Power flexible resource regulation and control method and device, computer equipment and storage medium

Abstract

The invention relates to a regulation and control method and device for power flexible resources, computer equipment and a storage medium. Relates to the technical field of power regulation. The method comprises the following steps: constructing a dynamic characteristic model of each flexible resource based on pre-acquired real-time operation data of each flexible resource in the power source load storage system; constructing a resource aggregation model of the power source load storage system based on the dynamic characteristic model; determining a regulation and control scheme of each flexible resource according to the resource aggregation model under the condition that a scheduling instruction for the power source load storage system is received; sending an adjustment instruction corresponding to the regulation and control scheme to the corresponding flexible resource, and receiving execution condition information of each flexible resource for the related adjustment instruction; and updating the resource aggregation model according to the execution condition information, and determining an updated regulation and control scheme of each flexible resource according to the updated resource aggregation model. By adopting the method, a uniform dynamic aggregation framework oriented to multiple types of resources can be obtained.

Description

Technical Field

[0001] The present application relates to the technical field of power regulation and control, and particularly to a method, device, computer device, and storage medium for regulating flexible power resources. Background Art

[0002] In a power system, with the large-scale access of renewable energy and the diversification of electricity loads, how to effectively aggregate and coordinate the operation of dispersed flexibility resources has become an issue to be studied.

[0003] Currently, in a distributed power system, flexible resources such as distributed generation, controllable loads, and energy storage devices are widely distributed, forming a "source-load-storage" cluster. However, the above resources have characteristics such as diverse properties, dynamic changes, and high dispersion. Existing methods include aggregation methods for specific types of resources, such as virtual power plant technology, demand response aggregation, etc., but they all perform static aggregation on specific types of resources and cannot perform dynamic aggregation of multiple types of resources. Therefore, the current power flexible resource regulation method lacks a unified dynamic aggregation framework for multiple types of resources. Summary of the Invention

[0004] Based on this, in view of the technical problem that the current power flexible resource regulation method lacks a unified dynamic aggregation framework for multiple types of resources, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for regulating flexible power resources.

[0005] In a first aspect, the present application provides a method for regulating flexible power resources, including:

[0006] Based on the real-time operation data of each flexible resource in a power source-load-storage system obtained in advance, construct a dynamic characteristic model for each flexible resource; wherein, the dynamic characteristic model includes the response speed, adjustment range, and duration of each flexible resource;

[0007] Based on the dynamic characteristic model, construct a resource aggregation model for the power source-load-storage system;

[0008] When receiving a scheduling instruction for the power source-load-storage system, determine a regulation plan for each flexible resource according to the resource aggregation model; the regulation plan for the flexible resource includes adjustment instructions for each flexible resource;

[0009] Send the adjustment instructions corresponding to the regulation plan to the corresponding flexible resources, and receive the execution status information of each flexible resource for the relevant adjustment instructions;

[0010] Update the resource aggregation model according to the execution information, and determine the updated regulation scheme for each flexible resource according to the updated resource aggregation model.

[0011] In one embodiment, constructing the resource aggregation model of the power source-load-storage system based on each of the dynamic characteristic models includes: collecting historical operation data associated with the real-time operation data for a preset historical duration based on a pre-constructed hierarchical data acquisition architecture; the hierarchical data acquisition architecture includes edge computing devices configured to match each flexible resource, and a time series database for storing the historical operation data is deployed on the edge computing devices; constructing a dynamic characteristic modeling method based on the historical operation data, and determining time series characteristics associated with each flexible resource based on the time series database; obtaining output data of the dynamic characteristic model based on the dynamic characteristic modeling method and the time series characteristics, and constructing the resource aggregation model in the power source-load-storage system based on the output data. In one embodiment,

[0012] In one embodiment, before determining the regulation scheme for each flexible resource according to the resource aggregation model when receiving a scheduling instruction for the power source-load-storage system, it further includes: constructing a resource configuration environment model and initializing a deep deterministic policy gradient network; constructing an experience pool for the power source-load-storage system, obtaining a first state transition sample associated with the resource allocation process of the flexible resource from the experience pool, and training the deep deterministic policy gradient network based on at least part of the first state transition samples.

[0013] In one embodiment, the deep deterministic policy gradient network includes a value network and a policy network; the training steps of training the deep deterministic policy gradient network include: obtaining a target value based on the value network, obtaining a resource allocation action based on the policy network; calculating a temporal difference error and updating the parameters of the value network; calculating a policy gradient and updating the parameters of the policy network.

[0014] In one embodiment, after training the deep deterministic policy gradient network, it further includes: executing the resource allocation action generated by the policy network, collecting a second state transition sample associated with the resource allocation process of the flexible resource, and adding the second state transition sample to the experience pool; updating and training the deep deterministic policy gradient network based on at least part of the samples in the updated experience pool, and updating the resource allocation action generated by the corresponding policy network after the updated training, until a trained deep deterministic policy gradient network is obtained when it is monitored that the deep deterministic policy gradient network reaches a convergence state or the number of training rounds of training the deep deterministic policy gradient network reaches a preset number of rounds.

[0015] In one embodiment, after determining the regulation schemes of the flexible resources according to the resource aggregation model, the method further includes: constructing a distributed execution framework for the power source-load-storage system; the distributed execution framework includes a central control node and multiple flexible resources; deploying a resource configuration optimization algorithm on the central control node, and deploying local resource management agents on each of the flexible resources; wherein, the central control node collects the resource usage and task queue status of the flexible resources at a preset period to generate a global resource view; generates a resource configuration instruction set based on the global resource view and the resource configuration optimization algorithm corresponding to the central control node; obtains sub-instruction sets of each flexible resource based on the central control node and the resource configuration instruction set, and sends the sub-instruction sets to the corresponding flexible resources.

[0016] In one embodiment, updating the aggregation model includes: updating the parameters corresponding to the dynamic characteristics of each flexible resource, and adjusting the resource aggregation strategy.

[0017] In a second aspect, the present application further provides a regulation device for power flexible resources, including:

[0018] A model construction module, configured to construct a dynamic characteristic model of each flexible resource based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance; wherein, the dynamic characteristic model includes the response speed, regulation range and duration of each flexible resource;

[0019] The model construction module is further configured to construct a resource aggregation model of the power source-load-storage system based on the dynamic characteristic model;

[0020] A scheme determination module, configured to determine the regulation scheme of each flexible resource according to the resource aggregation model when receiving a scheduling instruction for the power source-load-storage system; the regulation scheme of the flexible resource includes the regulation instructions of each flexible resource;

[0021] An information receiving module, configured to send the regulation instructions corresponding to the regulation scheme to the corresponding flexible resources, and receive the execution status information of each flexible resource for the relevant regulation instructions;

[0022] The scheme determination module is further configured to update the resource aggregation model according to the execution status information, and determine the updated regulation scheme of each flexible resource according to the updated resource aggregation model.

[0023] In a third aspect, the present application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0024] Based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance, construct a dynamic characteristic model for each of the flexible resources; wherein, the dynamic characteristic model includes the response speed, adjustment range, and duration of each of the flexible resources;

[0025] Based on the dynamic characteristic model, construct a resource aggregation model for the power source-load-storage system;

[0026] When a scheduling instruction for the power source-load-storage system is received, determine a regulation scheme for each of the flexible resources according to the resource aggregation model; the regulation scheme for the flexible resources includes adjustment instructions for each of the flexible resources;

[0027] Send the adjustment instructions corresponding to the regulation scheme to the corresponding flexible resources, and receive the execution status information of each flexible resource for the relevant adjustment instructions;

[0028] According to the execution status information, update the resource aggregation model, and determine an updated regulation scheme for each of the flexible resources according to the updated resource aggregation model.

[0029] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0030] Based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance, construct a dynamic characteristic model for each of the flexible resources; wherein, the dynamic characteristic model includes the response speed, adjustment range, and duration of each of the flexible resources;

[0031] Based on the dynamic characteristic model, construct a resource aggregation model for the power source-load-storage system;

[0032] When a scheduling instruction for the power source-load-storage system is received, determine a regulation scheme for each of the flexible resources according to the resource aggregation model; the regulation scheme for the flexible resources includes adjustment instructions for each of the flexible resources;

[0033] Send the adjustment instructions corresponding to the regulation scheme to the corresponding flexible resources, and receive the execution status information of each flexible resource for the relevant adjustment instructions;

[0034] Update the resource aggregation model according to the execution information, and determine the updated regulation scheme for each flexible resource according to the updated resource aggregation model.

[0035] In a fifth aspect, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0036] Based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance, construct a dynamic characteristic model for each flexible resource; wherein, the dynamic characteristic model includes the response speed, adjustment range, and duration of each flexible resource;

[0037] Based on the dynamic characteristic model, construct a resource aggregation model for the power source-load-storage system;

[0038] When receiving a scheduling instruction for the power source-load-storage system, determine the regulation scheme for each flexible resource according to the resource aggregation model; the regulation scheme for the flexible resource includes the adjustment instruction for each flexible resource;

[0039] Send the adjustment instruction corresponding to the regulation scheme to the corresponding flexible resource, and receive the execution situation information of each flexible resource for the relevant adjustment instruction;

[0040] Update the resource aggregation model according to the execution situation information, and determine the updated regulation scheme for each flexible resource according to the updated resource aggregation model.

[0041] The above-mentioned control method, device, computer equipment, storage medium, and computer program product for flexible power resources have the following beneficial effects during the control of flexible power resources: First, based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance, a dynamic characteristic model of each flexible resource is constructed; among them, the dynamic characteristic model includes the response speed, adjustment range, and duration of each flexible resource; then, based on the dynamic characteristic model, a resource aggregation model of the power source-load-storage system is constructed; then, when receiving a scheduling instruction for the power source-load-storage system, the control scheme for each flexible resource is determined according to the resource aggregation model; the control scheme for the flexible resource includes the adjustment instructions for each flexible resource; then, the adjustment instructions corresponding to the control scheme are sent to the corresponding flexible resource, and the execution status information of each flexible resource for the relevant adjustment instructions is received; finally, according to the execution status information, the resource aggregation model is updated, and the updated control scheme for each flexible resource is determined according to the updated resource aggregation model. In the above process, by constructing a dynamic characteristic model, the response speed, adjustment range, and duration of each flexible resource can be understood in real time, so as to quickly formulate a suitable control scheme when receiving a scheduling instruction; the application of the dynamic characteristic model enables the system to adjust according to real-time data and flexibly respond to changes in power demand; by receiving the execution status information of each flexible resource for the adjustment instructions, the effect of the control scheme can be understood in time, and the resource aggregation model can be updated according to the feedback; therefore, the above method improves the operation efficiency, flexibility, and reliability of the power source-load-storage system while forming a unified dynamic aggregation framework for multiple types of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0043] Figure 1 It is a schematic flowchart of the control method for flexible power resources in an embodiment;

[0044] Figure 2 It is a schematic flowchart of the control steps for flexible power resources in an embodiment;

[0045] Figure 3 It is a detailed schematic flowchart of the control method for flexible power resources in another embodiment;

[0046] Figure 4 It is a structural block diagram of the control device for flexible power resources in an embodiment;

[0047] Figure 5 Internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] In an exemplary embodiment, as Figure 1 shown, a method for regulating power flexible resources is provided, including the following steps S102 to S110. Among them:

[0050] Step S102: Based on the real-time operation data of each flexible resource in the pre-acquired power source-load-storage system, construct a dynamic characteristic model for each flexible resource; wherein, the dynamic characteristic model includes the response speed, adjustment range, and duration of each flexible resource.

[0051] Among them, the real-time operation data refers to the dynamic data collected during the operation of each flexible resource in the power source-load-storage system, which can be used to reflect the actual operation status and performance of the flexible resource; the dynamic characteristic model can be a mathematical representation of the behavior of each flexible resource under different operating conditions; the response speed refers to the reaction time of the flexible resource to the adjustment instruction; the adjustment range refers to the power range that the flexible resource can adjust; the duration refers to the time that the flexible resource can maintain the adjustment state.

[0052] Optionally, the real-time operation data may include information such as current, voltage, power, and load change.

[0053] Step S104: Based on the dynamic characteristic model, construct a resource aggregation model for the power source-load-storage system.

[0054] Among them, the resource aggregation model is constructed based on the dynamic characteristic model and plays a role in coordinated scheduling in the power source-load-storage system.

[0055] Step S106: When receiving a scheduling instruction for the power source-load-storage system, determine the regulation scheme for each flexible resource according to the resource aggregation model; the regulation scheme for the flexible resource includes the adjustment instructions for each flexible resource.

[0056] Among them, the regulation scheme is a specific operation plan formulated according to the resource aggregation model, including the adjustment instructions for each flexible resource; the adjustment instruction is a specific operation command, which can be used to indicate how each flexible resource should adjust its output or load.

[0057] Specifically, the adjustment instructions need to match the dynamic characteristics of each flexible resource to ensure its effectiveness and safety.

[0058] Step S108: Send the adjustment instructions corresponding to the regulation plan to the corresponding flexible resources, and receive the execution status information of each flexible resource for the relevant adjustment instructions.

[0059] Among them, the execution status information refers to the execution result feedback by the flexible resource after receiving the adjustment instruction, which may include whether the execution is successful, the actual response time, the output power, etc.

[0060] Step S110: Update the resource aggregation model according to the execution status information, and determine the updated regulation plan for each flexible resource according to the updated resource aggregation model.

[0061] Among them, updating the resource aggregation model means updating the resource aggregation model according to the execution status information, which means continuously optimizing the resource aggregation model during the system operation to reflect the latest status and performance of the flexible resources; the updated regulation plan refers to a regulation plan that is re-formulated based on the updated resource aggregation model and can better adapt to the current operation status of the power source-load-storage system.

[0062] In the above-mentioned regulation method for power flexible resources, during the regulation process of power flexible resources, the following beneficial effects are achieved: First, based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance, a dynamic characteristic model of each flexible resource is constructed; among them, the dynamic characteristic model includes the response speed, adjustment range, and duration of each flexible resource; then, based on the dynamic characteristic model, a resource aggregation model of the power source-load-storage system is constructed; then, when receiving a scheduling instruction for the power source-load-storage system, determine the regulation plan for each flexible resource according to the resource aggregation model; the regulation plan for the flexible resource includes the adjustment instruction for each flexible resource; then send the adjustment instruction corresponding to the regulation plan to the corresponding flexible resource, and receive the execution status information of each flexible resource for the relevant adjustment instruction; finally, update the resource aggregation model according to the execution status information, and determine the updated regulation plan for each flexible resource according to the updated resource aggregation model. In the above process, by constructing a dynamic characteristic model, the response speed, adjustment range, and duration of each flexible resource can be understood in real time, so as to quickly formulate a suitable regulation plan when receiving a scheduling instruction; the application of the dynamic characteristic model enables the system to adjust according to real-time data and flexibly respond to changes in power demand; by receiving the execution status information of each flexible resource for the adjustment instruction, the effect of the regulation plan can be understood in time, and the resource aggregation model can be updated according to the feedback; therefore, the above method improves the operation efficiency, flexibility, and reliability of the power source-load-storage system while forming a unified dynamic aggregation framework for multiple types of resources.

[0063] In an exemplary embodiment, such as Figure 2As shown, it includes steps S202 to S206, where: Step S202, based on a pre-constructed hierarchical data acquisition architecture, collect historical operation data associated with the real-time operation data for a preset historical duration; the hierarchical data acquisition architecture includes edge computing devices configured to match with each flexible resource, and a time series database for storing historical operation data is deployed on the edge computing devices; Step S204, based on the historical operation data, construct a dynamic characteristic modeling method, and determine the time series characteristics associated with each flexible resource based on the time series database; Step S206, based on the dynamic characteristic modeling method and the time series characteristics, obtain the output data of the dynamic characteristic model, and construct a resource aggregation model in the power source-load-storage system based on the output data.

[0064] Among them, the hierarchical data acquisition architecture refers to a design structure that allows for the effective collection and processing of data at multiple levels; historical operation data refers to the information related to real-time operation data recorded within a preset historical duration, which can be used to analyze past operation states and performance; edge computing devices refer to a form of distributed computing that can perform computing and data processing at the edge location close to the data source; a time series database refers to a database used for storing and retrieving time series data; the dynamic characteristic modeling method refers to a systematic method for constructing a model that can obtain the behavioral characteristics of flexible resources under different conditions by analyzing historical operation data; time series characteristics refer to the characteristics with time attributes extracted from time series data, which can include information such as trends, periodicity, peaks, and drops.

[0065] Specifically, the hierarchical data acquisition architecture generally includes a data source, a data processing layer, and a data storage layer; the dynamic characteristic modeling method generally includes steps such as data processing, feature extraction, and model establishment.

[0066] In this embodiment, through the hierarchical data acquisition architecture, the rapid transmission and analysis of data can be promoted. In addition, the hierarchical data acquisition architecture can be specifically designed for different flexible resources to ensure the accuracy and efficiency of data acquisition.

[0067] In an exemplary embodiment, before determining the regulation scheme for each flexible resource according to the resource aggregation model when receiving a scheduling instruction for the power source-load-storage system, it further includes:

[0068] Construct a resource configuration environment model and initialize a deep deterministic policy gradient network; construct an experience pool for the power source-load-storage system, obtain a first state transition sample associated with the resource allocation process of the flexible resource from the experience pool, and train the deep deterministic policy gradient network based on at least part of the first state transition samples.

[0069] Among them, the Deep Deterministic Policy Gradient (DDPG) network can be a method that combines deep learning and reinforcement learning, suitable for dealing with problems in continuous action spaces. It learns policies through two main components: a deep neural network for estimating policies (actions), and another for evaluating the value function of actions; the experience pool refers to a database used to store experience samples obtained during interactions with the environment; the state transition sample refers to the state change process recorded during the operation of the system.

[0070] Specifically, the experience pool can be used to store information such as the states, actions, rewards, and next states of multiple time steps associated with flexible resources; the state transition sample can include the current state, the actions taken, the rewards obtained, and the new state transferred to.

[0071] In this embodiment, through the construction of the experience pool and the collection of samples, dynamic decision support for the power source-load-storage system is realized, providing a data-driven basis for optimizing resource allocation and scheduling strategies.

[0072] Furthermore, in one embodiment, the Deep Deterministic Policy Gradient (DDPG) network includes a value network and a policy network; the training steps for training the Deep Deterministic Policy Gradient (DDPG) network include:

[0073] Obtaining a target value based on the value network, and obtaining a resource allocation action based on the policy network; calculating the temporal difference error and updating the parameters of the value network; calculating the policy gradient and updating the parameters of the policy network.

[0074] Among them, the value network is a neural network structure used to evaluate the expected return or value of taking a specific action in a given state; the target value refers to the return or benefit target expected to be obtained in a specific state; the policy network refers to a neural network that can be used to directly learn the mapping from a certain state to action selection; the resource allocation action refers to the specific operation instructions determined according to the output of the policy network in the power source-load-storage system; the temporal difference error refers to the difference in reinforcement learning by comparing the currently estimated value with the actually obtained reward, which can also be added with the future estimated value.

[0075] Specifically, the value network can calculate the value of each state by learning the state and reward relationships in historical data, and thus provide a basis for optimizing decisions; the target value is usually obtained by estimating the expected value of future rewards; the resource allocation action can be to allocate a specific power to a certain flexible resource, start or stop a certain device, etc.; the policy gradient can be calculated by the influence of the parameters of the policy network on the expected return.

[0076] In this embodiment, through multiple steps such as calculating the target value, selecting actions, evaluating values, and updating network parameters, it aims to continuously improve the resource allocation strategy of the power source-load-storage system, ultimately realizing the scheduling and utilization efficiency of flexible resources. Moreover, through the combination of the value network and the policy network, it can effectively cope with complex dynamic environments.

[0077] In one embodiment, after training the deep deterministic policy gradient network, it further includes:

[0078] Execute the resource allocation actions generated by the policy network, collect the second state transition samples related to the resource allocation process of flexible resources, and add the second state transition samples to the experience pool; based on at least some of the samples in the updated experience pool, update the training of the deep deterministic policy gradient network, and update the resource allocation actions generated by the corresponding policy network after the updated training until, when it is monitored that the deep deterministic policy gradient network reaches the convergence state or the number of training rounds of training the deep deterministic policy gradient network reaches the preset number of rounds, obtain the trained deep deterministic policy gradient network.

[0079] Among them, the second state transition sample refers to the new state transition information recorded by the system after implementing the resource allocation action. These samples include the current state, the executed action, the obtained reward, and the updated new state; updating the training of the deep deterministic policy gradient network refers to retraining the deep deterministic policy gradient network by means of reinforcement learning according to the samples in the updated experience pool; updating the corresponding policy network after the updated training refers to, after training is completed and the parameters of the policy network are updated, according to the latest resource allocation actions output by the new policy network; the deep deterministic policy gradient network reaching the convergence state refers to monitoring the learning curve of the policy network, and when its performance tends to be stable and there is no obvious improvement; the trained deep deterministic policy gradient network refers to the version of the deep deterministic policy gradient network obtained after the training process, with optimized policy capabilities, and can make more effective resource allocation decisions under specific states.

[0080] In this embodiment, by executing the policy, collecting new state transition samples, updating the experience pool, and training the deep deterministic policy gradient network, the resource allocation strategy in the power source-load-storage system is continuously optimized; and by emphasizing the cyclicity and dynamic adaptability of the training, it is ensured that the trained deep deterministic policy gradient network can achieve the best decision-making support and resource allocation capabilities in complex environments; by monitoring the convergence state and controlling the number of training rounds, overfitting or insufficient learning can also be effectively prevented.

[0081] Furthermore, in one embodiment, after determining the regulation schemes of each flexible resource according to the resource aggregation model, it further includes:

[0082] Build a distributed execution framework for the power source-load-storage system; the distributed execution framework includes a central control node and multiple flexible resources; deploy a resource configuration optimization algorithm on the central control node and a local resource management agent on each flexible resource; wherein, the central control node collects the resource usage and task queue status of the flexible resources at a preset period to generate a global resource view; based on the global resource view and the resource configuration optimization algorithm corresponding to the central control node, generate a resource configuration instruction set; based on the central control node and the resource configuration instruction set, obtain the sub-instruction sets of each flexible resource and send the sub-instruction sets to the corresponding flexible resources.

[0083] Among them, the central control node refers to the core part of the system, responsible for overall coordination and management, usually including a resource configuration optimization algorithm and data processing and decision-making capabilities, used to generate a global resource view and monitor the running status of the entire system; the flexible resources can be used as execution nodes, referring to each operation unit in the distributed execution framework, and each node is responsible for specific resource management, task execution, and data collection; the resource usage refers to the current usage status and historical consumption data of resources such as electricity, storage, and equipment on each flexible resource; the task queue status refers to the current task list to be processed by each flexible resource, which can include the priority of the task, resource requirements, and expected completion time, etc.; the global resource view is a comprehensive view of the system generated by the central control node based on the information collected from each flexible resource; the resource configuration instruction set is a specific instruction set generated by the central control node based on the global resource view and the optimization algorithm, which can include specific operation instructions for each flexible resource, such as resource allocation, task assignment, etc.; the sub-instruction set can be refined from the resource configuration instruction set of the central control node and is a specific execution instruction for each flexible resource.

[0084] In this embodiment, through the cooperation of the central control node and multiple flexible resources, effective resource management and optimized configuration are achieved; based on periodic data collection and analysis, a comprehensive global resource view and corresponding instruction sets can be generated to support more flexible and intelligent decision-making, thereby improving the performance and reliability of the power source-load-storage system.

[0085] In one embodiment, updating the aggregation model includes:

[0086] Update the parameters corresponding to the dynamic characteristics of each flexible resource and adjust the resource aggregation strategy.

[0087] Among them, the parameters corresponding to the dynamic characteristics refer to the specific numerical values and indicators used to describe and quantify the dynamic characteristics, such as the power output range, start-up and shutdown times, adjustable load levels, etc.; adjustment refers to modifying and optimizing the resource aggregation strategy according to the latest dynamic characteristic parameters and system requirements.

[0088] In this embodiment, by periodically updating the dynamic characteristic parameters, the power source-load-storage can better reflect the actual capabilities and status of flexible resources, thereby effectively adjusting the flexible resource aggregation strategy, not only improving the utilization efficiency of flexible resources, but also enhancing the adaptability of the power source-load-storage system to environmental changes.

[0089] This application provides a method for regulating power flexible resources. To better understand the process of the above method for regulating power flexible resources, in combination with Figure 3 As shown below, the specific process of the method for regulating power flexible resources of this application is elaborated in detail, including the following steps:

[0090] Step S302: Obtain the real-time operation data of each flexible resource in the distributed source-load-storage cluster (power source-load-storage system), and establish a dynamic characteristic model for each flexible resource according to the real-time operation data.

[0091] Among them, flexible resources include adjustable power generation equipment, controllable load equipment, and energy storage equipment; the dynamic characteristic model includes the response speed, adjustment range, and duration of flexible resources.

[0092] Furthermore, obtaining and modeling the flexible resources in the distributed source-load-storage cluster also includes: constructing a hierarchical data acquisition architecture based on edge computing, deploying edge computing devices at each flexible resource node in the distributed source-load-storage cluster; using the MQTT (Message Queuing Telemetry Transport) protocol to realize the communication between the edge computing devices and the flexible resources; deploying a time series database on the edge computing devices to store high-frequency data for the most recent seven days; implementing a data anomaly detection algorithm based on a sliding window to dynamically adjust the window size for different types of data; developing a data compression algorithm, using difference compression for continuously changing and stable data, and using wavelet transform compression for fluctuating data.

[0093] Based on the data obtained from the hierarchical data acquisition architecture, a dynamic characteristic modeling method driven by deep learning is constructed. A neural network structure with multi-channel input is designed to receive different types of time-series features. A three-layer stacked Peephole LSTM (Peephole Long Short-Term Memory) is constructed, with each layer containing 128 neurons. A multi-head self-attention mechanism is added after the LSTM layer, with the number of heads set to eight. A residual connection is added between the LSTM (Long Short-Term Memory) layer and the attention layer. Two fully connected layers are used, containing 64 and 32 neurons respectively, and the ReLU (Rectified Linear Unit) activation function is adopted. The Huber (Huber Lose) loss function is used, and the δ parameter is set to one. The Adam optimizer is used, and the cosine annealing strategy is adopted. L2 regularization is applied, and the Dropout rate is set to 0.3 after the LSTM layer and 0.5 after the fully connected layer.

[0094] Based on the output of the dynamic characteristic modeling method, construct a flexibility resource aggregation model based on a graph neural network, abstract the distributed source-load-storage cluster as a dynamic graph G(V, E, A), where V is the node set representing flexible resources, E is the edge set representing the relationships between resources, and A is the adjacency matrix; design a dynamic adjacency matrix generation mechanism to dynamically update the adjacency matrix A based on geographical location, electrical connection, and historical interaction data; construct a spatio-temporal graph attention network ST-GAT (Spatio-Temporal Graph Attention Network), including a graph attention layer GAT (Graph Attention Layer) and a temporal convolutional network TCN (Temporal Convolutional Network) with causal convolution; use dilated convolution in the TCN, set the convolution kernel size to three, and the dilation rate to [1, 2, 4, 8]; design a gated fusion unit to learn the gating parameters through a one-by-one convolution and adaptively integrate spatial and temporal features; construct multiple ST-GAT modules to capture dynamic characteristics at different time scales; implement a contrastive learning strategy, design a data augmentation method and construct positive and negative sample pairs; adopt the InfoNCE (Information Noise Contrastive Estimation) loss function, set the temperature parameter τ to 0.07 initially and adaptively adjust it during the training process; integrate the knowledge distillation technique, design the distillation loss as a weighted sum of KL (Kullback-Leibler) divergence and MSE (Mean Squared Error), with the weight ratio of 0.7 to 0.3; implement progressive distillation, first only distill the shallow features, and then gradually incorporate the deep features, with five epochs (training cycles) in each stage; adopt the Lookahead optimizer, select Adam as the internal optimizer, set the synchronization period K to five, and the slow weight step size α to 0.5; use the One Cycle learning rate strategy, and determine the maximum learning rate through the LR Range Test (Learning Rate Range Test).

[0095] Step S304, based on the dynamic characteristic model, construct a flexibility resource aggregation model for the distributed source-load-storage cluster.

[0096] Among them, the flexibility resource aggregation model is used to characterize the overall flexibility regulation ability of the cluster.

[0097] Step S306, receive the power grid scheduling instruction, determine the required flexibility regulation target of the cluster according to the power grid scheduling instruction, use the flexibility resource aggregation model to calculate the optimal resource allocation plan to meet the flexibility regulation target, send the regulation instruction to the corresponding flexible resources, and monitor the execution status of each flexible resource in real time.

[0098] Among them, the optimal resource allocation plan includes adjustment instructions for each flexible resource.

[0099] Furthermore, the resource allocation environment model includes a state space, an action space, and a reward function. The state space represents the current resource utilization status and the task queue status. The action space represents executable resource allocation operations. The reward function is used to evaluate the resource allocation effect. Initialize the deep deterministic policy gradient network, which includes a policy network and a value network. The policy network is used to generate resource allocation actions, and the value network is used to evaluate the value of the state-action pair. Adopt an experience replay mechanism to construct an experience pool, which is used to store state transition samples during the resource allocation process. Each sample includes the current state, the executed action, the obtained reward, and the next state. Randomly sample a batch of state transition samples from the experience pool and use the samples to train the deep deterministic policy gradient network. The training process includes the following sub-steps: a) Calculate the target Q value using the value network; b) Generate exploratory actions using the policy network; c) Calculate the temporal difference error and update the value network parameters; d) Calculate the policy gradient and update the policy network parameters.

[0100] Further, execute the resource allocation actions generated by the policy network in the resource allocation environment, observe the environmental feedback, obtain new state transition samples, and add the new state transition samples to the experience pool. Repeat the execution of the training process and the execution process until the deep deterministic policy gradient network converges or reaches the preset number of training rounds. Use the trained deep deterministic policy gradient network to generate the optimal resource allocation plan according to the current resource utilization status and the task queue status. Convert the optimal resource allocation plan into specific resource scheduling instructions, including virtual machine migration instructions and container reallocation instructions. Send the resource scheduling instructions to the relevant computing nodes to achieve dynamic optimization of resource allocation. Monitor the resource allocation effect and collect performance metric data, including resource utilization rate, task completion time, and system throughput. Based on the performance metric data, dynamically adjust the reward function to balance the resource utilization efficiency and the task processing performance. Periodically retrain the deep deterministic policy gradient network to adapt to the changing resource requirements and task characteristics.

[0101] Step S308: Dynamically adjust the flexible resource aggregation model according to the execution situation. Based on the updated flexible resource aggregation model, continuously optimize the resource allocation plan to achieve dynamic aggregation and coordinated control of the flexible resources in the distributed source-load-storage cluster.

[0102] Among them, the dynamic adjustment may include updating the dynamic characteristic parameters of each flexible resource and adjusting the resource aggregation strategy.

[0103] Furthermore, it is also necessary to build a distributed execution framework, which includes a central control node and multiple flexible resources. The central control node is responsible for global resource configuration decisions, and the flexible resources are responsible for local resource scheduling. The resource configuration optimization algorithm is deployed on the central control node. The resource configuration optimization algorithm is based on a deep reinforcement learning model and is used to generate a globally optimal resource configuration plan. The local resource management agent is deployed on each flexible resource. The local resource management agent includes a resource monitoring module, a task scheduling module, and a resource adjustment module. The central control node periodically collects the resource usage status and task queue information of each flexible resource to form a global resource view. Based on the global resource view, the central control node runs the resource configuration optimization algorithm to generate a resource configuration instruction set, which includes virtual machine migration instructions and container reallocation instructions. The central control node decomposes the resource configuration instruction set into sub-instruction sets for each flexible resource and sends the sub-instruction sets to the corresponding flexible resources. The local resource management agents of each flexible resource receive the sub-instruction sets and perform the following operations: The resource monitoring module real-time collects the local resource usage and task execution status. The task scheduling module adjusts the local task allocation strategy according to the sub-instruction sets. The resource adjustment module performs virtual machine migration or container reallocation operations. While performing the resource adjustment operation, the local resource management agent continuously monitors the adjustment process and collects key performance indicators, including resource utilization rate, task response time, and throughput. If an abnormal situation or performance degradation is detected, the local resource management agent triggers a quick response mechanism to perform local resource fine-tuning and reports the abnormality to the central control node. The central control node aggregates the execution feedback and performance data of each flexible resource for global analysis and evaluates the resource configuration effect. Based on the global analysis results, the central control node dynamically adjusts the parameters of the resource configuration optimization algorithm, including the reward function coefficient and the exploration strategy. The central control node maintains a global resource configuration historical database to record the decision-making process and execution effect of each resource configuration. Using the global resource configuration historical database, the resource configuration optimization algorithm is periodically trained and optimized offline to improve its adaptability to long-term resource demand changes. A visual monitoring interface is built to display the global resource configuration status, the resource utilization rate of each flexible resource, and the task execution situation in real time, supporting administrators to perform manual intervention and policy adjustment.

[0104] In addition, in one embodiment, dynamically updating the model and continuously optimizing the configuration further includes: initializing system parameters, including setting the optimization time domain, discretizing the time interval, the initial system state vector, and the optimization objective weight coefficients; based on the updated flexibility resource aggregation model, performing rolling prediction for a number of future time steps to obtain prediction results, where the prediction results include renewable energy output, load demand, demand response potential, and grid electricity price; updating system constraint conditions according to the prediction results, including power balance constraints, energy storage system constraints, demand response constraints, and grid interaction constraints; constructing a multi-objective optimization function, where the multi-objective optimization function includes three sub-objectives: operating cost, renewable energy utilization rate, and system stability, and each sub-objective is weighted by a weight coefficient; using a mixed-integer linear programming solver to solve the multi-objective optimization function to obtain optimization results, where the optimization results include the charge and discharge plan of the energy storage system, the demand response scheduling plan, and the grid interaction plan; only executing the first time step in the optimization results and sending corresponding control instructions to the energy storage system, demand response resources, and the grid; monitoring the execution effect, recording the actual system state, and calculating the prediction error; updating the prediction model parameters based on the actual system state and the prediction error; rolling the time window forward by one time step, updating the initial state, discarding the data of the executed time step, and adding a new time step at the end of the prediction time domain; returning to the rolling prediction step and restarting the prediction, optimization, and execution loop; periodically evaluating the optimization effect, calculating the actual operation indicators, and comparing and analyzing the reasons for the differences with the optimization objectives; according to the evaluation results, dynamically adjusting the optimization strategy, including updating the weight coefficients of the objective function, optimizing the structure of the prediction model, and adjusting the parameters of the mixed-integer linear programming solver.

[0105] Furthermore, in one embodiment, a flexible resource dynamic aggregation system for a distributed source-load-storage cluster is provided, including: a first unit configured to obtain real-time operation data of each flexible resource in the distributed source-load-storage cluster, where the flexible resources include adjustable power generation devices, controllable load devices, and energy storage devices; establish a dynamic characteristic model for each flexible resource according to the real-time operation data, where the dynamic characteristic model includes the response speed, adjustment range, and duration of the flexible resource; construct a flexible resource aggregation model for the distributed source-load-storage cluster based on the dynamic characteristic model, where the flexible resource aggregation model is used to characterize the overall flexibility regulation ability of the cluster; a second unit configured to receive a power grid dispatch instruction, determine the flexibility regulation target required by the cluster according to the power grid dispatch instruction; calculate an optimal resource allocation plan that meets the flexibility regulation target by using the flexible resource aggregation model, where the optimal resource allocation plan includes adjustment instructions for each flexible resource; send the adjustment instructions to the corresponding flexible resources and monitor the execution status of each flexible resource in real time; a third unit configured to dynamically adjust the flexible resource aggregation model according to the execution status, including updating the dynamic characteristic parameters of each flexible resource and adjusting the resource aggregation strategy; continuously optimize the resource allocation plan based on the updated flexible resource aggregation model to achieve dynamic aggregation and coordinated control of the flexible resources in the distributed source-load-storage cluster.

[0106] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless specifically stated in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least some of the steps or stages in other steps or other steps.

[0107] Based on the same inventive concept, an embodiment of the present application also provides a control device for power flexible resources for implementing the control method of the power flexible resources described above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the control device for power flexible resources provided below can refer to the limitations on the control method of power flexible resources in the above text, and will not be repeated here.

[0108] In an exemplary embodiment, as Figure 4As shown, a regulation device for flexible power resources is provided, including: a model construction module 401, a scheme determination module 402, and an information receiving module 403, where:

[0109] The model construction module 401 is used to construct a dynamic characteristic model of each flexible resource based on the real-time operation data of each flexible resource in the power source-load-storage system obtained in advance; among them, the dynamic characteristic model includes the response speed, regulation range, and duration of each flexible resource.

[0110] The model construction module 401 is also used to construct a resource aggregation model of the power source-load-storage system based on the dynamic characteristic model.

[0111] The scheme determination module 402 is used to determine the regulation scheme of each flexible resource according to the resource aggregation model when receiving a scheduling instruction for the power source-load-storage system; the regulation scheme of the flexible resource includes the regulation instructions of each flexible resource.

[0112] The information receiving module 403 is used to send the regulation instructions corresponding to the regulation scheme to the corresponding flexible resources and receive the execution status information of each flexible resource for the relevant regulation instructions.

[0113] The scheme determination module 402 is also used to update the resource aggregation model according to the execution status information and determine the updated regulation scheme of each flexible resource according to the updated resource aggregation model.

[0114] Furthermore, in one embodiment, the model construction module 401 is also used to collect historical operation data associated with the real-time operation data for a preset historical duration based on the pre-constructed hierarchical data acquisition architecture; the hierarchical data acquisition architecture includes edge computing devices matched with each flexible resource, and the edge computing devices are deployed with time series databases for storing historical operation data; based on the historical operation data, a dynamic characteristic modeling method is constructed, and the time series characteristics associated with each flexible resource are determined based on the time series database; based on the dynamic characteristic modeling method and the time series characteristics, the output data of the dynamic characteristic model is obtained, and a resource aggregation model in the power source-load-storage system is constructed based on the output data.

[0115] Furthermore, in one embodiment, the information receiving module 403 is also used to construct a resource configuration environment model and initialize a deep deterministic policy gradient network; construct an experience pool of the power source-load-storage system, obtain a first state transition sample associated with the resource allocation process of the flexible resource from the experience pool, and train the deep deterministic policy gradient network based on at least part of the first state transition samples.

[0116] Further, in one embodiment, the information receiving module 403 is further configured to obtain a target value based on the value network, obtain a resource allocation action based on the policy network; calculate a temporal difference error, and update the parameters of the value network; calculate a policy gradient, and update the parameters of the policy network.

[0117] Further, in one embodiment, the information receiving module 403 is further configured to execute the resource allocation action generated by the policy network, collect a second state transition sample of the resource allocation process associated with the flexible resource, and add the second state transition sample to the experience pool; based on at least some of the samples in the updated experience pool, update the training of the deep deterministic policy gradient network, and update the resource allocation action generated by the corresponding policy network after the updated training is performed, until a trained deep deterministic policy gradient network is obtained when it is monitored that the deep deterministic policy gradient network reaches a convergence state, or the number of training rounds for training the deep deterministic policy gradient network reaches a preset number of rounds.

[0118] Further, in one embodiment, the solution determination module 402 is further configured to construct a distributed execution framework for the power source-load-storage system; the distributed execution framework includes a central control node and multiple flexible resources; a resource configuration optimization algorithm is deployed on the central control node, and a local resource management agent is deployed on each flexible resource; wherein, the central control node collects the resource usage and task queue status of the flexible resources according to a preset period to generate a global resource view; based on the global resource view and the resource configuration optimization algorithm corresponding to the central control node, generate a resource configuration instruction set; based on the central control node and the resource configuration instruction set, obtain a sub-instruction set for each flexible resource, and send the sub-instruction set to the corresponding flexible resource.

[0119] Further, in one embodiment, the solution determination module 402 is further configured to update the parameters corresponding to the dynamic characteristics of each flexible resource, and adjust the resource aggregation strategy.

[0120] Each module in the above power flexible resource control device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor in the computer device in hardware form or independent of the processor, or stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0121] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as Figure 5As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the regulation data of the flexible power resources. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for regulating flexible power resources.

[0122] Those skilled in the art can understand that Figure 5 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0123] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0124] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0125] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0126] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0127] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0128] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0129] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for regulating flexible power resources, characterized in that: The method comprises: Based on the real-time operation data of each flexible resource in the power source-load-storage system acquired in advance, a dynamic characteristic model of each flexible resource is constructed; wherein the dynamic characteristic model includes the response speed, adjustment range and duration of each flexible resource; Based on the dynamic characteristic model, construct a resource aggregation model of the power source-load-storage system; Upon receiving a dispatch instruction for the power source-load-storage system, determining a control scheme for each of the flexible resources according to the resource aggregation model; the control scheme for the flexible resources includes an adjustment instruction for each of the flexible resources; Sending the adjustment instruction corresponding to the control scheme to the corresponding flexible resource, and receiving the execution status information of each flexible resource for the relevant adjustment instruction; The resource aggregation model is updated according to the execution status information, and an updated control scheme of each of the flexible resources is determined according to the updated resource aggregation model.

2. The method according to claim 1, characterized in that The resource aggregation model of the power source-load-storage system is constructed based on each of the dynamic characteristic models, including: Based on a pre-built hierarchical data collection architecture, historical operation data of a preset historical duration associated with the real-time operation data is collected; the hierarchical data collection architecture includes an edge computing device matched with each of the flexible resources, and the edge computing device is deployed with a time series database for storing the historical operation data; Based on the historical operation data, a dynamic characteristic modeling method is constructed, and based on the time series database, a time series feature associated with each of the flexible resources is determined; Based on the dynamic characteristic modeling method and the timing characteristics, the output data of the dynamic characteristic model is obtained, and the resource aggregation model in the power source-load-storage system is constructed based on the output data.

3. The method according to claim 1, characterized in that In the case of receiving a dispatch instruction for the power source-load-storage system, before determining the control scheme of each flexible resource according to the resource aggregation model, the method further includes: Build a resource configuration environment model and initialize a deep deterministic policy gradient network; An experience pool of the power source-load-storage system is constructed, and a first state transition sample of a resource allocation process associated with the flexible resource is obtained from the experience pool, and the deep deterministic policy gradient network is trained based on at least part of the first state transition sample.

4. The method according to claim 3, characterized in that The deep deterministic policy gradient network includes a value network and a policy network; The training step of training the deep deterministic policy gradient network includes: Acquire a target value based on the value network, and acquire a resource allocation action based on the policy network; Calculating the time difference error and updating the parameters of the value network; Calculate the policy gradient and update the parameters of the policy network.

5. The method according to claim 4, characterized in that After the training of the deep deterministic policy gradient network, the method further includes: Executing the resource allocation action generated by the policy network, collecting a second state transition sample of the resource allocation process associated with the flexible resource, and adding the second state transition sample to the experience pool; Based on at least part of the samples in the updated experience pool, the deep deterministic policy gradient network is updated and trained, and the resource allocation action generated by the corresponding policy network after the updated training is updated and executed, until the deep deterministic policy gradient network is monitored to reach a convergence state, or the training rounds of the deep deterministic policy gradient network reach a preset round, the deep deterministic policy gradient network that has been trained is obtained.

6. The method according to claim 1, characterized in that After determining the control scheme of each flexible resource according to the resource aggregation model, the method further includes: Constructing a distributed execution framework of the power source-load-storage system; the distributed execution framework includes a central control node and multiple flexible resources; Deploy a resource configuration optimization algorithm on the central control node, and deploy a local resource management agent on each of the flexible resources; wherein the central control node collects resource usage and task queue status of the flexible resources according to a preset period to generate a global resource view; Generate a resource configuration instruction set based on the global resource view and the resource configuration optimization algorithm corresponding to the central control node; Based on the central control node and the resource configuration instruction set, a sub-instruction set of each flexible resource is obtained, and the sub-instruction set is sent to each corresponding flexible resource.

7. The method according to claim 1, characterized in that The updating of the aggregation model comprises: The parameters corresponding to the dynamic characteristics of each of the flexible resources are updated, and the resource aggregation strategy is adjusted.

8. A control device for flexible power resources, characterized in that: The device comprises: A model building module, used to build a dynamic characteristic model of each flexible resource based on the real-time operation data of each flexible resource in the power source, load and storage system acquired in advance; wherein the dynamic characteristic model includes the response speed, adjustment range and duration of each flexible resource; The model building module is further used to build a resource aggregation model of the power source-load-storage system based on the dynamic characteristic model; A scheme determination module is used to determine the control scheme of each of the flexible resources according to the resource aggregation model when receiving a dispatch instruction for the power source-load-storage system; the control scheme of the flexible resources includes an adjustment instruction of each of the flexible resources; An information receiving module, used for sending the adjustment instruction corresponding to the control scheme to the corresponding flexible resource, and receiving the execution status information of each flexible resource for the relevant adjustment instruction; The scheme determination module is further used to update the resource aggregation model according to the execution status information, and determine the updated control scheme of each flexible resource according to the updated resource aggregation model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images