CN104133143B - A kind of Guangdong power system diagnostic system and method calculating platform based on Hadoop cloud - Google Patents

Abstract

The present invention provides a kind of Guangdong power system diagnostic system and method calculating platform based on Hadoop cloud, and this system includes voltage transformer, current transformer, transmitter, terminal unit, modem, optical fiber and multiple host computer；The method includes: the most each voltage and current gathered on power network line also transmits terminal afield；When the grid collapses, obtain electrical network electric quantity information the NameNode node being delivered in Hadoop cluster, be redistributed to DateNode node and carry out parallel processing, carry out electric network failure diagnosis and show.The present invention makes full use of the electric parameters fault message of power network line, result calculating current, the small echo signaling rate of voltage, small echo fault rate and small echo aberration rate after being reconstructed by Wavelet transformation based on multiresolution analysis, it is fused into preliminary trusted distribution function, carry out faulty grids sets of lines accurately to divide after the preliminary classification of non-faulting power network line collection, finally determine faulty grids sets of lines, it is possible to relatively accurately judge faulty grids circuit.

Description

Power grid line fault diagnosis system and method based on Hadoop cloud computing platform

Technical Field

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

Background

Fault diagnosis of an electrical grid is a common problem for electrical power systems. At present, many researches on power grid fault diagnosis at home and abroad are carried out, and the power grid fault diagnosis system has unique characteristics in the aspect of diagnosis. The method mainly comprises a fault diagnosis method based on an expert system, a fault diagnosis method based on an artificial neural network, a fault diagnosis method based on a fuzzy set theory, a fault diagnosis method based on a Petri network theory, a fault diagnosis method based on an MAS theory and the like. However, most of the diagnosis platforms are processing platforms of a master station and a computer, and these methods have their own defects, so that when a fault occurs in the current large and complex power grid structure, the system becomes overwhelmed and even crashes when massive data of the dispatching center rapidly emerge. Therefore, the perfect algorithm can not be effectively executed, so that the fault can not be diagnosed accurately and quickly in time, and the power supply can not be recovered as soon as possible.

Disclosure of Invention

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

The technical scheme of the invention is as follows:

a power grid line fault diagnosis system based on a Hadoop cloud computing platform comprises a voltage transformer, a current transformer, a transmitter, a remote terminal, a modem, an optical fiber and a plurality of upper computers;

the input end of the voltage transformer and the input end of the current transformer are respectively connected to a power grid line, the output end of the voltage transformer and the output end of the current transformer are respectively connected to the input end of the transmitter, the output end of the transmitter is connected to the input end of a remote terminal, and the remote terminal is communicated with a plurality of upper computers through a modem;

a plurality of upper computers form a Hadoop cluster, wherein one upper computer is used as a NameNode node, and the rest are DateNode nodes; hbase, HDFS and Map/Reduce are arranged in each upper computer, and a Hadoop cluster is a Hadoop cloud computing platform.

The NameNode node is provided with a data monitoring unit.

The power grid fault diagnosis method adopting the power grid line fault diagnosis system based on the Hadoop cloud computing platform comprises the following steps:

step 1: the voltage transformer and the current transformer respectively collect the voltage and the current on each power grid line in real time and transmit the voltage and the current to a remote terminal through the transmitter;

step 2: when a power grid fails, a remote terminal acquires power grid electrical quantity information and transmits the electrical quantity information to NameNode nodes in a Hadoop cluster through a modem, and the NameNode nodes distribute the electrical quantity information of each power grid line to HDFS (high level frequency synthesizer) in each DateNode node for file storage and then store the electrical quantity information to Hbase in each DateNode node;

and step 3: Map/Reduce in each DateNode node parallelly processes the electrical quantity information of the power grid line stored by Hbase, and carries out power grid fault diagnosis;

step 3.1: performing multi-resolution analysis wavelet transform on the electrical quantity information of each power grid line and extracting fault characteristics including wavelet signal rate, wavelet fault rate and wavelet variation rate;

step 3.1.1: performing multi-resolution analysis wavelet transformation on the electric quantity information of each power grid line respectively to obtain reconstructed wavelet transformation results of the current and the voltage of each power grid line:

D_ic1,D_ic2……D_iclas a result of wavelet transformation of the current signals of the individual grid lines, D_iv1,D_iv2……D_ivlThe result of wavelet transformation of the voltage signal of each network line, where l represents the number of sampling points of the signal, D_ic1,D_ic2……D_ickFor the result of the wavelet transformation of the current signal before the fault in the individual network lines, D_ic(k+1),D_ic(k+2)……D_iclThe wavelet transformation result of the current signal after each power grid line fault is obtained; d_iv1,D_iv2……D_ivkAs a result of wavelet transformation of the voltage signal before the fault of the respective grid line, D_iv(k+1),D_iv(k+2)……D_ivlThe wavelet transformation result of the voltage signal after each power grid line fault is obtained;

step 3.1.2, the singular value characteristic matrix Λ of the current wavelet transform coefficient matrix of the ith power grid line is obtained by calculating the wavelet transform coefficient matrix corresponding to the obtained wavelet transform result through the singular value decomposition theory_ic＝diag(λ_1c,λ_2c,……λ_pc) Singular value feature matrix Λ of sum voltage wavelet transform coefficient matrix_iv＝diag(λ_1v,λ_2v,……λ_pv) The singular value feature matrix represents the basic features of the wavelet transform coefficient matrix, and p is the minimum order value of square matrixes on the left side and the right side of the singular value feature matrix;

step 3.1.3: calculating 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 the voltage signal of the power grid line respectively;

the wavelet signal rate of the current signal of the ith power grid line after the fault occurs:s_icthe intensity degree W of the current signal energy of the ith power grid line is represented_icThe arithmetic mean value of the current signal distribution of the ith power grid line and the arithmetic mean value of the current signal distribution of any power grid line under the m scaleWherein, the wavelet signal energy distribution of the current signal of any power grid line under the j scale is as follows:D_jc(k)the wavelet transformation result of the current signal at the k moment under the j scale is represented;

voltage signal wavelet signal rate of ith power grid line after fault occurrence:s_ivthe intensity degree W of the voltage signal energy of the ith power grid line is represented_ivRepresenting the arithmetic mean value of the voltage signal energy distribution of the ith power grid line; the arithmetic mean of the voltage signal energy distribution of any power grid line under m scales:wavelet signal energy distribution of voltage signals of any power grid line under the j scale:D_jv(k)the wavelet transformation result of the voltage signal at the k moment under the j scale is represented;

step 3.1.4: calculating a current signal wavelet fault rate and a voltage signal wavelet fault rate after a fault occurs according to a current signal and a voltage signal of a power grid line;

wavelet fault rate of current signal of ith power grid line after fault occurrenceV_icRepresenting the degree of change of the amplitude of the i-th grid line current signal before and after the fault:wherein, $\left\{\begin{array}{c}{M}_{\mathrm{icq}}=\max \{D_{\mathrm{ic}1},D_{\mathrm{ic}2}......D_{\mathrm{ick}}\}\\ M_{\mathrm{ich}}=\max \{D_{\mathrm{ic}(k+1)},D_{\mathrm{ic}(k+2)}......D_{\mathrm{icl}}\}\end{array}\right.,$ M_icqcurrent signal { D) representing the ith grid line before fault_ic1,D_ic2……D_ickMaximum value of wavelet transform result, M_ichCurrent signal { D) representing the ith grid line after fault_ic(k+1),D_ic(k+2)……D_iclThe maximum value of the wavelet transform result;

wavelet failure rate of voltage signal of ith power grid line after failure V_ivRepresenting the degree of change of the amplitude of the current signal of the ith network line before and after the fault, wherein $\left\{\begin{array}{c}{M}_{\mathrm{ivq}}=\max \{D_{\mathrm{iv}1},D_{\mathrm{iv}2}......D_{\mathrm{ivk}}\}\\ M_{\mathrm{ivh}}=\max \{D_{\mathrm{iv}(k+1)},D_{\mathrm{iv}(k+2)}......D_{\mathrm{ivl}}\}\end{array}\right.,$ M_ivqVoltage signal { D) representing the ith grid line before fault_iv1,D_iv2……D_ivkMaximum value of wavelet transform result, M_ivhVoltage signal { D) representing the ith grid line after fault_iv(k+1),D_iv(k+2)……D_ivlThe maximum value of the wavelet transform result;

step 3.1.5: calculating 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 the voltage signal of the power grid line;

(1) wavelet variation rate of current signal of ith power grid line after fault occursB_icThe arithmetic mean value of diagonal elements of a singular value feature matrix of a wavelet transformation coefficient matrix of the current signal of the ith power grid line is used; arithmetic mean of diagonal elements of singular value feature matrix of wavelet transform coefficient matrix of current signal of any power grid line $B_c=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{ic}}}{p}};$

(2) Wavelet variation rate of voltage signal of ith power grid line after fault occursB_ivThe arithmetic mean value of diagonal elements of a singular value feature matrix of a wavelet transformation coefficient matrix of the voltage signal of the ith power grid line is used; arithmetic mean of diagonal elements of singular value feature matrix of wavelet transform coefficient matrix of voltage signal of any power grid line $B_v=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{iv}}}{p}};$

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

fault support of ith grid line $y_i=\sqrt{\frac{{\mathrm{\ρ}}_s\sqrt{\frac{{s_{\mathrm{ic}}}^2+{s_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_x\sqrt{\frac{{x_{\mathrm{ic}}}^2+{x_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_m\sqrt{\frac{{m_{\mathrm{ic}}}^2+{m_{\mathrm{iv}}}^2}{2}}}{3}}(i=1...n),$ Fault support factor rho_s、ρ_x、ρ_mAre all constants;

step 3.3: establishing a primary credible distribution function according to the fault support degree of each power grid line;

preliminary trustworthiness distribution function of ith grid lineWhereinTheta is uncertainty, and theta is 0.1-0.3;

step 3.4: performing primary classification on a fault power grid line set and a non-fault power grid line set on each power grid line through a Stirling formula, then accurately dividing the fault power grid line set and the non-fault power grid line set by adopting a fuzzy C mean value clustering method, and finally determining a fault power grid line set;

and 4, step 4: and each DateNode node transmits the determined fault power grid line set to the NameNode node to display the power grid line fault diagnosis result.

Has the advantages that:

the invention firstly proposes that the traditional computing mode based on one computer of the platform for power grid fault diagnosis is changed into the computing mode based on the Hadoop cloud computing platform, so that compared with the traditional computing mode, the Hadoop cloud computing platform provides the most reliable and safe data storage center, and users do not worry about troubles such as data loss, virus invasion and the like; the requirement on equipment of a user side is minimum, and the use is convenient; taking parallel computing as a core, scheduling computing task allocation and computing resources as required, and providing complete data processing services from data import integration processing, computing model setting to computing result output, multi-form presentation and the like; by adopting a distributed storage system, data are mutually prepared, and quick backup and recovery are realized, so that each data processing and calculation model is supported, and the calculation requirements of different fields and different characteristics are met; multiple copies are fault-tolerant, data is safe and carefree, mass storage is realized, and space is unlimited; simple configuration, a complete platform, and no need of spending a large amount of time to set up and maintain a computing environment; computing and storage resources are used in a service mode and are taken as required, and the utilization rate of physical machine hardware can be effectively improved generally on the premise of not adding new computing capacity through a virtualization technology.

The method has the advantages of high efficiency, low cost, good reliability and the like in the process of diagnosing the fault of the power grid line, thereby having certain practical significance.

The method of the invention fully utilizes the electric quantity fault information of the power grid lines, calculates the wavelet signal rate, the wavelet fault rate and the wavelet variation rate of current and voltage through the result of wavelet change reconstruction based on multi-resolution analysis, analyzes fault data in multiple angles, fuses the fault data into a primary credible distribution function for judgment and analysis, preliminarily classifies fault power grid line sets and non-fault power grid line sets of each power grid line through a Stirling formula, then accurately divides the fault power grid line sets and the non-fault power grid line sets by adopting a fuzzy C mean value clustering method, finally determines the fault power grid line sets, and can accurately judge the fault power grid lines. And the method is used for Hadoop cloud computing, so that the diagnosis process is quicker, the more complex the power grid is, the more power grid lines are, and the more obvious the advantages are compared with the traditional computing mode.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only examples described in the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

FIG. 1 is a structural block diagram of a power grid line fault diagnosis system based on a Hadoop cloud computing platform according to an embodiment of the present invention;

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

FIG. 3 is a flow chart of multi-resolution analysis of wavelet transforms and fault feature extraction according to an embodiment of the present invention;

FIG. 4 is a diagram of a conventional Hadoop cloud computing platform architecture;

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

FIG. 6 is a graph of a fault waveform and wavelet decomposition analysis of L4 in accordance with an embodiment of the present invention;

(a) is the sampled current waveform of L4;

(b) is a sampled voltage waveform diagram of L4;

(c) is a sampled current wavelet decomposition waveform of L4;

(d) is a sampled voltage wavelet decomposition waveform of L4;

FIG. 7 is a graph of a fault waveform and wavelet decomposition analysis of L6 in accordance with an embodiment of the present invention;

(a) is the sampled current waveform of L6;

(b) is a sampled voltage waveform diagram of L6;

(c) is a sampled current wavelet decomposition waveform of L6;

(d) is a sampled voltage wavelet decomposition waveform of L6;

FIG. 8 is a graph of a fault waveform and wavelet decomposition analysis of L8 in accordance with an embodiment of the present invention;

(a) is the sampled current waveform of L8;

(b) is a sampled voltage waveform diagram of L8;

(c) is a sampled current wavelet decomposition waveform of L8;

(d) is a sampled voltage wavelet decomposition waveform of L8.

Detailed Description

In order to make those skilled in the art better understand the technical solution of the present invention, the technical solution in the embodiment of the present invention will be clearly and completely described below with reference to the drawings in the embodiment of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

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

the specific types of the voltage transformer, the current transformer and the transmitter are specifically designed according to the voltage grade and the current magnitude;

the modem adopts a GQ-DOT02C2 RS485 serial port optical transceiver;

the optical fiber adopts GYSTY53/GYTY53 double-sheath single-armor layer stranded optical fiber;

the upper computer adopts a Lenovo (Lenovo) Yangtian M3320n-00 desktop computer;

the input end of the voltage transformer and the input end of the current transformer are respectively connected to a power grid line, the output end of the voltage transformer and the output end of the current transformer are respectively connected to the input end of the transmitter, the output end of the transmitter is connected to the input end of a remote terminal, and the remote terminal is communicated with a plurality of upper computers through a modem;

the structure of the conventional Hadoop cloud computing platform is shown in FIG. 4, and it can be seen from the figure that the Hadoop cloud computing platform comprises an HDFS (Hadoop distributed file system), an Hbase (Hbase), a Map/Reduce (Map/Reduce), an operating system (windows) of a cluster computer, virtualization software (Cygwin; JDK; Ant; TotoseSVN; Eclipse) and a server cluster.

HDFS is a distributed file system capable of running on a cluster of generally inexpensive hardware infrastructure that stores large amounts of data in a streamed data access pattern. HDFS has high fault tolerance and is capable of handling large amounts of data. The HDFS adopts a master/slave mode and consists of a NameNode node and a plurality of DateNode nodes. Since the HDFS file is divided into a plurality of data blocks and is accessed in the data nodes, the HDFS is very suitable for large data set applications.

Hbase is a distributed database in a Hadoop environment. The distributed file system HDFS is used as a support, and real-time reading and writing and random access to a large data set can be provided. The Hbase database is capable of handling very large tables, capable of handling more than 10 hundred million lines of data by ordinary computers, and has data tables consisting of hundreds of outlying elements.

The operating system of the node computer in the Hadoop cluster is a Windows operating system. The virtualization software refers to Cygwin, JDK, Ant, TotoseSVN, Eclipse, and only in this environment can the writing and modification of Hadoop code be performed.

The server cluster refers to a computer of each node running on the whole Hadoop platform, and comprises a NameNode node and a plurality of DateNode nodes.

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

The power grid fault diagnosis method adopting the power grid line fault diagnosis system based on the Hadoop cloud computing platform comprises the following steps as shown in FIG. 2:

step 1: the voltage transformer and the current transformer respectively collect the voltage and the current on each power grid line in real time and transmit the voltage and the current to a remote terminal through the transmitter;

step 2: when a power grid fails, a remote terminal acquires power grid electrical quantity information and transmits the electrical quantity information to NameNode nodes in a Hadoop cluster through a modem, and the NameNode nodes distribute the electrical quantity information of each power grid line to HDFS (high level frequency synthesizer) in each DateNode node for file storage and then store the electrical quantity information to Hbase in each DateNode node;

and step 3: Map/Reduce in each DateNode node parallelly processes the electrical quantity information of the power grid line stored by Hbase, and carries out power grid fault diagnosis;

step 3.1: performing multi-resolution analysis wavelet transform on the electrical quantity information of each power grid line and extracting fault characteristics including wavelet signal rate, wavelet fault rate and wavelet variation rate; as shown in fig. 3:

step 3.1.1: performing multi-resolution analysis wavelet transformation on the electric quantity information of each power grid line respectively to obtain reconstructed wavelet transformation results of the current and the voltage of each power grid line:

D_ic1,D_ic2……D_iclas a result of wavelet transformation of the current signals of the individual grid lines, D_iv1,D_iv2……D_ivlThe result of wavelet transformation of the voltage signal of each network line, where l represents the number of sampling points of the signal, D_ic1,D_ic2……D_ickFor the result of the wavelet transformation of the current signal before the fault in the individual network lines, D_ic(k+1),D_ic(k+2)……D_iclThe wavelet transformation result of the current signal after each power grid line fault is obtained; d_iv1,D_iv2……D_ivkAs a result of wavelet transformation of the voltage signal before the fault of the respective grid line, D_iv(k+1),D_iv(k+2)……D_ivlThe wavelet transformation result of the voltage signal after each power grid line fault is obtained;

step 3.1.2, the singular value characteristic matrix Λ of the current wavelet transform coefficient matrix of the ith power grid line is obtained by calculating the wavelet transform coefficient matrix corresponding to the obtained wavelet transform result through the singular value decomposition theory_ic＝diag(λ_1c,λ_2c,……λ_pc) Singular value feature matrix Λ of sum voltage wavelet transform coefficient matrix_iv＝diag(λ_1v,λ_2v,……λ_pv)，λ_1c,λ_2c,……λ_pcDiagonal elements, lambda, of a singular value feature matrix representing a current wavelet transform coefficient matrix of the ith grid line_1v,λ_2v,……λ_pvTo representDiagonal elements of a singular value feature matrix of a voltage wavelet transform coefficient matrix of the ith power grid line, wherein the singular value feature matrix represents basic features of the wavelet transform coefficient matrix, and p is the minimum order value of square matrixes on the left side and the right side of the singular value feature matrix;

step 3.1.3: calculating 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 the voltage signal of the power grid line respectively;

the wavelet signal rate of the current signal of the ith power grid line after the fault occurs:s_icthe intensity degree W of the current signal energy of the ith power grid line is represented_icThe arithmetic mean value of the current signal distribution of the ith power grid line and the arithmetic mean value of the current signal distribution of any power grid line under the m scaleWherein, the wavelet signal energy distribution of the current signal of any power grid line under the j scale is as follows: $S_{\mathrm{jc}}=\underset{k}{\mathrm{\Σ}}\left|D_{\mathrm{jc}\left(k\right)}\right|(j=1......m,k=1......l),$ D_jc(k)；

voltage signal wavelet signal rate of ith power grid line after fault occurrence:s_ivthe intensity degree W of the voltage signal energy of the ith power grid line is represented_ivRepresenting the arithmetic mean value of the voltage signal energy distribution of the ith power grid line; the arithmetic mean of the voltage signal energy distribution of any power grid line under m scales:wavelet signal energy distribution of voltage signals of any power grid line under the j scale:

step 3.1.4: calculating a current signal wavelet fault rate and a voltage signal wavelet fault rate after a fault occurs according to a current signal and a voltage signal of a power grid line;

wavelet fault rate of current signal of ith power grid line after fault occurrenceV_icRepresenting the degree of change of the amplitude of the i-th grid line current signal before and after the fault:wherein, $\left\{\begin{array}{c}{M}_{\mathrm{icq}}=\max \{D_{\mathrm{ic}1},D_{\mathrm{ic}2}......D_{\mathrm{ick}}\}\\ M_{\mathrm{ich}}=\max \{D_{\mathrm{ic}(k+1)},D_{\mathrm{ic}(k+2)}......D_{\mathrm{icl}}\}\end{array}\right.,$ M_icqcurrent signal { D) representing the ith grid line before fault_ic1,D_ic2……D_ickMaximum value of wavelet transform result, M_ichCurrent signal { D) representing the ith grid line after fault_ic(k+1),D_ic(k+2)……D_iclThe maximum value of the wavelet transform result;

wavelet failure rate of voltage signal of ith power grid line after failure V_ivRepresenting the degree of change of the amplitude of the voltage signal of the ith network line before and after the fault, wherein $\left\{\begin{array}{c}{M}_{\mathrm{ivq}}=\max \{D_{\mathrm{iv}1},D_{\mathrm{iv}2}......D_{\mathrm{ivk}}\}\\ M_{\mathrm{ivh}}=\max \{D_{\mathrm{iv}(k+1)},D_{\mathrm{iv}(k+2)}......D_{\mathrm{ivl}}\}\end{array}\right.,$ M_ivqVoltage signal { D) representing the ith grid line before fault_iv1,D_iv2……D_ivkMaximum value of wavelet transform result, M_ivhVoltage signal { D) representing the ith grid line after fault_iv(k+1),D_iv(k+2)……D_ivlThe maximum value of the wavelet transform result;

step 3.1.5: calculating 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 the voltage signal of the power grid line;

(1) wavelet variation rate of current signal of ith power grid line after fault occursB_icThe arithmetic mean value of diagonal elements of a singular value feature matrix of a wavelet transformation coefficient matrix of the current signal of the ith power grid line is used; arithmetic mean of diagonal elements of singular value feature matrix of wavelet transform coefficient matrix of current signal of any power grid line $B_c=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{ic}}}{p}};$

(2) Wavelet variation rate of voltage signal of ith power grid line after fault occursB_ivThe arithmetic mean value of diagonal elements of a singular value feature matrix of a wavelet transformation coefficient matrix of the voltage signal of the ith power grid line is used; arithmetic mean of diagonal elements of singular value feature matrix of wavelet transform coefficient matrix of voltage signal of any power grid line $B_v=\sqrt{\frac{\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\λ}}_{\mathrm{iv}}}{p}};$

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

fault support of ith grid line $y_i=\sqrt{\frac{{\mathrm{\ρ}}_s\sqrt{\frac{{s_{\mathrm{ic}}}^2+{s_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_x\sqrt{\frac{{x_{\mathrm{ic}}}^2+{x_{\mathrm{iv}}}^2}{2}}+{\mathrm{\ρ}}_m\sqrt{\frac{{m_{\mathrm{ic}}}^2+{m_{\mathrm{iv}}}^2}{2}}}{3}}(i=1...n),$ Fault support factor rho_s、ρ_x、ρ_mAre all constants, in general, p_s＝0.3，ρ_x＝1.9，ρ_m＝0.8；

Step 3.3: establishing a primary credible distribution function according to the fault support degree of each power grid line;

preliminary trustworthiness distribution function of ith grid lineWherein theta is uncertainty, and theta is 0.1-0.3;

step 3.4: performing primary classification on a fault power grid line set and a non-fault power grid line set on each power grid line through a Stirling formula, then accurately dividing the fault power grid line set and the non-fault power grid line set by adopting a fuzzy C mean value clustering method, and finally determining a fault power grid line set;

(1) the initial credible distribution function of each power grid line is used as the fault probability of the corresponding power grid line, and the Stalllin function value is obtained from the fault probability of the power grid line to divide the initial power grid line into two categories, namely a fault power grid line set phi₁NAND fault power grid line set phi₂(ii) a If the function value of Stirling is satisfiedWherein, depending on the size of the faulty grid, the faulty grid line set Φ is assigned₁In that, the remaining grid lines are grouped into a non-faulty grid line set Φ₂In (1).

(2) Establishing a target function of the fuzzy C-means clustering method by using a membership matrix U of the fuzzy C-means clustering method;

the membership matrix U of the fuzzy C-means clustering method has element values between 0 and 1, but the membership matrix U has normalization regulation, namely all elements U of the membership matrix U_fThe sum of i is always equal to 1:

$\underset{f=1}{\overset{c}{\mathrm{\Σ}}}{u}_{\mathrm{fi}}=1,\∀i=1...n,c=2$

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

$J(U,c_1,...,c_c)=\underset{f=1}{\overset{c}{\mathrm{\Σ}}}{J}_i=\underset{f=1}{\overset{c}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{u}_{\mathrm{fi}}^{\mathrm{\η}}{d}_{\mathrm{fi}}^2,(c=2)$

in the formula u_fi is between 0 and 1; c. C_fTo blur the cluster centers of the group f, d_fi＝||c_f-w_iAnd η∈ [1, ∞) ] is a weighted index where η is 1.5 < η < 2.5.

(3) Obtaining the necessary condition that the target function of the fuzzy C-means clustering method reaches the minimum value:

$c_f=\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{u}_{\mathrm{fi}}^{\mathrm{\&eta;}}{w}_i}{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{u}_{\mathrm{fi}}^{\mathrm{\&eta;}}}$ and $u_{\mathrm{fi}}=\frac{1}{\underset{k=1}{\overset{c}{\mathrm{\&Sigma;}}}{(d_{\mathrm{fi}}/d_{\mathrm{ki}})}^{2/(\mathrm{\&eta;}-1)}}$

(4) determining a clustering center C according to the necessary condition that the target function of the fuzzy C-means clustering method reaches the minimum value_iAnd a membership matrix U;

firstly, initializing a membership matrix U by using a random number between 0 and 1 to satisfy the formula

Secondly, utilizeCalculating f cluster centers c_f；

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

If the change amount of the scalar value is less than the threshold value gamma, the iteration is stopped;

finally, byCalculating a new membership matrix, and recalculating f clustering centers c_f。

After the membership matrix U is solved, if U_1i＞u_2iIf the grid line belongs to the fault grid line set, otherwise, the grid line belongs to the non-fault grid line set.

And 4, step 4: and each DateNode node transmits the determined fault power grid line set to the NameNode node to display the power grid line fault diagnosis result.

In the present embodiment, a fault diagnosis is performed on a grid line as shown in fig. 5. The figure comprises 11 buses (B1-B11), 10 lines (L1-L10) and 20 breaker switches (CB 1-CB 20). When the current L4 breaks down, main protection actions at two sides of a line L4 break away circuit breakers CB6 and CB14 after the fault occurs; a second backup protection malfunction of the line L8 trips a circuit breaker CB 17; meanwhile, the two sides of the line L6 are protected to trip the circuit breakers CB8 and CB 9;

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

The fault diagnosis analysis data table is as follows:

TABLE 1 Fault diagnosis Process data analysis Table

When the calculation is 2, η is 2, and as can be seen from the data calculated in the table, u is₁₁＞u₂₁,u₁₂＜u₂₂,u₁₃＜u₂₃The fault line is L4, and the embodiment is enough to illustrate that the invention is reliable and practical and can be used for fault diagnosis of large-scale power grids.

The invention provides a fault diagnosis system and method of a power grid line based on a Hadoop cloud computing platform, aiming at the background that mass data need to be processed and a fault decision needs to be made in time when a large and complex power grid fails. The main diagnosis algorithm is also based on wavelet transformation of multi-resolution analysis and fuzzy C-means clustering algorithm, and faults are quickly and accurately diagnosed. The Hadoop cloud computing platform is low in construction cost, can process a large amount of data in parallel, and is extremely high in efficiency, so that huge loss caused by power grid faults can be saved as far as possible.

The foregoing is directed to embodiments of the present application and it is noted that numerous modifications and adaptations may be made by those skilled in the art without departing from the principles of the present application and are intended to be within the scope of the present application.

Claims (1)

1. A power grid fault diagnosis method adopts a power grid line fault diagnosis system based on a Hadoop cloud computing platform, and comprises a voltage transformer, a current transformer, a transmitter, a remote terminal, a modem, an optical fiber and a plurality of upper computers;

the input end of the voltage transformer and the input end of the current transformer are respectively connected to a power grid line, the output end of the voltage transformer and the output end of the current transformer are respectively connected to the input end of the transmitter, the output end of the transmitter is connected to the input end of a remote terminal, and the remote terminal is communicated with a plurality of upper computers through a modem;

a plurality of upper computers form a Hadoop cluster, wherein one upper computer is used as a NameNode node, and the rest are DateNode nodes; hbase, HDFS and Map/Reduce are arranged in each upper computer, and a Hadoop cluster is a Hadoop cloud computing platform; the method is characterized in that: the method comprises the following steps:

step 1: the voltage transformer and the current transformer respectively collect the voltage and the current on each power grid line in real time and transmit the voltage and the current to a remote terminal through the transmitter;

step 2: when a power grid fails, a remote terminal acquires power grid electrical quantity information and transmits the electrical quantity information to NameNode nodes in a Hadoop cluster through a modem, and the NameNode nodes distribute the electrical quantity information of each power grid line to HDFS (high level frequency synthesizer) in each DateNode node for file storage and then store the electrical quantity information to Hbase in each DateNode node;

and step 3: Map/Reduce in each DateNode node parallelly processes the electrical quantity information of the power grid line stored by Hbase, and carries out power grid fault diagnosis;

step 3.1: performing multi-resolution analysis wavelet transform on the electrical quantity information of each power grid line and extracting fault characteristics including wavelet signal rate, wavelet fault rate and wavelet variation rate;

step 3.1.1: performing multi-resolution analysis wavelet transformation on the electric quantity information of each power grid line respectively to obtain reconstructed wavelet transformation results of the current and the voltage of each power grid line:

D_ic1，D_ic2……D_iclas a result of wavelet transformation of the current signals of the individual grid lines, D_iv1，D_iv2……D_ivlThe result of wavelet transformation of the voltage signal of each network line, where l represents the number of sampling points of the signal, D_ic1，D_ic2……D_ickFor the result of the wavelet transformation of the current signal before the fault in the individual network lines, D_ic(k+1)，D_ic(k+2)……D_iclThe wavelet transformation result of the current signal after each power grid line fault is obtained; d_iv1，D_iv2……D_ivkFor small variations of the voltage signal before a fault in the individual network linesChanging the result, D_iv(k+1)，D_iv(k+2)……D_ivlThe wavelet transformation result of the voltage signal after each power grid line fault is obtained;

step 3.1.2: calculating a singular value characteristic matrix of a current wavelet transform coefficient matrix and a singular value characteristic matrix of a voltage wavelet transform coefficient matrix of the ith power grid line by using a singular value decomposition theory according to the obtained wavelet transform coefficient matrix corresponding to the wavelet transform result, wherein the singular value characteristic matrix represents the basic characteristics of the wavelet transform coefficient matrix;

step 3.1.3: calculating 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 the voltage signal of the power grid line respectively;

the wavelet signal rate of the current signal of the ith power grid line after the fault occurs:s_icthe intensity degree W of the current signal energy of the ith power grid line is represented_icThe arithmetic mean value of the current signal distribution of the ith power grid line and the arithmetic mean value of the current signal distribution of any power grid line under the m scaleWherein, the wavelet signal energy distribution of the current signal of any power grid line under the j scale is as follows:

voltage signal wavelet signal rate of ith power grid line after fault occurrence:s_ivthe intensity degree W of the voltage signal energy of the ith power grid line is represented_ivRepresenting the arithmetic mean value of the voltage signal energy distribution of the ith power grid line; the arithmetic mean of the voltage signal energy distribution of any power grid line under m scales:wavelet signal energy distribution of voltage signals of any power grid line under the j scale:

step 3.1.4: calculating a current signal wavelet fault rate and a voltage signal wavelet fault rate after a fault occurs according to a current signal and a voltage signal of a power grid line;

wavelet fault rate of current signal of ith power grid line after fault occurrenceV_icRepresenting the degree of change of the amplitude of the i-th grid line current signal before and after the fault:wherein,M_icqcurrent signal { D) representing the ith grid line before fault_ic1，D_ic2……D_ickMaximum value of wavelet transform result, M_ichCurrent signal { D) representing the ith grid line after fault_ic(k+1)，D_ic(k+2)……D_iclThe maximum value of the wavelet transform result;

wavelet failure rate of voltage signal of ith power grid line after failure V_ivRepresenting the degree of change of the amplitude of the current signal of the ith network line before and after the fault, whereinM_ivqVoltage signal { D) representing the ith grid line before fault_iv1，D_iv2……D_ivkMaximum value of wavelet transform result, M_ivhVoltage signal { D) representing the ith grid line after fault_iv(k+1)，D_iv(k+2)……D_ivlThe maximum value of the wavelet transform result;

step 3.1.5: calculating 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 the voltage signal of the power grid line;

(1) wavelet variation rate of current signal of ith power grid line after fault occursB_icThe average value of diagonal elements of a singular value feature matrix of a wavelet transformation coefficient matrix of the current signal of the ith power grid line is obtained; arithmetic mean of diagonal elements of singular value feature matrix of wavelet transform coefficient matrix of current signal of any power grid line

(2) Wavelet variation rate of voltage signal of ith power grid line after fault occursB_ivThe average value of diagonal elements of a singular value feature matrix of a wavelet transformation coefficient matrix of the voltage signal of the ith power grid line is obtained; arithmetic mean of diagonal elements of singular value feature matrix of wavelet transform coefficient matrix of voltage signal of any power grid line

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

fault support of ith grid lineFault support factor rho_s、ρ_x、ρ_mAre all constants;

step 3.3: establishing a primary credible distribution function according to the fault support degree of each power grid line;

preliminary trustworthiness distribution function of ith grid lineWherein theta is uncertainty, and theta is 0.1-0.3;

step 3.4: performing primary classification on a fault power grid line set and a non-fault power grid line set on each power grid line through a Stirling formula, then accurately dividing the fault power grid line set and the non-fault power grid line set by adopting a fuzzy C mean value clustering method, and finally determining a fault power grid line set;

and 4, step 4: and each DateNode node transmits the determined fault power grid line set to the NameNode node to display the power grid line fault diagnosis result.