CN110705859A - Evaluation method of operating status of medium and low voltage distribution network based on PCA-self-organizing neural network - Google Patents
Evaluation method of operating status of medium and low voltage distribution network based on PCA-self-organizing neural networkDownload PDFInfo
- Publication number
- CN110705859A CN110705859ACN201910912141.3ACN201910912141ACN110705859ACN 110705859 ACN110705859 ACN 110705859ACN 201910912141 ACN201910912141 ACN 201910912141ACN 110705859 ACN110705859 ACN 110705859A
- Authority
- CN
- China
- Prior art keywords
- index
- distribution network
- evaluation
- layer
- self
- 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
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
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
- 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
- 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
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Translated fromChineseDescription
Translated fromChinese技术领域technical field
本发明涉及中低压配电网运行状态评估技术领域，具体涉及一种基于PCA-自组织神经网络的中低压配电网运行状态评估方法。The invention relates to the technical field of evaluation of the operation state of a medium and low voltage distribution network, in particular to a method for evaluating the operation state of a medium and low voltage distribution network based on a PCA-self-organizing neural network.
背景技术Background technique
近年来，作为智能电网的重要组成部分，配电网智能化也成为智能电网发展的新趋势，且由于中低压配电网是电力系统与用户直接相连的关键环节，其运行状态直接影响到国民生活和经济。因此，构建一套科学有效、快速精确的中低压配电网运行状态评估体系是当务之急。In recent years, as an important part of the smart grid, the intelligence of the distribution network has also become a new trend in the development of the smart grid, and because the medium and low voltage distribution network is the key link between the power system and the user, its operation status directly affects the national life and economy. Therefore, it is imperative to build a scientific, effective, fast and accurate operating status evaluation system for medium and low voltage distribution networks.
我国配电网有分布范围广，线路多且排布混乱，设备种类和数量繁多，自动化水平较低等问题，导致配电网运行数据采集困难，而且目前对配电网评估的研究大部分停留在高、中压配电网层面，对中低压配电网特性的评估较少。然而，随着用户对供电可靠性要求的不断提高以及智能电网的不断发展，配电网的健康状况及其对配电网系统的影响逐渐得到了重视。而对配电网实施合理的评估，将有助于判定配电网的运行状态，有针对性的对配电网进行改造，使配电网的供电可靠性、电能质量、电网资产利用率等方面得到改善。my country's distribution network has problems such as wide distribution, many lines and chaotic arrangements, many types and quantities of equipment, and low automation level, which make it difficult to collect distribution network operation data, and most of the current research on distribution network evaluation has stayed At the level of high and medium voltage distribution networks, there are few evaluations of the characteristics of medium and low voltage distribution networks. However, with the continuous improvement of users' requirements for power supply reliability and the continuous development of smart grids, the health status of the distribution network and its impact on the distribution network system have gradually received attention. A reasonable evaluation of the distribution network will help to determine the operation status of the distribution network, and to transform the distribution network in a targeted manner, so as to improve the reliability of power supply, power quality, and utilization of power grid assets. aspects have been improved.
配电网是一种直接面向用户的终端网络，对维持供电可靠性和地区经济平稳发展有着极其重要的作用。目前我国的配电网在电能质量、供电可靠性上相较于西方发达国家仍有不小的差距。近年来，电网运行状态评估系统成为智能电网的重要组成部分，其对提高调度效率，避免故障发生与扩大，提高供电可靠性具有重要意义。而配电网作为连接输电网和用户的重要环节，其运行状态的好坏直接影响着用户用电的可靠性和电能质量，与人民生活水平和国民经济发展息息相关。为了满足用户日益增长的用电需求，需要提供一种安全可靠、优质经济的中低压配电网运行状态评估方法。Distribution network is a terminal network directly facing users, which plays an extremely important role in maintaining the reliability of power supply and the stable development of regional economy. At present, my country's distribution network still has a big gap compared with western developed countries in terms of power quality and power supply reliability. In recent years, the power grid operation status evaluation system has become an important part of the smart grid, which is of great significance to improve the dispatching efficiency, avoid the occurrence and expansion of faults, and improve the reliability of power supply. As an important link connecting the transmission network and users, the distribution network directly affects the reliability and power quality of users' electricity consumption, and is closely related to people's living standards and the development of the national economy. In order to meet the increasing electricity demand of users, it is necessary to provide a safe, reliable, high-quality and economical operation status evaluation method for medium and low voltage distribution networks.
现有技术中涉及中低压配电网运行状态评估存在的弊端有：The disadvantages of the prior art involving the evaluation of the operation status of the medium and low voltage distribution network are:
1)、在中低压配电网运行状态评估中，配电网状态影响因素多导致评估指标数目众多，且暂时没有统一合理的方法来筛选评估指标，亟需一种客观合理的方法来筛选评估指标。1) In the evaluation of the operation status of the medium and low voltage distribution network, there are many factors affecting the distribution network state, resulting in a large number of evaluation indicators, and there is no unified and reasonable method to screen the evaluation indicators. An objective and reasonable method is urgently needed to screen and evaluate index.
2)、中低压配电网结构复杂，设备繁多，自动化水平较低，现有的SCADA等系统的运行数据存在采集困难、精度较差、数据不完整等特点。因此，在对中低压配电网运行状态进行评估时，传统评估流程有较大误差，亟需对异常数据进行辨识和清理。2) The structure of the medium and low voltage distribution network is complex, the equipment is numerous, and the automation level is low. The operation data of the existing SCADA and other systems have the characteristics of difficulty in collection, poor accuracy, and incomplete data. Therefore, when evaluating the operating status of the medium and low voltage distribution network, the traditional evaluation process has large errors, and it is urgent to identify and clean up abnormal data.
发明内容SUMMARY OF THE INVENTION
本发明的目的在于提供一种基于PCA-自组织神经网络的中低压配电网运行状态评估方法，从安全性、可靠性、优质性、经济性多方面对中低压配电网进行各方面的综合评估，该方法能够为中低压配电网运行状态控制和管理提供有效的技术支持和参考意见。The purpose of the present invention is to provide a method for evaluating the operation state of medium and low voltage distribution network based on PCA-self-organizing neural network, which can evaluate the medium and low voltage distribution network in various aspects from the aspects of safety, reliability, quality and economy. Comprehensive evaluation, this method can provide effective technical support and reference for the control and management of the operating state of the medium and low voltage distribution network.
本发明采取的技术方案为：The technical scheme adopted in the present invention is:
基于PCA-自组织神经网络的中低压配电网运行状态评估方法，包括以下步骤：The method for evaluating the operation status of medium and low voltage distribution network based on PCA-self-organizing neural network includes the following steps:
步骤1：根据配电网运行状态基础数据和设备参数，采用基于改进的主成分分析法PCA特征提取算法，提取出最能反映配电网状态的指标变量，并构建层次分析评估指标体系；Step 1: According to the basic data and equipment parameters of the operation state of the distribution network, the PCA feature extraction algorithm based on the improved principal component analysis method is used to extract the index variables that can best reflect the state of the distribution network, and build an AHP evaluation index system;
步骤2：针对计算指标变量所需数据，进行基于自组织神经网络的数据清理；Step 2: Perform data cleaning based on the self-organizing neural network for the data required for calculating the index variables;
步骤3：对清理过的配电网数据进行统计分析，计算指标层单项指标值和单项指标得分。然后利用熵组合权重算法，计算出各个指标层指标的综合权重，最后根据指标层单项指标得分和综合权重，逐层向上计算中间层和目标层的指标得分；Step 3: Statistical analysis is performed on the cleaned distribution network data, and the single index value and single index score of the index layer are calculated. Then use the entropy combination weight algorithm to calculate the comprehensive weight of each index layer index, and finally calculate the index score of the intermediate layer and the target layer layer by layer according to the single index score and comprehensive weight of the index layer;
步骤4：将评估得分划分评估健康等级，再根据评估健康等级从上层指标到下层指标找出配电网运行薄弱环节。Step 4: Divide the evaluation scores into evaluation health levels, and then find out the weak links in the operation of the distribution network from the upper-level indicators to the lower-level indicators according to the evaluation health levels.
本发明一种基于PCA-自组织神经网络的中低压配电网运行状态评估方法，技术效果如下：The present invention is a method for evaluating the operation state of a medium and low voltage distribution network based on a PCA-self-organizing neural network, and the technical effects are as follows:
1：本发明方法考虑了异常数据对配电网运行状态评估的影响，通过自组织神经网络的数据清理算法对配电网运行异常数据进行辨识及清除，根据算例分析表明自组织神经网络建模简单，收敛速度快且精度较高，能有效降低因异常数据带来的评估误差。1: The method of the present invention considers the influence of abnormal data on the evaluation of the operation state of the distribution network, and identifies and clears the abnormal data of the operation of the distribution network through the data cleaning algorithm of the self-organizing neural network. The model is simple, the convergence speed is fast and the accuracy is high, which can effectively reduce the evaluation error caused by abnormal data.
2：本发明方法采用一种改进的主成分分析法(PCA)来构建状态评估指标体系，采用熵组合权重法，来计算各评估指标的主客观综合权重，从安全性、可靠性、优质性和经济性方面对配电网进行各方面的综合评估，形成了一套科学、有效的中低压配电网运行状态评估体系。2: The method of the present invention adopts an improved principal component analysis (PCA) to construct a state evaluation index system, and adopts the entropy combined weight method to calculate the subjective and objective comprehensive weights of each evaluation index. Comprehensive evaluation of all aspects of the distribution network in terms of economic and economic aspects has formed a scientific and effective evaluation system for the operation status of medium and low voltage distribution networks.
3：本发明方法能够为中低压配电网运行状态控制和管理提供有效的技术支持和参考意见。3: The method of the present invention can provide effective technical support and reference opinions for the control and management of the operation state of the medium and low voltage distribution network.
附图说明Description of drawings
图1为本发明评估方法的流程图。FIG. 1 is a flow chart of the evaluation method of the present invention.
具体实施方式Detailed ways
如图1所示，基于PCA-自组织神经网络的中低压配电网运行状态评估方法，包括以下步骤：As shown in Figure 1, the method for evaluating the operation status of medium and low voltage distribution network based on PCA-self-organizing neural network includes the following steps:
步骤1：根据配电网运行状态基础数据和设备参数，采用基于改进的主成分分析法PCA特征提取算法，提取出最能反映配电网状态的指标变量，并构建层次分析评估指标体系。基础数据包括10kV线路电流、台区关口电压、10kV母线电压等。Step 1: According to the basic data and equipment parameters of the operation state of the distribution network, the PCA feature extraction algorithm based on the improved principal component analysis method is used to extract the index variables that can best reflect the state of the distribution network, and build an AHP evaluation index system. The basic data includes 10kV line current, station gate voltage, 10kV busbar voltage, etc.
设备参数包括主变额定容量、配变额定容量、开关故障信息等。Equipment parameters include main transformer rated capacity, distribution transformer rated capacity, switch fault information, etc.
步骤2：针对计算指标变量所需数据，进行基于自组织神经网络的数据清理；Step 2: Perform data cleaning based on the self-organizing neural network for the data required for calculating the index variables;
步骤3：对清理过的配电网数据进行统计分析，计算指标层单项指标值和单项指标得分。然后利用熵组合权重算法，计算出各个指标层指标的综合权重，最后根据指标层单项指标得分和综合权重，逐层向上计算中间层和目标层的指标得分；Step 3: Statistical analysis is performed on the cleaned distribution network data, and the single index value and single index score of the index layer are calculated. Then use the entropy combination weight algorithm to calculate the comprehensive weight of each index layer index, and finally calculate the index score of the intermediate layer and the target layer layer by layer according to the single index score and comprehensive weight of the index layer;
步骤4：将评估得分划分评估健康等级，再根据评估健康等级从上层指标到下层指标找出配电网运行薄弱环节。Step 4: Divide the evaluation scores into evaluation health levels, and then find out the weak links in the operation of the distribution network from the upper-level indicators to the lower-level indicators according to the evaluation health levels.
所述步骤1中，基于改进的主成分分析法PCA的特征提取算法，将配电网诸多评估指标进行分类，在每一个大类中利用主成分分析筛选出最能反映配电网状态的指标变量，然后利用层次分析法构建评评估指标体系，包括以下步骤：In the step 1, based on the feature extraction algorithm of the improved principal component analysis method PCA, many evaluation indexes of the distribution network are classified, and in each category, the principal component analysis is used to screen out the indexes that can best reflect the state of the distribution network. variables, and then use the analytic hierarchy process to construct an evaluation index system, including the following steps:
步骤1.1、分别选取配电网状态评估指标参量，对各个评估指标参量进行量化，构建评估指标量化矩阵，即：X＝[X1,X2,X3]；Step 1.1. Select the distribution network state evaluation index parameters respectively, quantify each evaluation index parameter, and construct an evaluation index quantification matrix, namely: X=[X1 , X2 , X3 ];
其中：X1表示历年来发生的事故统计中对应评估指标参量的百分比；Among them: X1 represents the percentage of the corresponding evaluation index parameters in the accident statistics that have occurred over the years;
X2表示历年来产生严重缺陷统计中对应评估指标参量的百分比；X2 represents the percentage of the corresponding evaluation index parameters in the statistics of serious defects generated over the years;
X3表示历年来产生一般缺陷统计中对应评估指标参量的百分比；X3 represents the percentage of the corresponding evaluation index parameters in the statistics of general defects generated over the years;
并在进行主成分分析之前先消除量纲的影响，采用原始数据标准化：
其中：zij为标准化后的变量值，xij为实际变量值，sj为标准差，
依据协方差原理，对指标变量进行标准化变换后，变量协方差矩阵即为其相关系数矩阵，标准化变换后的相关的协方差系数是等价的。According to the covariance principle, after the standardization transformation of the index variable, the variable covariance matrix is its correlation coefficient matrix, and the relevant covariance coefficients after standardized transformation are equivalent.
步骤1.2、解相关系数矩阵，并求出相关系数矩阵的特征值及特征向量，按照特征值大小排序λ1≥λ2≥λ3≥…λp，其中，对应于每个特征值λi的特征向量为αi，||αi||＝1；然后计算累计方差贡献率：
步骤1.3、求取主成分载荷：
分析选取的m个主成分，计算主成分中状态指标参量的重要度
所述步骤2中，在基于改进的主成分分析法PCA特征提取算法，得出评估所需的指标变量后，针对计算指标变量所需数据进行自组织神经网络特征映射，将离群的神经元中包含的组清除，完成异常数据清除；In the step 2, after obtaining the index variables required for the evaluation based on the improved PCA feature extraction algorithm, the self-organizing neural network feature mapping is performed for the data required for calculating the index variables, and the outlier neurons are mapped. The group contained in the clearing, completes the abnormal data clearing;
基于自组织神经网络的数据清理算法的模型建立步骤如下：The model building steps of the data cleaning algorithm based on self-organizing neural network are as follows:
1)、将当前输入模式向量和竞争中的各个神经元对应的内星向量进行归一化如下：1), normalize the current input pattern vector and the inner star vector corresponding to each neuron in the competition as follows:
式中：(j＝1,2,3,...m)；为归一化样本，X为输入模式向量；j为神经元编号；In the formula: (j=1,2,3,...m); is the normalized sample, X is the input pattern vector; j is the neuron number;
2)、每当获得一个输入模式向量时，所有神经元对应的内星向量都将与其进行相似性对比，然后将比较得最相似的内星向量列为竞争神经元，而要使两个向量最为相似，其点积就要最大，即欧式距离最小：2) Whenever an input pattern vector is obtained, the inner star vectors corresponding to all neurons will be compared for their similarity, and then the most similar inner star vector will be listed as a competing neuron, and the two vectors should be used. The most similar, its dot product is the largest, that is, the Euclidean distance is the smallest:
式中：(j＝1,2,3,...m)；
3)、网络输出与权值调整：当最佳匹配单元向输入向量调整时，最佳匹配单元中的量会随着时间和距离而降低，神经元的更新公式为：Ab(s+1)＝Ab(s)+f(u,b,s)β(s)[X(t)-Ab(s)]；3) Network output and weight adjustment: When the best matching unit is adjusted to the input vector, the amount in the best matching unit will decrease with time and distance. The update formula of the neuron is: Ab (s+1 )=Ab (s)+f(u,b,s)β(s)[X(t)-Ab (s)];
式中：t为训练样本的指数；X(t)为输入向量；s为步长指数；β(s)为单调递减的学习系数；u为输入向量的最佳匹配单元指数；f(u,b,s)为步长为s时神经元u和神经元b之间距离的临近函数；T为训练样本的大小。In the formula: t is the index of the training sample; X(t) is the input vector; s is the step size index; β(s) is the monotonically decreasing learning coefficient; u is the best matching unit index of the input vector; f(u, b, s) is the proximity function of the distance between neuron u and neuron b when the step size is s; T is the size of the training sample.
所述步骤3中，对清理过的配电网数据进行统计分析，计算指标层单项指标值和单项指标得分，然后利用熵组合权重算法计算出各个指标层指标的综合权重，最后根据指标层单项指标得分和综合权重逐层向上计算中间层和目标层的指标得分：In the step 3, statistical analysis is performed on the cleaned distribution network data, the single index value and the single index score of the index layer are calculated, and then the comprehensive weight of each index layer index is calculated by using the entropy combination weight algorithm, and finally the single index layer is calculated according to the index layer. Index scores and comprehensive weights Calculate the index scores of the intermediate layer and the target layer layer by layer:
式中：y(k+1)表示层次分析中第k+1层指标X(k+1)的评分；
所述步骤4中，根据计算中低压配电网运行状态评估的各层次指标得分，并将评估得分划分为四个健康等级，分别为：健康、比较健康、一般和不健康。In the step 4, according to the calculation of each level index score of the operation state evaluation of the medium and low voltage distribution network, the evaluation score is divided into four health levels, namely: healthy, relatively healthy, average and unhealthy.
实施例：Example:
对广西省某地区的配电网作为评估对象做详细仿真计算和分析：A detailed simulation calculation and analysis of the distribution network in a certain area of Guangxi Province as the evaluation object:
1.构建状态评估指标体系：1. Build a state evaluation index system:
在中低压配电网运行状态评估中，评估指标的数量过多会影响评估的效率和精度，同时也会对数据清理算法带来不便。因此，利用基于改进主成分分析(PCA)的特征提取算法进行指标提取和删除冗余指标变量能够提高评估效率。首先预选取影响中低压配电网运行状态的若干个单项评估指标，然后对预选取指标进行基于改进主成分分析的特征提取，构建量化矩阵并计算特征值和方差贡献率，得到出电压合格率和变压器负载率的累计方差贡献率分别为25.13％和63.56％，即满足所选主成分的累计方差贡献率在85％以上的条件，选择以上两个指标作为主成分，求取主成分载荷并通过指标参量的重要度公式计算得到各单项评估指标重要度，将各单项评估指标重要度归一化到[0,1]，本发明选取重要度大于0.5的单项评估指标，筛选得到6个中低压配电网运行状态评估单项指标，并对其建立层次分析指标体系如下表1所示：In the evaluation of the operation status of the medium and low voltage distribution network, too many evaluation indicators will affect the efficiency and accuracy of the evaluation, and it will also bring inconvenience to the data cleaning algorithm. Therefore, using the feature extraction algorithm based on improved principal component analysis (PCA) to extract indicators and delete redundant indicator variables can improve the evaluation efficiency. Firstly, several single evaluation indicators that affect the operation status of the medium and low voltage distribution network are preselected, and then the preselected indicators are extracted based on the improved principal component analysis, and the quantitative matrix is constructed and the eigenvalue and variance contribution rate are calculated to obtain the voltage qualification rate. The cumulative variance contribution rate of and transformer load rate are 25.13% and 63.56% respectively, that is, the condition that the cumulative variance contribution rate of the selected principal component is more than 85%, the above two indicators are selected as the principal component, and the principal component load is calculated and calculated. The importance of each individual evaluation index is calculated by the importance formula of the index parameter, and the importance of each individual evaluation index is normalized to [0, 1]. The single index of low-voltage distribution network operation status evaluation is established, and the analytic hierarchy process index system is established for it, as shown in Table 1 below:
表1层次分析指标体系表Table 1 AHP index system table
2.异常数据辨识与清理：2. Abnormal data identification and cleaning:
首先，根据现场监测数据进行编号，对内星权向量和输入模式向量进行归一化处理；然后通过计算欧氏距离来获得最相似的向量，找出获胜节点；最后调整获胜域内节的权值，建立6*6大小的竞争层网络，对自组织神经网络进行重复学习训练后，将原始数据归为36个神经元中，并以此进行数据辨识和清理。选取3972组配电网实时监测数据，对监测数据进行自组织神经网络学习训练，每个神经元中包含了一定量的监测数据，临近神经元之间的欧式距离越远，则说明该神经元的监测数据不健康，需要进行清理。仿真可得3号神经元，15号神经元和33号神经元与临近的神经元之间欧氏距离远，需将其进行清理，需要剔除的异常数据编号如下表2所示：First, the numbering is performed according to the field monitoring data, and the inner star weight vector and the input pattern vector are normalized; then the most similar vector is obtained by calculating the Euclidean distance, and the winning node is found; finally, the weight of the inner node in the winning domain is adjusted. , establish a 6*6 competitive layer network, and after repeated learning and training of the self-organizing neural network, the original data is classified into 36 neurons, and the data is identified and cleaned. 3972 groups of real-time monitoring data of distribution network are selected, and self-organizing neural network learning and training is carried out on the monitoring data. Each neuron contains a certain amount of monitoring data. The farther the Euclidean distance between adjacent neurons is, the more The monitoring data is unhealthy and needs to be cleaned up. The simulation results show that the Euclidean distance between neuron 3, neuron 15, and neuron 33 is far from the adjacent neurons. It needs to be cleaned up. The number of abnormal data that needs to be eliminated is shown in Table 2 below:
表2异常数据编号表Table 2 Abnormal data number table
3.实际配电网运行状态评估：3. Evaluation of the actual distribution network operation status:
(1)：确定评估指标权重：根据AHP-Delphi法计算六个单项评估指标的主观权重Wj,j＝1,2,...,n。然后根据熵权法计算各个单项评估指标的客观权重wj,j＝1,2,...,n,最后利用熵组合权重法计算出六个单项评估指标的综合权重，所得各单项指标的主客观权重如下表3所示：(1): Determine the weights of the evaluation indicators: Calculate the subjective weights Wj of the six individual evaluation indicators according to the AHP-Delphi method, j=1,2,...,n. Then, the objective weights wj , j=1,2,...,n of each single evaluation index are calculated according to the entropy weight method, and finally the comprehensive weights of the six single evaluation indexes are calculated by the entropy combined weight method. The subjective and objective weights are shown in Table 3 below:
表3各单项指标的主客观权重表Table 3 Subjective and objective weights of individual indicators
求得各个单项指标权重后，向上层计算指标体系中间层的指标权重，然后整理可得中低压配电网运行状态评估指标体系的权重因子如下表4所示：After obtaining the weight of each single index, the index weight of the middle layer of the index system is calculated to the upper layer, and then the weight factors of the evaluation index system of the operation state of the medium and low voltage distribution network can be obtained as shown in Table 4 below:
表4低压配电网运行状态评估指标体系的权重因子Table 4 Weighting factors of low-voltage distribution network operation status evaluation index system
(2)：计算评估评分评级并对比：由基于自组织神经网络异常数据辨识和清理的算例分析可知，3号、15号、33号神经元中数据异常，需被剔除。选取剩余任一神经元中的数据与33号神经元中的数据进行状态评估，并结合单项指标模型，单项指标评分公式以及权重因子进行计算，最终得出该地区的中低压配电网运行状态综合评估得分对比如下表5所示：(2): Calculate the evaluation score and compare it: According to the example analysis based on the identification and cleaning of abnormal data of self-organizing neural network, it can be seen that the abnormal data in neurons No. 3, No. 15, and No. 33 need to be eliminated. Select the data in any remaining neuron and the data in neuron No. 33 to evaluate the state, and combine the single index model, single index scoring formula and weight factor to calculate, and finally obtain the operating state of the medium and low voltage distribution network in the region The comprehensive evaluation score comparison is shown in Table 5 below:
表5中低压配电网运行状态综合评估得分对比表Table 5. Comparison of scores for comprehensive evaluation of operating status of medium and low voltage distribution networks
由上表5可以看出，广西省该地区的配电网运行状态的整体评分为71.59分，评估健康等级为比较健康。其中，由应被剔除的33号神经元数据的评估得分较20号神经元的评估得分高，且在优质性方面远高于正常数据的评估得分，经过调查分析，该地区电压合格率一般，特别是低电压现象较为普遍，达不到较为健康的水平，证明了使用未经清理过的异常数据进行状态评估时有较大偏差，进而证明基于自组织神经网络的数据清理算法进行数据辨识和清理能够有效提升配电网运行状态评估的精度，真实反映配电网的实际运行状态。It can be seen from Table 5 above that the overall score of the operation status of the distribution network in this area of Guangxi Province is 71.59 points, and the assessment health level is relatively healthy. Among them, the evaluation score of neuron No. 33 that should be excluded is higher than that of neuron No. 20, and its high quality is much higher than that of normal data. After investigation and analysis, the voltage pass rate in this area is average, In particular, the phenomenon of low voltage is relatively common and cannot reach a relatively healthy level, which proves that there is a large deviation when using uncleaned abnormal data for state evaluation, and further proves that the data cleaning algorithm based on self-organizing neural network can perform data identification and analysis. Cleaning can effectively improve the accuracy of distribution network operating state assessment, and truly reflect the actual operating state of the distribution network.
Claims (5)
Translated fromChinese
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910912141.3ACN110705859A (en) | 2019-09-25 | 2019-09-25 | Evaluation method of operating status of medium and low voltage distribution network based on PCA-self-organizing neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910912141.3ACN110705859A (en) | 2019-09-25 | 2019-09-25 | Evaluation method of operating status of medium and low voltage distribution network based on PCA-self-organizing neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110705859Atrue CN110705859A (en) | 2020-01-17 |
Family
ID=69196361
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910912141.3APendingCN110705859A (en) | 2019-09-25 | 2019-09-25 | Evaluation method of operating status of medium and low voltage distribution network based on PCA-self-organizing neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110705859A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112270485A (en)* | 2020-11-03 | 2021-01-26 | 国网浙江省电力有限公司绍兴供电公司 | Principal component factor analysis method for comprehensive evaluation of distribution network uninterrupted operation |
| CN113392958A (en)* | 2021-06-02 | 2021-09-14 | 湖南大学 | Parameter optimization and application method and system of fuzzy neural network FNN |
| CN114384348A (en)* | 2021-10-20 | 2022-04-22 | 国网宁夏电力有限公司检修公司 | A method, medium and system for monitoring the health margin of key components of a converter valve |
| CN115345392A (en)* | 2021-05-12 | 2022-11-15 | 中国石油化工股份有限公司 | A distribution network abnormal situation analysis and key index screening method |
| CN115905891A (en)* | 2022-12-19 | 2023-04-04 | 上海交通大学 | PMU data-based power distribution network operation mode and key influence factor identification method |
| CN116405187A (en)* | 2023-04-21 | 2023-07-07 | 石家庄铁道大学 | Distributed node intrusion situation sensing method based on block chain |
| CN118585940A (en)* | 2024-07-30 | 2024-09-03 | 中国电建集团昆明勘测设计研究院有限公司 | Comprehensive determination method, device and storage medium for swelling potential of expansive soil |
| WO2025019255A1 (en)* | 2023-07-14 | 2025-01-23 | Schlumberger Technology Corporation | Methods for confidence assessment with feature importance in data driven algorithms |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599530A (en)* | 2016-10-31 | 2017-04-26 | 北京千安哲信息技术有限公司 | Processing method and device for detection data |
| CN110110977A (en)* | 2019-04-19 | 2019-08-09 | 广西电网有限责任公司钦州供电局 | It is a kind of based on improve principal component-step analysis power distribution network appraisal procedure |
- 2019
- 2019-09-25CNCN201910912141.3Apatent/CN110705859A/enactivePending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599530A (en)* | 2016-10-31 | 2017-04-26 | 北京千安哲信息技术有限公司 | Processing method and device for detection data |
| CN110110977A (en)* | 2019-04-19 | 2019-08-09 | 广西电网有限责任公司钦州供电局 | It is a kind of based on improve principal component-step analysis power distribution network appraisal procedure |
Non-Patent Citations (1)
| Title |
|---|
| 吴胜聪等: "基于自组织神经网络的火电厂健康状态数据提取算法", 《电工电气》* |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112270485A (en)* | 2020-11-03 | 2021-01-26 | 国网浙江省电力有限公司绍兴供电公司 | Principal component factor analysis method for comprehensive evaluation of distribution network uninterrupted operation |
| CN115345392A (en)* | 2021-05-12 | 2022-11-15 | 中国石油化工股份有限公司 | A distribution network abnormal situation analysis and key index screening method |
| CN113392958A (en)* | 2021-06-02 | 2021-09-14 | 湖南大学 | Parameter optimization and application method and system of fuzzy neural network FNN |
| CN114384348A (en)* | 2021-10-20 | 2022-04-22 | 国网宁夏电力有限公司检修公司 | A method, medium and system for monitoring the health margin of key components of a converter valve |
| CN114384348B (en)* | 2021-10-20 | 2024-12-03 | 国网宁夏电力有限公司检修公司 | A method, medium and system for monitoring health margin of key components of a converter valve |
| CN115905891A (en)* | 2022-12-19 | 2023-04-04 | 上海交通大学 | PMU data-based power distribution network operation mode and key influence factor identification method |
| CN116405187A (en)* | 2023-04-21 | 2023-07-07 | 石家庄铁道大学 | Distributed node intrusion situation sensing method based on block chain |
| CN116405187B (en)* | 2023-04-21 | 2024-04-09 | 石家庄铁道大学 | Distributed node intrusion situation sensing method based on block chain |
| WO2025019255A1 (en)* | 2023-07-14 | 2025-01-23 | Schlumberger Technology Corporation | Methods for confidence assessment with feature importance in data driven algorithms |
| CN118585940A (en)* | 2024-07-30 | 2024-09-03 | 中国电建集团昆明勘测设计研究院有限公司 | Comprehensive determination method, device and storage medium for swelling potential of expansive soil |
| CN118585940B (en)* | 2024-07-30 | 2024-10-11 | 中国电建集团昆明勘测设计研究院有限公司 | Comprehensive discrimination method, device and storage medium for expansive soil expansive potential |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110705859A (en) | Evaluation method of operating status of medium and low voltage distribution network based on PCA-self-organizing neural network | |
| CN107516170B (en) | A differential self-healing control method based on equipment failure probability and power grid operation risk | |
| CN106251059B (en) | Cable state evaluation method based on probabilistic neural network algorithm | |
| WO2021073462A1 (en) | 10 kv static load model parameter identification method based on similar daily load curves | |
| CN104700321B (en) | A kind of power transmission and transformation equipment state operation trend analysis method | |
| CN108320043A (en) | A kind of distribution network equipment state diagnosis prediction method based on electric power big data | |
| CN110175789A (en) | Quan Fangfa is assigned based on fuzzy hierarchy and the power distribution network diagnostic assessment index weights of CRITIC method | |
| CN109670676A (en) | Distributing net platform region method for prewarning risk and system based on Support Vector data description | |
| CN114519514B (en) | Low-voltage transformer area reasonable line loss value measuring and calculating method, system and computer equipment | |
| CN111429034A (en) | Method for predicting power distribution network fault | |
| CN109934455A (en) | A cost assessment method for power grid engineering | |
| CN113189418A (en) | Topological relation identification method based on voltage data | |
| CN109002933A (en) | Distribution wire route based on ReliefF and t-SNE becomes relational model optimization method | |
| CN111612296B (en) | Method for quantitatively configuring online monitoring device of power equipment of converter station | |
| CN115545514A (en) | Health degree evaluation-based differentiated operation and maintenance method and device for power distribution fusion equipment | |
| CN113011084B (en) | Intelligent identification method for power transmission line fault reasons based on correlation vector machine | |
| CN114091549A (en) | Equipment fault diagnosis method based on deep residual error network | |
| CN110378610B (en) | Identification method of weak links in distribution network based on user importance and equipment operating status | |
| CN114386753A (en) | Equipment risk comprehensive analysis early warning method based on main transformer load condition | |
| CN117540344A (en) | Distribution transformer anomaly monitoring and operation trend prediction method, equipment and medium | |
| CN115577978A (en) | Power distribution network target investment decision element weight coefficient measuring and calculating method | |
| CN116953381A (en) | Aging failure rate assessment method of distribution transformers based on recovery-aggregation-decomposition | |
| CN116090872A (en) | Power distribution area health state evaluation method | |
| CN107947156A (en) | Based on the electric network fault critical clearing time method of discrimination for improving Softmax recurrence | |
| CN115600928A (en) | OAO-RVM-Based Evaluation Method of GIS Potty Insulator State |
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:20200117 |