A computing power resource allocation method and system for power grid operation

Abstract

The present disclosure provides a computing power resource allocation method and system for power grid operation, involving digital twin technology, the method comprising: performing computing power performance analysis on multiple distributed computing power nodes, and performing performance identification according to the computing power performance analysis results; performing visual simulation modeling on multiple distributed computing power nodes with performance identifications, and generating a computing power node twin model; performing task level evaluation on real-time processing tasks, and adding real-time processing tasks to the task processing sequence according to the task level evaluation results; inputting the task processing sequence into the computing power node twin model for computing power node matching, and generating a computing power node matching solution sequence; and performing task processing for power grid operation based on the computing power node matching solution sequence. The method can solve the technical problem that the unreasonable computing power resource allocation leads to low power grid data processing efficiency, and can improve the rationality and accuracy of computing power resource allocation under the condition of limited computing power resources.

Description

Technical Field

[0001] The present disclosure relates to digital twin technology, and more specifically, to a computing resource allocation method and system for power grid operation. Background Art

[0002] As the proportion of renewable energy power generation continues to increase, the position of renewable energy in the power industry will continue to strengthen. Since renewable energy power generation has strong uncertainty, it is necessary to use a series of power digitalization technologies such as perception, prediction, control, and scheduling to offset the fluctuations caused by renewable energy grid connection to power grid operation, which requires powerful computing and analysis capabilities.

[0003] At present, the allocation of computing power resources for power grid operation does not take into account the different actual operating status of the power grid and the amount of data that needs to be processed in different regions and time periods, resulting in unreasonable allocation of computing power resources, low efficiency in power grid operation data processing, and affecting the smooth operation of the power grid.

[0004] The shortcomings of the existing method for allocating computing power resources for power grid operation are that the power grid data processing efficiency is low due to unreasonable allocation of computing power resources. Summary of the invention

[0005] Therefore, in order to solve the above technical problems, the technical solutions adopted by the embodiments of the present disclosure are as follows:

[0006] A method for allocating computing resources for power grid operation, comprising the following steps: obtaining a task dispatching center and multiple distributed computing nodes in a power grid operation area; performing computing performance analysis on the multiple distributed computing nodes, and performing performance identification on the multiple distributed computing nodes according to the computing performance analysis results; based on digital twin technology, performing visual simulation modeling on multiple distributed computing nodes with performance identifications to generate a computing node twin model of the power grid operation area; obtaining real-time processing tasks received by the task dispatching center, and performing task level evaluation on the real-time processing tasks according to a preset task evaluation rule table, and adding the real-time processing tasks to a task processing sequence according to the task level evaluation results; inputting the task processing sequence into the computing node twin model to perform computing node matching to generate a computing node matching scheme sequence; and performing task processing for power grid operation based on the computing node matching scheme sequence.

[0007] A computing power resource allocation system for power grid operation, comprising: a distributed computing power node acquisition module, the distributed computing power node acquisition module is used to acquire a task dispatch center and multiple distributed computing power nodes in a power grid operation area; a computing power performance analysis module, the computing power performance analysis module is used to perform computing power performance analysis on multiple distributed computing power nodes, and to perform performance identification on the multiple distributed computing power nodes according to the computing power performance analysis results; a computing power node twin model generation module, the computing power node twin model generation module is used to perform visual simulation modeling on multiple distributed computing power nodes with performance identifications based on digital twin technology, and generate a computing power node twin model of the power grid operation area; A task level evaluation module, the task level evaluation module is used to obtain the real-time processing tasks received by the task scheduling center, and perform task level evaluation on the real-time processing tasks according to a preset task evaluation rule table, and add the real-time processing tasks to the task processing sequence according to the task level evaluation results; a computing power node matching scheme sequence generation module, the computing power node matching scheme sequence generation module is used to input the task processing sequence into the computing power node twin model for computing power node matching, and generate a computing power node matching scheme sequence; a power grid operation task processing module, the power grid operation task processing module is used to perform power grid operation task processing based on the computing power node matching scheme sequence.

[0008] Due to the adoption of the above technical method, the present disclosure has achieved the following technical advances compared with the prior art:

[0009] (1) It can solve the technical problem in the existing technology that the unreasonable allocation of computing power resources leads to low efficiency of power grid data processing. By generating a sequence of computing power node matching solutions to process power grid operation tasks, it can improve the rationality and accuracy of computing power resource allocation when computing power resources are limited, thereby improving the efficiency of power grid data processing and ensuring the stable and safe operation of the power grid.

[0010] (2) By using digital twin technology to simulate and model multiple distributed computing nodes in the power grid operation area, the task processing delay under different computing resource allocation conditions can be accurately predicted, thereby improving the efficiency and accuracy of generating computing node matching solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for describing the embodiments are briefly introduced below.

[0012] Figure 1 A flowchart of a computing resource allocation method for power grid operation is provided for this application;

[0013] Figure 2A connection diagram for performing computing power performance analysis on multiple distributed computing power nodes in a computing power resource allocation method for power grid operation is provided for this application;

[0014] Figure 3 A structural diagram of a computing resource allocation system for power grid operation is provided for this application. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0016] Based on the above description, if Figure 1 As shown, the present disclosure provides a computing resource allocation method for power grid operation, comprising:

[0017] Obtain the task dispatch center and multiple distributed computing nodes in the power grid operation area;

[0018] The method provided in this application is used to combine digital twin technology to reasonably allocate the computing power resources required during the operation of the power grid, thereby improving the efficiency of power grid operation data processing. The method provided in this application is specifically implemented in a computing power resource allocation system for power grid operation.

[0019] First, obtain the task dispatch center and multiple distributed computing power nodes in the power grid operation area, where the power grid operation area refers to the target power grid coverage area to be allocated computing power resources, and the task dispatch center refers to the data tasks that need to be processed for power grid operation in the power grid operation area, which is mainly responsible for coordinating, controlling and monitoring the operation status of the power system to ensure the stable, safe and efficient operation of the power system. A distributed computing power node refers to an independent computing unit in the power grid operation area, which realizes data interaction with the task dispatch center through signal transmission. For example, computers, servers, mobile devices, etc. can all be used as distributed computing power nodes. By obtaining the task dispatch center and multiple distributed computing power nodes, support is provided for the next step of computing power performance analysis in the power grid operation area.

[0020] Performing a computing power performance analysis on a plurality of distributed computing power nodes, and performing a performance identification on the plurality of distributed computing power nodes according to the computing power performance analysis results;

[0021] like Figure 2 As shown, in one embodiment, it also includes:

[0022] Obtaining performance indicator parameters of multiple distributed computing power nodes, wherein the performance indicator parameters include hardware performance parameters and bandwidth parameters;

[0023] Randomly selecting a first computing power node from multiple distributed computing power nodes, and obtaining a first hardware performance parameter and a first bandwidth parameter of the first computing power node;

[0024] Performing data processing performance analysis on the first hardware performance parameter to obtain a first unit data processing amount of the first computing power node;

[0025] Performing data transmission performance analysis on the first bandwidth parameter to obtain a first unit data transmission volume of the first computing power node;

[0026] A first computing power performance analysis result of a first computing power node is generated based on the first unit data processing amount and the first unit data transmission amount, and the first computing power performance analysis result includes a first unit data processing delay.

[0027] A computing power performance analysis is performed on multiple distributed computing power nodes. First, performance indicator parameters are extracted for the multiple distributed computing power nodes to obtain performance indicator parameters of the multiple distributed computing power nodes, wherein the performance indicator parameters include hardware performance parameters and bandwidth parameters. The hardware performance parameters refer to performance indicators of distributed computing power nodes for processing data, including hardware parameters such as CPU, memory, and storage devices; the bandwidth parameters are used to characterize the rate at which distributed computing power nodes transmit data, including uplink bandwidth and downlink bandwidth. The uplink bandwidth refers to the rate at which data is sent from a user device to a network, and the downlink bandwidth refers to the rate at which data is sent from a network to a user device.

[0028] Then, a computing power node is randomly selected from multiple distributed computing power nodes as the first computing power node, wherein the first computing power node refers to any one of the multiple distributed computing power nodes, and the first hardware performance parameter and the first bandwidth parameter of the first computing power node are obtained. According to the first hardware performance parameter, the data processing performance analysis of the first computing power node is performed, wherein the data processing performance analysis refers to calculating the current hardware performance parameters of the first computing power node to evaluate the ability and efficiency of the first computing power node to process data. The first unit data processing amount of the first computing power node is obtained, and the first unit data processing amount refers to the data processing capacity per unit time, wherein the unit time can be set by technicians in this field based on actual conditions, for example: 0.1 seconds. The first unit data processing amount is used to characterize the efficiency of the first computing power node in processing data, wherein the larger the first unit data processing amount, the higher the efficiency of the first computing power node in processing data.

[0029] A data transmission performance analysis is performed on the first computing power node according to the first bandwidth parameter. The data transmission performance analysis refers to calculating the data transmission capability of the first computing power node, including the uplink data transmission capability and the downlink data transmission capability, obtaining a first unit uplink data transmission volume and a first unit downlink data transmission volume of the first computing power node, and forming a first unit data transmission volume according to the first unit uplink data transmission volume and the first unit downlink data transmission volume, wherein the unit data transmission volume refers to the data transmission capability per unit time, and the larger the unit data transmission volume, the higher the data transmission efficiency of the computing power node.

[0030] The unit data processing delay of the first computing power node is calculated based on the first unit data processing volume and the first unit data transmission volume, wherein the unit data processing delay refers to the time required for unit data from receiving the data to be processed from the task scheduling center to sending the completed processing data during the data processing process of the first computing power node, which is used to characterize the speed and efficiency of the first node in processing data, obtain the first unit data processing delay, and use the first unit data processing delay as the first computing power performance analysis result of the first computing power node.

[0031] Then, computing power performance analysis is performed on multiple distributed computing power nodes in turn to obtain computing power performance analysis results of multiple distributed computing power nodes, and computing power performance marking is performed on multiple distributed computing power nodes according to the obtained computing power performance analysis results to obtain multiple distributed computing power nodes with computing power performance markings. At the same time, it also provides data support for the next step of building a digital twin model of the computing power nodes.

[0032] Based on digital twin technology, multiple distributed computing nodes with performance identifiers are visualized and simulated to generate a twin model of computing nodes in the power grid operation area.

[0033] Digital twin technology is a concept that connects an actual physical object, process or system with its digitized virtual representation. Through digital twins, a virtual model corresponding to the actual object or process can be created in a computer environment to simulate, analyze, optimize and monitor the actual situation.

[0034] Obtain the location coordinates and basic information of multiple distributed computing nodes, including equipment type, hardware performance parameters, bandwidth parameters and other information; input the location coordinates, basic information and performance identifiers of multiple distributed computing nodes into the digital twin 3D simulation software for visual simulation modeling, where the digital twin 3D simulation software includes Ansys TwinBuilder, Siemens Simcenter, COMSOL Multiphysics, etc. Technical personnel in this field can select the appropriate 3D simulation software according to the actual situation to generate a computing node twin model of the power grid operation area. By constructing a computing node twin model based on digital twin technology, it provides support for the next step of simulating the allocation of computing resources, which can improve the accuracy of the simulation allocation of computing resources.

[0035] Acquire the real-time processing tasks received by the task scheduling center, and perform task level evaluation on the real-time processing tasks according to a preset task evaluation rule table, and add the real-time processing tasks into the task processing sequence according to the task level evaluation results;

[0036] In one embodiment, it further includes:

[0037] Obtaining preset evaluation indicators, wherein the evaluation indicators include time indicators, importance indicators, and risk indicators;

[0038] Evaluate the real-time processing task based on the preset evaluation index to obtain a time evaluation result, an importance evaluation result, and a risk evaluation result;

[0039] Performing weighted calculation on the time evaluation result, the importance evaluation result, and the risk evaluation result according to preset indicator weights to obtain a real-time task evaluation result;

[0040] The real-time task evaluation result is input into the preset task evaluation rule table for task level matching to generate a task level evaluation result.

[0041] Acquire the real-time processing tasks received by the task scheduling center, wherein the real-time processing tasks refer to data tasks that need to be processed during the operation of the power grid, including real-time operation monitoring data of power grid voltage, generators, substations, transmission lines, etc.

[0042] Obtain preset evaluation indicators, wherein the preset evaluation indicators include a time indicator, an importance indicator, and a risk indicator. The time indicator refers to the time urgency of the real-time processing task, wherein the higher the time urgency, the greater the time indicator; the importance indicator refers to the importance of the real-time processing task to the current operation of the entire power grid, wherein the greater the importance, the greater the importance indicator; the risk indicator refers to the risk that may be caused when the real-time processing task is not processed, wherein the greater the risk caused, the greater the risk indicator.

[0043] Then, the real-time processing task is evaluated according to the preset evaluation index to obtain the time evaluation result, importance evaluation result and risk evaluation result of the real-time processing task. The preset index weight is obtained, and the preset index weight refers to the weight ratio of the time index, importance index and risk index. The greater the influence of any index on the stability of power grid operation, the greater the weight ratio of the index. The specific weight setting method can be obtained by weight calculation through the existing coefficient of variation method. The coefficient of variation method is a weighting method commonly used by technicians in this field and will not be described in detail here.

[0044] The time evaluation result, the importance evaluation result, and the risk evaluation result are weighted according to the preset indicator weights to obtain a weighted calculation result, and the weighted calculation result is used as the real-time task evaluation result.

[0045] Obtain a preset task evaluation rule table, wherein the preset task evaluation rule table can be customized by a technician in this field according to actual conditions, wherein the larger the task evaluation result, the higher the task level, for example: when the task evaluation result is between 50-60, it is a secondary task; when the task evaluation result is between 60-70, it is a tertiary task. Then input the real-time task evaluation result into the preset task evaluation rule table for task level matching, and obtain the task level of the real-time task evaluation result. By determining the task level of the real-time evaluation result, support is provided for the next step of task processing sorting.

[0046] Then, according to the task level of the real-time processing task, the real-time processing task is added to the task processing sequence. In the task processing sequence, the higher the task level, the higher the task processing ranking, that is, the power grid operation data processing tasks with high computing resource processing task levels are preferentially called. By constructing a task processing sequence, tasks with higher importance can be processed preferentially, improving the rationality of power grid operation task processing.

[0047] Input the task processing sequence into the computing power node twin model to perform computing power node matching, and generate a computing power node matching solution sequence;

[0048] In one embodiment, it further includes:

[0049] Selecting a first task to be processed in the task processing sequence;

[0050] Decomposing the first task to be processed to generate a plurality of first subtasks;

[0051] The data processing amount of the multiple first subtasks is calculated, and the multiple first subtasks are labeled according to the data processing amount calculation results to obtain multiple first subtasks with data processing amount labels.

[0052] The task processing sequence is input into the computing power node twin model for computing power node matching. First, the first task to be processed is selected in the task processing sequence. The first task to be processed is the highest ranked task to be processed in the task processing sequence, that is, it has the highest priority level.

[0053] Then, the first task to be processed is decomposed, and the task decomposition refers to decomposing the first task to be processed into multiple concurrent subtasks that can be processed simultaneously according to the preset task decomposition principle, and obtaining multiple first subtasks. The data processing volume is calculated for the multiple first subtasks, and the data volume calculation refers to obtaining the amount of data that needs to be processed by the first subtask, obtaining the data processing volume of the multiple first subtasks, and identifying the multiple first subtasks according to the data processing volume, and obtaining multiple first subtasks with data processing volume identification. By calculating the data volume of the first subtask, support is provided for the next step of computing power resource allocation.

[0054] In one embodiment, it further includes:

[0055] Obtaining real-time data processing information of multiple distributed computing nodes in the computing node twin model;

[0056] The plurality of distributed computing nodes are screened according to the real-time data processing information to obtain a plurality of remaining distributed computing nodes, wherein the remaining distributed computing nodes refer to computing nodes with remaining data processing space.

[0057] Randomly matching the first subtasks with the remaining distributed computing nodes to generate a plurality of first computing node matching solutions;

[0058] Optimize the multiple first computing power node matching solutions, generate a first optimal computing power node matching solution, and add it to the computing power node matching solution sequence.

[0059] The real-time data processing information of multiple distributed computing nodes in the computing node twin model is obtained, and the real-time data processing information includes the computing performance analysis results and real-time data processing space of the computing nodes. Then, the computing nodes with remaining data processing space in the multiple distributed computing nodes are screened and extracted according to the real-time data processing information, and the computing nodes with remaining data processing space are marked as remaining distributed computing nodes, so as to obtain multiple remaining distributed computing nodes.

[0060] Then, a first subtask to be processed is randomly selected from the plurality of first subtasks, and then the randomly selected first subtask is assigned to any one of the plurality of remaining distributed computing nodes to obtain a plurality of first computing node matching solutions. Then, the first computing node matching solutions are optimized, and the optimization refers to finding a matching solution with the smallest overall data processing delay among the first computing node matching solutions, and obtaining the first optimal computing node matching solution according to the optimization result.

[0061] Then, computing power nodes are matched for multiple tasks in the task processing sequence in order, multiple computing power node matching schemes are generated, and the multiple computing power node matching schemes are added to the computing power node matching scheme sequence in the task processing order to generate a computing power node matching scheme sequence.

[0062] By optimizing the computing power node matching solution, the computing power node matching solution with the smallest processing delay for the current task can be obtained, thereby improving the efficiency of the current task processing.

[0063] In one embodiment, it further includes:

[0064] Randomly select a first matching scheme from multiple first computing power node matching schemes;

[0065] Inputting the first matching scheme into the computing power node twin model to perform computing power node matching;

[0066] Performing data processing delay prediction according to the data processing amount of the first subtask in the first matching solution and the unit data processing delay of the matched computing power node, generating a data processing delay prediction result of the first matching solution;

[0067] Sequentially obtain multiple data processing delay prediction results of multiple matching solutions of the first computing power nodes;

[0068] The first computing power node matching scheme corresponding to the minimum delay value among the multiple data processing delay prediction results is used as the first optimal computing power node matching scheme.

[0069] A first matching scheme is randomly selected from multiple first computing power node matching schemes, the first matching scheme being any one of the multiple first computing power node matching schemes, and then the first matching scheme is input into the computing power node twin model for computing power node matching to obtain multiple matching computing power nodes. And the data processing delay of the data processing amount of the first subtask is predicted according to the unit data processing delay of the matching computing power node, the data processing delay refers to the time required for the current computing power node to process the first subtask, and the data processing delay prediction results of multiple computing power nodes are obtained, and the processing delay prediction result with the largest value among the multiple data processing delay prediction results is used as the data processing delay prediction result of the first matching scheme, that is, since multiple subtasks are processed simultaneously by multiple computing power nodes, the task processing time required by the computing power node with the longest task processing time is used as the task processing time of the computing power node matching scheme.

[0070] Then, the data processing delay prediction results in the multiple first computing power node matching solutions are calculated in turn to obtain multiple data processing delay prediction results of the multiple first computing power node matching solutions. And the first computing power node matching solution corresponding to the minimum delay value among the multiple data processing delay prediction results, that is, the matching solution with the shortest data processing time, is used as the first optimal computing power node matching solution.

[0071] The task processing of power grid operation is performed based on the computing power node matching scheme sequence.

[0072] In one embodiment, it further includes:

[0073] When processing tasks for power grid operation, the subtask data processing time of each distributed computing power node is collected;

[0074] Obtain a preset data processing time threshold, and when the subtask data processing time of the distributed computing power node meets the preset data processing time threshold, mark the distributed computing power node as an abnormal computing power node;

[0075] At this time, the subtask allocation of the abnormal computing power node is suspended, and the computing power node matching scheme sequence is updated according to the remaining computing power nodes.

[0076] The task processing sequence of the power grid operation is processed according to the computing power node matching scheme sequence. When the task processing sequence is updated, the computing power node matching scheme sequence needs to be adjusted at any time according to the real-time task processing sequence, thereby improving the accuracy of computing power resource allocation.

[0077] When processing tasks for power grid operation according to the computing power node matching scheme sequence, the subtask data processing time of each distributed computing power node is obtained, and a preset data processing time threshold is obtained, wherein each distributed computing power node corresponds to a data processing time threshold, and the task processing time threshold can be set according to the data processing delay prediction result of the computing power node, for example: the data processing time threshold is set to 1.5 times the data processing delay prediction result.

[0078] Then, the subtask data processing time of the distributed computing power node is judged according to the data processing time threshold. When the subtask data processing time of the distributed computing power node is greater than the data processing time threshold, the corresponding distributed computing power node is marked as an abnormal computing power node. At this time, the subtask allocation of the abnormal computing power node is first suspended, and the computing power node matching plan is regenerated according to the remaining computing power nodes.

[0079] The above method solves the technical problem in the prior art that unreasonable allocation of computing resources leads to low efficiency of power grid data processing. When computing resources are limited, the rationality and accuracy of computing resource allocation can be improved, thereby improving the efficiency of power grid data processing and ensuring stable and safe operation of the power grid.

[0080] In one embodiment, Figure 3 The invention provides a computing resource allocation system for power grid operation, including a distributed computing node acquisition module, a computing performance analysis module, a computing node twin model generation module, a task level evaluation module, a computing node matching scheme sequence generation module, and a power grid operation task processing module, wherein:

[0081] A distributed computing power node acquisition module, which is used to acquire a task dispatch center and multiple distributed computing power nodes in a power grid operation area;

[0082] A computing power performance analysis module, which is used to perform computing power performance analysis on multiple distributed computing power nodes and to identify the performance of the multiple distributed computing power nodes according to the computing power performance analysis results;

[0083] A computing power node twin model generation module, wherein the computing power node twin model generation module is used to perform visual simulation modeling on multiple distributed computing power nodes with performance identifiers based on digital twin technology to generate a computing power node twin model of a power grid operation area;

[0084] A task level evaluation module, the task level evaluation module is used to obtain the real-time processing tasks received by the task scheduling center, and perform task level evaluation on the real-time processing tasks according to a preset task evaluation rule table, and add the real-time processing tasks into the task processing sequence according to the task level evaluation results;

[0085] A computing power node matching solution sequence generation module, wherein the computing power node matching solution sequence generation module is used to input the task processing sequence into the computing power node twin model to perform computing power node matching and generate a computing power node matching solution sequence;

[0086] A power grid operation task processing module, wherein the power grid operation task processing module is used to perform power grid operation task processing based on the computing power node matching scheme sequence.

[0087] In one embodiment, the system further comprises:

[0088] A performance indicator parameter acquisition module, which is used to obtain performance indicator parameters of multiple distributed computing power nodes, wherein the performance indicator parameters include hardware performance parameters and bandwidth parameters;

[0089] A first parameter acquisition module, wherein the first parameter acquisition module is used to randomly select a first computing power node from a plurality of distributed computing power nodes, and acquire a first hardware performance parameter and a first bandwidth parameter of the first computing power node;

[0090] a data processing performance analysis module, the data processing performance analysis module being used to perform data processing performance analysis on the first hardware performance parameter to obtain a first unit data processing amount of the first computing power node;

[0091] a data transmission performance analysis module, wherein the data transmission performance analysis module is used to perform data transmission performance analysis on the first bandwidth parameter to obtain a first unit data transmission volume of the first computing power node;

[0092] A first computing power performance analysis result generating module is used to generate a first computing power performance analysis result of a first computing power node based on the first unit data processing amount and the first unit data transmission amount, and the first computing power performance analysis result includes a first unit data processing delay.

[0093] In one embodiment, the system further comprises:

[0094] A preset evaluation index acquisition module, which is used to acquire preset evaluation indexes, wherein the evaluation indexes include time index, importance index, and risk index;

[0095] A real-time processing task evaluation module, which is used to evaluate the real-time processing task based on the preset evaluation index to obtain a time evaluation result, an importance evaluation result, and a risk evaluation result;

[0096] A real-time task evaluation result obtaining module, wherein the real-time task evaluation result obtaining module is used to perform weighted calculation on the time evaluation result, the importance evaluation result, and the risk evaluation result according to preset indicator weights to obtain a real-time task evaluation result;

[0097] The task level evaluation result generation module is used to input the real-time task evaluation result into the preset task evaluation rule table for task level matching to generate a task level evaluation result.

[0098] In one embodiment, the system further comprises:

[0099] A first pending task selection module, the first pending task selection module being used to select a first pending task in the task processing sequence;

[0100] A task decomposition module, wherein the task decomposition module is used to decompose the first task to be processed to generate a plurality of first subtasks;

[0101] A data processing volume calculation module is used to calculate the data processing volume of multiple first subtasks, and identify the multiple first subtasks according to the data processing volume calculation results to obtain multiple first subtasks with data processing volume identifications.

[0102] In one embodiment, the system further comprises:

[0103] A real-time data processing information acquisition module, which is used to obtain real-time data processing information of multiple distributed computing power nodes in the computing power node twin model;

[0104] A remaining distributed computing power node acquisition module, wherein the remaining distributed computing power node acquisition module is used to screen multiple distributed computing power nodes according to the real-time data processing information to obtain multiple remaining distributed computing power nodes, wherein the remaining distributed computing power nodes refer to computing power nodes with remaining data processing space,

[0105] A first computing power node matching solution generation module, the first computing power node matching solution generation module is used to randomly match the plurality of the first subtasks with the plurality of the remaining distributed computing power nodes to generate a plurality of first computing power node matching solutions;

[0106] The first optimal computing power node matching solution generation module is used to optimize multiple first computing power node matching solutions, generate a first optimal computing power node matching solution, and add it to the computing power node matching solution sequence.

[0107] In one embodiment, the system further comprises:

[0108] A first matching solution selection module, the first matching solution selection module is used to randomly select a first matching solution from a plurality of first computing power node matching solutions;

[0109] A computing power node matching module, wherein the computing power node matching module is used to input the first matching scheme into the computing power node twin model to perform computing power node matching;

[0110] A data processing delay prediction module, the data processing delay prediction module is used to perform data processing delay prediction according to the data processing amount of the first subtask in the first matching scheme and the unit data processing delay of the matched computing power node, and generate a data processing delay prediction result of the first matching scheme;

[0111] A data processing delay prediction result acquisition module, wherein the data processing delay prediction result acquisition module is used to sequentially obtain multiple data processing delay prediction results of multiple first computing power node matching solutions;

[0112] The first optimal computing power node matching solution obtaining module is used to use the first computing power node matching solution corresponding to the minimum delay value in the multiple data processing delay prediction results as the first optimal computing power node matching solution.

[0113] In one embodiment, the system further comprises:

[0114] A subtask data processing time acquisition module, wherein the subtask data processing time acquisition module is used to collect the subtask data processing time of each distributed computing power node when performing task processing for power grid operation;

[0115] An abnormal computing power node marking module, wherein the abnormal computing power node marking module is used to obtain a preset data processing time threshold, and when the subtask data processing time of a distributed computing power node meets the preset data processing time threshold, the distributed computing power node is marked as an abnormal computing power node;

[0116] A computing power node matching scheme sequence update module is used to suspend the subtask allocation of the abnormal computing power node at this time, and update the computing power node matching scheme sequence according to the remaining computing power nodes.

[0117] In summary, compared with the prior art, the embodiments of the present disclosure have the following technical effects:

[0118] (1) By generating a computing power node matching scheme sequence to process the tasks of power grid operation, the rationality and accuracy of computing power resource allocation can be improved when computing power resources are limited, thereby improving the efficiency of power grid data processing and ensuring the stable and safe operation of the power grid.

[0119] (2) By constructing a task processing sequence, tasks with higher importance can be processed first, thereby improving the rationality of power grid operation task processing. By optimizing the computing power node matching solution, the computing power node matching solution with the minimum processing delay for the current task can be obtained, thereby improving the efficiency of the current task processing.

[0120] (3) By using digital twin technology to simulate and model multiple distributed computing nodes in the power grid operation area, the task processing delay under different computing resource allocation conditions can be accurately predicted, thereby improving the efficiency and accuracy of generating computing node matching solutions.

[0121] The above-described embodiments only express several implementation methods of the present disclosure, but they cannot be understood as limiting the scope of the invention patent. Therefore, without departing from the scope of the present disclosure concept as defined by the attached claims, a person of ordinary skill in the art may make various types of substitutions, modifications and changes, and these substitutions, modifications and changes all belong to the protection scope of the present disclosure.

Claims

1. A computing resource allocation method for power grid operation, characterized in that: The method comprises: Obtain the task dispatch center and multiple distributed computing nodes in the power grid operation area; Performing a computing power performance analysis on a plurality of distributed computing power nodes, and performing a performance identification on the plurality of distributed computing power nodes according to the computing power performance analysis results; Based on digital twin technology, multiple distributed computing nodes with performance identifiers are visualized and simulated to generate a twin model of computing nodes in the power grid operation area. Acquire the real-time processing tasks received by the task scheduling center, and perform task level evaluation on the real-time processing tasks according to a preset task evaluation rule table, and add the real-time processing tasks into the task processing sequence according to the task level evaluation results; Input the task processing sequence into the computing power node twin model to perform computing power node matching, and generate a computing power node matching solution sequence; Performing task processing for power grid operation based on the computing power node matching scheme sequence; Wherein, the step of inputting the task processing sequence into the computing power node twin model for computing power node matching further includes: Selecting a first task to be processed in the task processing sequence; Decomposing the first task to be processed to generate a plurality of first subtasks; Calculating the data processing amount of the plurality of first subtasks, and marking the plurality of first subtasks according to the data processing amount calculation results, to obtain a plurality of first subtasks with data processing amount marks; The performing computing power performance analysis on the multiple distributed computing power nodes also includes: Obtaining performance indicator parameters of multiple distributed computing power nodes, wherein the performance indicator parameters include hardware performance parameters and bandwidth parameters; Randomly selecting a first computing power node from multiple distributed computing power nodes, and obtaining a first hardware performance parameter and a first bandwidth parameter of the first computing power node; Performing data processing performance analysis on the first hardware performance parameter to obtain a first unit data processing amount of the first computing power node; Performing data transmission performance analysis on the first bandwidth parameter to obtain a first unit data transmission volume of the first computing power node; Generate a first computing power performance analysis result of a first computing power node based on the first unit data processing amount and the first unit data transmission amount, wherein the first computing power performance analysis result includes a first unit data processing delay; Obtaining real-time data processing information of multiple distributed computing nodes in the computing node twin model; The plurality of distributed computing nodes are screened according to the real-time data processing information to obtain a plurality of remaining distributed computing nodes, wherein the remaining distributed computing nodes refer to computing nodes with remaining data processing space. Randomly matching the first subtasks with the remaining distributed computing nodes to generate a plurality of first computing node matching solutions; Optimizing the plurality of first computing power node matching solutions, generating a first optimal computing power node matching solution, and adding the first optimal computing power node matching solution to the computing power node matching solution sequence; The optimizing the plurality of first computing power node matching schemes to generate a first optimal computing power node matching scheme also includes: Randomly selecting a first matching scheme from a plurality of matching schemes for the first computing power nodes; Inputting the first matching scheme into the computing power node twin model to perform computing power node matching; Performing data processing delay prediction according to the data processing amount of the first subtask in the first matching solution and the unit data processing delay of the matched computing power node, generating a data processing delay prediction result of the first matching solution; Sequentially obtain multiple data processing delay prediction results of multiple matching solutions of the first computing power nodes; The first computing power node matching scheme corresponding to the minimum delay value among the multiple data processing delay prediction results is used as the first optimal computing power node matching scheme.

2. The method according to claim 1, characterized in that The task level evaluation of the real-time processing task is performed according to the preset task evaluation rule table, and further includes: Obtaining preset evaluation indicators, wherein the evaluation indicators include time indicators, importance indicators, and risk indicators; Evaluate the real-time processing task based on the preset evaluation index to obtain a time evaluation result, an importance evaluation result, and a risk evaluation result; Performing weighted calculation on the time evaluation result, the importance evaluation result, and the risk evaluation result according to preset indicator weights to obtain a real-time task evaluation result; The real-time task evaluation result is input into the preset task evaluation rule table for task level matching to generate a task level evaluation result.

3. The method according to claim 1, characterized in that Also includes: When processing tasks for power grid operation, the subtask data processing time of each distributed computing power node is collected; Obtain a preset data processing time threshold, and when the subtask data processing time of the distributed computing power node meets the preset data processing time threshold, mark the distributed computing power node as an abnormal computing power node; At this time, the subtask allocation of the abnormal computing power node is suspended, and the computing power node matching scheme sequence is updated according to the remaining computing power nodes.

4. A computing resource allocation system for power grid operation, characterized in that: The system is used to perform the steps of any one of the methods for allocating computing resources for power grid operation as described in claims 1-3, the system comprising: A distributed computing power node acquisition module, which is used to acquire a task dispatch center and multiple distributed computing power nodes in a power grid operation area; A computing power performance analysis module, which is used to perform computing power performance analysis on multiple distributed computing power nodes and to identify the performance of the multiple distributed computing power nodes according to the computing power performance analysis results; A computing power node twin model generation module, wherein the computing power node twin model generation module is used to perform visual simulation modeling on multiple distributed computing power nodes with performance identifiers based on digital twin technology to generate a computing power node twin model of a power grid operation area; A task level evaluation module, the task level evaluation module is used to obtain the real-time processing tasks received by the task scheduling center, and perform task level evaluation on the real-time processing tasks according to a preset task evaluation rule table, and add the real-time processing tasks into the task processing sequence according to the task level evaluation results; A computing power node matching solution sequence generation module, wherein the computing power node matching solution sequence generation module is used to input the task processing sequence into the computing power node twin model to perform computing power node matching and generate a computing power node matching solution sequence; A power grid operation task processing module, wherein the power grid operation task processing module is used to perform power grid operation task processing based on the computing power node matching scheme sequence.

Images