, and set hyperparameters such as learning rate, discount factor, and entropy coefficient.

[0084] Step 34: Define the reward function:

[0085] In the formula: As a reward for peak shaving; Reward voltage usage and punish voltage exceeding limits; encourage voltage stability. As a service reward, penalties will be imposed for battery swapping service failures caused by regulatory adjustments; As an efficiency incentive, priority is given to using Class A / B battery swapping cabinets to improve control efficiency.

[0086] Step 35: Execute the strategy and collect experience: In each control cycle, the agent outputs an action (power adjustment instruction) based on the current state, observes the next state and reward after execution, and stores the experience (state, action, reward, next state) in the replay buffer.

[0087] Step 36, Network Training and Update: Sample a batch of data from the experience replay buffer and perform gradient updates on the policy network and value network respectively, using a soft update method to synchronize the target network parameters.

[0088] Step 37: Determine if the termination condition is met: If the maximum number of training steps is reached or the policy converges (average reward stabilizes), then stop training; otherwise, return to step C5 to continue learning.

[0089] Step 38: Output the optimal scheduling strategy: After training, deploy the optimal strategy to each battery swapping cabinet controller to achieve adaptive power regulation based on local conditions, while meeting the goals of peak shaving and voltage stability.

[0090] In summary, the present invention provides a method for real-time optimization and scheduling of power distribution networks by aggregating a large number of dispersed battery swapping cabinets to form large-scale schedulable resources. This fully taps the scheduling potential of idle fully charged batteries and rechargeable capacity in the battery swapping cabinets, making this type of infrastructure, which originally only served electric bicycle battery swapping, an important supplementary force for peak shaving in the power distribution network. In actual scheduling, peak shaving and valley filling operations can be accurately realized, effectively offsetting the operational pressure on the power distribution network caused by the inherent volatility and uncertainty of renewable energy. This significantly improves the adaptability and adjustment flexibility of the power distribution network to various load fluctuations, reduces wind and solar curtailment, and improves the overall energy utilization efficiency.

[0091] Example 2 according to Figure 4 As shown, the present invention also provides a system for real-time optimization and scheduling of power distribution networks involving a cluster of battery swapping cabinets, comprising: The schedulable modeling and grouping module 1 is used to model the schedulable capability of the battery swapping cabinet based on the battery swapping demand prediction curve, and to cluster and dynamically group the battery swapping cabinets according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. The optimized scheduling model module 2 is used to establish an optimized scheduling model for the battery swapping cabinets to participate in the real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources, and to obtain a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. SAC algorithm scheduling module 3, based on the optimized scheduling model and scheduling constraint framework, uses the soft actor-commentator algorithm to quickly generate the charging power of each battery swapping cabinet, and performs real-time optimized scheduling of the battery swapping cabinet cluster participating in the distribution network.

[0092] Example 3 The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, such as a real-time optimization scheduling program for a battery swapping cabinet cluster participating in a power distribution network.

[0093] When the processor executes the computer program, it implements the steps of the above-mentioned method for the real-time optimization scheduling of the power distribution network by the battery swapping cabinet cluster, for example: The schedulable capability of battery swapping cabinets is modeled based on the battery swapping demand forecast curve, and the battery swapping cabinets are clustered and dynamically grouped according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. Based on the battery swapping cabinet cluster resources, an optimized scheduling model for battery swapping cabinets to participate in real-time peak shaving of the distribution network is established, resulting in a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. Based on the aforementioned optimized scheduling model and scheduling constraint framework, the soft actor-commentator algorithm is used to quickly generate the charging power of each battery swapping cabinet, enabling the battery swapping cabinet cluster to participate in the real-time optimized scheduling of the power distribution network.

[0094] Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system, for example: The schedulable modeling and grouping module 1 is used to model the schedulable capability of the battery swapping cabinet based on the battery swapping demand prediction curve, and to cluster and dynamically group the battery swapping cabinets according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. The optimized scheduling model module 2 is used to establish an optimized scheduling model for the battery swapping cabinets to participate in the real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources, and to obtain a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. SAC algorithm scheduling module 3, based on the optimized scheduling model and scheduling constraint framework, uses the soft actor-commentator algorithm to quickly generate the charging power of each battery swapping cabinet, and performs real-time optimized scheduling of the battery swapping cabinet cluster participating in the distribution network.

[0095] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the mobile terminal.

[0096] For example, the computer program can be divided into a schedulable modeling grouping module 1, an optimized scheduling model module 2, and a SAC algorithm scheduling module 3; The specific functions of each module are as follows: The schedulable modeling and grouping module 1 is used to model the schedulable capability of the battery swapping cabinet based on the battery swapping demand prediction curve, and to cluster and dynamically group the battery swapping cabinets according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. The optimized scheduling model module 2 is used to establish an optimized scheduling model for the battery swapping cabinets to participate in the real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources, and to obtain a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. SAC algorithm scheduling module 3, based on the optimized scheduling model and scheduling constraint framework, uses the soft actor-commentator algorithm to quickly generate the charging power of each battery swapping cabinet, and performs real-time optimized scheduling of the battery swapping cabinet cluster participating in the distribution network.

[0097] The mobile terminal can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The mobile terminal may include, but is not limited to, a processor and memory.

[0098] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the mobile terminal, connecting various parts of the mobile terminal via various interfaces and lines.

[0099] The memory can be used to store the computer program and / or module. The processor implements various functions of the mobile terminal by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0100] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function (such as sound playback, image playback, etc.); the data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). Furthermore, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disks, RAM, plug-in hard disks, SmartMediaCards (SMC), Secure Digital (SD) cards, FlashCards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0101] Example 4 The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method for real-time optimization scheduling of a battery swapping cabinet cluster in a power distribution network.

[0102] If the modules / units integrated in the mobile terminal are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0103] Based on this understanding, all or part of the processes in the above method can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of the above-described aggregated reinforcement learning resource scheduling method. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form.

[0104] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0105] It should be noted that the content contained in the computer-readable medium may be appropriately added to or subtracted from the content as required by the legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium may not include electrical carrier signals and telecommunication signals.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for real-time optimization scheduling of power distribution networks involving a cluster of battery swapping cabinets, characterized in that, include: The schedulable capability of battery swapping cabinets is modeled based on the battery swapping demand forecast curve, and the battery swapping cabinets are clustered and dynamically grouped according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. Based on the battery swapping cabinet cluster resources, an optimized scheduling model for battery swapping cabinets to participate in real-time peak shaving of the distribution network is established, resulting in a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. Based on the aforementioned optimized scheduling model and scheduling constraint framework, the soft actor-commentator algorithm is used to quickly generate the charging power of each battery swapping cabinet, enabling the battery swapping cabinet cluster to participate in the real-time optimized scheduling of the power distribution network.

2. The method for real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets according to claim 1, characterized in that, The schedulable capacity of battery swapping cabinets is modeled based on the battery swapping demand forecast curve, and the cabinets are dynamically grouped according to the schedulable capacity, specifically including: Modeling the dispatchable capacity of the battery swapping cabinet, the modeling process includes calculating the battery swapping demand capacity, defining the minimum backup power capacity, setting power and energy constraints, and calculating the valley filling capacity and peak shaving capacity in different scenarios. Clustering and dynamic grouping of distributed battery swapping cabinets are performed, including standardization of battery swapping prediction curves, clustering analysis based on K-means algorithm, extraction of swapping trigger time, scheduling capability scoring, and adaptive level grouping.

3. The method for real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets according to claim 2, characterized in that, The formula for calculating the battery swapping capacity demand is as follows: in, This corresponds to the battery swapping capacity requirement; ; The number of batteries required by the battery swapping cabinet user at any given time; The formula for defining the minimum reserve capacity is as follows: in, For battery swapping cabinet i exist t Minimum backup power capacity at any given time; The formulas for setting the capacity and power constraints of the battery swapping cabinet are as follows: in, For battery swapping cabinet i exist t The capacity of a given moment; For this battery swapping cabinet t The charging power at any given time; The charging coefficient of the battery swapping cabinet; The battery swapping cabinet's power capacity meets the following constraints: in, The formula represents the total number of batteries in the battery swapping cabinet. It clarifies the upper limit of the battery swapping cabinet's charge capacity at any given time, as well as the lower limit of the charge capacity that does not affect the quality of the battery swapping service at any given time. The charging power of the battery swapping cabinet meets the following constraints: In the formula, This formula defines the maximum charging power of the battery swapping cabinet. t The upper limit of charging power at any given time is limited by the amount of electricity that can be charged at that time. The maximum rated charging power limit is limited by the lower limit of the charging power, which is limited by the battery swapping capacity required by the user at that moment. Both valley filling capacity and peak shaving capacity are based on and The size relationship is calculated by dividing the range of values, including: when At that time, the power swapping cabinet was in t Peak capacity at any time The range is: when At that time, the power swapping cabinet was in t Peak capacity at any time The range is: The battery swapping cabinet can respond to peak shaving tasks by reducing or stopping the default charging power to empty batteries. At this time, the lower limit of the charging power of the battery swapping cabinet at time t becomes 0. when At time t, the peak-shaving capacity of the power swapping cabinet is... The range is: when At time t, the peak-shaving capacity of the power swapping cabinet is... The range is: 。 4. The method for real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets according to claim 2, characterized in that, The standardization process for the battery swapping prediction curve specifically includes: Battery swapping prediction curve After normalization, the processing formula is as follows: Similarity is measured using Euclidean distance and clustered using the K-means algorithm. The specific formula for Euclidean distance is as follows: Using volatility indicators Extract the set of group switching trigger times by combining thresholds; Scheduling capability is scored using a function, expressed as follows: in, Indicates battery swapping cabinet i At any moment t The score of regulatory ability, ∈[0.5,1.5] represents the service guarantee weighting factor, used to measure the degree to which future demand compresses the current adjustable space, and is taken as an empirical value. =1; Based on the scoring results, the battery swapping cabinets are classified into three categories at each time point: Category A, Category B, and Category C. The conditions for class A are met. The Class A battery swapping cabinet currently has a large amount of adjustable power resources and the rigid demand for service in the next period is relatively small. It has sufficient charging adjustment space and can take priority in undertaking peak shaving tasks. The conditions for class B are met. The Class B battery swapping cabinet still has a certain degree of controllability during the current period; The conditions for class C are met. The C-type battery swapping cabinet currently has a small number of empty batteries, or the demand will be high in the next period. It is necessary to prioritize basic battery swapping services, and the ability to participate in regulation is limited. Among them, the scoring threshold and scoring threshold The dynamic classification boundary is as follows: in, , These are the current regulatory capacity score sequences. The mean and standard deviation.

5. A method for real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets according to claim 2, characterized in that, The process of establishing an optimized scheduling model for battery swapping cabinets to participate in real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources includes: The objective function of the optimized scheduling model for the participation of the battery swapping cabinet in real-time peak shaving of the distribution network is set as the average voltage deviation of each node in the distribution network. The constraints include the charging power constraints of the battery swapping cabinet, the power flow conservation constraints of the distribution network, and the safe operation constraints of the distribution network. The charging power constraint of the battery swapping cabinet meets the following conditions: in, The formula represents the total number of batteries in the battery swapping cabinet. It clarifies the upper limit of the battery swapping cabinet's charge capacity at any given time, as well as the lower limit of the charge capacity that does not affect the quality of the battery swapping service at any given time. in, This formula defines the maximum charging power of the battery swapping cabinet. t The upper limit of charging power at any given time is limited by the amount of electricity that can be charged at that time. The maximum rated charging power limit is limited by the lower limit of the charging power, which is limited by the battery swapping capacity required by the user at that moment. The power flow conservation constraints of the distribution network satisfy: in, , For nodes Flow to Node Active and reactive power; , For nodes Load at the location; , For connecting nodes With nodes Power loss on the branch; , This is the impedance of the branch; For nodes Voltage; The aforementioned power distribution network security constraints satisfy: in, For this branch road The maximum safe current it can carry; , These are the minimum and maximum voltages allowed by the distribution network, respectively.

6. The method for real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets according to claim 1, characterized in that, The specific process of using the soft actor-commentator algorithm to quickly generate the charging power of each battery swapping station based on the optimized scheduling model and scheduling constraint framework, and then performing real-time optimized scheduling of the battery swapping station cluster in the distribution network is as follows: S1, obtain system operation and environmental status data such as battery swapping demand forecast, battery status, node voltage, and voltage sensitivity coefficient calculated by dendritic network; S2. Construct the state space and action space. The state space contains information about the remaining power of the battery swapping cabinet, and the action space contains the charging power adjustment amount for each battery swapping cabinet. And is subject to adjustable capacity constraints; S3, initialize the various networks and experience replay buffers of the soft actor-critic algorithm, and set hyperparameters such as the learning rate; S4, Define the reward function S5 executes the strategy and collects experience data such as status and actions, storing it in the replay buffer; S6, sampled data is used to perform gradient updates on the policy network and value network, and the target network parameters are softly updated; S7: If the maximum number of training steps is reached or the strategy converges, training stops, the optimal scheduling strategy is output and deployed to the battery swapping cabinet controller.

7. A method for real-time optimization scheduling of a power distribution network by a cluster of battery swapping cabinets according to claim 6, characterized in that, The reward function is as follows: in, As a reward for peak shaving; Reward voltage usage and punish voltage exceeding limits; encourage voltage stability. As a service reward, penalties will be imposed for battery swapping service failures caused by regulatory adjustments; As an efficiency incentive, priority is given to using Class A / B battery swapping cabinets to improve control efficiency.

8. A system for real-time optimization and scheduling of power distribution networks involving a cluster of battery swapping cabinets, characterized in that, include: The schedulable modeling and grouping module is used to model the schedulable capability of the battery swapping cabinet based on the battery swapping demand prediction curve, and to cluster and dynamically group the battery swapping cabinets according to the schedulable capability to obtain battery swapping cabinet cluster resources with differentiated scheduling potential. The optimized scheduling model module is used to establish an optimized scheduling model for the battery swapping cabinets to participate in the real-time peak shaving of the distribution network based on the battery swapping cabinet cluster resources, and to obtain a scheduling constraint framework that takes into account both the operational safety of the distribution network and the quality of battery swapping services. The SAC algorithm scheduling module, based on the optimized scheduling model and scheduling constraint framework, uses the soft actor-commentator algorithm to quickly generate the charging power of each battery swapping cabinet, and performs real-time optimized scheduling of the battery swapping cabinet cluster participating in the distribution network.

9. A mobile terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for real-time optimization scheduling of power distribution network participation by the battery swapping cabinet cluster as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for real-time optimization scheduling of the power distribution network by the battery swapping cabinet cluster as described in any one of claims 1-7.

Images