Virtual power plant scheduling system and method based on cloud edge collaboration

Abstract

The invention discloses a virtual power plant scheduling system and method based on cloud edge collaboration, and relates to the field of virtual power plants. The method comprises the steps of obtaining an edge cluster capable of being used for processing calculation tasks and the calculation tasks needing to be processed at the current moment, counting calculation resources of the edge cluster, constructing a resource matrix by utilizing the obtained calculation resources, counting calculation resource demands of the calculation tasks, and constructing a demand matrix by utilizing the obtained calculation resource demands. And constructing a demand matching model of the computing resources and the computing tasks, allocating the computing tasks needing to be processed to the edge cluster for processing by utilizing the model, and when any computing task is allocated to the edge cluster, updating the resource matrix and the demand matrix, so that when a large-scale task is faced, the computing task can be allocated to the edge cluster, and the computing task can be allocated to the edge cluster. The computing resources are quickly and uniformly distributed, the edge computing resources are fully utilized, and the scheduling reliability of the virtual power plant cloud edge collaborative system is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of virtual power plants, in particular to a virtual power plant scheduling system and method based on cloud-edge collaboration. BACKGROUND

[0002] With the rapid development of energy transformation and distributed energy, virtual power plants (VPPs) have gradually become an important part of the power system as an effective means of integrating distributed energy resources (DERs). Virtual power plants aggregate a large number of distributed energy units (such as wind power, photovoltaic power, energy storage systems, etc.) to achieve flexible scheduling and optimized management of energy.

[0003] Currently, the operation of virtual power plants faces many challenges, especially in terms of scheduling strategies. Cloud and edge servers need to be collaboratively controlled in terms of computing and storage resources while performing daily scheduling and control tasks. However, virtual power plants usually contain a large number of distributed energy units, and the operating characteristics of different units differ greatly, with different time scale scheduling requirements. For example, the output of photovoltaic power units is greatly affected by weather conditions, while energy storage systems need to consider the charging and discharging cycle and power limit. Therefore, at the same time, virtual power plants need to handle numerous computing tasks, including power prediction, optimized scheduling, real-time control, etc.

[0004] Traditional virtual power plant scheduling methods mainly rely on centralized cloud computing, which concentrates all tasks in the cloud for processing. In the face of large-scale distributed energy, this approach can easily lead to computing resource overload, and cannot fully utilize edge computing resources, resulting in uneven distribution of computing resources. In addition, in the centralized scheduling mode, if computing tasks cannot be allocated to corresponding computing resources in a timely manner, faults or network delays will occur, severely affecting the scheduling reliability of the entire system. Therefore, we propose a virtual power plant scheduling system and method based on cloud-edge collaboration. SUMMARY

[0005] The main purpose of the present application is to provide a virtual power plant scheduling system and method based on cloud-edge collaboration, which can effectively solve the problems in the background art.

[0006] To achieve the above-mentioned purpose, the technical scheme adopted by the present application is,

[0007] A virtual power plant scheduling method based on cloud-edge collaboration, comprising:

[0008] Step 1: Obtain the edge cluster available for processing computing tasks at the current time t and the computing tasks that need to be processed represents the i-th edge cluster available for processing computing tasks at the current time t; This represents the j-th computational task that needs to be processed at time t.

[0009] The computational tasks include at least one of the following: power flow calculation, short-circuit calculation, transient stability analysis, small-disturbance stability analysis, optimal power flow calculation, reliability assessment, power market analysis, and edge computing tasks.

[0010] Step 2: Calculate the computing resources of the edge cluster and construct a resource matrix R using the obtained computing resources. t ,in, This represents the nth computing resource of the mth edge cluster available for processing computing tasks at time t; i = 1, 2, ..., m;

[0011] The computing resources include at least one of the following: CPU resources, memory resources, storage resources, and network resources.

[0012] Step 3: Calculate the computing resource requirements of the computing task, and construct a requirement matrix D using the obtained computing resource requirements. t ,in, This represents the nth computational resource requirement of the Jth computational task that needs to be processed at time t; j = 1, 2, ..., J;

[0013] Step 4: Construct a demand matching model between the computing resources and the computing tasks. Using this model, allocate the j-th computing task to the edge cluster for processing. The expression for the demand matching model is:

[0014]

[0015] In the formula, This represents the f-th computing resource of the edge cluster assigned the k-th computing task at time t. θ represents the f-th computational resource requirement of the k-th assigned computational task at time t; kf F represents the weight of the f-th computational resource requirement of the k-th assigned computational task; k F represents the load rate of the edge cluster assigned the k-th computing task before the task was assigned; k ' represents the load rate of the edge cluster assigned the k-th computing task after task allocation; F max This indicates the load rate threshold for the edge cluster;

[0016] The load rate F of the edge cluster assigned the kth computing task before the task was assigned k The calculation formula is:

[0017]

[0018] In the formula, The maximum value in the calculation result is represented as The maximum value in the calculation result is represented as kf The fth computing resource of the edge cluster allocated with the kth computing task is represented as

[0019] The load rate F of the edge cluster allocated with the kth computing task after task allocation is represented as k The calculation formula of the maximum value is as follows:

[0020]

[0021] In the formula, The maximum value in the calculation result is represented as The maximum value in the calculation result is represented as

[0022] Step five: updating the resource matrix and the demand matrix when any computing task is allocated to the edge cluster.

[0023] A virtual power plant scheduling system based on cloud-edge collaboration, comprising:

[0024] An edge cluster acquisition module is configured to acquire edge clusters available for processing computing tasks at a current time t In the formula, The ith edge cluster available for processing computing tasks is represented as

[0025] A computing task acquisition module is configured to acquire computing tasks to be processed at a current time t

[0026] A resource matrix construction module is configured to count computing resources of the edge clusters and construct a resource matrix by using the acquired computing resources.

[0027] A demand matrix construction module is configured to count computing resource demands of the computing tasks and construct a demand matrix by using the acquired computing resource demands.

[0028] A demand matching module is configured to construct a demand matching model of the computing resources and the computing tasks, and allocate the jth computing task to be processed to the edge cluster for processing by using the model.

[0029] A matrix updating module is configured to dynamically update the resource matrix and the demand matrix when any computing task is allocated to the edge cluster.

[0030] The system further comprises a memory, a processor and a computer program stored on the memory and executable on the processor.

[0031] The present application has the following advantages,

[0032] Compared with the prior art, by acquiring edge clusters available for processing computing tasks and computing tasks needing to be processed at the current moment, counting the computing resources of the edge clusters, constructing a resource matrix by using the acquired computing resources, counting the computing resource requirements of the computing tasks, constructing a demand matrix by using the acquired computing resource requirements, constructing a demand matching model of the computing resources and the computing tasks, allocating the computing tasks needing to be processed to the edge clusters for processing by using the model, and updating the resource matrix and the demand matrix when any computing task is allocated to the edge clusters, the computing resources can be quickly and evenly allocated when facing large-scale tasks, the edge computing resources are fully utilized, and the scheduling reliability of the virtual power plant cloud edge collaborative system is improved. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 A structure schematic diagram of the virtual power plant scheduling method based on cloud edge collaboration according to the present application;

[0034] Figure 2 A structure schematic diagram of the virtual power plant scheduling system based on cloud edge collaboration according to the present application;

[0035] Figure 3 An implementation flow schematic diagram of the virtual power plant scheduling method based on cloud edge collaboration according to the present application. DETAILED DESCRIPTION

[0036] The present application will be further described below in combination with specific embodiments, wherein the drawings are only used for exemplary description, and the representations are only schematic diagrams, not physical diagrams, and cannot be understood as limitations on the present application. In order to better illustrate the specific embodiments of the present application, some components of the drawings will be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0037] As shown in Figures 1-2 , the specific implementation flow of the technical scheme of the present application includes the following steps:

[0038] Step 1: acquiring edge clusters available for processing computing tasks and computing tasks needing to be processed at the current moment t represents the i-th edge cluster available for processing computing tasks at the current moment t; represents the j-th computing task needing to be processed at the current moment t; wherein the computing tasks include power flow calculation, short-circuit calculation, transient stability analysis, small disturbance stability analysis, optimal power flow calculation, reliability evaluation, power market analysis, and edge computing tasks.

[0039] Step 2: counting the computing resources of the edge clusters, and constructing a resource matrix R by using the acquired computing resources​t wherein, represents the nth item of computing resource of the mth edge cluster available for processing the computing task at the current time t; i = 1, 2, …, m; wherein the computing resource includes at least one of CPU resource, memory resource, storage resource and network resource.

[0040] Step 3: Statistic the computing resource requirement of the computing task, and construct the demand matrix D by using the obtained computing resource requirement t wherein, represents the nth item of computing resource requirement of the jth computing task to be processed at the current time t; j = 1, 2, …, J;

[0041] It should be noted that the power grid computing and analysis task is an important part of power system operation, planning and control, and there are various types of tasks, each of which has different requirements for computing resources. In this embodiment, the power grid computing and analysis task and its specific requirements for computing resources are mentioned:

[0042] 1) Power flow calculation

[0043] Task description: Power flow calculation is the most basic analysis task in power grid, which is used to determine the steady-state operating state of power grid under given operating conditions, including node voltage, branch power, etc.

[0044] Computing resource requirement:

[0045] CPU resource: moderate computing intensity, usually requires multi-core processor to support fast calculation of large-scale power grid.

[0046] Memory resource: moderate memory requirement, depends on the size of power grid (number of nodes). Large-scale power grid may require higher memory capacity.

[0047] Storage resource: low storage requirement, mainly used for storing power grid topology and operating parameters.

[0048] Network resource: low network requirement, mainly used for data input and result output.

[0049] 2) Short circuit calculation

[0050] Task description: Short circuit calculation is used to evaluate the current level of power grid when short circuit fault occurs, to ensure the correct action of protection devices.

[0051] Computing resource requirement:

[0052] CPU resource: moderate computing intensity, needs to quickly process complex electrical equations.

[0053] Memory resources: Moderate memory requirements, proportional to the grid size.

[0054] Storage resources: Low storage requirements, mainly for storing fault scenarios and related parameters.

[0055] Network resources: Low network requirements, mainly for data exchange.

[0056] 3) Transient stability analysis

[0057] Task description: Transient stability analysis is used to assess the dynamic behavior of the power grid after large disturbances (such as faults, generator trips) to ensure that the system can recover to a stable state.

[0058] Computational resource requirements:

[0059] CPU resources: High computational intensity, requiring high-performance processors or clusters, as transient analysis involves dynamic simulation of time steps.

[0060] Memory resources: High memory requirements, especially for transient simulation of large-scale power grids.

[0061] Storage resources: Moderate storage requirements for storing intermediate results and historical data during simulation.

[0062] Network resources: Moderate network requirements for distributed computing and data synchronization.

[0063] 4) Small signal stability analysis

[0064] Task description: Small signal stability analysis is used to assess the dynamic behavior of the power grid under small disturbances, usually through eigenvalue analysis to determine the stability of the system.

[0065] Computational resource requirements:

[0066] CPU resources: High computational intensity, requiring high-performance computing resources to handle eigenvalue problems.

[0067] Memory resources: High memory requirements, especially for large-scale matrix operations.

[0068] Storage resources: Low storage requirements, mainly for storing power grid parameters and eigenvalue results.

[0069] Network resources: Low network requirements.

[0070] 5) Optimal power flow calculation

[0071] Task description: Optimal power flow calculation is used to optimize the operating state of the power grid to minimize operating costs or losses while meeting various constraints.

[0072] Computational resource requirements:

[0073] CPU resources: High computational intensity, requiring high-performance processors or optimized algorithms.

[0074] Memory resources: High memory demand, especially for large-scale grid optimization problems.

[0075] Storage resources: Moderate storage demand for storing optimization models and results.

[0076] Network resources: Low network demand.

[0077] 6) Reliability assessment

[0078] Task description: Reliability assessment is used to analyze the reliability and power supply capability of the grid under different fault scenarios, often involving methods such as Monte Carlo simulation.

[0079] Computational resource requirements:

[0080] CPU resources: High computational intensity, especially for large-scale Monte Carlo simulations.

[0081] Memory resources: High memory demand for storing fault scenarios and reliability indicators.

[0082] Storage resources: High storage demand for storing large amounts of simulation results and reliability data.

[0083] Network resources: Moderate network demand for distributed computing.

[0084] 7) Distributed energy access analysis

[0085] Task description: Analyze the operational characteristics of distributed energy (such as solar, wind, and energy storage) after accessing the grid, including power fluctuations and voltage support.

[0086] Computational resource requirements:

[0087] CPU resources: Moderate computational intensity, requiring support for dynamic simulation and optimization.

[0088] Memory resources: Moderate memory demand, depending on the number of distributed energy sources and the size of the grid.

[0089] Storage resources: Moderate storage demand for storing distributed energy operation data.

[0090] Network resources: Moderate network demand for real-time data exchange.

[0091] 8) Power market analysis

[0092] Task description: Analyze supply and demand balance, price prediction, and trading strategies in the electricity market.

[0093] Computational resource requirements:

[0094] CPU resources: Medium computing intensity, need to support complex economic models and optimization algorithms.

[0095] Memory resources: Medium memory requirement, for storing market data and model parameters.

[0096] Storage resources: High storage requirement, for storing historical transaction data and market prediction results.

[0097] Network resources: High network requirement, for real-time data interaction and market information update.

[0098] 9) Edge computing tasks

[0099] Task description: Data processing and real-time analysis at the edge nodes of the power grid (such as substations, distributed energy terminals), such as fault detection, line loss analysis, etc.

[0100] Computing resource requirements:

[0101] CPU resources: Low to medium computing intensity, depending on task complexity.

[0102] Memory resources: Low to medium memory requirement.

[0103] Storage resources: Low storage requirement, mainly for local data caching.

[0104] Network resources: Low network requirement, mainly for communication with the cloud or higher nodes.

[0105] Step 4: Build a demand matching model for computing resources and computing tasks, and use the model to assign the jth computing task to be processed to the edge cluster for processing. The expression of the demand matching model is:

[0106]

[0107] where, represents the fth computing resource of the edge cluster assigned to the kth computing task at time t; represents the fth computing resource requirement of the kth assigned computing task at time t; θ kf represents the weight of the fth computing resource requirement of the kth assigned computing task; F k represents the load rate of the edge cluster assigned to the kth computing task before task assignment; F k represents the load rate of the edge cluster assigned to the kth computing task after task assignment; F max represents the load rate threshold of the edge cluster;

[0108] The demand matching model can allocate edge clusters with the most suitable computing resources to the computing tasks that need to be processed. This ensures both the rapid allocation of computing tasks and the conservation of computing resources, thereby keeping the system running smoothly and efficiently.

[0109] The load rate F of the edge cluster assigned the kth computing task before the task was assigned k The calculation formula is:

[0110]

[0111] In the formula, Indicated as taking The maximum value in the calculation results; Cr kf This represents the f-th computing resource of the edge cluster assigned the k-th computing task;

[0112] The load factor F of the edge cluster assigned the kth computing task after task allocation k The formula for calculating ' is:

[0113]

[0114] In the formula, Indicated as taking The maximum value in the calculation results.

[0115] It should be noted that the computing resource load rate is an indicator that measures the resource utilization of a computing system. It reflects the ratio between the current usage and the maximum available resources. For a cluster consisting of multiple computing nodes, which is an edge cluster in this scheme, the cluster load rate can be calculated by taking the maximum load rate of each node. The calculation formula is: Cluster load rate = max{node 1 load rate, node 2 load rate, ..., node r load rate}; where the node r load rate is the resource load rate of the r-th node, which can be the CPU, memory, storage, or network load rate.

[0116] The method for calculating load factor may differ depending on the computing resource. Below are some common methods for calculating load factor for computing resources:

[0117] CPU load

[0118] CPU load rate represents the ratio between the current CPU usage and the maximum available CPU. The formula is: CPU load rate = Current CPU usage / Maximum CPU usage; Maximum CPU usage is typically 100%, and the current CPU usage can be obtained through tools or APIs provided by the operating system.

[0119] Memory load rate

[0120] Memory load rate represents the proportion between current memory usage and total available memory. The calculation formula is: memory load rate = current memory usage / total memory capacity; where total memory capacity is the total amount of memory installed by the system, and current memory usage can be obtained through the tools or API provided by the operating system;

[0121] Storage load rate

[0122] Storage load rate represents the proportion between current storage space usage and total available storage space. The calculation formula is: storage load rate = current storage usage / total storage capacity; where current storage usage can be obtained through the tools or API provided by the operating system, and total storage capacity is the total capacity of the storage device installed by the system.

[0123] Network load rate

[0124] Network load rate represents the proportion between current network bandwidth usage and total available network bandwidth. The calculation formula is: network load rate = current network usage / total network bandwidth; where current network usage can be obtained through the tools or API provided by the operating system, and total network bandwidth is the maximum transmission rate of the network device.

[0125] In addition, for the determination of the load rate threshold F max of the edge cluster, it can be classified and determined according to the load rate of each computing resource. The load rate of the computing resource can be divided into different load levels according to its numerical range. The following is the common load level classification:

[0126] Low load, load rate in the range of 0%-30%;

[0127] Definition: Low resource utilization, indicating that the current resource has a large idle capacity.

[0128] Application scenario: Suitable for processing low-priority tasks or resource maintenance.

[0129] Medium load, load rate in the range of 30%-70%;

[0130] Definition: Moderate resource utilization, indicating that the resource is in a relatively reasonable usage state.

[0131] Application scenario: Suitable for processing routine tasks, with high resource utilization but still some expansion capacity.

[0132] High load, load rate in the range of 70%-100%;

[0133] Definition: High resource utilization, close to or reaching full load state.

[0134] Application scenario: Suitable for processing high-priority tasks, but attention should be paid to resource overload risk.

[0135] Overload, load rate > 100%;

[0136] Definition: Resource usage exceeds maximum capacity, usually indicates unreasonable resource allocation or excessive tasks.

[0137] Application scenario: need to adjust task allocation or increase resources urgently to avoid system failure. Among them, the actual usage and the maximum available amount can be defined according to specific resources (such as CPU, memory, storage, etc.). Avoid high load or overload operation to reduce the risk of system failure. Generally, it is considered that the running efficiency of computing resources is the highest when the load rate is in the middle and low range, so the load rate threshold F max is set to 80%-90%, at this time, the running efficiency of the edge cluster is best.

[0138] Step 5: When any computing task is assigned to the edge cluster, update the resource matrix and demand matrix, as shown in the implementation flowchart Figure 3 If there is no computing task assigned to the edge cluster, it may be a computing task cancellation or withdrawal at this time, and further judgment needs to be made on whether there is a computing task. If there is a computing task, repeat step 4, if not, go back to step 1, enter the next period of circulation, and so on.

[0139] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A virtual power plant scheduling method based on cloud-edge collaboration, characterized in that, include: Step 1: Obtain the edge clusters available for processing computing tasks at time t. and the computational tasks that need to be processed This represents the i-th edge cluster that can be used to process computing tasks; This represents the j-th computational task that needs to be processed. Step 2: Calculate the computing resources of the edge cluster and construct a resource matrix R using the obtained computing resources. t ,in, This represents the nth computing resource of the m-th edge cluster that can be used to process computing tasks; i = 1, 2, ..., m; Step 3: Calculate the computing resource requirements of the computing task, and construct a requirement matrix D using the obtained computing resource requirements. t ,in, This represents the nth computational resource requirement of the Jth computational task that needs to be processed; j = 1, 2, ..., J; Step 4: Construct a demand matching model between the computing resources and the computing tasks. Using this model, allocate the j-th computing task to the edge cluster for processing. The expression for the demand matching model is: In the formula, This represents the f-th computing resource of the edge cluster assigned the k-th computing task at time t; θ represents the f-th computational resource requirement of the k-th assigned computational task; kf F represents the weight of the f-th computational resource requirement of the k-th assigned computational task; k F represents the load rate of the edge cluster assigned the k-th computing task before the task was assigned; k ' represents the load rate of the edge cluster assigned the k-th computing task after task allocation; F max This represents the load rate threshold for the edge cluster.

2. The virtual power plant scheduling method based on cloud-edge collaboration according to claim 1, characterized in that, The method further includes: Step 5: When any of the computing tasks is assigned to the edge cluster, update the resource matrix and the demand matrix.

3. The virtual power plant scheduling method based on cloud-edge collaboration according to claim 1, characterized in that, The computational tasks include at least one of the following: power flow calculation, short-circuit calculation, transient stability analysis, small-disturbance stability analysis, optimal power flow calculation, reliability assessment, power market analysis, and edge computing tasks.

4. The virtual power plant scheduling method based on cloud-edge collaboration according to claim 1, characterized in that, The computing resources include at least one of the following: CPU resources, memory resources, storage resources, and network resources.

5. The virtual power plant scheduling method based on cloud-edge collaboration according to claim 1, characterized in that, The load rate F of the edge cluster assigned the kth computing task before the task was assigned k The calculation formula is: In the formula, Indicated as taking The maximum value in the calculation results; Cr kf This represents the f-th computing resource of the edge cluster assigned to the k-th computing task.

6. The virtual power plant scheduling method based on cloud-edge collaboration according to claim 1, characterized in that, The load factor F of the edge cluster assigned the kth computing task after task allocation k The formula for calculating ' is: In the formula, Indicated as taking The maximum value in the calculation results.

7. A virtual power plant dispatching system based on cloud-edge collaboration, characterized in that, The system is applied to the cloud-edge collaborative virtual power plant scheduling method according to any one of claims 1-6, comprising: The edge cluster acquisition module is used to acquire edge clusters that can be used to process computing tasks at the current time t. in, This represents the i-th edge cluster that can be used to process computing tasks; The computation task acquisition module is used to acquire the computation tasks that need to be processed at the current time t. in, This represents the j-th computational task that needs to be processed. The resource matrix construction module is used to statistically analyze the computing resources of the edge cluster and construct a resource matrix R using the acquired computing resources. t ,in, This represents the nth computing resource of the m-th edge cluster that can be used to process computing tasks; i = 1, 2, ..., m; The demand matrix construction module is used to statistically analyze the computing resource requirements of the computing task and construct a demand matrix D using the obtained computing resource requirements. t ,in, This represents the nth computational resource requirement of the Jth computational task that needs to be processed; j = 1, 2, ..., J; The demand matching module is used to construct a demand matching model between the computing resources and the computing tasks. Using this model, the j-th computing task that needs to be processed is allocated to the edge cluster for processing. The expression of the demand matching model is: In the formula, This represents the f-th computing resource of the edge cluster assigned the k-th computing task at time t; θ represents the f-th computational resource requirement of the k-th assigned computational task; kf F represents the weight of the f-th computational resource requirement of the k-th assigned computational task; k F represents the load rate of the edge cluster assigned the k-th computing task before the task was assigned; k ' represents the load rate of the edge cluster assigned the k-th computing task after task allocation; F max This indicates the load rate threshold for the edge cluster; The matrix update module is used to dynamically update the resource matrix and the demand matrix when any of the computing tasks is assigned to the edge cluster.

8. A virtual power plant dispatching system based on cloud-edge collaboration according to claim 7, characterized in that, The system further includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, is able to implement the steps of the cloud-edge collaborative virtual power plant scheduling method according to any one of claims 1-6.

Images