CN104133143A - Power grid line fault diagnosis system and method based on Hadoop cloud computing platform - Google Patents

Abstract

The present invention provides a grid line fault diagnosis system and method based on the Hadoop cloud computing platform. The system includes a voltage transformer, a current transformer, a transmitter, a remote terminal, a modem, an optical fiber, and a plurality of upper computers; the method includes: Collect the voltage and current on the power grid line in real time and transmit them to remote terminals; when the power grid fails, obtain the electrical quantity information of the power grid and send it to the NameNode node in the Hadoop cluster, and then distribute it to the DateNode node for parallel processing to diagnose the power grid fault and display. The present invention makes full use of the electric quantity fault information of the power grid line, and calculates the wavelet signal rate, wavelet failure rate and wavelet variation rate of current and voltage through the reconstructed results of wavelet change based on multi-resolution analysis, and fuses them into a preliminary reliable distribution function. Preliminary classification of the faulty grid line set and non-faulty grid line set is carried out, and then the faulty grid line set is finally determined, so that the faulty grid line can be judged more accurately.

Description

A system and method for power grid line fault diagnosis based on Hadoop cloud computing platform

technical field

The invention belongs to the technical field of power grid fault diagnosis, in particular to a power grid line fault diagnosis system and method based on a Hadoop cloud computing platform.

Background technique

Fault diagnosis of power grid is an old-fashioned problem for power system. At present, there are many researches on power grid fault diagnosis at home and abroad, and all of them have their unique characteristics in diagnosis. There are mainly the following methods: fault diagnosis method based on expert system, fault diagnosis method based on artificial neural network, fault diagnosis method based on fuzzy set theory, fault diagnosis method based on Petri net theory, fault diagnosis method based on MAS theory, etc. . The diagnosis platform is mostly a processing platform of a computer in the main station. These methods have their own defects. Massive data, the system will become overwhelmed, or even collapse. No matter how perfect the algorithm is, it cannot be effectively implemented, and it is impossible to diagnose faults in a timely, accurate and fast manner, and restore power supply as soon as possible.

Contents of the invention

Aiming at the deficiencies in the prior art, the present invention provides a power grid line fault diagnosis system and method based on Hadoop cloud computing platform.

Technical scheme of the present invention is:

A power grid line fault diagnosis system based on the Hadoop cloud computing platform, including voltage transformers, current transformers, transmitters, remote terminals, modems, optical fibers and multiple upper computers;

The input terminal of the voltage transformer and the input terminal of the current transformer are respectively connected to the grid line, the output terminal of the voltage transformer and the output terminal of the current transformer are respectively connected to the input terminal of the transmitter, and the output terminal of the transmitter is connected to the remote The input terminal of the terminal, the remote terminal communicates with multiple host computers through a modem;

Multiple host computers form a Hadoop cluster, one of which is the NameNode node, and the rest are DateNode nodes; each host computer is equipped with Hbase, HDFS and Map/Reduce, and the Hadoop cluster is the Hadoop cloud computing platform.

The NameNode node is provided with a data monitoring unit.

Adopt the grid fault diagnosis method of the power grid line fault diagnosis system based on Hadoop cloud computing platform, comprise the following steps:

Step 1: The voltage transformer and current transformer respectively collect the voltage and current on each grid line in real time, and transmit them to the remote terminal through the transmitter;

Step 2: When the power grid fails, the remote terminal obtains the electrical quantity information of the grid, and transmits the electrical quantity information to the NameNode node in the Hadoop cluster through the modem, and the NameNode node distributes the electrical quantity information of each grid line to each DateNode node HDFS stores files, and then stores them in Hbase in each DateNode node;

Step 3: Map/Reduce in each DateNode node processes the electrical quantity information of the grid line stored in Hbase in parallel, and performs grid fault diagnosis;

Step 3.1: Perform multi-resolution analysis wavelet transform on the electrical quantity information of each grid line and extract fault features, including wavelet signal rate, wavelet failure rate and wavelet variation rate;

Step 3.1.1: Perform multi-resolution analysis wavelet transform on the electrical quantity information of each grid line respectively, and obtain the reconstructed wavelet transform results of the current and voltage of each grid line:

D _ic1 , D _ic2 ... D _icl is the wavelet transform result of the current signal of each grid line, D _iv1 , D _iv2 ... D _ivl is the wavelet transform result of the voltage signal of each grid line, where l represents the number of sampling points of the signal D _ic1 , D _ic2 ... D _ick is the wavelet transform result of the current signal before each power grid line fault, D _ic(k+1) , D _ic(k+2) ... D _icl is the current signal wavelet transformation result after each power grid line fault The wavelet transform result of the current signal; D _iv1 , D _iv2 ... D _ivk is the wavelet transform result of the voltage signal before each grid line fault, D _iv(k+1) , D _iv(k+2) ... D _ivl is Wavelet transform results of voltage signals after each power grid line fault;

Step 3.1.2: The wavelet transform coefficient matrix corresponding to the obtained wavelet transform result is calculated by singular value decomposition theory to obtain the singular value characteristic matrix Λ _ic =diag(λ _1c ,λ _2c , ...λ _pc ) and the singular value characteristic matrix Λ iv of the voltage wavelet transform coefficient matrix Λ _iv =diag(λ _1v ,λ _2v ,...λ _pv ), the singular value characteristic matrix represents the basic characteristics of the wavelet transform coefficient matrix, and p is the singular value The minimum value of the order of the square matrix on the left and right sides of the characteristic matrix;

Step 3.1.3: Calculate the wavelet signal rate of the current signal and the wavelet signal rate of the voltage signal after the fault occurs according to the current signal and voltage signal of the grid line respectively;

Wavelet signal rate of the current signal of the i-th grid line after the fault occurs: s _ic represents the intensity of the current signal energy of the i-th power grid line, W _ic represents the arithmetic mean value of the current signal distribution of the i-th power grid line, and the arithmetic mean value of the current signal distribution of any power grid line under the m scale Among them, the wavelet signal energy distribution of the current signal of any grid line at the j scale: D _jc(k) represents the wavelet transform result of the current signal at time k at the jth scale;

The wavelet signal rate of the voltage signal of the i-th grid line after the fault occurs: s _iv represents the intensity of the voltage signal energy of the i-th power grid line, W _iv represents the arithmetic mean value of the voltage signal energy distribution of the i-th power grid line; the arithmetic mean value of the voltage signal energy distribution of any power grid line under m scales: The wavelet signal energy distribution of the voltage signal of any grid line at the j scale: D _jv(k) represents the wavelet transform result of the voltage signal at time k at the jth scale;

Step 3.1.4: Calculate the current signal wavelet failure rate and voltage signal wavelet failure rate after the fault occurs according to the current signal and voltage signal of the grid line;

The wavelet failure rate of the current signal of the i-th grid line after the fault occurs V _ic represents the change degree of the current signal of the i-th grid line before and after the fault: in, $\left\{\begin{array}{c}{m}_{\mathrm{icq}}=\max \{{\mathrm{D.}}_{\mathrm{ic}1},{\mathrm{D.}}_{\mathrm{ic}2}......{\mathrm{D.}}_{\mathrm{ick}}\}\\ m_{\mathrm{ich}}=\max \{{\mathrm{D.}}_{\mathrm{ic}(k+1)},{\mathrm{D.}}_{\mathrm{ic}(k+2)}......{\mathrm{D.}}_{\mathrm{icl}}\}\end{array}\right.,$ M _icq represents the maximum value of the wavelet transformation result of the current signal {D _ic1 , D _ic2 ... D _ick } of the i-th grid line before the fault, and M _ich represents the current signal of the i-th grid line after the fault {D _{ic( k+1)} ,D _ic(k+2) ……D _icl }The maximum value of the wavelet transform result;

The wavelet failure rate of the voltage signal of the i-th grid line after the fault occurs V _iv represents the change degree of the current signal of the i-th grid line before and after the fault, where $\left\{\begin{array}{c}{m}_{\mathrm{ivq}}=\max \{{\mathrm{D.}}_{\mathrm{iv}1},{\mathrm{D.}}_{\mathrm{iv}2}......{\mathrm{D.}}_{\mathrm{ivk}}\}\\ m_{\mathrm{ivh}}=\max \{{\mathrm{D.}}_{\mathrm{iv}(k+1)},{\mathrm{D.}}_{\mathrm{iv}(k+2)}......{\mathrm{D.}}_{\mathrm{ivl}}\}\end{array}\right.,$ M _ivq represents the maximum value of the wavelet transformation result of the voltage signal {D _iv1 , D _iv2 ... D _ivk } of the i-th power grid line before the fault, and M _ivh represents the voltage signal of the i-th power grid line after the fault {D _{iv( k+1)} ,D _iv(k+2) ……D _ivl }The maximum value of the wavelet transform result;

Step 3.1.5: Calculate the wavelet variation rate of the current signal and the wavelet variation rate of the voltage signal after the fault occurs according to the current signal and voltage signal of the grid line;

(1) The wavelet variation rate of the current signal of the i-th grid line after the fault occurs B _ic is the arithmetic mean of the diagonal elements of the singular value eigenmatrix of the wavelet transform coefficient matrix of the current signal of the i-th grid line; Arithmetic mean $B_c=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{ic}}}{p}};$

(2) The wavelet variation rate of the voltage signal of the i-th grid line after the fault occurs B _iv is the arithmetic mean of the diagonal elements of the singular value eigenmatrix of the wavelet transform coefficient matrix of the voltage signal of the i-th grid line; Arithmetic mean $B_v=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{iv}}}{p}};$

Step 3.2: Calculate the fault support degree of each grid line according to the wavelet signal rate, wavelet failure rate and wavelet variation rate;

The fault support degree of the i-th grid line ${\mathrm{the\; y}}_i=\sqrt{\frac{{\mathrm{\ρ}}_{\mathrm{the\; s}}\sqrt{\frac{{{\mathrm{the\; s}}_{\mathrm{ic}}}^2+{{\mathrm{the\; s}}_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_x\sqrt{\frac{{x_{\mathrm{ic}}}^2+{x_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_m\sqrt{\frac{{m_{\mathrm{ic}}}^2+{m_{\mathrm{iv}}}^2}{2}}}{3}}(i=1...\mathrm{no}),$ Fault support coefficients ρ _s , ρ _x , ρ _m are all constants;

Step 3.3: Establish a preliminary trust distribution function according to the fault support degree of each grid line;

Preliminary trusted distribution function of the i-th grid line Among them, θ is the uncertainty, θ=0.1～0.3;

Step 3.4: Preliminarily classify the faulty grid line set and non-faulty grid line set for each grid line by Stirling formula, and then use the fuzzy C-means clustering method to accurately divide the faulty grid line set and non-faulty grid line set , and finally determine the faulty grid line set;

Step 4: Each DateNode node transmits the determined set of faulty grid lines to the NameNode node for grid line fault diagnosis result display.

Beneficial effect:

The present invention proposes for the first time that the grid fault diagnosis platform is changed from the traditional calculation mode based on a computer to the calculation mode of the Hadoop-based cloud computing platform, so that compared with the traditional calculation mode, the Hadoop cloud computing platform provides the most reliable, The safest data storage center, users no longer have to worry about data loss, virus intrusion, etc.; the minimum requirements for client equipment, easy to use; with parallel computing as the core, on-demand scheduling of computing tasks and computing resources, and provide data from the Import and integrate processing, calculation model setting to calculation result output, multi-form display and other complete data processing services; use distributed storage system, data mutual backup, fast backup and recovery, support various data processing, calculation models, to meet different fields, Computing requirements with different characteristics; multiple copies of fault tolerance, data security and worry-free, massive storage, unlimited space; simple configuration, complete platform, no need to spend a lot of time building and maintaining the computing environment; using computing and storage resources as a service, according to Need to take, through virtualization technology, even without adding new computing power, it can usually effectively improve the utilization rate of physical machine hardware.

The proposal of the present invention has many benefits such as high efficiency, low cost, and good reliability in the grid line fault diagnosis process, so it has certain practical significance.

The method of the present invention makes full use of the electrical quantity fault information of the power grid line, and calculates the wavelet signal rate, wavelet failure rate and wavelet variation rate of current and voltage through the reconstructed results of wavelet changes based on multi-resolution analysis, and analyzes faults from multiple angles The data are fused into a preliminary trustworthy distribution function for judgment and analysis, and the preliminary classification of faulty grid line sets and non-faulty grid line sets is carried out for each grid line through the Stirling formula, and then the fuzzy C-means clustering method is used to carry out The precise division of the faulty grid line set and the non-faulty grid line set finally determines the faulty grid line set, which can accurately determine the faulty grid line. And applying this method to Hadoop cloud computing makes the diagnosis process faster. The more complex the power grid is, the more power grid lines there are, and the advantages compared with the traditional computing mode are more obvious.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some examples recorded in the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without creative work.

Fig. 1 is the structural block diagram of the grid line fault diagnosis system based on Hadoop cloud computing platform of the embodiment of the present invention;

Fig. 2 is the flowchart of the power grid line fault diagnosis method based on Hadoop cloud computing platform according to the embodiment of the present invention

Fig. 3 is the multi-resolution analysis wavelet transform of the specific embodiment of the present invention and extracts fault feature flowchart;

Fig. 4 is a traditional Hadoop cloud computing platform structure diagram;

Fig. 5 is the analysis diagram based on Hadoop cloud computing platform of the embodiment of the present invention;

Fig. 6 is the fault waveform and wavelet decomposition analysis diagram of L4 in the specific embodiment of the present invention;

(a) is the sampling current waveform diagram of L4;

(b) is the sampling voltage waveform diagram of L4;

(c) is the sampling current wavelet decomposition waveform diagram of L4;

(d) is the wavelet decomposition waveform diagram of the sampling voltage of L4;

Fig. 7 is the fault waveform and wavelet decomposition analysis diagram of L6 in the specific embodiment of the present invention;

(a) is the sampling current waveform diagram of L6;

(b) is the sampling voltage waveform diagram of L6;

(c) is the sampling current wavelet decomposition waveform diagram of L6;

(d) is the wavelet decomposition waveform diagram of the sampling voltage of L6;

Fig. 8 is the fault waveform and wavelet decomposition analysis diagram of L8 in the specific embodiment of the present invention;

(a) is the sampling current waveform diagram of L8;

(b) is the sampling voltage waveform diagram of L8;

(c) is the sampling current wavelet decomposition waveform diagram of L8;

(d) is the wavelet decomposition waveform diagram of the sampling voltage of L8.

Detailed ways

In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

A power grid line fault diagnosis system based on the Hadoop cloud computing platform, as shown in Figure 1, includes a voltage transformer, a current transformer, a transmitter, a remote terminal, a modem, an optical fiber, and a plurality of upper computers;

The specific models of voltage transformers, current transformers, and transmitters should be specified according to the voltage level and current size;

The modem adopts GQ-DOT02C2 RS485 serial port optical transceiver;

The optical fiber adopts GYSTY53/GYTY53 double-sheathed single-armoured stranded optical fiber;

The upper computer adopts Lenovo Yangtian M3320n-00 desktop computer;

The input terminal of the voltage transformer and the input terminal of the current transformer are respectively connected to the grid line, the output terminal of the voltage transformer and the output terminal of the current transformer are respectively connected to the input terminal of the transmitter, and the output terminal of the transmitter is connected to the remote The input terminal of the terminal, the remote terminal communicates with multiple host computers through a modem;

The traditional Hadoop cloud computing platform structure is shown in Figure 4. It can be seen from the figure that the Hadoop cloud computing platform includes HDFS, Hbase, Map/Reduce, cluster computer operating system (windows) and virtualization software (Cygwin; JDK; Ant ; TortoiseSVN; Eclipse), server cluster.

HDFS is a distributed file system that can run on clusters composed of common and cheap hardware infrastructures. It stores massive data in a streaming data access mode. HDFS is highly fault tolerant and capable of handling large amounts of data. HDFS adopts master/slave mode, which consists of a NameNode node and multiple DateNode nodes. HDFS files are divided into multiple data blocks for access in data nodes, so HDFS is very suitable for the application of large data sets.

Hbase is a distributed database in Hadoop environment. Based on the distributed file system HDFS, it can provide real-time read and write and random access to large data sets. The Hbase database can handle very large tables, and can process more than 1 billion rows of data through ordinary computers, and there are data tables composed of hundreds of outer column elements.

The operating system of the node computers in the Hadoop cluster is the Windows operating system. Virtualization software refers to Cygwin, JDK, Ant, TortoiseSVN, Eclipse, and only in this environment can the Hadoop code be written and modified.

The server cluster refers to the computers of each node running on the entire Hadoop platform, including a NameNode node and multiple DateNode nodes.

In this embodiment, a plurality of upper computers form a Hadoop cluster, one of which is used as a NameNode node, and the rest are DateNode nodes; each upper computer is provided with Hbase, HDFS and Map/Reduce, and the Hadoop cluster is the Hadoop cloud computing platform. The NameNode node is provided with a data monitoring unit.

The power grid fault diagnosis method using the power grid line fault diagnosis system based on the Hadoop cloud computing platform, as shown in Figure 2, includes the following steps:

Step 1: The voltage transformer and current transformer respectively collect the voltage and current on each grid line in real time, and transmit them to the remote terminal through the transmitter;

Step 2: When the power grid fails, the remote terminal obtains the electrical quantity information of the grid, and transmits the electrical quantity information to the NameNode node in the Hadoop cluster through the modem, and the NameNode node distributes the electrical quantity information of each grid line to each DateNode node HDFS stores files, and then stores them in Hbase in each DateNode node;

Step 3: Map/Reduce in each DateNode node processes the electrical quantity information of the grid line stored in Hbase in parallel, and performs grid fault diagnosis;

Step 3.1: Perform multi-resolution analysis wavelet transform on the electrical quantity information of each grid line and extract fault features, including wavelet signal rate, wavelet failure rate and wavelet variation rate; as shown in Figure 3:

Step 3.1.1: Perform multi-resolution analysis wavelet transform on the electrical quantity information of each grid line respectively, and obtain the reconstructed wavelet transform results of the current and voltage of each grid line:

D _ic1 , D _ic2 ... D _icl is the wavelet transform result of the current signal of each grid line, D _iv1 , D _iv2 ... D _ivl is the wavelet transform result of the voltage signal of each grid line, where l represents the number of sampling points of the signal D _ic1 , D _ic2 ... D _ick is the wavelet transform result of the current signal before each power grid line fault, D _ic(k+1) , D _ic(k+2) ... D _icl is the current signal wavelet transformation result after each power grid line fault The wavelet transform result of the current signal; D _iv1 , D _iv2 ... D _ivk is the wavelet transform result of the voltage signal before each grid line fault, D _iv(k+1) , D _iv(k+2) ... D _ivl is Wavelet transform results of voltage signals after each power grid line fault;

Step 3.1.2: The wavelet transform coefficient matrix corresponding to the obtained wavelet transform result is calculated by singular value decomposition theory to obtain the singular value characteristic matrix Λ _ic =diag(λ _1c ,λ _2c , ...λ _pc ) and the singular value characteristic matrix Λ _{iv of the voltage wavelet transformation coefficient matrix Λ iv} =diag(λ _1v ,λ _2v ,...λ _pv ), λ _1c ,λ _2c ,...λ _pc represent the The diagonal elements of the singular value characteristic matrix of the current wavelet transform coefficient matrix, λ _1v , λ _2v , ... λ _pv represent the diagonal elements of the singular value characteristic matrix of the voltage wavelet transform coefficient matrix of the i-th grid line, the singular value characteristic The matrix represents the basic characteristics of the wavelet transform coefficient matrix, and p is the minimum value of the order of the square matrix on the left and right sides of the singular value characteristic matrix;

Step 3.1.3: Calculate the wavelet signal rate of the current signal and the wavelet signal rate of the voltage signal after the fault occurs according to the current signal and voltage signal of the grid line respectively;

Wavelet signal rate of the current signal of the i-th grid line after the fault occurs: s _ic represents the intensity of the current signal energy of the i-th power grid line, W _ic represents the arithmetic mean value of the current signal distribution of the i-th power grid line, and the arithmetic mean value of the current signal distribution of any power grid line under the m scale Among them, the wavelet signal energy distribution of the current signal of any grid line at the j scale: $S_{\mathrm{jc}}=\underset{k}{\mathrm{\Σ}}\left|{\mathrm{D.}}_{\mathrm{jc}\left(k\right)}\right|(j=1......m,k=1......l),$ _Djc(k) ;

The wavelet signal rate of the voltage signal of the i-th grid line after the fault occurs: s _iv represents the intensity of the voltage signal energy of the i-th power grid line, W _iv represents the arithmetic mean value of the voltage signal energy distribution of the i-th power grid line; the arithmetic mean value of the voltage signal energy distribution of any power grid line under m scales: The wavelet signal energy distribution of the voltage signal of any grid line at the j scale:

Step 3.1.4: Calculate the current signal wavelet failure rate and voltage signal wavelet failure rate after the fault occurs according to the current signal and voltage signal of the grid line;

The wavelet failure rate of the current signal of the i-th grid line after the fault occurs V _ic represents the change degree of the current signal of the i-th grid line before and after the fault: in, $\left\{\begin{array}{c}{m}_{\mathrm{icq}}=\max \{{\mathrm{D.}}_{\mathrm{ic}1},{\mathrm{D.}}_{\mathrm{ic}2}......{\mathrm{D.}}_{\mathrm{ick}}\}\\ m_{\mathrm{ich}}=\max \{{\mathrm{D.}}_{\mathrm{ic}(k+1)},{\mathrm{D.}}_{\mathrm{ic}(k+2)}......{\mathrm{D.}}_{\mathrm{icl}}\}\end{array}\right.,$ M _icq represents the maximum value of the wavelet transformation result of the current signal {D _ic1 , D _ic2 ... D _ick } of the i-th grid line before the fault, and M _ich represents the current signal of the i-th grid line after the fault {D _{ic( k+1)} ,D _ic(k+2) ……D _icl }The maximum value of the wavelet transform result;

The wavelet failure rate of the voltage signal of the i-th grid line after the fault occurs V _iv represents the change degree of the voltage signal of the i-th grid line before and after the fault, where $\left\{\begin{array}{c}{m}_{\mathrm{ivq}}=\max \{{\mathrm{D.}}_{\mathrm{iv}1},{\mathrm{D.}}_{\mathrm{iv}2}......{\mathrm{D.}}_{\mathrm{ivk}}\}\\ m_{\mathrm{ivh}}=\max \{{\mathrm{D.}}_{\mathrm{iv}(k+1)},{\mathrm{D.}}_{\mathrm{iv}(k+2)}......{\mathrm{D.}}_{\mathrm{ivl}}\}\end{array}\right.,$ M _ivq represents the maximum value of the wavelet transformation result of the voltage signal {D _iv1 , D _iv2 ... D _ivk } of the i-th power grid line before the fault, and M _ivh represents the voltage signal of the i-th power grid line after the fault {D _{iv( k+1)} ,D _iv(k+2) ……D _ivl }The maximum value of the wavelet transform result;

Step 3.1.5: Calculate the wavelet variation rate of the current signal and the wavelet variation rate of the voltage signal after the fault occurs according to the current signal and voltage signal of the grid line;

(1) The wavelet variation rate of the current signal of the i-th grid line after the fault occurs B _ic is the arithmetic mean of the diagonal elements of the singular value eigenmatrix of the wavelet transform coefficient matrix of the current signal of the i-th grid line; Arithmetic mean $B_c=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{ic}}}{p}};$

(2) The wavelet variation rate of the voltage signal of the i-th grid line after the fault occurs B _iv is the arithmetic mean of the diagonal elements of the singular value eigenmatrix of the wavelet transform coefficient matrix of the voltage signal of the i-th grid line; Arithmetic mean $B_v=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{iv}}}{p}};$

Step 3.2: Calculate the fault support degree of each grid line according to the wavelet signal rate, wavelet failure rate and wavelet variation rate;

The fault support degree of the i-th grid line ${\mathrm{the\; y}}_i=\sqrt{\frac{{\mathrm{\ρ}}_{\mathrm{the\; s}}\sqrt{\frac{{{\mathrm{the\; s}}_{\mathrm{ic}}}^2+{{\mathrm{the\; s}}_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_x\sqrt{\frac{{x_{\mathrm{ic}}}^2+{x_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_m\sqrt{\frac{{m_{\mathrm{ic}}}^2+{m_{\mathrm{iv}}}^2}{2}}}{3}}(i=1...\mathrm{no}),$ Fault support coefficients ρ _s , ρ _x , ρ _m are all constants, generally ρ _s = 0.3, ρ _x = 1.9, ρ _m = 0.8;

Step 3.3: Establish a preliminary trust distribution function according to the fault support degree of each grid line;

Preliminary trusted distribution function of the i-th grid line Among them, θ is the uncertainty, θ=0.1～0.3;

Step 3.4: Preliminarily classify the faulty grid line set and non-faulty grid line set for each grid line by Stirling formula, and then use the fuzzy C-means clustering method to accurately divide the faulty grid line set and non-faulty grid line set , and finally determine the faulty grid line set;

(1) The initial trust distribution function of each grid line is used as the failure probability of the corresponding grid line, and the initial grid line is divided into two categories by calculating the Stirling function value for the failure probability of the grid line, and the faulty grid line set Φ ₁ and non-fault grid line set Φ ₂ ; if the Stirling function value is satisfied Among them, ε depends on the scale of the faulty grid, and it is classified into the faulty grid line set _Φ1 , and the remaining grid lines are classified into the non-faulty grid line set _Φ2 .

(2) Utilize the membership matrix U of the fuzzy C-means clustering method to establish the objective function of the fuzzy C-means clustering method;

The value of the elements in the membership matrix U of the fuzzy C-means clustering method is between 0 and 1, but the membership degree matrix U has a normalization regulation, that is, the sum of all elements u _f i of the membership degree matrix U is always equal to 1:

$\underset{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; c</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; fi</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&ForAll;</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}$

Then, the objective function of the fuzzy C-means clustering method is:

$\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; c</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; c</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; fi</span>}}^{\mathrm{<span\; class="notranslate">\; \&eta;</span>}}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; the\; fi</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In the formula, u _f i is between 0 and 1; c _f is the cluster center of fuzzy group f, d _fi =||c _f -w _i || is the distance between the fth cluster center and the ith data point Euclidean distance; and η∈[1,∞) is a weighted index, where η takes a value of 1.5<η<2.5.

(3) Obtain the necessary conditions for the objective function of the fuzzy C-means clustering method to reach the minimum value:

$c_f=\frac{\underset{i=1}{\overset{\mathrm{no}}{\mathrm{\&Sigma;}}}{u}_{\mathrm{the\; fi}}^{\mathrm{\&eta;}}{w}_i}{\underset{i=1}{\overset{\mathrm{no}}{\mathrm{\&Sigma;}}}{u}_{\mathrm{the\; fi}}^{\mathrm{\&eta;}}}$ and $u_{\mathrm{the\; fi}}=\frac{1}{\underset{k=1}{\overset{c}{\mathrm{\&Sigma;}}}{(d_{\mathrm{the\; fi}}/d_{\mathrm{the\; ki}})}^{2/(\mathrm{\&eta;}-1)}}$

(4) According to the necessary condition that the objective function of the fuzzy C-means clustering method reaches the minimum value, determine the cluster center c _i and the membership matrix U;

First, use random numbers between 0 and 1 to initialize the membership degree matrix U so that it satisfies the formula

Second, use Calculate f cluster centers c _f ;

Then, calculate the objective function value of the fuzzy C-means clustering method, if the objective function value is relative to the objective function value in the previous iteration

If the change in the value of the scalar function is less than the threshold γ=0.001, stop the iteration;

Finally, use Calculate a new membership matrix, and recalculate f cluster centers c _f .

After obtaining the membership degree matrix U, if u _1i >u _2i , then the grid line belongs to the set of faulty grid lines, and if not, it belongs to the set of non-faulty grid lines.

Step 4: Each DateNode node transmits the determined set of faulty grid lines to the NameNode node for grid line fault diagnosis result display.

In this embodiment, fault diagnosis is performed on the grid line shown in FIG. 5 . In the figure, there are 11 busbars (B1~B11), 10 lines (L1~L10), and 20 circuit breaker switches (CB1~CB20). Now that L4 is faulty, the main protection on both sides of the line L4 will trip the circuit breakers CB6 and CB14 after the fault; the second backup protection of the line L8 will malfunction and trip the circuit breaker CB17; at the same time, the protections on both sides of the line L6 will trip and the circuit breaker will be tripped Device CB8, CB9;

Figure 6, Figure 7, and Figure 8 are the fault waveforms and wavelet decomposition analysis diagrams of L4, L6, and L8 respectively.

The fault diagnosis analysis data table is as follows:

Table 1 Data analysis table of fault diagnosis process

When calculating ε=2, η=2, it can be seen from the calculated data in the table that u ₁₁ >u ₂₁ , u ₁₂ <u ₂₂ , u ₁₃ <u ₂₃ , and the fault line is L4. This embodiment is enough to illustrate this The invention is reliable and practical, and can be used for fault diagnosis of large-scale power grids.

The present invention is aimed at the background that massive data needs to be processed and fault decisions need to be made in time when a huge and complex power grid fails, and proposes a fault diagnosis system and method for power grid lines based on Hadoop cloud computing platform. Its main diagnosis algorithm is still based on multi-resolution analysis wavelet transform and fuzzy C-means clustering algorithm, which can diagnose faults quickly and accurately. Due to the low construction cost of the Hadoop cloud computing platform, it can process a large amount of data in parallel with high efficiency, which can save the huge loss caused by the power grid failure as much as possible.

The above description is only the specific implementation of the present application. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present application, some improvements and modifications can also be made. It should be regarded as the protection scope of this application.

Claims (3)

1. A grid line fault diagnosis system based on Hadoop cloud computing platform, is characterized in that: comprise voltage transformer, current transformer, transmitter, remote terminal, modem, optical fiber and a plurality of upper computers; The input terminal of the voltage transformer and the input terminal of the current transformer are respectively connected to the grid line, the output terminal of the voltage transformer and the output terminal of the current transformer are respectively connected to the input terminal of the transmitter, and the output terminal of the transmitter is connected to the remote The input terminal of the terminal, the remote terminal communicates with multiple host computers through a modem; Multiple host computers form a Hadoop cluster, one of which is the NameNode node, and the rest are DateNode nodes; each host computer is equipped with Hbase, HDFS and Map/Reduce, and the Hadoop cluster is the Hadoop cloud computing platform. 2. the grid line fault diagnosis system based on Hadoop cloud computing platform according to claim 1, is characterized in that: described NameNode node is provided with data monitoring unit. 3. adopt the grid fault diagnosis method of the power grid line fault diagnosis system based on Hadoop cloud computing platform claimed in claim 1, it is characterized in that: comprise the following steps: Step 1: The voltage transformer and current transformer respectively collect the voltage and current on each grid line in real time, and transmit them to the remote terminal through the transmitter; Step 2: When the power grid fails, the remote terminal obtains the electrical quantity information of the grid, and transmits the electrical quantity information to the NameNode node in the Hadoop cluster through the modem, and the NameNode node distributes the electrical quantity information of each grid line to each DateNode node HDFS stores files, and then stores them in Hbase in each DateNode node; Step 3: Map/Reduce in each DateNode node processes the electrical quantity information of the grid line stored in Hbase in parallel, and performs grid fault diagnosis; Step 3.1: Perform multi-resolution analysis wavelet transform on the electrical quantity information of each grid line and extract fault features, including wavelet signal rate, wavelet failure rate and wavelet variation rate; Step 3.1.1: Perform multi-resolution analysis wavelet transform on the electrical quantity information of each grid line respectively, and obtain the reconstructed wavelet transform results of the current and voltage of each grid line: D _ic1 , D _ic2 ... D _icl is the wavelet transform result of the current signal of each grid line, D _iv1 , D _iv2 ... D _ivl is the wavelet transform result of the voltage signal of each grid line, where l represents the number of sampling points of the signal D _ic1 , D _ic2 ... D _ick is the wavelet transform result of the current signal before each power grid line fault, D _ic(k+1) , D _ic(k+2) ... D _icl is the current signal wavelet transformation result after each power grid line fault The wavelet transform result of the current signal; D _iv1 , D _iv2 ... D _ivk is the wavelet transform result of the voltage signal before each grid line fault, D _iv(k+1) , D _iv(k+2) ... D _ivl is Wavelet transform results of voltage signals after each power grid line fault; Step 3.1.2: The wavelet transform coefficient matrix corresponding to the obtained wavelet transform result is calculated by singular value decomposition theory to obtain the singular value characteristic matrix of the current wavelet transform coefficient matrix and the singular value characteristic matrix of the voltage wavelet transform coefficient matrix of the i-th grid line , the singular value feature matrix represents the basic features of the wavelet transform coefficient matrix; Step 3.1.3: Calculate the wavelet signal rate of the current signal and the wavelet signal rate of the voltage signal after the fault occurs according to the current signal and voltage signal of the grid line respectively; Wavelet signal rate of the current signal of the i-th grid line after the fault occurs: s _ic represents the intensity of the current signal energy of the i-th power grid line, W _ic represents the arithmetic mean value of the current signal distribution of the i-th power grid line, and the arithmetic mean value of the current signal distribution of any power grid line under the m scale Among them, the wavelet signal energy distribution of the current signal of any grid line at the j scale: $S_{\mathrm{jc}}=\underset{k}{\mathrm{\&Sigma;}}\left|{\mathrm{D.}}_{\mathrm{jc}\left(k\right)}\right|(j=1......m,k=1......l);$ The wavelet signal rate of the voltage signal of the i-th grid line after the fault occurs: s _iv represents the intensity of the voltage signal energy of the i-th power grid line, W _iv represents the arithmetic mean value of the voltage signal energy distribution of the i-th power grid line; the arithmetic mean value of the voltage signal energy distribution of any power grid line under m scales: The wavelet signal energy distribution of the voltage signal of any grid line at the j scale: Step 3.1.4: Calculate the current signal wavelet failure rate and voltage signal wavelet failure rate after the fault occurs according to the current signal and voltage signal of the grid line; The wavelet failure rate of the current signal of the i-th grid line after the fault occurs V _ic represents the change degree of the current signal of the i-th grid line before and after the fault: in, $\left\{\begin{array}{c}{m}_{\mathrm{icq}}=\max \{{\mathrm{D.}}_{\mathrm{ic}1},{\mathrm{D.}}_{\mathrm{ic}2}......{\mathrm{D.}}_{\mathrm{ick}}\}\\ m_{\mathrm{ich}}=\max \{{\mathrm{D.}}_{\mathrm{ic}(k+1)},{\mathrm{D.}}_{\mathrm{ic}(k+2)}......{\mathrm{D.}}_{\mathrm{icl}}\}\end{array}\right.,$ M _icq represents the maximum value of the wavelet transformation result of the current signal {D _ic1 , D _ic2 ... D _ick } of the i-th grid line before the fault, and M _ich represents the current signal of the i-th grid line after the fault {D _{ic( k+1)} ,D _ic(k+2) ……D _icl }The maximum value of the wavelet transform result; The wavelet failure rate of the voltage signal of the i-th grid line after the fault occurs V _iv represents the change degree of the current signal of the i-th grid line before and after the fault, where $\left\{\begin{array}{c}{m}_{\mathrm{ivq}}=\max \{{\mathrm{D.}}_{\mathrm{iv}1},{\mathrm{D.}}_{\mathrm{iv}2}......{\mathrm{D.}}_{\mathrm{ivk}}\}\\ m_{\mathrm{ivh}}=\max \{{\mathrm{D.}}_{\mathrm{iv}(k+1)},{\mathrm{D.}}_{\mathrm{iv}(k+2)}......{\mathrm{D.}}_{\mathrm{ivl}}\}\end{array}\right.,$ M _ivq represents the maximum value of the wavelet transformation result of the voltage signal {D _iv1 , D _iv2 ... D _ivk } of the i-th power grid line before the fault, and M _ivh represents the voltage signal of the i-th power grid line after the fault {D _{iv( k+1)} ,D _iv(k+2) ……D _ivl }The maximum value of the wavelet transform result; Step 3.1.5: Calculate the wavelet variation rate of the current signal and the wavelet variation rate of the voltage signal after the fault occurs according to the current signal and voltage signal of the grid line; (1) The wavelet variation rate of the current signal of the i-th grid line after the fault occurs B _ic is the average value of the diagonal elements of the singular value eigenmatrix of the wavelet transform coefficient matrix of the current signal of the i-th grid line; average value $B_c=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{\mathrm{ic}}}{p}};$ (2) The wavelet variation rate of the voltage signal of the i-th grid line after the fault occurs B _iv is the mean value of the diagonal elements of the singular value eigenmatrix of the wavelet transform coefficient matrix of the voltage signal of the i-th grid line; average value $B_v=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{\mathrm{iv}}}{p}};$ Step 3.2: Calculate the fault support degree of each grid line according to the wavelet signal rate, wavelet failure rate and wavelet variation rate; The fault support degree of the i-th grid line ${\mathrm{the\; y}}_i=\sqrt{\frac{{\mathrm{\&rho;}}_{\mathrm{the\; s}}\sqrt{\frac{{{\mathrm{the\; s}}_{\mathrm{ic}}}^2+{{\mathrm{the\; s}}_{\mathrm{iv}}}^2}{2}}+{\mathrm{\&rho;}}_x\sqrt{\frac{{x_{\mathrm{ic}}}^2+{x_{\mathrm{iv}}}^2}{2}}+{\mathrm{\&rho;}}_m\sqrt{\frac{{m_{\mathrm{ic}}}^2+{m_{\mathrm{iv}}}^2}{2}}}{3}}(i=1...\mathrm{no}),$ Fault support coefficients ρ _s , ρ _x , ρ _m are all constants; Step 3.3: Establish a preliminary trust distribution function according to the fault support degree of each grid line; Preliminary trusted distribution function of the i-th grid line Among them, θ is the uncertainty, θ=0.1～0.3; Step 3.4: Preliminarily classify the faulty grid line set and non-faulty grid line set for each grid line by Stirling formula, and then use the fuzzy C-means clustering method to accurately divide the faulty grid line set and non-faulty grid line set , and finally determine the set of faulty grid lines; Step 4: Each DateNode node transmits the determined set of faulty grid lines to the NameNode node for grid line fault diagnosis result display.