CN117370022A - Computing power resource allocation method and system for power grid operation - Google Patents
Computing power resource allocation method and system for power grid operation Download PDFInfo
- Publication number
- CN117370022A CN117370022A CN202311389251.9A CN202311389251A CN117370022A CN 117370022 A CN117370022 A CN 117370022A CN 202311389251 A CN202311389251 A CN 202311389251A CN 117370022 A CN117370022 A CN 117370022A
- Authority
- CN
- China
- Prior art keywords
- task
- computing
- node
- power
- data processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000013468 resource allocation Methods 0.000 title claims abstract description 25
- 238000012545 processing Methods 0.000 claims abstract description 217
- 238000004364 calculation method Methods 0.000 claims abstract description 98
- 238000011156 evaluation Methods 0.000 claims abstract description 88
- 238000004458 analytical method Methods 0.000 claims abstract description 47
- 238000005516 engineering process Methods 0.000 claims abstract description 12
- 238000005094 computer simulation Methods 0.000 claims abstract description 10
- 230000000007 visual effect Effects 0.000 claims abstract description 8
- 230000005540 biological transmission Effects 0.000 claims description 25
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 description 7
- 230000004075 alteration Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 238000011144 upstream manufacturing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The disclosure provides a computing power resource allocation method and system for power grid operation, and relates to a digital twin technology, wherein the method comprises the following steps: performing calculation performance analysis on the plurality of distributed calculation nodes, and performing performance identification according to calculation performance analysis results; performing visual simulation modeling on a plurality of distributed computing nodes with performance identifiers to generate a computing node twin model; performing task grade evaluation on the real-time processing task, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result; inputting the task processing sequence into a computing power node twin model to perform computing power node matching, and generating a computing power node matching scheme sequence; and performing task processing of power grid operation based on the power node matching scheme sequence. The method can solve the technical problem of lower power grid data processing efficiency caused by unreasonable calculation force resource allocation, and can improve the rationality and accuracy of calculation force resource allocation under the condition of limited calculation force resources.
Description
Technical Field
The present disclosure relates to digital twinning technology, and more particularly, to a method and system for computing power resource allocation for grid operation.
Background
With the continuous improvement of the power generation ratio of the new energy, the position of the new energy in the power industry is continuously strengthened. Because the new energy power generation has strong uncertainty, a series of electric power digital technical means such as sensing, prediction, control and scheduling are needed to counteract fluctuation of the new energy grid connection on the operation of the power grid, and strong calculation and analysis capability is needed.
At present, when the power grid operation computing resources are distributed, the actual operation state of the power grid and the data quantity to be processed are different under the conditions of different areas and different time periods, so that the computing resources are distributed unreasonably, the power grid operation data processing efficiency is low, and the stable operation of the power grid is influenced.
The existing power grid operation computing power resource allocation method has the following defects: and the power grid data processing efficiency is lower due to unreasonable calculation power resource allocation.
Disclosure of Invention
Therefore, in order to solve the above technical problems, the technical solution adopted in the embodiments of the present disclosure is as follows:
a computing power resource allocation method for grid operation, comprising the steps of: acquiring a task scheduling center and a plurality of distributed computing nodes of a power grid operation area; performing calculation performance analysis on a plurality of distributed calculation nodes, and performing performance identification on the plurality of distributed calculation nodes according to calculation performance analysis results; based on a digital twin technology, performing visual simulation modeling on a plurality of distributed computing nodes with performance identifications to generate a computing node twin model of a power grid operation area; acquiring a real-time processing task received by the task scheduling center, performing task grade evaluation on the real-time processing task according to a preset task evaluation rule table, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result; inputting the task processing sequence into the computing power node twin model to perform computing power node matching, and generating a computing power node matching scheme sequence; and performing task processing of power grid operation based on the power calculation node matching scheme sequence.
A computing power resource allocation system for grid operation, comprising: the distributed computing node acquisition module is used for acquiring a task scheduling center and a plurality of distributed computing nodes of a power grid operation area; the power calculation performance analysis module is used for carrying out power calculation performance analysis on a plurality of distributed power calculation nodes and carrying out performance identification on the distributed power calculation nodes according to the power calculation performance analysis result; the power calculation node twin model generation module is used for carrying out visual simulation modeling on a plurality of distributed power calculation nodes with performance identifiers based on a digital twin technology to generate a power calculation node twin model of a power grid operation area; the task grade evaluation module is used for acquiring the real-time processing task received by the task scheduling center, evaluating the task grade of the real-time processing task according to a preset task evaluation rule table, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result; the power calculation node matching scheme sequence generation module is used for inputting the task processing sequence into the power calculation node twin model to perform power calculation node matching and generate a power calculation node matching scheme sequence; and the power grid operation task processing module is used for performing power grid operation task processing based on the power calculation node matching scheme sequence.
By adopting the technical method, compared with the prior art, the technical progress of the present disclosure has the following points:
(1) The method can solve the technical problem that the power grid data processing efficiency is low due to unreasonable power resource allocation in the prior art, and can improve the rationality and accuracy of power resource allocation under the condition of limited power resources by generating a power node matching scheme sequence to perform task processing of power grid operation, thereby improving the power grid data processing efficiency and guaranteeing the stable and safe operation of the power grid.
(2) The digital twin technology is used for carrying out simulation modeling on a plurality of distributed computing nodes in the power grid operation area, so that the task processing time delay under the condition of different computing resource allocation can be accurately predicted, and the efficiency and the accuracy of computing node matching scheme generation can be improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings that are used in the description of the embodiments will be briefly described below.
Fig. 1 is a schematic flow chart of a method for distributing computing power resources for grid operation;
fig. 2 is a schematic connection diagram of performing a computing power performance analysis on a plurality of distributed computing nodes in a computing power resource allocation method for power grid operation;
fig. 3 is a schematic structural diagram of a computing power resource distribution system for grid operation.
Detailed Description
The technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure. All other embodiments, which can be made by one of ordinary skill in the art without inventive effort, based on the embodiments in this disclosure are intended to be within the scope of this disclosure.
Based on the above description, as shown in fig. 1, the present disclosure provides a computing power resource allocation method for grid operation, including:
acquiring a task scheduling center and a plurality of distributed computing nodes of a power grid operation area;
the method is used for reasonably distributing the computational power resources required in the power grid operation process by combining the digital twin technology, so that the efficiency of power grid operation data processing is improved.
Firstly, a task scheduling center and a plurality of distributed computing nodes of a power grid operation area are obtained, wherein the power grid operation area refers to a target power grid coverage area to be subjected to computing resource allocation, the task scheduling center refers to a data task which is used for executing and distributing power grid operation and needs to be processed in the power grid operation area, and the task scheduling center is mainly responsible for coordinating, controlling and monitoring the operation state of a power system, so that the stable, safe and efficient operation of the power system is ensured. The distributed computing node refers to an independent computing unit in the power grid operation area, and the distributed computing node and the task scheduling center realize data interaction in a signal transmission mode, for example: computers, servers, mobile devices, etc. may act as distributed computing nodes. And by acquiring the task scheduling center and the plurality of distributed computing nodes, support is provided for computing performance analysis of the next step of power grid operation area.
Performing calculation performance analysis on a plurality of distributed calculation nodes, and performing performance identification on the plurality of distributed calculation nodes according to calculation performance analysis results;
as shown in fig. 2, in one embodiment, further includes:
acquiring performance index parameters of a plurality of distributed computing nodes, wherein the performance index parameters comprise hardware performance parameters and bandwidth parameters;
randomly selecting a first computing node from a plurality of distributed computing nodes, and acquiring a first hardware performance parameter and a first bandwidth parameter of the first computing node;
performing data processing performance analysis on the first hardware performance parameter to obtain a first unit data processing amount of the first computing node;
performing data transmission performance analysis on the first bandwidth parameter to obtain a first unit data transmission quantity of the first computing node;
generating a first computational power performance analysis result of a first computational power node based on the first unit data throughput and the first unit data throughput, the first computational power performance analysis result including a first unit data processing delay.
Performing calculation performance analysis on a plurality of distributed calculation nodes, firstly, extracting performance index parameters of the plurality of distributed calculation nodes to obtain the performance index parameters of the plurality of distributed calculation nodes, wherein the performance index parameters comprise hardware performance parameters and bandwidth parameters, and the hardware performance parameters refer to performance indexes of the distributed calculation nodes for processing data, wherein the hardware performance parameters comprise hardware parameters such as a CPU, a memory and storage equipment; the bandwidth parameter is used to characterize a rate at which the distributed computing node transmits data, and includes an upstream bandwidth and a downstream bandwidth, where the upstream bandwidth refers to a rate at which data is transmitted from the user device to the network, and the downstream bandwidth refers to a rate at which data is transmitted from the network to the user device.
And then randomly selecting a power node from the plurality of distributed power computing nodes as a first power node, wherein the first power node refers to any power node in the plurality of distributed power computing nodes, and acquiring a first hardware performance parameter and a first bandwidth parameter of the first power node. And carrying out data processing performance analysis on the first computing power node according to the first hardware performance parameter, wherein the data processing performance analysis refers to the step of calculating the current hardware performance parameter of the first computing power node to evaluate the capability and the efficiency of the first computing power node for processing data. Obtaining a first unit data throughput of the first computing node, the first unit data throughput referring to a data processing capacity per unit time, wherein the unit time can be set by a person skilled in the art based on practical situations, for example: 0.1 seconds. Wherein the first unit data throughput is used to characterize an efficiency of the first computing node to process the data, wherein the greater the first unit data throughput, the greater the efficiency of the first computing node to process the data.
And carrying out data transmission performance analysis on the first computing node according to the first bandwidth parameter, wherein the data transmission performance analysis refers to calculating the data transmission capacity of the first computing node, the data transmission capacity comprises uplink data transmission capacity and downlink data transmission capacity, a first unit uplink data transmission capacity and a first unit downlink data transmission capacity of the first computing node are obtained, the first unit data transmission capacity is formed according to the first unit uplink data transmission capacity and the first unit downlink data transmission capacity, the unit data transmission capacity refers to the data transmission capacity in unit time, and the larger the unit data transmission capacity is, the higher the data transmission efficiency of the computing node is represented.
And calculating the unit data processing time delay of the first computing node according to the first unit data processing amount and the first unit data transmission amount, wherein the unit data processing time delay refers to the time required from the time of receiving the to-be-processed data of the task scheduling center to the time of sending the unit data between the time of receiving the to-be-processed data and the time of sending the processed data, which is used for representing the speed and the efficiency of processing the data by the first node, obtaining the first unit data processing time delay, and taking the first unit data processing time delay as a first computing performance analysis result of the first computing node.
And then sequentially carrying out calculation performance analysis on the plurality of distributed calculation nodes to obtain calculation performance analysis results of the plurality of distributed calculation nodes, carrying out calculation performance marking on the plurality of distributed calculation nodes according to the obtained calculation performance analysis results to obtain a plurality of distributed calculation nodes with calculation performance marking, and simultaneously providing data support for digital twin model building of the calculation nodes in the next step.
Based on a digital twin technology, performing visual simulation modeling on a plurality of distributed computing nodes with performance identifications to generate a computing node twin model of a power grid operation area;
digital twinning is a concept that connects a real physical object, process or system with its digitized virtual representation. Through digital twinning, a virtual model corresponding to an actual object or process can be created in a computer environment to simulate, analyze, optimize and monitor the actual situation.
Acquiring position coordinates and basic information of a plurality of distributed computing nodes, wherein the position coordinates and the basic information comprise information such as equipment types, hardware performance parameters, bandwidth parameters and the like; the position coordinates, the basic information and the performance identification of the distributed power calculation nodes are input into digital twin three-dimensional simulation software to carry out visual simulation modeling, wherein the digital twin three-dimensional simulation software comprises Ansys Twin Builder, siemens Simcenter, COMSOL Multiphysics and the like, and a person skilled in the art can select the adaptive three-dimensional simulation software according to actual conditions to generate a power calculation node twin model of a power grid operation area. By constructing the computational power node twin model based on the digital twin technology, support is provided for the next step of simulated distribution of computational power resources, and the accuracy of the simulated distribution of the computational power resources can be improved.
Acquiring a real-time processing task received by the task scheduling center, performing task grade evaluation on the real-time processing task according to a preset task evaluation rule table, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result;
in one embodiment, further comprising:
acquiring a preset evaluation index, wherein the evaluation index comprises a time index, an importance index and a risk index;
evaluating 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;
weighting calculation is carried out on the time evaluation result, the importance evaluation result and the risk evaluation result according to preset index weight, and a real-time task evaluation result is obtained;
and inputting the real-time task evaluation result into the preset task evaluation rule table to perform task grade matching, and generating a task grade evaluation result.
And acquiring a real-time processing task received by the task scheduling center, wherein the real-time processing task refers to a data task to be processed in the running process of the power grid, and comprises real-time running monitoring data of power grid voltage, a generator, a transformer substation, a power transmission line and the like.
Acquiring preset evaluation indexes, wherein the preset evaluation indexes comprise time indexes, importance indexes and risk indexes, the time indexes refer to time urgency of a real-time processing task, and the time indexes are larger as the time urgency is higher; the importance index refers to the importance degree of a real-time processing task on the current whole power grid operation, wherein the importance index is larger as the importance degree is larger; the risk index refers to a risk possibly caused when a real-time processing task is not processed, wherein the larger the risk is, the larger the risk index is.
And then, evaluating the real-time processing task according to the preset evaluation index to obtain a time evaluation result, an importance evaluation result and a risk evaluation result of the real-time processing task. The preset index weight is obtained, wherein the preset index weight refers to the weight proportion of a time index, an importance index and a risk index, the greater the influence degree of which index on the running stability of the power grid is, the greater the weight proportion of the index is, and the specific weight setting method can be obtained by carrying out weight calculation through the existing variation coefficient method, and the variation coefficient method is a common weighting method for a person skilled in the art and is not explained here.
And carrying out weighted calculation on the time evaluation result, the importance evaluation result and the risk evaluation result according to the preset index weight to obtain a weighted calculation result, and taking the weighted calculation result as a real-time task evaluation result.
Acquiring a preset task evaluation rule table, wherein the preset task evaluation rule table can be set by a person skilled in the art in a self-defined manner according to actual conditions, and the larger the task evaluation result is, the higher the task grade is, for example: setting the task evaluation result as a secondary task when the task evaluation result is between 50 and 60; and the task evaluation result is three-level tasks when the task evaluation result is between 60 and 70. And then inputting the real-time task evaluation result into the preset task evaluation rule table to perform task grade matching, and obtaining the task grade 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 sequencing.
And adding the real-time processing task into a task processing sequence according to the task grade of the real-time processing task, wherein in the task processing sequence, the higher the task grade is, the more front the task processing order is, namely, the higher the power grid operation data processing task with the higher the computing power resource processing task grade is preferentially called. By constructing the task processing sequence, tasks with higher importance degree can be processed preferentially, and the rationality of power grid operation task processing is improved.
Inputting the task processing sequence into the computing power node twin model to perform computing power node matching, and generating a computing power node matching scheme sequence;
in one embodiment, further comprising:
selecting a first task to be processed from the task processing sequence;
performing task decomposition on the first task to be processed to generate a plurality of first subtasks;
and calculating the data processing amount of the plurality of first subtasks, and identifying the plurality of first subtasks according to the data processing amount calculation result to obtain a plurality of first subtasks with the data processing amount identification.
Inputting the task processing sequence into the calculation node twin model to perform calculation node matching, and firstly, selecting a first task to be processed from the task processing sequence, wherein the first task to be processed is the task to be processed with the highest order in the task processing sequence, namely, the task with the highest priority processing level.
And then performing task decomposition on the first task to be processed, wherein the task decomposition refers to decomposing the first task to be processed into a plurality of concurrent subtasks which can be processed simultaneously according to a preset task decomposition principle, so as to obtain a plurality of first subtasks. And calculating the data processing amount of the plurality of first subtasks, wherein the data calculation means that the data amount required to be processed by the first subtasks is obtained, the data processing amount of the plurality of first subtasks is obtained, the plurality of first subtasks are identified according to the data processing amount, and the plurality of first subtasks with the data processing amount identification are obtained. By calculating the data amount of the first subtask, support is provided for the next step of computing power resource allocation.
In one embodiment, further comprising:
acquiring real-time data processing information of a plurality of distributed computing nodes in the computing node twin model;
screening the distributed computing nodes according to the real-time data processing information to obtain a plurality of residual distributed computing nodes, wherein the residual distributed computing nodes are computing nodes with residual data processing space,
performing task random matching on the plurality of first subtasks and the plurality of residual distributed computing nodes to generate a plurality of first computing node matching schemes;
and optimizing the plurality of first computing force node matching schemes, generating a first optimal computing force node matching scheme, and adding the first optimal computing force node matching scheme into a computing force node matching scheme sequence.
And acquiring real-time data processing information of a plurality of distributed computing nodes in the computing node twin model, wherein the real-time data processing information comprises computing performance analysis results of the computing nodes and a real-time data processing space. And screening and extracting the computing nodes with the residual data processing space in the plurality of distributed computing nodes according to the real-time data processing information, marking the computing nodes with the residual data processing space as residual distributed computing nodes, and obtaining a plurality of residual distributed computing nodes.
And then randomly selecting a first subtask to be processed from the plurality of first subtasks, and then distributing the randomly selected first subtask to any one of a plurality of remaining distributed computing nodes to obtain a plurality of first computing node matching schemes. And then optimizing the first computing force node matching scheme, wherein optimizing means that a matching scheme with the smallest overall data processing time delay in the first computing force node matching scheme is found, and the first optimal computing force node matching scheme is obtained according to an optimizing result.
And then sequentially carrying out calculation node matching on a plurality of tasks in the task processing sequence according to the task processing sequence to generate a plurality of calculation node matching schemes, and adding the calculation node matching schemes into the calculation node matching scheme sequence according to the task processing sequence to generate a calculation node matching scheme sequence.
The computing power node matching scheme with the minimum processing time delay of the current task can be obtained by optimizing the computing power node matching scheme, so that the processing efficiency of the current task can be improved.
In one embodiment, further comprising:
randomly selecting a first matching scheme from a plurality of first computing node matching schemes;
inputting the first matching scheme into the computing node twin model to perform computing node matching;
performing data processing time delay prediction according to the data processing amount of the first subtask in the first matching scheme and the unit data processing time delay of the matched computing node, and generating a data processing time delay prediction result of the first matching scheme;
sequentially obtaining a plurality of data processing time delay prediction results of a plurality of first computing power node matching schemes;
and taking a first computing force node matching scheme corresponding to the minimum time delay value in the data processing time delay prediction results as the first optimal computing force node matching scheme.
Randomly selecting a first matching scheme from the plurality of first computing node matching schemes, wherein the first matching scheme is any one scheme of the plurality of first computing node matching schemes, and then inputting the first matching scheme into the computing node twin model to perform computing node matching to obtain a plurality of matching computing nodes. And carrying out data processing time delay prediction on the data processing amount of the first subtask according to the unit data processing time delay of the matched computing node, wherein the data processing time delay refers to the time required by the current computing node to process the first subtask, the data processing time delay prediction results of a plurality of computing nodes are obtained, the processing time delay prediction result with the largest numerical value in the data processing time delay prediction results is used as the data processing time delay prediction result of the first matching scheme, namely, as the plurality of subtasks are processed by the plurality of computing nodes at the same time, the task processing time length required by the computing node with the longest task processing time is used as the task processing time length of the computing node matching scheme.
And then sequentially calculating the data processing time delay prediction results in the plurality of first computing power node matching schemes to obtain a plurality of data processing time delay prediction results of the plurality of first computing power node matching schemes. And taking a first computing power node matching scheme corresponding to the minimum time delay value in the data processing time delay prediction results, namely a matching scheme with the minimum data processing time length, as the first optimal computing power node matching scheme.
And performing task processing of power grid operation based on the power calculation node matching scheme sequence.
In one embodiment, further comprising:
when the task processing of the power grid operation is carried out, collecting subtask data processing time of each distributed computing node;
acquiring a preset data processing time threshold, and marking the distributed computing node as an abnormal computing node when the subtask data processing time of the distributed computing node meets the preset data processing time threshold;
at the moment, the subtask allocation of the abnormal computing power node is suspended, and the computing power node matching scheme sequence is updated according to the rest computing power nodes.
And performing task processing on the task processing sequence operated by the power grid according to the computing power node matching scheme sequence, and when the task processing sequence is updated, adjusting the computing power node matching scheme sequence at any time according to the real-time task processing sequence, thereby improving the accuracy of computing power resource allocation.
When performing task processing of power grid operation according to a computing power node matching scheme sequence, acquiring subtask data processing time of each distributed computing power node, and acquiring a preset data processing time threshold, wherein each distributed computing power node corresponds to one data processing time threshold, and the task processing time threshold can be set according to a data processing time delay prediction result of the computing power node, for example: the data processing time threshold is set to be 1.5 times of the data processing time delay prediction result.
And judging the subtask data processing time of the distributed computing node according to the data processing time threshold, and marking the corresponding distributed computing node as an abnormal computing node when the subtask data processing time of the distributed computing node is greater than the data processing time threshold. At this time, subtask allocation of the abnormal computing power node is firstly paused, and a computing power node matching scheme is regenerated according to the rest computing power nodes.
The method solves the technical problem of low power grid data processing efficiency caused by unreasonable power resource allocation in the prior art, and can improve the rationality and accuracy of power resource allocation under the condition of limited power resources, thereby improving the power grid data processing efficiency and ensuring the stable and safe operation of the power grid.
In one embodiment, as shown in fig. 3, there is provided a computing power resource distribution system for power grid operation, including a distributed computing power node acquisition module, a computing power performance analysis module, a computing power node twin model generation module, a task level evaluation module, a computing power node matching scheme sequence generation module, a power grid operation task processing module, wherein:
the distributed computing node acquisition module is used for acquiring a task scheduling center and a plurality of distributed computing nodes of a power grid operation area;
the power calculation performance analysis module is used for carrying out power calculation performance analysis on a plurality of distributed power calculation nodes and carrying out performance identification on the distributed power calculation nodes according to the power calculation performance analysis result;
the power calculation node twin model generation module is used for carrying out visual simulation modeling on a plurality of distributed power calculation nodes with performance identifiers based on a digital twin technology to generate a power calculation node twin model of a power grid operation area;
the task grade evaluation module is used for acquiring the real-time processing task received by the task scheduling center, evaluating the task grade of the real-time processing task according to a preset task evaluation rule table, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result;
the power calculation node matching scheme sequence generation module is used for inputting the task processing sequence into the power calculation node twin model to perform power calculation node matching and generate a power calculation node matching scheme sequence;
and the power grid operation task processing module is used for performing power grid operation task processing based on the power calculation node matching scheme sequence.
In one embodiment, the system further comprises:
the performance index parameter acquisition module is used for acquiring performance index parameters of the distributed computing nodes, wherein the performance index parameters comprise hardware performance parameters and bandwidth parameters;
the first parameter acquisition module is used for randomly selecting a first computing node from a plurality of distributed computing nodes and acquiring a first hardware performance parameter and a first bandwidth parameter of the first computing node;
the data processing performance analysis module is used for performing data processing performance analysis on the first hardware performance parameter to obtain a first unit data processing amount of the first computing node;
the data transmission performance analysis module is used for carrying out data transmission performance analysis on the first bandwidth parameter to obtain a first unit data transmission quantity of the first computing node;
the first power performance analysis result generation module is used for generating a first power performance analysis result of a first power node based on the first unit data processing amount and the first unit data transmission amount, and the first power performance analysis result comprises first unit data processing time delay.
In one embodiment, the system further comprises:
the system comprises a preset evaluation index acquisition module, a control module and a control module, wherein the preset evaluation index acquisition module is used for acquiring preset evaluation indexes, wherein the evaluation indexes comprise time indexes, importance indexes and risk indexes;
the real-time processing task evaluation module is used for evaluating 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;
the real-time task evaluation result obtaining module is used for carrying out weighted calculation on the time evaluation result, the importance evaluation result and the risk evaluation result according to preset index weight to obtain a real-time task evaluation result;
the task grade evaluation result generation module is used for inputting the real-time task evaluation result into the preset task evaluation rule table to perform task grade matching and generating a task grade evaluation result.
In one embodiment, the system further comprises:
the first task to be processed selecting module is used for selecting a first task to be processed from the task processing sequence;
the task decomposition module is used for decomposing the task of the first task to be processed to generate a plurality of first subtasks;
the data processing amount calculation module is used for calculating the data processing amount of the plurality of first subtasks and identifying the plurality of first subtasks according to the data processing amount calculation result to obtain a plurality of first subtasks with data processing amount identification.
In one embodiment, the system further comprises:
the real-time data processing information acquisition module is used for acquiring real-time data processing information of a plurality of distributed computing nodes in the computing node twin model;
a residual distributed computing node obtaining module, which is used for screening a plurality of distributed computing nodes according to the real-time data processing information to obtain a plurality of residual distributed computing nodes, wherein the residual distributed computing nodes refer to computing nodes with residual data processing space,
the first computing power node matching scheme generation module is used for carrying out task random matching on a plurality of first subtasks and a plurality of residual distributed computing power nodes to generate a plurality of first computing power node matching schemes;
the first optimal force node matching scheme generation module is used for optimizing a plurality of first force node matching schemes, generating a first optimal force node matching scheme and adding the first optimal force node matching scheme into a force node matching scheme sequence.
In one embodiment, the system further comprises:
the first matching scheme selection module is used for randomly selecting a first matching scheme from a plurality of first computing force node matching schemes;
the computing force node matching module is used for inputting the first matching scheme into the computing force node twin model to perform computing force node matching;
the data processing time delay prediction module is used for performing data processing time delay prediction according to the data processing amount of the first subtask in the first matching scheme and the unit data processing time delay of the matched computing node, and generating a data processing time delay prediction result of the first matching scheme;
the data processing time delay prediction result obtaining module is used for sequentially obtaining a plurality of data processing time delay prediction results of a plurality of first computing power node matching schemes;
the first optimal force node matching scheme obtaining module is used for taking a first force node matching scheme corresponding to a time delay minimum value in the data processing time delay prediction results as the first optimal force node matching scheme.
In one embodiment, the system further comprises:
the subtask data processing time acquisition module is used for acquiring the subtask data processing time of each distributed computing node when the task processing of the power grid operation is carried out;
the abnormal computing power node marking module is used for acquiring a preset data processing time threshold, and marking the distributed computing power node as an abnormal computing power node when the subtask data processing time of the distributed computing power node meets the preset data processing time threshold;
and the computing power node matching scheme sequence updating module is used for suspending subtask allocation of abnormal computing power nodes at the moment and updating the computing power node matching scheme sequence according to the rest computing power nodes.
In summary, compared with the prior art, the embodiments of the present disclosure have the following technical effects:
(1) The power grid operation task processing is performed by generating the power node matching scheme sequence, so that the rationality and the accuracy of power resource allocation can be improved under the condition of limited power resources, the power grid data processing efficiency is improved, and the stable and safe operation of the power grid is ensured.
(2) By constructing the task processing sequence, tasks with higher importance degree can be preferentially processed, the rationality of power grid operation task processing is improved, and the computing node matching scheme with the smallest time delay in the current task processing can be obtained by optimizing the computing node matching scheme, so that the efficiency of the current task processing can be improved.
(3) The digital twin technology is used for carrying out simulation modeling on a plurality of distributed computing nodes in the power grid operation area, so that the task processing time delay under the condition of different computing resource allocation can be accurately predicted, and the efficiency and the accuracy of computing node matching scheme generation can be improved.
The above examples merely represent a few embodiments of the present disclosure and are not to be construed as limiting the scope of the invention. Accordingly, various alterations, modifications and variations may be made by those having ordinary skill in the art without departing from the scope of the disclosed concept as defined by the following claims and all such alterations, modifications and variations are intended to be included within the scope of the present disclosure.
Claims (8)
1. A method of computing power resource allocation for grid operation, the method comprising:
acquiring a task scheduling center and a plurality of distributed computing nodes of a power grid operation area;
performing calculation performance analysis on a plurality of distributed calculation nodes, and performing performance identification on the plurality of distributed calculation nodes according to calculation performance analysis results;
based on a digital twin technology, performing visual simulation modeling on a plurality of distributed computing nodes with performance identifications to generate a computing node twin model of a power grid operation area;
acquiring a real-time processing task received by the task scheduling center, performing task grade evaluation on the real-time processing task according to a preset task evaluation rule table, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result;
inputting the task processing sequence into the computing power node twin model to perform computing power node matching, and generating a computing power node matching scheme sequence;
and performing task processing of power grid operation based on the power calculation node matching scheme sequence.
2. The method of claim 1, wherein the performing a computational power performance analysis on a plurality of distributed computational power nodes further comprises:
acquiring performance index parameters of a plurality of distributed computing nodes, wherein the performance index parameters comprise hardware performance parameters and bandwidth parameters;
randomly selecting a first computing node from a plurality of distributed computing nodes, and acquiring a first hardware performance parameter and a first bandwidth parameter of the first computing node;
performing data processing performance analysis on the first hardware performance parameter to obtain a first unit data processing amount of the first computing node;
performing data transmission performance analysis on the first bandwidth parameter to obtain a first unit data transmission quantity of the first computing node;
generating a first computational power performance analysis result of a first computational power node based on the first unit data throughput and the first unit data throughput, the first computational power performance analysis result including a first unit data processing delay.
3. The method of claim 1, wherein said performing task level evaluation on said real-time processing task according to a preset task evaluation rule table, further comprises:
acquiring a preset evaluation index, wherein the evaluation index comprises a time index, an importance index and a risk index;
evaluating 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;
weighting calculation is carried out on the time evaluation result, the importance evaluation result and the risk evaluation result according to preset index weight, and a real-time task evaluation result is obtained;
and inputting the real-time task evaluation result into the preset task evaluation rule table to perform task grade matching, and generating a task grade evaluation result.
4. The method of claim 1, wherein said inputting the task processing sequence into the computing node twinning model for computing node matching further comprises, before:
selecting a first task to be processed from the task processing sequence;
performing task decomposition on the first task to be processed to generate a plurality of first subtasks;
and calculating the data processing amount of the plurality of first subtasks, and identifying the plurality of first subtasks according to the data processing amount calculation result to obtain a plurality of first subtasks with the data processing amount identification.
5. The method as recited in claim 4, further comprising:
acquiring real-time data processing information of a plurality of distributed computing nodes in the computing node twin model;
screening the distributed computing nodes according to the real-time data processing information to obtain a plurality of residual distributed computing nodes, wherein the residual distributed computing nodes are computing nodes with residual data processing space,
performing task random matching on the plurality of first subtasks and the plurality of residual distributed computing nodes to generate a plurality of first computing node matching schemes;
and optimizing the plurality of first computing force node matching schemes, generating a first optimal computing force node matching scheme, and adding the first optimal computing force node matching scheme into a computing force node matching scheme sequence.
6. The method of claim 5, wherein optimizing the plurality of first force node matching schemes generates a first optimal force node matching scheme, further comprising:
randomly selecting a first matching scheme from a plurality of first computing node matching schemes;
inputting the first matching scheme into the computing node twin model to perform computing node matching;
performing data processing time delay prediction according to the data processing amount of the first subtask in the first matching scheme and the unit data processing time delay of the matched computing node, and generating a data processing time delay prediction result of the first matching scheme;
sequentially obtaining a plurality of data processing time delay prediction results of a plurality of first computing power node matching schemes;
and taking a first computing force node matching scheme corresponding to the minimum time delay value in the data processing time delay prediction results as the first optimal computing force node matching scheme.
7. The method as recited in claim 1, further comprising:
when the task processing of the power grid operation is carried out, collecting subtask data processing time of each distributed computing node;
acquiring a preset data processing time threshold, and marking the distributed computing node as an abnormal computing node when the subtask data processing time of the distributed computing node meets the preset data processing time threshold;
at the moment, the subtask allocation of the abnormal computing power node is suspended, and the computing power node matching scheme sequence is updated according to the rest computing power nodes.
8. A computing power resource allocation system for grid operation, characterized by the steps for performing any one of the methods for computing power resource allocation for grid operation as claimed in claims 1-7, the system comprising:
the distributed computing node acquisition module is used for acquiring a task scheduling center and a plurality of distributed computing nodes of a power grid operation area;
the power calculation performance analysis module is used for carrying out power calculation performance analysis on a plurality of distributed power calculation nodes and carrying out performance identification on the distributed power calculation nodes according to the power calculation performance analysis result;
the power calculation node twin model generation module is used for carrying out visual simulation modeling on a plurality of distributed power calculation nodes with performance identifiers based on a digital twin technology to generate a power calculation node twin model of a power grid operation area;
the task grade evaluation module is used for acquiring the real-time processing task received by the task scheduling center, evaluating the task grade of the real-time processing task according to a preset task evaluation rule table, and adding the real-time processing task into a task processing sequence according to a task grade evaluation result;
the power calculation node matching scheme sequence generation module is used for inputting the task processing sequence into the power calculation node twin model to perform power calculation node matching and generate a power calculation node matching scheme sequence;
and the power grid operation task processing module is used for performing power grid operation task processing based on the power calculation node matching scheme sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311389251.9A CN117370022A (en) | 2023-10-25 | 2023-10-25 | Computing power resource allocation method and system for power grid operation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311389251.9A CN117370022A (en) | 2023-10-25 | 2023-10-25 | Computing power resource allocation method and system for power grid operation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117370022A true CN117370022A (en) | 2024-01-09 |
Family
ID=89403758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311389251.9A Pending CN117370022A (en) | 2023-10-25 | 2023-10-25 | Computing power resource allocation method and system for power grid operation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117370022A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103617305A (en) * | 2013-10-22 | 2014-03-05 | 芜湖大学科技园发展有限公司 | Self-adaptive electric power simulation cloud computing platform job scheduling algorithm |
| CN114168313A (en) * | 2021-09-02 | 2022-03-11 | 国网北京市电力公司 | Computing power dispatching system |
| CN114741200A (en) * | 2022-04-27 | 2022-07-12 | 国网智能电网研究院有限公司 | Data center station-oriented computing resource allocation method and device and electronic equipment |
| CN115622862A (en) * | 2022-10-18 | 2023-01-17 | 浪潮通信信息系统有限公司 | Computing power network system based on digital twins and computing power processing method |
| CN115794407A (en) * | 2022-12-15 | 2023-03-14 | 中国电信股份有限公司 | Computing resource allocation method and device, electronic equipment and nonvolatile storage medium |
| CN116541176A (en) * | 2023-05-24 | 2023-08-04 | 中国电信股份有限公司北京研究院 | Optimization method and optimization device for computing power resource allocation, electronic equipment and medium |
| CN116614385A (en) * | 2023-05-24 | 2023-08-18 | 浪潮通信信息系统有限公司 | Service scheduling path planning method, device and equipment based on digital twin |
-
2023
- 2023-10-25 CN CN202311389251.9A patent/CN117370022A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103617305A (en) * | 2013-10-22 | 2014-03-05 | 芜湖大学科技园发展有限公司 | Self-adaptive electric power simulation cloud computing platform job scheduling algorithm |
| CN114168313A (en) * | 2021-09-02 | 2022-03-11 | 国网北京市电力公司 | Computing power dispatching system |
| CN114741200A (en) * | 2022-04-27 | 2022-07-12 | 国网智能电网研究院有限公司 | Data center station-oriented computing resource allocation method and device and electronic equipment |
| CN115622862A (en) * | 2022-10-18 | 2023-01-17 | 浪潮通信信息系统有限公司 | Computing power network system based on digital twins and computing power processing method |
| CN115794407A (en) * | 2022-12-15 | 2023-03-14 | 中国电信股份有限公司 | Computing resource allocation method and device, electronic equipment and nonvolatile storage medium |
| CN116541176A (en) * | 2023-05-24 | 2023-08-04 | 中国电信股份有限公司北京研究院 | Optimization method and optimization device for computing power resource allocation, electronic equipment and medium |
| CN116614385A (en) * | 2023-05-24 | 2023-08-18 | 浪潮通信信息系统有限公司 | Service scheduling path planning method, device and equipment based on digital twin |
Non-Patent Citations (1)
| Title |
|---|
| 贾庆民等: "确定性算力网络研究", 《通信学报》, vol. 43, no. 10, 31 October 2022 (2022-10-31), pages 55 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111045828A (en) | Distributed edge calculation method based on distribution network area terminal and related device | |
| CN104901823A (en) | Method and device for generating alarm threshold value, and method and device for monitoring service performance index | |
| CN104981782A (en) | Resource control system, control pattern generation device, control device, resource control method and program | |
| CN112261120B (en) | Cloud-side cooperative task unloading method and device for power distribution internet of things | |
| CN116361377B (en) | Load prediction system, method and medium based on industrial Internet of things service platform | |
| CN116244159B (en) | Training duration prediction method and device, multi-heterogeneous computing equipment and medium | |
| CN112819258A (en) | Bank branch to store customer quantity prediction method and device | |
| CN114491388A (en) | Data processing method, device, equipment and storage medium of power system | |
| CN115438964A (en) | Random planning method and equipment for power transmission network and storage medium | |
| CN111445065A (en) | Energy consumption optimization method and system for refrigeration group control of data center | |
| CN108132840A (en) | Resource regulating method and device in a kind of distributed system | |
| Ma et al. | Virtual machine migration techniques for optimizing energy consumption in cloud data centers | |
| CN117370022A (en) | Computing power resource allocation method and system for power grid operation | |
| CN111864728B (en) | Important equipment identification method and system for reconfigurable power distribution network | |
| Zheng et al. | A monte-carlo approach for full-ahead stochastic dag scheduling | |
| CN115186891A (en) | Transient simulation analysis decision method, device and equipment for power grid harmonic oscillation | |
| CN109446020B (en) | Dynamic evaluation method and device of cloud storage system | |
| CN116089021B (en) | Deep learning-oriented large-scale load mixed part scheduling method, device and medium | |
| CN111542078B (en) | Core network control plane elastic resource allocation method in NFV environment | |
| CN116776993B (en) | Quantum bit scheduling method, device, equipment and storage medium | |
| Niewiadomska-Szynkiewicz et al. | Application of a Java-based framework to parallel simulation of large-scale systems | |
| Scherer et al. | PRACTISE--Demonstrating a Neural Network Based Framework for Robust Prediction of Data Center Workload | |
| Al-Sinayyid et al. | Maximizing the processing rate for streaming applications in Apache Storm | |
| CN114414938B (en) | Dynamic response method and system for power distribution network faults | |
| Ma et al. | A Data Collection Technology of Virtual Reality Simulation System Based on Load Balancing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |