CN108802565B - Medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning - Google Patents
Medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning Download PDFInfo
- Publication number
- CN108802565B CN108802565B CN201810401025.0A CN201810401025A CN108802565B CN 108802565 B CN108802565 B CN 108802565B CN 201810401025 A CN201810401025 A CN 201810401025A CN 108802565 B CN108802565 B CN 108802565B
- Authority
- CN
- China
- Prior art keywords
- distribution network
- machine learning
- disconnection
- data
- fault detection
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a machine learning-based method for detecting disconnection and non-grounding faults of a medium-voltage distribution network, which comprises the following steps of: s1, extracting daily electric quantity data from the distribution transformer level; s2, preprocessing the extracted data, including performing a trend removing item and periodic processing on the time sequence; and S3, classifying the preprocessed data by adopting a random forest algorithm based on feature selection to obtain a fault detection result. Compared with the prior art, the method is based on the power distribution and utilization information system data and the machine learning algorithm, a pure data driving method is used, a set of detection method for intelligently detecting the disconnection of the power distribution network is provided, through multiple comparison, the classification accuracy of the method is high, and parameters of main standards for judging the disconnection of the power distribution network can be selected.
Description
Technical Field
The invention relates to a power distribution network fault diagnosis technology, in particular to a medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning.
Background
The power system is the core part of the urban infrastructure, and the medium voltage power grid is the center for bearing loads and is an important component of the power distribution network of the power system. The coverage area is wide, and the required power resources can be directly provided for cities and areas thereof. The importance of the medium voltage distribution network as an important component of an urban power supply system is self-evident. Because a medium-voltage distribution network transformer substation usually adopts a neutral point ungrounded mode, when the conditions that the wires on two sides of a fracture are not grounded after a single-phase line of a line is broken or the wires on the non-power side fall to the ground and the like occur, no obvious fault characteristic is generated, and the fault cannot be detected through the existing relay protection device in the transformer substation.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a method for detecting the disconnection and non-grounding faults of a medium voltage distribution network based on machine learning.
The purpose of the invention can be realized by the following technical scheme:
a medium voltage distribution network disconnection and non-grounding fault detection method based on machine learning comprises the following steps:
s1, extracting daily electric quantity data from the distribution transformer level;
s2, preprocessing the extracted data, including performing a trend removing item and periodic processing on the time sequence;
and S3, classifying the preprocessed data by adopting a random forest algorithm based on feature selection to obtain a fault detection result.
Preferably, the daily electricity consumption data is extracted in step S1, and then electricity consumption data that does not meet the preprocessing requirement is screened out.
Preferably, the step S2 specifically includes:
s21, extracting a time sequence of time periods corresponding to three weeks before each transformer;
s22, taking the average value of each window corresponding to the time points in the previous three-week time sequence as a template;
s23, subtracting the template from the original time sequence to obtain a preprocessed time sequence;
and S24, performing stability test.
Preferably, the feature selection-based random forest algorithm specifically includes:
and (3) arranging the features in a descending order by using the variable importance measurement result of the random forest, taking the top p% as a feature subset, randomly classifying a training set and a testing set, and calculating the accuracy after multiple averaging, and the mean value and the standard deviation of the AUC value.
Preferably, the proportion of the randomly-divided training set to the test set is: 70% as training set and 30% as test set.
Preferably, the daily electricity consumption data includes: voltage, current, active power and reactive power of the distribution transformer.
Compared with the prior art, the method is based on the power distribution and utilization information system data and the machine learning algorithm, a pure data driving method is used, a set of detection method for intelligently detecting the disconnection of the power distribution network is provided, through multiple comparison, the classification accuracy of the method is high, and parameters of main standards for judging the disconnection of the power distribution network can be selected.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.
As shown in fig. 1, a method for detecting disconnection and non-grounding faults of a medium voltage distribution network based on machine learning includes the following steps:
s1, extracting daily electric quantity data from the distribution transformer level;
s2, preprocessing the extracted data, including performing a trend removing item and periodic processing on the time sequence;
and S3, classifying the preprocessed data by adopting a random forest algorithm based on feature selection to obtain a fault detection result.
And (S1) extracting the daily electricity consumption data and screening out electricity consumption data which do not meet the preprocessing requirement.
Step S2 specifically includes:
s21, extracting a time sequence of time periods corresponding to three weeks before each transformer;
s22, taking the average value of each window corresponding to the time points in the previous three-week time sequence as a template;
s23, subtracting the template from the original time sequence to obtain a preprocessed time sequence;
and S24, performing stability test.
The random forest algorithm based on feature selection specifically comprises the following steps: the method comprises the following steps of utilizing variable importance measurement results of random forests to arrange the features in a descending order, taking the first p% as a feature subset, then randomly dividing a training set and a testing set, taking 70% as the training set and taking 30% as the testing set, and calculating the accuracy after multiple averaging, and the mean value and the standard deviation of AUC values, wherein the specific process is as follows:
inputting: an original data set S;
and (3) outputting: the classification accuracy, the mean value and the standard deviation of the AUC values and the importance Rank of the corresponding feature set;
step1, initializing, and setting the repeated training times N1And N2,Fori=1:N1;
Step2, randomly extracting K training subsets in a bootstrap with a set back in a starting data set S, and constructing K independent classification regression trees according to the K training subsets, wherein the complement of each training subset is marked as out-of-bag (OOB);
step3 random drawing m in each node of each tree for m featurestryA feature, generating a featureCalculating the Kini index of each feature, and selecting the feature with the minimum Kini index in the feature subset as a splitting feature;
step4, growing each tree to the maximum extent without pruning;
step5, averaging the importance of each feature, arranging the importance in descending order to obtain Rank, taking the top p% as a feature subset, and Forj being 1: N2;
Step6, separating 70% of training sets and 30% of testing sets after randomly scrambling the data set S, and repeating Step 2-4;
step7, forming the K trees into a random forest, and determining the classification result according to the voting of a tree classifier;
step8, calculating the importance of each feature and the accuracy and AUC value of the model in the test set;
step9 mean and standard deviation of calculated accuracy and AUC values.
Wherein, usually, m is takentry=floor(log2(m)+1。
The daily electricity consumption data includes: voltage, current, active power and reactive power of the distribution transformer. The efficiency of the active and reactive power data is low and the 10kV feeder current typically only collects one phase, so the relevant characteristic quantities of the active and reactive power and its rate of change are not considered.
Examples
In the embodiment, the power distribution information system data of 2016 to 2017 in a certain area in east China are used for diagnosing the branch line disconnection and non-grounding fault of the medium-voltage distribution network. Current, power, voltage A, B, C three-phase power data is extracted from the distribution transformer hierarchy. The reason for this is to accurately position the branches of the distribution transformer. After the power consumption data which do not meet the preprocessing requirement are screened out, 32 line breaking records are obtained, and the three-phase power consumption data correspond to 604 distribution transformers. Meanwhile, three-phase electricity consumption data of one circle before the disconnection of the 604 distribution transformers is extracted to be used as comparison of normal electricity consumption.
For data preprocessing, as can be seen from the time sequence of the three-phase power utilization data, a certain trend and periodicity exist. To eliminate these two effects, the time series is processed with a detrending term and periodically. After the pre-treatment, a smooth time series is obtained.
Parameters in the random forest algorithm selected based on the characteristics are respectively 500 ntree and 7 mtry. Through multiple experiments, the average importance of the characteristics is found to be less sensitive to the values of ntree and mtry, and the influence of parameter change on the model is weaker.
The result of feature selection shows that the voltage characteristic has better disconnection detection capability. Therefore, Logistic regression, SVC and the random forest algorithm described above are used below to compare the classification results.
First, the AUC, Accuracy (ACC), sensitivity (TPR) and specificity (TNR) of the three classifier results were investigated.
70% of the data from the sample set were randomly drawn as a training set and 30% as a test set, and repeated 1000 times to obtain an average of the predicted results. The classification results were measured from AUC values, Accuracy (ACC), sensitivity (TPR) and specificity (TNR) of the three classifications. 70% of the data from the sample set were randomly drawn as a training set and 30% as a test set, and repeated 1000 times to obtain an average of the predicted results.
Claims (4)
1. A medium voltage distribution network disconnection ungrounded fault detection method based on machine learning is characterized by comprising the following steps:
s1, extracting daily electric quantity data from the distribution transformer level;
s2, preprocessing the extracted data, including performing a trend removing item and periodic processing on the time sequence;
s3, classifying the preprocessed data by using a random forest algorithm based on feature selection to obtain a fault detection result;
the step S2 specifically includes:
s21, extracting a time sequence of time periods corresponding to three weeks before each transformer;
s22, taking the average value of each window corresponding to the time points in the previous three-week time sequence as a template;
s23, subtracting the template from the original time sequence to obtain a preprocessed time sequence;
s24, carrying out stability test;
the random forest algorithm based on feature selection specifically comprises the following steps:
and (3) arranging the features in a descending order by using the variable importance measurement result of the random forest, taking the top p% as a feature subset, randomly classifying a training set and a testing set, and calculating the accuracy after multiple averaging, and the mean value and the standard deviation of the AUC value.
2. The machine learning-based disconnection and non-grounding fault detection method for the medium-voltage distribution network, according to claim 1, is characterized in that after daily electricity consumption data are extracted in the step S1, electricity consumption data which do not meet preprocessing requirements are screened out.
3. The machine learning-based method for detecting the disconnection and non-grounding fault of the medium voltage distribution network according to claim 1, wherein the proportion of the randomly-divided training set to the test set is as follows: 70% as training set and 30% as test set.
4. The machine learning-based disconnection and non-grounding fault detection method for the medium-voltage distribution network, according to claim 1, is characterized in that the daily electricity consumption data comprises: voltage, current, active power and reactive power of the distribution transformer.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810401025.0A CN108802565B (en) | 2018-04-28 | 2018-04-28 | Medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810401025.0A CN108802565B (en) | 2018-04-28 | 2018-04-28 | Medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108802565A CN108802565A (en) | 2018-11-13 |
CN108802565B true CN108802565B (en) | 2020-09-01 |
Family
ID=64093518
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810401025.0A Active CN108802565B (en) | 2018-04-28 | 2018-04-28 | Medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108802565B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110119755A (en) * | 2019-03-22 | 2019-08-13 | 国网浙江省电力有限公司信息通信分公司 | Electricity method for detecting abnormality based on Ensemble learning model |
CN114626487B (en) * | 2022-05-16 | 2023-09-05 | 南昌工程学院 | Linear transformation relation checking method based on random forest classification algorithm |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2915632A1 (en) * | 2007-04-24 | 2008-10-31 | Transfix Toulon Sa Sa Soc Nouv | Low voltage front link protecting method, involves setting electric screen on low voltage link, and connecting electric screen with three-phase disconnection and independent stiker |
WO2011117495A1 (en) * | 2010-03-22 | 2011-09-29 | Peugeot Citroën Automobiles SA | Non-intrusive device for diagnosis, by vectorial comparison, of operating situation(s) in at least one communication network |
CN103136587A (en) * | 2013-03-07 | 2013-06-05 | 武汉大学 | Power distribution network operating state classification recognition method based on support vector machine |
CN103308823A (en) * | 2013-05-23 | 2013-09-18 | 国家电网公司 | Method for locating single-phase disconnection non-ground fault of power distribution network |
CN105866615A (en) * | 2016-06-16 | 2016-08-17 | 山东大学 | 10kV power distribution line disconnection fault judging method based on three-phase voltage current |
CN106569093A (en) * | 2016-11-02 | 2017-04-19 | 南方电网科学研究院有限责任公司 | Fault discrimination method for same-tower multi-loop power transmission line |
CN107015113A (en) * | 2017-04-11 | 2017-08-04 | 国网河南省电力公司电力科学研究院 | The power distribution network broken string recognition methods compared is mutated based on forward-order current |
-
2018
- 2018-04-28 CN CN201810401025.0A patent/CN108802565B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2915632A1 (en) * | 2007-04-24 | 2008-10-31 | Transfix Toulon Sa Sa Soc Nouv | Low voltage front link protecting method, involves setting electric screen on low voltage link, and connecting electric screen with three-phase disconnection and independent stiker |
WO2011117495A1 (en) * | 2010-03-22 | 2011-09-29 | Peugeot Citroën Automobiles SA | Non-intrusive device for diagnosis, by vectorial comparison, of operating situation(s) in at least one communication network |
CN103136587A (en) * | 2013-03-07 | 2013-06-05 | 武汉大学 | Power distribution network operating state classification recognition method based on support vector machine |
CN103308823A (en) * | 2013-05-23 | 2013-09-18 | 国家电网公司 | Method for locating single-phase disconnection non-ground fault of power distribution network |
CN105866615A (en) * | 2016-06-16 | 2016-08-17 | 山东大学 | 10kV power distribution line disconnection fault judging method based on three-phase voltage current |
CN106569093A (en) * | 2016-11-02 | 2017-04-19 | 南方电网科学研究院有限责任公司 | Fault discrimination method for same-tower multi-loop power transmission line |
CN107015113A (en) * | 2017-04-11 | 2017-08-04 | 国网河南省电力公司电力科学研究院 | The power distribution network broken string recognition methods compared is mutated based on forward-order current |
Non-Patent Citations (2)
Title |
---|
基于改进的AdaBoost算法的中压配电网断线不接地故障检测;郭乃网;《电测与仪表》;20190825;第56卷(第16期);第1-6,18页 * |
基于机器学习的中压配电网断线不接地故障检测;苏运;《电测与仪表》;20190110;第56卷(第1期);第96-101页 * |
Also Published As
Publication number | Publication date |
---|---|
CN108802565A (en) | 2018-11-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhao et al. | Data-driven correction approach to refine power curve of wind farm under wind curtailment | |
CN108376982B (en) | Load phase sequence identification method and device | |
CN109507535B (en) | Method and device for predicting operation stage and operation life of transformer substation grounding grid | |
CN105894177B (en) | Power equipment operation risk analysis and evaluation method based on decision tree algorithm | |
CN107422266B (en) | Fault diagnosis method and device for high-capacity battery energy storage system | |
CN104865499A (en) | Super-high voltage direct-current power transmission line region internal and external fault identification method | |
CN106469356A (en) | Transmission facility state evaluation based on multidimensional data and risk analysis method and device | |
Rezgui et al. | A smart algorithm for the diagnosis of short-circuit faults in a photovoltaic generator | |
Zhang et al. | A weighted random forest approach to improve predictive performance for power system transient stability assessment | |
CN108802565B (en) | Medium-voltage power distribution network disconnection ungrounded fault detection method based on machine learning | |
CN106953379A (en) | A kind of balance control method and device of energy-storage battery state-of-charge in parallel | |
CN107679089A (en) | A kind of cleaning method for electric power sensing data, device and system | |
CN104331532A (en) | Power transformer state evaluation method based on rough set-cloud model | |
CN108847682A (en) | A kind of grid-connected performance estimating method of photo-voltaic power generation station based on operation data and system | |
CN110348683A (en) | The main genetic analysis method, apparatus equipment of electrical energy power quality disturbance event and storage medium | |
CN113922412A (en) | Panorama evaluation method and system for new energy multi-station short circuit ratio, storage medium and computing device | |
CN116128467A (en) | Power utilization system maintenance method, device, computer equipment and storage medium | |
CN112633315A (en) | Electric power system disturbance classification method | |
Serrano-Luján et al. | Case of study: Photovoltaic faults recognition method based on data mining techniques | |
CN107368979B (en) | Method and device for predicting disturbance event of weather-sensitive regional power distribution network | |
CN104484828A (en) | High-voltage electrical appliance operation quality evaluation method and system based on logistic regression | |
CN116502149A (en) | Low-voltage power distribution network user-transformation relation identification method and system based on current characteristic conduction | |
Hu et al. | Operational reliability evaluation method based on big data technology | |
CN116879672A (en) | Intelligent identification method for total fault of area based on time convolution network | |
Qiu et al. | Comprehensive evaluation of power quality based on improved TOPSIS-RSR method |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |