CN-122001786-A - Power network monitoring method and device, electronic equipment, storage medium and product

CN122001786ACN 122001786 ACN122001786 ACN 122001786ACN-122001786-A

Abstract

The embodiment of the application discloses a power network monitoring method, a device, electronic equipment, a storage medium and a product. The method comprises the steps of respectively obtaining configuration description files corresponding to at least two Sketch algorithms, generating resource consumption analysis reports corresponding to the Sketch algorithms based on the configuration description files, obtaining power network monitoring demand information, generating mathematical optimization models for describing task unloading and resource allocation based on the power network monitoring demand information and the resource consumption analysis reports, solving the mathematical optimization models, determining unloading decision schemes of the at least two Sketch algorithms, converting the unloading decision schemes into hardware configuration information, and controlling a data processing unit DPU to execute the Sketch power network monitoring tasks based on the hardware configuration information. The scheme can control the DPU to conduct adaptive efficient power network monitoring based on the unloading decision scheme.

Inventors

MEI CHUNHUA
HU JU
WEN JUNLI
YANG CHENLIANG
LIANG GUOPENG
WU WENPENG

Assignees

中国南方电网有限责任公司

Dates

Publication Date: 20260508
Application Date: 20260210

Claims (10)

1. A method of power network monitoring, the method comprising: Responding to the triggering of a power network monitoring event, respectively acquiring configuration description files corresponding to at least two Sketch algorithms, and generating a resource consumption analysis report corresponding to the Sketch algorithms based on the configuration description files; Acquiring power network monitoring demand information, and generating a mathematical optimization model for describing task unloading and resource allocation based on the power network monitoring demand information and each resource consumption analysis report; solving the mathematical optimization model, and determining an unloading decision scheme of at least two Sketch algorithms; and converting the offloading decision scheme into hardware configuration information, and controlling a Data Processing Unit (DPU) to execute a Sketch power network monitoring task based on the hardware configuration information.
2. The method of claim 1, wherein generating a resource consumption analysis report corresponding to the sktch algorithm based on the configuration profile comprises: Analyzing the configuration description file, and extracting target meta-information from the configuration description file, wherein the target meta-information comprises data structure meta-information, operation logic meta-information and data stream meta-information; converting the target meta-information into a standardized intermediate representation; Determining respective hardware deployment configurations for executing each operation in the standardized intermediate representation based on hardware configuration parameters of the data processing unit DPU, and simulating execution of the corresponding operation under each hardware deployment configuration to generate a simulation execution result; And generating a resource consumption analysis report corresponding to the Sketch algorithm based on the simulation execution result.
3. The method of claim 1, further comprising, prior to generating a mathematical optimization model describing task offloading and resource allocation based on the power network monitoring demand information and the respective resource consumption analysis reports: Acquiring hardware configuration parameters of the DPU; Generating a mathematical optimization model for describing task offloading and resource allocation based on the power network monitoring demand information and the respective resource consumption analysis reports, comprising: determining at least one decision variable and at least one constraint condition based on each of the resource consumption analysis report and the hardware configuration parameters, and generating a set of decision variables based on at least one of the decision variables, and generating a set of constraints based on at least one constraint condition; Constructing an objective function based on each resource consumption analysis report, the power network monitoring demand information and the hardware configuration parameters; A mathematical optimization model describing task offloading and resource allocation is generated based on the set of decision variables, the set of constraints, and the objective function.
4. A method according to claim 3, wherein constructing an objective function based on each of the resource consumption analysis report, the power network monitoring demand information, and the hardware configuration parameters comprises: constructing at least two optimization targets based on each resource consumption analysis report, the power network monitoring demand information and the hardware configuration parameters; And integrating at least two optimization targets based on a weighted chebyshev method to generate an objective function.
5. The method of claim 1, wherein the mathematical optimization model comprises a set of decision variables, a set of constraints, and an objective function; Solving the mathematical optimization model to determine an unloading decision scheme of at least two Sketch algorithms, wherein the unloading decision scheme comprises the following steps: analyzing a decision variable set and the constraint condition set in the mathematical optimization model based on a preset core-memory allocation strategy, and determining initial deployment schemes of at least two Sketch algorithms; Constructing a neighborhood action set, wherein the neighborhood action set comprises at least one atomic operation for changing a deployment scheme; And traversing each atomic operation in the neighborhood action set, updating the initial deployment scheme by using the current atomic operation, generating a new deployment scheme, determining the function value of a mathematical function in the mathematical optimization model under the new deployment scheme until the function value meets a preset condition, and generating an unloading decision scheme of at least two sktch algorithms.
6. The method according to claim 1, characterized in that in controlling the data processing unit DPU to perform a Sketch power network monitoring task based on the hardware configuration information, further comprising: acquiring running state information of the Sketch power network monitoring task in real time; If the running state information does not meet the preset running condition, updating the mathematical optimization model based on the running state information, and returning to execute the process of solving the mathematical optimization model, and determining an unloading decision scheme of at least two Sketch algorithms until the running state information meets the preset running condition.
7. An electrical power network monitoring device, comprising: the resource consumption analysis report generation module is used for responding to the triggering of the power network monitoring event, respectively acquiring configuration description files corresponding to at least two Sketch algorithms and generating a resource consumption analysis report corresponding to the Sketch algorithms based on the configuration description files; the mathematical optimization model generation module is used for acquiring the power network monitoring demand information and generating a mathematical optimization model for describing task unloading and resource allocation based on the power network monitoring demand information and each resource consumption analysis report; The unloading decision scheme determining module is used for solving the mathematical optimization model and determining unloading decision schemes of at least two Sketch algorithms; And the power network monitoring task execution module is used for converting the unloading decision scheme into hardware configuration information and controlling the data processing unit DPU to execute the Sketch power network monitoring task based on the hardware configuration information.
8. An electronic device, the electronic device comprising: At least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the power network monitoring method of any one of claims 1-6.
9. A computer readable storage medium, characterized in that the computer readable storage medium stores computer instructions for causing a processor to implement the power network monitoring method of any one of claims 1-6 when executed.
10. A computer program product, characterized in that the computer program product comprises a computer program which, when executed by a processor, implements the power network monitoring method according to any one of claims 1-6.

Description

Power network monitoring method and device, electronic equipment, storage medium and product Technical Field The present application relates to the field of network measurement and monitoring technologies, and in particular, to a method and apparatus for monitoring an electric power network, an electronic device, a storage medium, and a product. Background With the development of smart power grids and energy Internet, the services borne by the power network are increasingly complex and various, including protection control, metering acquisition, video inspection and the like, and the services have extremely high requirements on real-time sensing of network states and accurate identification of abnormal flow. Under the background, the flow measurement technology based on the Sketch becomes a key technology for realizing high-precision network monitoring due to the characteristics of small occupied memory and high inquiry speed. The sktch can efficiently estimate statistics such as flow size and frequency in large-scale traffic through hash mapping and counter arrays. In recent years, the technical development in the field has two major trends, namely, the heterogeneous evolution of the Sketch algorithm, the occurrence of a plurality of variants such as Count-Min, count Sketch, hyperLogLog and the like to adapt to different measurement targets, but the difficulty of compatibility and unified analysis in deployment is caused, and the unloading of a measurement platform to special hardware (such as a data processing unit (Data Processing Unit, DPU) intelligent network card) to release CPU resources and promote the processing speed. DPUs have multi-core, high bandwidth, and programmability, but also have strict resource constraints (e.g., on-chip memory, limited computational units). The current focus of research is on how to efficiently and synergistically deploy diverse sktech algorithms on resource-constrained DPUs, thereby enabling low-overhead, high-precision real-time network measurements. However, the prior art mostly optimizes a single sktech in isolation or performs a simple static resource allocation, lacking a systematic solution that enables a globally optimal adaptation between algorithm heterogeneity and hardware constraints. Disclosure of Invention The embodiment of the application provides a power network monitoring method, a device, electronic equipment, a storage medium and a product, which can control a DPU to carry out adaptive high-efficiency power network monitoring based on an unloading decision scheme. According to an aspect of the present application, there is provided a power network monitoring method comprising: Responding to the triggering of a power network monitoring event, respectively acquiring configuration description files corresponding to at least two Sketch algorithms, and generating a resource consumption analysis report corresponding to the Sketch algorithms based on the configuration description files; Acquiring power network monitoring demand information, and generating a mathematical optimization model for describing task unloading and resource allocation based on the power network monitoring demand information and each resource consumption analysis report; solving the mathematical optimization model, and determining an unloading decision scheme of at least two Sketch algorithms; and converting the offloading decision scheme into hardware configuration information, and controlling a Data Processing Unit (DPU) to execute a Sketch power network monitoring task based on the hardware configuration information. According to an aspect of the present application, there is provided an electrical power network monitoring apparatus, the apparatus comprising: the resource consumption analysis report generation module is used for responding to the triggering of the power network monitoring event, respectively acquiring configuration description files corresponding to at least two Sketch algorithms and generating a resource consumption analysis report corresponding to the Sketch algorithms based on the configuration description files; the mathematical optimization model generation module is used for acquiring the power network monitoring demand information and generating a mathematical optimization model for describing task unloading and resource allocation based on the power network monitoring demand information and each resource consumption analysis report; The unloading decision scheme determining module is used for solving the mathematical optimization model and determining unloading decision schemes of at least two Sketch algorithms; And the power network monitoring task execution module is used for converting the unloading decision scheme into hardware configuration information and controlling the data processing unit DPU to execute the Sketch power network monitoring task based on the hardware configuration information. According to another aspect of the present application, t