+

CN114936808A - A cloud-side collaborative task management system and method for substation fault detection - Google Patents

A cloud-side collaborative task management system and method for substation fault detection Download PDF

Info

Publication number
CN114936808A
CN114936808A CN202210855759.2A CN202210855759A CN114936808A CN 114936808 A CN114936808 A CN 114936808A CN 202210855759 A CN202210855759 A CN 202210855759A CN 114936808 A CN114936808 A CN 114936808A
Authority
CN
China
Prior art keywords
data
cloud
module
task
fault
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
Application number
CN202210855759.2A
Other languages
Chinese (zh)
Inventor
陈页
杨穷千
黄丹丹
刘萌萌
李振廷
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Lab
Original Assignee
Zhejiang Lab
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang Lab filed Critical Zhejiang Lab
Priority to CN202210855759.2A priority Critical patent/CN114936808A/en
Publication of CN114936808A publication Critical patent/CN114936808A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
    • H02J13/00002—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by monitoring
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
    • H02J13/00006—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by information or instructions transport means between the monitoring, controlling or managing units and monitored, controlled or operated power network element or electrical equipment
    • H02J13/00016—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by information or instructions transport means between the monitoring, controlling or managing units and monitored, controlled or operated power network element or electrical equipment using a wired telecommunication network or a data transmission bus
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
    • H02J13/00006—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by information or instructions transport means between the monitoring, controlling or managing units and monitored, controlled or operated power network element or electrical equipment
    • H02J13/00022—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by information or instructions transport means between the monitoring, controlling or managing units and monitored, controlled or operated power network element or electrical equipment using wireless data transmission
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • Power Engineering (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Development Economics (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Game Theory and Decision Science (AREA)
  • Health & Medical Sciences (AREA)
  • Educational Administration (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Remote Monitoring And Control Of Power-Distribution Networks (AREA)

Abstract

本发明公开了一种面向变电站故障检测的云边协同任务管理系统及方法，涉及电力系统故障检测的技术领域，包含依次连接的数据采集模块、数据解析模块、负载预测模块、任务调度模块、通信模块，所述的任务调度模块还连接有资源监控模块、故障检测模块、故障警报模块；本发明采用云边协同任务管理的方式，降低网络传输时延，利于紧急和时间敏感的任务顺利完成；采用负载预测方式，预测下一时刻系统负载情况，有利于提高资源利用效率；该发明结构简单、操作容易、便于推广。

Figure 202210855759

The invention discloses a cloud-side collaborative task management system and method for substation fault detection, relates to the technical field of power system fault detection, and includes a data acquisition module, a data analysis module, a load prediction module, a task scheduling module, a communication The task scheduling module is also connected with a resource monitoring module, a fault detection module, and a fault alarm module; the present invention adopts a cloud-side collaborative task management method to reduce network transmission delay and facilitate the smooth completion of urgent and time-sensitive tasks; The load prediction method is adopted to predict the system load situation at the next moment, which is beneficial to improve the resource utilization efficiency; the invention has simple structure, easy operation and convenient promotion.

Figure 202210855759

Description

一种面向变电站故障检测的云边协同任务管理系统及方法A cloud-side collaborative task management system and method for substation fault detection

技术领域technical field

本发明涉及电力系统故障检测的技术领域，具体涉及一种面向变电站故障检测的云边协同任务管理系统及方法。The invention relates to the technical field of power system fault detection, in particular to a cloud-edge collaborative task management system and method for substation fault detection.

背景技术Background technique

新型电力系统背景下，变电站内的智能监测设备众多，数据呈现形式也不尽相同，有温度、湿度传感器等采集的数字编码数据，有巡检摄像采集的图片数据，也有声音采集设备采集的声纹数据，形成了多源异构的变电站故障检测数据集。Under the background of the new power system, there are many intelligent monitoring devices in the substation, and the data presentation forms are also different. There are digitally encoded data collected by temperature and humidity sensors, image data collected by inspection cameras, and sound collected by sound collection equipment. The multi-source and heterogeneous substation fault detection data set is formed.

如何对海量数据进行处理与分析，实现数据价值，为电力系统的稳定运行提供数据支撑值得关注。分析各类数据需要的计算资源不同，变电站对不同类型故障切除的时间要求和紧急程度也有所不同，传统的将所有数据都上传到云端进行分析和处理，再将结果下发的方式可能由于通信拥塞等原因导致紧急故障无法及时切除，给变电站带来安全隐患。How to process and analyze massive data, realize the value of data, and provide data support for the stable operation of the power system deserves attention. Different computing resources are required to analyze various types of data, and substations have different time requirements and urgency for different types of fault removal. The traditional method of uploading all data to the cloud for analysis and processing, and then sending the results may be due to communication. Due to congestion and other reasons, emergency faults cannot be removed in time, which brings security risks to the substation.

发明内容SUMMARY OF THE INVENTION

针对现有技术的不足，本发明提供了一种面向变电站故障检测的云边协同任务管理系统及方法，解决了上述背景技术中提出的问题。In view of the deficiencies of the prior art, the present invention provides a cloud-side collaborative task management system and method for substation fault detection, which solves the problems raised in the above-mentioned background art.

为实现以上目的，本发明通过以下技术方案予以实现：一种面向变电站故障检测的云边协同任务管理系统，包含依次连接的数据采集模块、数据解析模块、负载预测模块、任务调度模块、通信模块，所述通信模块还连接所述数据采集模块，所述的任务调度模块还依次连接有资源监控模块、故障检测模块、故障警报模块；In order to achieve the above objects, the present invention is achieved through the following technical solutions: a cloud-side collaborative task management system for substation fault detection, comprising a data acquisition module, a data analysis module, a load prediction module, a task scheduling module, and a communication module connected in sequence , the communication module is also connected to the data acquisition module, and the task scheduling module is further connected to a resource monitoring module, a fault detection module, and a fault alarm module in sequence;

所述数据采集模块收集变电站的数据采集设备产生的数据；The data acquisition module collects data generated by the data acquisition equipment of the substation;

所述数据解析模块对采集到的数据做初步分类处理；The data analysis module performs preliminary classification processing on the collected data;

所述负载预测模块采用基于相似日的负载预测方法，以变电站所处海拔、变电站电压等级、日均气温、湿度和天气类型作为相似日评判依据，根据关联度大小，筛选出训练集和测试集；采用长短期记忆网络算法，结合变电站内电气设备的使用时间及故障率，以数据采集模块所采集数据规模作为调整参考，预测下一时刻的负载情况；The load prediction module adopts a load prediction method based on similar days, and uses the altitude of the substation, the voltage level of the substation, the daily average temperature, humidity and weather type as the evaluation basis for similar days, and filters out the training set and the test set according to the degree of correlation. ;Using long-term and short-term memory network algorithm, combined with the use time and failure rate of electrical equipment in the substation, and using the scale of data collected by the data acquisition module as an adjustment reference to predict the load situation at the next moment;

所述任务调度模块将故障检测任务按照所需资源和任务对时间的紧迫程度，采用群体智能算法分配给云端或边缘侧的计算节点；The task scheduling module uses a swarm intelligence algorithm to allocate the fault detection task to the computing nodes on the cloud or edge side according to the required resources and the urgency of the task to time;

所述资源监控模块包括资源监控器和监控报警器，所述资源监控器监控系统中云端和边缘侧计算节点的资源状态；当计算节点出现异常而无法执行当前任务时，所述监控报警器发出警报，并由任务调度模块将当前无法执行的任务重新分配给其他节点；The resource monitoring module includes a resource monitor and a monitoring alarm, the resource monitor monitors the resource status of the cloud and edge computing nodes in the system; when the computing node is abnormal and cannot perform the current task, the monitoring alarm sends out Alerts, and the task scheduling module reassigns tasks that cannot currently be executed to other nodes;

所述故障检测模块利用计算节点资源，对数据进行进一步解析，得出具体故障内容及故障发生点，分析可能的故障原因及故障处理措施，供运维人员参考；The fault detection module utilizes computing node resources to further analyze the data, obtain specific fault contents and fault occurrence points, analyze possible fault causes and fault handling measures, and provide reference for operation and maintenance personnel;

所述故障警报模块为故障发生后向运维人员发出警报，并将故障分析数据发送至云端备份，作为事后故障分析和下一时刻事前预测的历史数据；The fault alarm module sends an alarm to the operation and maintenance personnel after the fault occurs, and sends the fault analysis data to the cloud for backup, as the historical data of the fault analysis after the event and the prediction in advance at the next moment;

所述通信模块包括电力有线专网和5G电力虚拟专网，所述电力有线专网用于有线数据采集设备的数据传输、边缘计算节点之间的有线传输；所述5G电力虚拟专网用于移动数据采集设备的数据传输、边缘侧与云端计算节点之间及系统间各模块之间的无线传输，同时作为电力有线专网故障时的备用通信方式。The communication module includes a power wired private network and a 5G power virtual private network, and the power wired private network is used for data transmission of wired data acquisition equipment and wired transmission between edge computing nodes; the 5G power virtual private network is used for Data transmission of mobile data acquisition equipment, wireless transmission between edge side and cloud computing nodes, and between modules in the system, as a backup communication method when the power wired private network fails.

本发明还提供了一种面向变电站故障检测的云边协同任务管理方法，包括以下步骤：The present invention also provides a cloud-side collaborative task management method for substation fault detection, comprising the following steps:

S1：采集变电站内电气设备的数据，包括变压器、互感器、断路器、隔离开关和其他设备的数字编码数据、图像数据和声纹数据；S1: Collect data of electrical equipment in substations, including digitally encoded data, image data and voiceprint data of transformers, transformers, circuit breakers, disconnectors and other equipment;

S2：初步解析所收集的数据，按来源进行分类；S2: Preliminary analysis of the collected data, classified by source;

S3：采用基于相似日的负载预测方法，预测下一时刻的负载情况；S3: Use the load prediction method based on similar days to predict the load situation at the next moment;

S4：采用基于群体智能的任务管理方法，将故障检测任务分配到云端或边缘侧的适当计算节点；S4: Adopt the task management method based on swarm intelligence to assign fault detection tasks to appropriate computing nodes on the cloud or edge side;

S5：监测任务执行节点的资源使用情况，判断是否出现计算节点异常，若出现计算节点异常，则重新进行该任务调度；S5: Monitor the resource usage of the task execution node, determine whether there is an abnormality in the computing node, and re-schedule the task if there is an abnormality in the computing node;

S6：对节点所分配的任务进行进一步解析，判断是否存在故障，若存在，则分析具体故障内容及故障发生点；若不存在，则结束任务；S6: further analyze the task assigned by the node to determine whether there is a fault, if so, analyze the specific fault content and the point of failure; if not, end the task;

S7：根据故障检测结果判断是否需要发出警报，若需要，则立刻警报通知运维人员；若不需要，则结束任务；S7: Determine whether an alarm needs to be issued according to the fault detection result, if necessary, immediately notify the operation and maintenance personnel; if not, end the task;

S8：将故障信息上传至云端保存。S8: Upload the fault information to the cloud for storage.

作为优选，所述步骤S3包含以下子步骤：Preferably, the step S3 includes the following sub-steps:

S3.1：输入历史样本和待预测日的影响因素数据，所述影响因素数据包括变电站所处海拔、变电站电压等级、变电站内电气设备的使用时间、变电站内电气设备的综合故障率、日均气温、湿度和天气类型；S3.1: Input historical samples and influencing factor data of the day to be predicted. The influencing factor data include the altitude of the substation, the voltage level of the substation, the use time of the electrical equipment in the substation, the comprehensive failure rate of the electrical equipment in the substation, the daily average Air temperature, humidity and weather type;

S3.2：选取负载预测所需的数据集，包括训练集和测试集；S3.2: Select the data set required for load prediction, including training set and test set;

S3.3：采用长短期记忆网络算法进行时间序列预测，预测下一时刻的负载情况，包括下一时刻所需计算、存储和网络资源。S3.3: Use the long short-term memory network algorithm for time series prediction, and predict the load situation at the next moment, including the computing, storage and network resources required for the next moment.

作为优选，所述步骤S3.2中选取负载预测所需的数据集，包括以下步骤：Preferably, the data set required for load prediction is selected in the step S3.2, including the following steps:

S3.2.1：构造相似日影响因素矩阵，包括子序列和母序列；S3.2.1: Construct a matrix of similar daily influencing factors, including subsequences and parent sequences;

S3.2.2：对子序列和母序列做初值化处理；S3.2.2: Initialize the subsequence and the parent sequence;

S3.2.3：选取测试集，以待预测日的影响因素数据作为母序列，历史样本中随机选取指定比例的样本的影响因素数据作为子序列，计算各子序列与母序列间的关联度，选取关联度从高到低排序前

Figure 100002_DEST_PATH_IMAGE001
位的样本作为测试集，并选取关联度最高的样本作为训练标签日；S3.2.3: Select the test set, take the influencing factor data of the day to be forecasted as the parent sequence, and randomly select the influencing factor data of a specified proportion of samples from the historical sample as the subsequence, calculate the correlation between each subsequence and the parent sequence, and select Relevance before sorting from high to low
Figure 100002_DEST_PATH_IMAGE001
The most relevant samples are selected as the test set, and the samples with the highest correlation are selected as the training label days;

S3.2.4：选取训练集，以训练标签日的影响因素数据作为母序列，历史样本中剩余的样本的影响因素数据作为子序列，计算各子序列与母序列间的关联度，选取关联度从高到低排序前

Figure 100002_DEST_PATH_IMAGE003
位的样本作为训练集。S3.2.4: Select the training set, take the influence factor data of the training label day as the parent sequence, and the influence factor data of the remaining samples in the historical sample as the subsequence, calculate the correlation degree between each subsequence and the parent sequence, and select the correlation degree from Before sorting high to low
Figure 100002_DEST_PATH_IMAGE003
bit samples as the training set.

作为优选，所述步骤S3.2.3中，历史样本中随机选取40% 的样本的影响因素数据作为子序列。Preferably, in the step S3.2.3, the influencing factor data of 40% of the historical samples are randomly selected as subsequences.

作为优选，所述步骤S4中的群体智能的任务管理方法，包含以下步骤：Preferably, the swarm intelligence task management method in the step S4 includes the following steps:

S4.1：输入步骤S2中经初步解析的数据和步骤S3中的下一时刻负载预测情况；S4.1: Input the data preliminarily parsed in step S2 and the load prediction situation at the next moment in step S3;

S4.2：计算云边协同任务评价指标；S4.2: Computing cloud-side collaborative task evaluation indicators;

S4.3：比较任务部署至云端计算节点的评价结果和部署至边缘侧计算节点的评价结果的大小，若将任务部署至云端的评价结果更小则将任务部署至云端，反之则部署至边缘侧；S4.3: Compare the evaluation result of the task deployed to the cloud computing node and the evaluation result deployed to the edge computing node. If the evaluation result of deploying the task to the cloud is smaller, deploy the task to the cloud, otherwise deploy it to the edge side;

S4.4：根据步骤S4.3的任务部署结果，采用群体智能算法将故障检测任务进一步分配到云端或边缘侧的适当计算节点。S4.4: According to the task deployment result in step S4.3, the swarm intelligence algorithm is used to further distribute the fault detection task to the appropriate computing nodes on the cloud or edge side.

作为优选，所述步骤S4.4的群体智能算法为蚁群优化算法，其包含以下步骤：Preferably, the swarm intelligence algorithm in step S4.4 is an ant colony optimization algorithm, which includes the following steps:

S4.4.1：初始化参数，将

Figure 100002_DEST_PATH_IMAGE004
只蚂蚁随机分配到
Figure 100002_DEST_PATH_IMAGE006
个计算节点，将出发节点加入第
Figure 100002_DEST_PATH_IMAGE007
只蚂蚁的禁忌表
Figure 100002_DEST_PATH_IMAGE008
中，设置迭代次数
Figure 100002_DEST_PATH_IMAGE009
；S4.4.1: Initialization parameters, the
Figure 100002_DEST_PATH_IMAGE004
ants are randomly assigned to
Figure 100002_DEST_PATH_IMAGE006
computing nodes, add the departure node to the
Figure 100002_DEST_PATH_IMAGE007
Taboo list of ants
Figure 100002_DEST_PATH_IMAGE008
, set the number of iterations
Figure 100002_DEST_PATH_IMAGE009
;

S4.4.2：计算在

Figure 100002_DEST_PATH_IMAGE011
时刻从节点
Figure 100002_DEST_PATH_IMAGE013
到节点
Figure DEST_PATH_IMAGE015
的概率
Figure DEST_PATH_IMAGE016
，选择新的计算节点，并将新的节点增加到禁忌表
Figure 807354DEST_PATH_IMAGE008
中；S4.4.2: Calculated at
Figure 100002_DEST_PATH_IMAGE011
time slave node
Figure 100002_DEST_PATH_IMAGE013
to node
Figure DEST_PATH_IMAGE015
The probability
Figure DEST_PATH_IMAGE016
, select a new computing node, and add the new node to the tabu list
Figure 807354DEST_PATH_IMAGE008
middle;

S4.4.3：更新节点

Figure DEST_PATH_IMAGE017
之间的残留信息；S4.4.3: Update Node
Figure DEST_PATH_IMAGE017
residual information between;

S4.4.4：以

Figure 442604DEST_PATH_IMAGE004
只蚂蚁中网络拓扑距离之和最小为目标选择最优计算节点；S4.4.4: with
Figure 442604DEST_PATH_IMAGE004
Only select the optimal computing node as the goal with the smallest sum of network topology distances among the ants;

S4.4.5：更新最优网络拓扑路径上的残留信息；S4.4.5: Update the residual information on the optimal network topology path;

S4.4.6：判断是否达到所设置的迭代次数，若未达到则重复该过程，反之得到最优任务分配方案。S4.4.6: Determine whether the set number of iterations is reached, if not, repeat the process, otherwise obtain the optimal task allocation scheme.

本发明采用云边协同的任务管理方式可以对不同类型的数据进行分层处理，很大程度上避免了由于通信拥塞等原因导致紧急故障无法及时切除，给变电站带来安全隐患情况。云计算擅长非实时、长周期、对计算能力要求特别高的数据处理与分析，边缘计算适用于实时、短周期的数据处理与分析，能较好地支撑本地业务的实时智能化决策与执行。将数据分析与处理任务按特征分配到云端和边缘侧节点执行，有利于合理利用资源，高效地对数据进行生命周期管理与价值挖掘。The present invention adopts the task management mode of cloud-side coordination, which can perform hierarchical processing on different types of data, and largely avoids the situation that emergency faults cannot be removed in time due to communication congestion and other reasons, which brings security risks to the substation. Cloud computing is good at data processing and analysis that is non-real-time, long-term, and requires particularly high computing power. Edge computing is suitable for real-time, short-cycle data processing and analysis, and can better support real-time intelligent decision-making and execution of local businesses. Allocating data analysis and processing tasks to the cloud and edge nodes according to their characteristics is conducive to rational utilization of resources and efficient life cycle management and value mining of data.

本发明采用云边协同任务管理的方式，降低了网络传输时延，有利于紧急和时间敏感的任务顺利完成；采用负载预测方式，预测下一时刻系统负载情况，利于提高资源利用效率；该发明结构简单、操作容易、便于推广。The invention adopts the cloud-side collaborative task management method to reduce the network transmission delay, which is conducive to the smooth completion of urgent and time-sensitive tasks; the method of load prediction is used to predict the system load situation at the next moment, which is beneficial to improve the efficiency of resource utilization; the invention The structure is simple, the operation is easy, and the promotion is convenient.

附图说明Description of drawings

图1为本发明的负荷聚类方法的流程图；Fig. 1 is the flow chart of the load clustering method of the present invention;

图2为本发明的负荷数据预处理的流程图；Fig. 2 is the flow chart of the load data preprocessing of the present invention;

图3为本发明的奇异值分解法对负荷数据降维的流程图；Fig. 3 is the flow chart of the dimensionality reduction of load data by singular value decomposition method of the present invention;

图4为本发明的考虑密度的改进K-means算法的流程图。FIG. 4 is a flow chart of the improved K-means algorithm considering the density of the present invention.

具体实施方式Detailed ways

下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments.

如图1所示，本发明的一种面向变电站故障检测的云边协同任务管理系统，包含依次连接的数据采集模块、数据解析模块、负载预测模块、任务调度模块、通信模块，所述的任务调度模块还连接有资源监控模块、故障检测模块、故障警报模块；As shown in Figure 1, a cloud-side collaborative task management system for substation fault detection of the present invention includes a data acquisition module, a data analysis module, a load prediction module, a task scheduling module, and a communication module that are connected in sequence. The scheduling module is also connected with a resource monitoring module, a fault detection module, and a fault alarm module;

所述数据采集模块收集变电站的数据采集设备产生的数据；The data acquisition module collects data generated by the data acquisition equipment of the substation;

所述数据解析模块对采集到的数据做初步分类处理；The data analysis module performs preliminary classification processing on the collected data;

所述负载预测模块采用基于相似日的负载预测方法，以变电站所处海拔、变电站电压等级、日均气温、湿度和天气类型作为相似日评判依据，根据关联度大小，筛选出训练集和测试集；采用长短期记忆网络算法，结合变电站内电气设备的使用时间及故障率，以数据采集模块所采集数据规模作为调整参考，预测下一时刻的负载情况；The load prediction module adopts a load prediction method based on similar days, and uses the altitude of the substation, the voltage level of the substation, the daily average temperature, humidity and weather type as the evaluation basis for similar days, and filters out the training set and the test set according to the degree of correlation. ;Using long-term and short-term memory network algorithm, combined with the use time and failure rate of electrical equipment in the substation, and using the scale of data collected by the data acquisition module as an adjustment reference to predict the load situation at the next moment;

所述任务调度模块将故障检测任务按照所需资源和任务对时间的紧迫程度，采用群体智能算法分配给云端或边缘侧的计算节点；The task scheduling module uses a swarm intelligence algorithm to allocate the fault detection task to the computing nodes on the cloud or edge side according to the required resources and the urgency of the task to time;

所述资源监控模块包括资源监控器和监控报警器，所述资源监控器监控系统中云端和边缘侧计算节点的资源状态；当计算节点出现异常而无法执行当前任务时，所述监控报警器发出警报，并由任务调度模块将当前无法执行的任务重新分配给其他节点；The resource monitoring module includes a resource monitor and a monitoring alarm, the resource monitor monitors the resource status of the cloud and edge computing nodes in the system; when the computing node is abnormal and cannot perform the current task, the monitoring alarm sends out Alerts, and the task scheduling module reassigns tasks that cannot currently be executed to other nodes;

所述故障检测模块利用计算节点资源，对数据进行进一步解析，得出具体故障内容及故障发生点，分析可能的故障原因及故障处理措施，供运维人员参考；The fault detection module utilizes computing node resources to further analyze the data, obtain specific fault contents and fault occurrence points, analyze possible fault causes and fault handling measures, and provide reference for operation and maintenance personnel;

所述故障警报模块为故障发生后向运维人员发出警报，并将故障分析数据发送至云端备份，作为事后故障分析和下一时刻事前预测的历史数据；The fault alarm module sends an alarm to the operation and maintenance personnel after the fault occurs, and sends the fault analysis data to the cloud for backup, as the historical data of the fault analysis after the event and the prediction in advance at the next moment;

所述通信模块包括电力有线专网和5G电力虚拟专网，所述电力有线专网用于有线数据采集设备的数据传输、边缘计算节点之间的有线传输；所述5G电力虚拟专网用于移动数据采集设备的数据传输、边缘侧与云端计算节点之间及系统间各模块之间的无线传输，同时作为电力有线专网故障时的备用通信方式。The communication module includes a power wired private network and a 5G power virtual private network, and the power wired private network is used for data transmission of wired data acquisition equipment and wired transmission between edge computing nodes; the 5G power virtual private network is used for Data transmission of mobile data acquisition equipment, wireless transmission between edge side and cloud computing nodes, and between modules in the system, as a backup communication method when the power wired private network fails.

本发明的一种面向变电站故障检测的云边协同任务管理方法，如图2所示，包括以下步骤：A cloud-side collaborative task management method for substation fault detection of the present invention, as shown in FIG. 2 , includes the following steps:

S1：采集变电站内电气设备的数据，包括变压器、互感器、断路器、隔离开关和其他设备的数字编码数据、图像数据和声纹数据；S1: Collect data of electrical equipment in substations, including digitally encoded data, image data and voiceprint data of transformers, transformers, circuit breakers, disconnectors and other equipment;

S2：初步解析所收集的数据，按来源进行分类；S2: Preliminary analysis of the collected data, classified by source;

S3：采用基于相似日的负载预测方法，预测下一时刻的负载情况；S3: Use the load prediction method based on similar days to predict the load situation at the next moment;

S4：采用基于群体智能的任务管理方法，将故障检测任务分配到云端或边缘侧的适当计算节点；S4: Adopt the task management method based on swarm intelligence to assign fault detection tasks to appropriate computing nodes on the cloud or edge side;

S5：监测任务执行节点的资源使用情况，判断是否出现计算节点异常，若出现计算节点异常，则重新进行该任务调度；S5: Monitor the resource usage of the task execution node, determine whether there is an abnormality in the computing node, and re-schedule the task if there is an abnormality in the computing node;

S6：对节点所分配的任务进行进一步解析，判断是否存在故障，若存在，则分析具体故障内容及故障发生点；若不存在，则结束任务；S6: further analyze the task assigned by the node to determine whether there is a fault, if so, analyze the specific fault content and the point of failure; if not, end the task;

S7：根据故障检测结果判断是否需要发出警报，若需要，则立刻警报通知运维人员；若不需要，则结束任务；S7: Determine whether an alarm needs to be issued according to the fault detection result, if necessary, immediately notify the operation and maintenance personnel; if not, end the task;

S8：将故障信息上传至云端保存。S8: Upload the fault information to the cloud for storage.

其中，所述步骤S3的基于相似日的负载预测方法，如图3所示，其包含以下子步骤：Wherein, the load prediction method based on similar days in step S3, as shown in FIG. 3, includes the following sub-steps:

S3.1：输入历史样本和待预测日的影响因素数据，所述影响因素数据包括变电站所处海拔、变电站电压等级、变电站内电气设备的使用时间、变电站内电气设备的综合故障率、日均气温、湿度和天气类型；S3.1: Input historical samples and influencing factor data of the day to be predicted. The influencing factor data include the altitude of the substation, the voltage level of the substation, the use time of the electrical equipment in the substation, the comprehensive failure rate of the electrical equipment in the substation, the daily average Air temperature, humidity and weather type;

S3.2：选取负载预测所需的数据集，包括训练集和测试集，具体如下：S3.2: Select the data set required for load prediction, including training set and test set, as follows:

S3.2.1：构造相似日影响因素矩阵，其中子序列

Figure DEST_PATH_IMAGE019
为第
Figure 160024DEST_PATH_IMAGE013
天的
Figure 208096DEST_PATH_IMAGE004
个影响因素组成的向量，
Figure DEST_PATH_IMAGE020
，
Figure 960151DEST_PATH_IMAGE006
和
Figure 313772DEST_PATH_IMAGE004
分别表示历史样本天数和影响因素数量；母序列
Figure DEST_PATH_IMAGE021
为待测试日或训练标签日的
Figure 767756DEST_PATH_IMAGE004
个影响因素组成的向量。S3.2.1: Construct a matrix of similar daily influence factors, in which subsequences
Figure DEST_PATH_IMAGE019
for the first
Figure 160024DEST_PATH_IMAGE013
God's
Figure 208096DEST_PATH_IMAGE004
A vector of influencing factors,
Figure DEST_PATH_IMAGE020
,
Figure 960151DEST_PATH_IMAGE006
and
Figure 313772DEST_PATH_IMAGE004
Respectively represent the number of historical sample days and the number of influencing factors; the parent series
Figure DEST_PATH_IMAGE021
for test day or training label day
Figure 767756DEST_PATH_IMAGE004
A vector of influencing factors.

S3.2.2：对子序列和母序列做初值化处理，即每个序列中的影响因素数据都除以序列中的第一个数据，处理后的子序列和母序列分别用

Figure DEST_PATH_IMAGE022
和
Figure DEST_PATH_IMAGE023
表示。S3.2.2: Initialize the subsequence and parent sequence, that is, the influencing factor data in each sequence is divided by the first data in the sequence, and the processed subsequence and parent sequence are respectively
Figure DEST_PATH_IMAGE022
and
Figure DEST_PATH_IMAGE023
express.

S3.2.3：选取测试集，以待预测日的影响因素数据作为母序列，历史样本中随机选取40%的样本的影响因素数据作为子序列，计算各子序列与母序列间的关联度

Figure DEST_PATH_IMAGE024
，选取关联度从高到低排序前
Figure DEST_PATH_IMAGE026
位的样本作为测试集，并选取关联度最高的样本作为训练标签日。S3.2.3: Select the test set, take the influencing factor data of the day to be predicted as the parent sequence, randomly select the influencing factor data of 40% of the samples in the historical sample as the subsequence, and calculate the correlation between each subsequence and the parent sequence
Figure DEST_PATH_IMAGE024
, select the correlation degree from high to low before sorting
Figure DEST_PATH_IMAGE026
The samples with the highest degree of correlation are selected as the test set, and the samples with the highest correlation are selected as the training label days.

关联度的计算方式为The correlation is calculated as

Figure DEST_PATH_IMAGE027
（1）
Figure DEST_PATH_IMAGE027
(1)

Figure DEST_PATH_IMAGE028
（2）
Figure DEST_PATH_IMAGE028
(2)

其中，

Figure DEST_PATH_IMAGE029
为第
Figure 622448DEST_PATH_IMAGE013
个子序列的第
Figure 822486DEST_PATH_IMAGE007
个影响因素与母序列的第
Figure 474571DEST_PATH_IMAGE007
个影响因素的关联度系数；
Figure DEST_PATH_IMAGE030
为初值化处理后的母序列与第
Figure 166584DEST_PATH_IMAGE013
个子序列的第
Figure 434754DEST_PATH_IMAGE007
个影响因素之差的绝对值，
Figure DEST_PATH_IMAGE031
为初值化处理后的母序列与第
Figure 145090DEST_PATH_IMAGE013
个子序列两极最小差绝对值，
Figure DEST_PATH_IMAGE032
为初值化处理后的母序列与第
Figure 512617DEST_PATH_IMAGE013
个子序列两极最大差绝对值；
Figure DEST_PATH_IMAGE033
为分辨系数，取值范围为（0,1），通常取0.5。in,
Figure DEST_PATH_IMAGE029
for the first
Figure 622448DEST_PATH_IMAGE013
the first subsequence
Figure 822486DEST_PATH_IMAGE007
The first influencing factor and the parent sequence
Figure 474571DEST_PATH_IMAGE007
The correlation coefficient of each influencing factor;
Figure DEST_PATH_IMAGE030
is the initialized mother sequence and the first
Figure 166584DEST_PATH_IMAGE013
the first subsequence
Figure 434754DEST_PATH_IMAGE007
The absolute value of the difference between the influencing factors,
Figure DEST_PATH_IMAGE031
is the initialized mother sequence and the first
Figure 145090DEST_PATH_IMAGE013
The absolute value of the minimum difference between the poles of the subsequences,
Figure DEST_PATH_IMAGE032
is the initialized mother sequence and the first
Figure 512617DEST_PATH_IMAGE013
The absolute value of the maximum difference between the poles of the subsequences;
Figure DEST_PATH_IMAGE033
is the resolution coefficient, the value range is (0,1), usually 0.5.

S3.2.4：选取训练集，以训练标签日的影响因素数据作为母序列，历史样本中剩余60%的样本的影响因素数据作为子序列，计算各子序列与母序列间的关联度

Figure DEST_PATH_IMAGE034
，选取关联度从高到低排序前
Figure 347718DEST_PATH_IMAGE003
位的样本作为训练集。S3.2.4: Select the training set, use the influence factor data of the training label day as the parent sequence, and the influence factor data of the remaining 60% of the historical samples as the subsequence, and calculate the correlation between each subsequence and the parent sequence
Figure DEST_PATH_IMAGE034
, select the correlation degree from high to low before sorting
Figure 347718DEST_PATH_IMAGE003
bit samples as the training set.

S3.3：采用长短期记忆网络算法进行时间序列预测，预测下一时刻的负载情况，包括下一时刻所需计算、存储和网络资源。S3.3: Use the long short-term memory network algorithm for time series prediction, and predict the load situation at the next moment, including the computing, storage and network resources required for the next moment.

长短期记忆网络单元中的遗忘门、输入门、输出门迭代过程的数学表达式为The mathematical expression of the iterative process of forgetting gate, input gate and output gate in the long short-term memory network unit is:

Figure DEST_PATH_IMAGE035
（3）
Figure DEST_PATH_IMAGE035
(3)

Figure DEST_PATH_IMAGE036
（4）
Figure DEST_PATH_IMAGE036
(4)

Figure DEST_PATH_IMAGE037
（5）
Figure DEST_PATH_IMAGE037
(5)

Figure DEST_PATH_IMAGE038
（6）
Figure DEST_PATH_IMAGE038
(6)

Figure DEST_PATH_IMAGE039
（7）
Figure DEST_PATH_IMAGE039
(7)

其中，

Figure DEST_PATH_IMAGE040
、
Figure DEST_PATH_IMAGE041
、
Figure DEST_PATH_IMAGE042
分别为遗忘门、输入门和输出门的输出，
Figure DEST_PATH_IMAGE043
、
Figure DEST_PATH_IMAGE044
和
Figure DEST_PATH_IMAGE045
分别为各门的输入，
Figure DEST_PATH_IMAGE046
和
Figure DEST_PATH_IMAGE048
分别为各门对应的权重和偏置量，
Figure DEST_PATH_IMAGE050
为Sigmoid激活函数。in,
Figure DEST_PATH_IMAGE040
,
Figure DEST_PATH_IMAGE041
,
Figure DEST_PATH_IMAGE042
are the outputs of the forget gate, the input gate and the output gate, respectively,
Figure DEST_PATH_IMAGE043
,
Figure DEST_PATH_IMAGE044
and
Figure DEST_PATH_IMAGE045
are the input of each gate, respectively.
Figure DEST_PATH_IMAGE046
and
Figure DEST_PATH_IMAGE048
are the weights and biases corresponding to each gate, respectively,
Figure DEST_PATH_IMAGE050
is the sigmoid activation function.

所述步骤S4中的群体智能的任务管理方法，如图4所示，包含以下步骤：The swarm intelligence task management method in the step S4, as shown in Figure 4, includes the following steps:

S4.1：输入步骤S2中经初步解析的数据和步骤S3中的下一时刻负载预测情况；S4.1: Input the data preliminarily parsed in step S2 and the load prediction situation at the next moment in step S3;

S4.2：计算云边协同任务评价指标

Figure DEST_PATH_IMAGE051
；S4.2: Computing Cloud-Edge Collaborative Task Evaluation Metrics
Figure DEST_PATH_IMAGE051
;

其中，

Figure DEST_PATH_IMAGE052
为任务传输到云端或边缘侧计算节点需要消耗的时间；
Figure DEST_PATH_IMAGE054
为任务的紧急程度，紧急程度分为重大、紧急、一般三类，本发明中的紧急程度按来源分类，重大故障数据来源包括变压器，紧急故障数据来源包括互感器、断路器，一般故障数据来源包括隔离开关和其他设备任务，部署至云端的该项计算值分别为0.3、0.5、1，任务部署至边缘侧的该项计算值分别为0.2、0.5、1；
Figure DEST_PATH_IMAGE056
为任务部署至云端或边缘侧的负载均衡度；in,
Figure DEST_PATH_IMAGE052
The time it takes to transmit tasks to the cloud or edge computing nodes;
Figure DEST_PATH_IMAGE054
is the urgency of the task, and the urgency is divided into three categories: major, urgent, and general. The urgency in the present invention is classified according to the source. The major fault data sources include transformers, the emergency fault data sources include transformers and circuit breakers, and the general fault data sources Including isolating switch and other equipment tasks, the calculated values of this item deployed to the cloud are 0.3, 0.5, and 1, respectively, and the calculated values of this item deployed to the edge side are 0.2, 0.5, and 1, respectively;
Figure DEST_PATH_IMAGE056
Load balancing for tasks deployed to the cloud or edge;

S4.3：比较任务部署至云端计算节点的评价结果

Figure DEST_PATH_IMAGE057
和部署至边缘侧计算节点的评价结果
Figure DEST_PATH_IMAGE058
大小，若将任务部署至云端的评价结果更小则将任务部署至云端，反之则部署至边缘侧；S4.3: Compare the evaluation results of tasks deployed to cloud computing nodes
Figure DEST_PATH_IMAGE057
and evaluation results deployed to edge computing nodes
Figure DEST_PATH_IMAGE058
If the evaluation result of deploying the task to the cloud is smaller, the task will be deployed to the cloud; otherwise, it will be deployed to the edge side;

其中，

Figure 39819DEST_PATH_IMAGE057
为第
Figure 342012DEST_PATH_IMAGE013
个任务部署至云端的评价结果，评价结果的计算方式为评价指标项之和；
Figure 208336DEST_PATH_IMAGE058
为第
Figure 343783DEST_PATH_IMAGE013
个任务部署至边缘侧的评价结果，评价结果的计算方式与云端同理；in,
Figure 39819DEST_PATH_IMAGE057
for the first
Figure 342012DEST_PATH_IMAGE013
The evaluation result of each task deployed to the cloud, and the calculation method of the evaluation result is the sum of the evaluation index items;
Figure 208336DEST_PATH_IMAGE058
for the first
Figure 343783DEST_PATH_IMAGE013
The evaluation results of a task deployed to the edge side, and the calculation method of the evaluation results is the same as that of the cloud;

S4.4：根据步骤S4.3的任务部署结果，采用群体智能算法将故障检测任务进一步分配到云端或边缘侧的适当计算节点，所述群体智能算法为蚁群优化算法，具体步骤如下：S4.4: According to the task deployment result of step S4.3, use a swarm intelligence algorithm to further allocate the fault detection task to the appropriate computing nodes on the cloud or edge side. The swarm intelligence algorithm is an ant colony optimization algorithm, and the specific steps are as follows:

S4.4.1：初始化参数，将

Figure 953756DEST_PATH_IMAGE004
只蚂蚁（即
Figure 107525DEST_PATH_IMAGE004
个任务）随机分配到
Figure 410331DEST_PATH_IMAGE006
个计算节点，将出发节点加入第
Figure 767494DEST_PATH_IMAGE007
只蚂蚁的禁忌表
Figure DEST_PATH_IMAGE059
中，设置迭代次数
Figure DEST_PATH_IMAGE060
；S4.4.1: Initialization parameters, the
Figure 953756DEST_PATH_IMAGE004
an ant (ie
Figure 107525DEST_PATH_IMAGE004
tasks) are randomly assigned to
Figure 410331DEST_PATH_IMAGE006
computing nodes, add the departure node to the
Figure 767494DEST_PATH_IMAGE007
Taboo list of ants
Figure DEST_PATH_IMAGE059
, set the number of iterations
Figure DEST_PATH_IMAGE060
;

S4.4.2：根据式（8）计算在

Figure DEST_PATH_IMAGE061
时刻从节点
Figure DEST_PATH_IMAGE062
到节点
Figure DEST_PATH_IMAGE063
的概率，选择新的计算节点，并将新的节点增加到禁忌表中；S4.4.2: Calculate according to formula (8) at
Figure DEST_PATH_IMAGE061
time slave node
Figure DEST_PATH_IMAGE062
to node
Figure DEST_PATH_IMAGE063
The probability of , select a new computing node, and add the new node to the tabu list;

Figure DEST_PATH_IMAGE064
（8）
Figure DEST_PATH_IMAGE064
(8)

其中，

Figure DEST_PATH_IMAGE065
为
Figure 899267DEST_PATH_IMAGE061
时刻在节点
Figure DEST_PATH_IMAGE066
之间的残留信息，在初始时刻该值为一常数；
Figure DEST_PATH_IMAGE067
为节点
Figure 990368DEST_PATH_IMAGE062
到节点
Figure 339440DEST_PATH_IMAGE063
的启发信息，该值为两节点之间网络拓扑距离的倒数；
Figure DEST_PATH_IMAGE069
和
Figure 964326DEST_PATH_IMAGE070
分别为残留信息的重要程度和启发信息的重要程度。in,
Figure DEST_PATH_IMAGE065
for
Figure 899267DEST_PATH_IMAGE061
moment at node
Figure DEST_PATH_IMAGE066
The residual information between , the value is a constant at the initial moment;
Figure DEST_PATH_IMAGE067
for the node
Figure 990368DEST_PATH_IMAGE062
to node
Figure 339440DEST_PATH_IMAGE063
The heuristic information of , the value is the reciprocal of the network topology distance between two nodes;
Figure DEST_PATH_IMAGE069
and
Figure 964326DEST_PATH_IMAGE070
are the importance of residual information and the importance of heuristic information, respectively.

S4.4.3：更新节点

Figure 57047DEST_PATH_IMAGE066
之间的残留信息；S4.4.3: Update Node
Figure 57047DEST_PATH_IMAGE066
residual information between;

Figure DEST_PATH_IMAGE072
（9）
Figure DEST_PATH_IMAGE072
(9)

Figure DEST_PATH_IMAGE073
（10）
Figure DEST_PATH_IMAGE073
(10)

其中，

Figure DEST_PATH_IMAGE074
为更新后的在节点
Figure 654250DEST_PATH_IMAGE066
之间的残留信息；
Figure 298858DEST_PATH_IMAGE033
为信息持久性；
Figure DEST_PATH_IMAGE075
为第
Figure 411039DEST_PATH_IMAGE007
只蚂蚁在节点
Figure 41872DEST_PATH_IMAGE066
之间的残留信息差值；
Figure DEST_PATH_IMAGE076
为信息素总浓度；
Figure DEST_PATH_IMAGE077
为第
Figure 824408DEST_PATH_IMAGE007
只蚂蚁走过的网络拓扑距离之和。in,
Figure DEST_PATH_IMAGE074
for the updated node
Figure 654250DEST_PATH_IMAGE066
residual information between;
Figure 298858DEST_PATH_IMAGE033
for information persistence;
Figure DEST_PATH_IMAGE075
for the first
Figure 411039DEST_PATH_IMAGE007
only ants in the node
Figure 41872DEST_PATH_IMAGE066
The residual information difference between;
Figure DEST_PATH_IMAGE076
is the total concentration of pheromone;
Figure DEST_PATH_IMAGE077
for the first
Figure 824408DEST_PATH_IMAGE007
The sum of the network topological distances traveled by an ant.

S4.4.4：以

Figure 515283DEST_PATH_IMAGE004
只蚂蚁中网络拓扑距离之和最小为目标选择最优计算节点；S4.4.4: with
Figure 515283DEST_PATH_IMAGE004
Only select the optimal computing node as the goal with the smallest sum of network topology distances among the ants;

Figure DEST_PATH_IMAGE078
（11）
Figure DEST_PATH_IMAGE078
(11)

S4.4.5：更新最优网络拓扑路径上的残留信息；S4.4.5: Update the residual information on the optimal network topology path;

Figure DEST_PATH_IMAGE079
（12）
Figure DEST_PATH_IMAGE079
(12)

其中，

Figure DEST_PATH_IMAGE080
和
Figure DEST_PATH_IMAGE081
分别为更新后的残留信息和原始残留信息；
Figure DEST_PATH_IMAGE082
为全局信息挥发系数。in,
Figure DEST_PATH_IMAGE080
and
Figure DEST_PATH_IMAGE081
are the updated residual information and the original residual information, respectively;
Figure DEST_PATH_IMAGE082
is the global information volatility coefficient.

S4.4.6：判断是否达到所设置的迭代次数，若未达到则重复该过程，反之得到最优任务分配方案。S4.4.6: Determine whether the set number of iterations is reached, if not, repeat the process, otherwise obtain the optimal task allocation scheme.

以上所述，仅为本发明较佳的具体实施方式，但本发明的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本发明揭露的技术范围内，根据本发明的技术方案及其发明构思加以等同替换或改变，都应涵盖在本发明的保护范围之内。The above description is only a preferred embodiment of the present invention, but the protection scope of the present invention is not limited to this. The equivalent replacement or change of the inventive concept thereof shall be included within the protection scope of the present invention.

Claims (7)

1. The utility model provides a transformer substation fault detection-oriented cloud limit collaborative task management system which characterized in that: the system comprises a data acquisition module, a data analysis module, a load prediction module, a task scheduling module and a communication module which are connected in sequence, wherein the communication module is also connected with the data acquisition module, and the task scheduling module is also connected with a resource monitoring module, a fault detection module and a fault alarm module in sequence;
the data acquisition module is used for collecting data generated by data acquisition equipment of the transformer substation;
the data analysis module carries out primary classification processing on the acquired data;
the load prediction module adopts a load prediction method based on similar days, takes the altitude of the transformer substation, the voltage level of the transformer substation, the daily average temperature, the humidity and the weather type as the basis of similar daily evaluation, and screens out a training set and a test set according to the degree of association; the method comprises the steps that a long-term and short-term memory network algorithm is adopted, the service time and the fault rate of electrical equipment in a transformer substation are combined, the scale of data collected by a data collection module is used as an adjustment reference, and the load condition at the next moment is predicted;
the task scheduling module distributes the fault detection task to computing nodes on the cloud side or the edge side by adopting a swarm intelligence algorithm according to the required resources and the time urgency degree of the task;
the resource monitoring module comprises a resource monitor and a monitoring alarm, and the resource monitor monitors resource states of cloud computing nodes and edge computing nodes in the system; when the computing node is abnormal and cannot execute the current task, the monitoring alarm gives an alarm, and the task scheduling module redistributes the task which cannot be executed at present to other nodes;
the fault detection module further analyzes the data by utilizing the computing node resources to obtain specific fault content and fault occurrence points, and analyzes possible fault reasons and fault processing measures for reference of operation and maintenance personnel;
the fault alarm module sends an alarm to operation and maintenance personnel after a fault occurs, and sends fault analysis data to the cloud backup to be used as historical data for post fault analysis and prediction in advance at the next moment;
the communication module comprises a power wired private network and a 5G power virtual private network, wherein the power wired private network is used for data transmission of wired data acquisition equipment and wired transmission among edge computing nodes; the 5G electric virtual private network is used for data transmission of the mobile data acquisition equipment, wireless transmission between the edge side and the cloud computing node and among modules among systems, and is used as a standby communication mode when the electric wired private network fails.
2. A cloud-edge collaborative task management method for substation fault detection is characterized by comprising the following steps:
s1: collecting data of electrical equipment in a transformer substation, including digital coding data, image data and sound pattern data of a transformer, a mutual inductor, a circuit breaker, a disconnecting switch and other equipment;
s2: preliminarily analyzing the collected data, and classifying according to sources;
s3: predicting the load condition at the next moment by adopting a load prediction method based on similar days;
s4: distributing the fault detection task to a proper computing node at a cloud end or an edge side by adopting a group intelligence-based task management method;
s5: monitoring the resource use condition of a task execution node, judging whether a computing node is abnormal or not, and if the computing node is abnormal, re-scheduling the task;
s6: further analyzing the tasks distributed by the nodes, judging whether faults exist, and if so, analyzing specific fault contents and fault occurrence points; if not, ending the task;
s7: judging whether an alarm needs to be sent out or not according to a fault detection result, and if so, immediately alarming to inform operation and maintenance personnel; if not, ending the task;
s8: and uploading the fault information to a cloud for storage.
3. The substation fault detection-oriented cloud-edge collaborative task management method according to claim 2, wherein the step S3 includes the following sub-steps:
s3.1: inputting historical samples and influence factor data of a day to be predicted, wherein the influence factor data comprises the altitude of a transformer substation, the voltage level of the transformer substation, the service time of electrical equipment in the transformer substation, the comprehensive failure rate of the electrical equipment in the transformer substation, the daily average air temperature, the humidity and the weather type;
s3.2: selecting a data set required by load prediction, wherein the data set comprises a training set and a testing set;
s3.3: and predicting the time sequence by adopting a long-term and short-term memory network algorithm, and predicting the load condition at the next moment, wherein the load condition comprises calculation, storage and network resources required at the next moment.
4. The substation fault detection-oriented cloud-edge collaborative task management method according to claim 3, wherein the step S3.2 of selecting a data set required for load prediction includes the steps of:
s3.2.1: constructing a similar daily influence factor matrix comprising a subsequence and a mother sequence;
s3.2.2: carrying out initialization processing on the subsequence and the mother sequence;
s3.2.3: selecting a test set, taking the influence factor data of a day to be predicted as a mother sequence, randomly selecting the influence factor data of a sample with a specified proportion from historical samples as subsequences, calculating the association degree between each subsequence and the mother sequence, and selecting the subsequences before the association degree is sorted from high to low
Figure DEST_PATH_IMAGE001
Taking the sample of the bits as a test set, and selecting the sample with the highest correlation degree as a training label day;
s3.2.4: selecting a training set, taking the influence factor data of a training label day as a mother sequence, taking the influence factor data of the rest samples in the historical samples as subsequences, calculating the association degree between each subsequence and the mother sequence, and selecting the subsequences before the association degree is ranked from high to low
Figure DEST_PATH_IMAGE002
Samples of bits are used as a training set.
5. The substation fault detection-oriented cloud-edge collaborative task management method according to claim 4, characterized in that: in step S3.2.3, the influence factor data of 40% of the historical samples are randomly selected as a subsequence.
6. The substation fault detection-oriented cloud-edge collaborative task management method according to claim 2, characterized in that: the group intelligent task management method in step S4 includes the following steps:
s4.1: inputting the preliminarily analyzed data in step S2 and the load prediction situation at the next time in step S3;
s4.2: calculating a cloud edge cooperative task evaluation index;
s4.3: comparing the evaluation results of the tasks deployed to the cloud computing nodes with the evaluation results of the tasks deployed to the edge computing nodes, deploying the tasks to the cloud if the evaluation results of the tasks deployed to the cloud are smaller, and deploying the tasks to the edge if the evaluation results of the tasks deployed to the edge computing nodes are smaller;
s4.4: and according to the task deployment result in the step S4.3, further distributing the fault detection task to a proper computing node on the cloud side or the edge side by adopting a swarm intelligence algorithm.
7. The substation fault detection-oriented cloud-edge collaborative task management method according to claim 6, wherein the swarm intelligence algorithm of step S4.4 is an ant colony optimization algorithm, and comprises the following steps:
s4.4.1: initializing parameters to be
Figure DEST_PATH_IMAGE003
Ants are randomly assigned to
Figure DEST_PATH_IMAGE004
A computing node joining the departure node to the first
Figure DEST_PATH_IMAGE005
Taboo watch of only ants
Figure DEST_PATH_IMAGE006
In, the number of iterations is set
Figure DEST_PATH_IMAGE007
；
S4.4.2: is calculated at
Figure DEST_PATH_IMAGE008
Time slave node
Figure DEST_PATH_IMAGE009
To the node
Figure DEST_PATH_IMAGE010
Probability of (2)
Figure DEST_PATH_IMAGE011
Selecting a new compute node and adding the new node to the tabu table
Figure DEST_PATH_IMAGE012
Performing the following steps;
s4.4.3: updating a node
Figure DEST_PATH_IMAGE013
Residual information in between;
s4.4.4: to be provided with
Figure 283489DEST_PATH_IMAGE003
Selecting an optimal computing node for the target with the minimum sum of network topological distances in the ants;
s4.4.5: updating residual information on the optimal network topology path;
s4.4.6: and judging whether the set iteration times are reached, if not, repeating the process, otherwise, obtaining an optimal task allocation scheme.
CN202210855759.2A 2022-07-21 2022-07-21 A cloud-side collaborative task management system and method for substation fault detection Pending CN114936808A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210855759.2A CN114936808A (en) 2022-07-21 2022-07-21 A cloud-side collaborative task management system and method for substation fault detection

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210855759.2A CN114936808A (en) 2022-07-21 2022-07-21 A cloud-side collaborative task management system and method for substation fault detection

Publications (1)

Publication Number Publication Date
CN114936808A true CN114936808A (en) 2022-08-23

Family

ID=82868181

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210855759.2A Pending CN114936808A (en) 2022-07-21 2022-07-21 A cloud-side collaborative task management system and method for substation fault detection

Country Status (1)

Country Link
CN (1) CN114936808A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117154945A (en) * 2023-09-07 2023-12-01 深圳市艾能聚科技有限公司 Intelligent monitoring method for power transformation and distribution stations based on cloud computing technology
CN117527535A (en) * 2023-11-06 2024-02-06 北京易华录信息技术股份有限公司 Cloud edge collaborative fusion deposit management method and system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190164087A1 (en) * 2017-11-30 2019-05-30 B.yond, Inc. Decomposing tasks through artificial intelligence chaining
CN112882383A (en) * 2021-01-11 2021-06-01 合肥工业大学 Substation monitoring and edge cloud cooperation-oriented prediction control scheduling method
CN113342510A (en) * 2021-08-05 2021-09-03 国能大渡河大数据服务有限公司 Water and power basin emergency command cloud-side computing resource cooperative processing method
CN113553146A (en) * 2021-04-25 2021-10-26 北京航空航天大学 A cloud-edge collaborative computing task merging scheduling method
CN113645587A (en) * 2021-10-12 2021-11-12 北京智盟信通科技有限公司 Transformer substation comprehensive monitoring analysis method and system based on edge Internet of things agent
CN114371925A (en) * 2022-01-13 2022-04-19 北京邮电大学 Edge cloud cooperative multi-task scheduling method and system for guaranteeing edge cloud load ratio

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190164087A1 (en) * 2017-11-30 2019-05-30 B.yond, Inc. Decomposing tasks through artificial intelligence chaining
CN112882383A (en) * 2021-01-11 2021-06-01 合肥工业大学 Substation monitoring and edge cloud cooperation-oriented prediction control scheduling method
CN113553146A (en) * 2021-04-25 2021-10-26 北京航空航天大学 A cloud-edge collaborative computing task merging scheduling method
CN113342510A (en) * 2021-08-05 2021-09-03 国能大渡河大数据服务有限公司 Water and power basin emergency command cloud-side computing resource cooperative processing method
CN113645587A (en) * 2021-10-12 2021-11-12 北京智盟信通科技有限公司 Transformer substation comprehensive monitoring analysis method and system based on edge Internet of things agent
CN114371925A (en) * 2022-01-13 2022-04-19 北京邮电大学 Edge cloud cooperative multi-task scheduling method and system for guaranteeing edge cloud load ratio

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
HAOYU WANG等: "HealthEdge: Task Scheduling for Edge Computing with Health Emergency and Human Behavior Consideration in Smart Homes", 《2017 IEEE INTERNATIONAL CONFERENCE ON BIG DATA》 *
王瑞等: "运用相似日和 LSTM 的短期负荷双向组合预测", 《电力系统及其自动化学报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117154945A (en) * 2023-09-07 2023-12-01 深圳市艾能聚科技有限公司 Intelligent monitoring method for power transformation and distribution stations based on cloud computing technology
CN117527535A (en) * 2023-11-06 2024-02-06 北京易华录信息技术股份有限公司 Cloud edge collaborative fusion deposit management method and system

Similar Documents

Publication Publication Date Title
CN109872003B (en) Object state prediction method, system, computer equipment and storage medium
CN110705873B (en) Power distribution network running state portrait analysis method
CN116073372A (en) Intelligent electricity utilization-based safety monitoring management system and method
CN110675070B (en) Criticality analysis method of intelligent microgrid equipment
CN106529708A (en) Distribution network planning system based on cloud platform
CN116451876A (en) Power distribution network fault prediction and active overhaul system based on artificial intelligence
WO2019196869A1 (en) Method for determining list of patrolling base stations, and patrolling apparatus
CN114936808A (en) A cloud-side collaborative task management system and method for substation fault detection
CN118316190B (en) Electric power system monitoring system based on Internet of things
CN113169709A (en) Solar power generation control system and method based on machine learning
CN115730749B (en) Power dispatching risk early warning method and device based on fusion power data
CN108964023A (en) A kind of busbar voltage situation short term prediction method and system for power grid
CN116089218A (en) Dynamic baseline alarm method and system based on Kubernetes historical data and trend analysis
CN113722303B (en) A visualization-based electric energy equipment and information operation and maintenance system
CN110658725A (en) Energy supervision and prediction system and method based on artificial intelligence
CN114689965B (en) Online monitoring method and system for power equipment based on embedded intelligent sensors
CN118508597B (en) Emergency power supply method, system and equipment for power system guarantee object
CN117353306B (en) Optical storage charge-discharge scheduling method, optical storage charge-discharge scheduling system, electronic equipment and storage medium
CN115598459A (en) Power failure prediction method for 10kV feeder line fault of power distribution network
CN120048080A (en) Intelligent hydrogeological environment monitoring method
CN115271561B (en) Emergency power supply management and control system and mobile emergency power supply scheduling method for load recovery
CN119761644A (en) A natural disaster power grid damage assessment method and system based on big data analysis
Xu et al. Enhancing power marketing audit through IoT and multi-sensor information fusion: A substation scenario analysis
Zheng et al. [Retracted] Application Based on Artificial Intelligence in Substation Operation and Maintenance Management
CN109359810A (en) A method for evaluating the operation state of power communication transmission network based on multi-strategy balance

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20220823

点击 这是indexloc提供的php浏览器服务，不要输入任何密码和下载