CN205622133U - GIS switchgear action state monitoring system - Google Patents

Abstract

The utility model provides a pair of GIS switchgear action state monitoring system, include: the system comprises a multi-source information acquisition unit, a GIS switching system and a control unit, wherein the multi-source information acquisition unit detects the working state of the GIS switching system and forms a detection signal, and the multi-source information acquisition unit conditions the detection signal into an electric analog signal; the information fusion unit is arranged in the computer terminal and used for receiving the electric analog signal, comparing the electric analog signal with a preset value and judging the working state of the GIS switching system; and the fault decision and prediction unit is used for alarming or predicting according to the comparison result of the electric analog signal and the preset value. Compared with the prior art, the beneficial effects of the utility model are as follows: the GIS fault diagnosis can be carried out based on the information acquired by various sensor signals, the diagnosis accuracy is improved, and the occurrence of false alarm is reduced.

Description

GIS switchgear action state monitoring system

Technical Field

The utility model discloses the electrical equipment field, more specifically relates to a judge electrical equipment GIS switchgear action state monitoring system.

Background

A Gas Insulated metal enclosed Switchgear (GIS) is used as a form of high-voltage distribution equipment, all primary equipment except a transformer in a transformer substation are organically combined into a whole through optimized design and enclosed in a metal shell, SF6 Gas is filled as an arc extinguishing and insulating medium to form an enclosed combined electrical appliance, and the highest distribution voltage can reach 1100 kV. The GIS overcomes many limitations of conventional open-type switchgear, has the advantages of small floor area, high reliability, strong safety, small maintenance workload and the like, enables high-voltage and ultrahigh-voltage power transmission and transformation to directly enter urban areas, and is widely used in recent years. With the continuous improvement of the GIS and the development requirement of the power system, the selection of the GIS for the high-voltage switch equipment becomes the development trend of the whole world. GIS is developing towards co-integration, miniaturization, intellectualization, and ultra-high pressure and high capacity. The GIS main components comprise a circuit breaker, a disconnecting switch, a grounding switch, a voltage transformer, a current transformer, a lightning arrester, a sleeve, a cable terminal, a bus, a shell, SF6 gas, an SF6 density monitoring device, a GIS insulator and the like. The circuit breaker, the isolating switch and the grounding switch are collectively called GIS switch equipment and are core elements of the GIS.

The operation state of the high-voltage GIS switch equipment directly influences the operation stability and the power supply reliability of the power system. Due to the totally-enclosed design of the GIS equipment, an operator cannot directly observe the state of the equipment, and whether the equipment is connected or disconnected in place is judged only according to a return signal of the auxiliary contact and on-site confirmation of the operator. After the switch knife switch is operated, due to various reasons, the situation that the switching-on and switching-off are successfully displayed in a monitoring background and a field, but the actual contact is not switched on and off in place can occur, so that a power grid safety event is caused, and considerable economic loss and serious social influence are caused.

The switch equipment is instantaneous equipment, and in normal operation, the mechanism is in a static state, and occasionally operates or acts in an accident, and the process is very short and high-speed, so that great difficulty is brought to monitoring. Past experience is that a regular shutdown maintenance system is established, faults cannot be found timely by the preventive maintenance system, blindness is high, and mechanical life of a switch is even reduced by excessive maintenance operation. According to the principle of selecting fault monitoring objects of the circuit breaker suggested by IEEE, time parameters, metal short-circuit time, total stroke, insertion stroke, overtravel, moving contact speed, opening and closing coil current, contact service life and protection action parameters in the opening and closing process of the circuit breaker are monitored, the parameters are analyzed in detail, a monitoring method and an analysis and judgment method are provided, and a processing scheme after data are out of limit is provided. However, the current technology is mostly directed to the measurement of the mechanical characteristics of the switchgear, and is based on indirect measurement, and the effectiveness and reliability of the technology still need to be improved.

In addition, the current devices (systems) for monitoring the state of power equipment including high-voltage circuit breakers can be roughly divided into: a centralized online monitoring system and a portable online monitoring system. Consistent with theoretical research, in the aspect of a high-voltage circuit breaker online state monitoring device (system), more situations are to monitor one or more aspects of mechanical characteristics, mechanical vibration, contact electrical life and insulating performance of a high-voltage circuit breaker, the working reliability and correctness of the monitoring device are still to be proved and continuously summarized and improved in practice, and the problems to be considered include: reliability, feasibility and economy. The above factors are also the main reasons for restricting the popularization and development of the state maintenance of the switchgear.

SUMMERY OF THE UTILITY MODEL

To the defect among the prior art, the utility model aims at providing an increase diagnostic accuracy, reduce the GIS switchgear action state monitoring system that the condition of wrong report takes place.

In order to solve the technical problem, the utility model provides a pair of GIS switchgear action state monitoring system, include: the system comprises a multi-source information acquisition unit, a GIS switching system and a control unit, wherein the multi-source information acquisition unit detects the working state of the GIS switching system and forms a detection signal, and the multi-source information acquisition unit conditions the detection signal into an electric analog signal; the information fusion unit is arranged in the computer terminal and used for receiving the electric analog signal, comparing the electric analog signal with a preset value and judging the working state of the GIS switching system; and the fault decision and prediction unit is used for alarming or predicting according to the comparison result of the electric analog signal and the preset value.

Preferably, a monitor and a signal conditioning circuit which are connected with each other are arranged in the multi-source information acquisition unit.

Preferably, the monitor comprises a relay control loop state detection module, a connection execution high-voltage switch state detection module, a mechanical mechanism state detection module and an air leakage detection module.

Preferably, the information fusion unit comprises a data acquisition board card, a fault extraction module, an information fusion module and a daily fault judgment module which are connected with each other; the data acquisition board card is communicated with the signal conditioning circuit.

Preferably, the fault decision and prediction unit comprises a fault state fuzzy feature library, a fault judgment module and a fault prediction module which are connected with each other; the fault state fuzzy feature library is communicated with the information fusion module, the fault judgment module is communicated with the fault state fuzzy feature library, the fault forecast module is respectively communicated with the daily fault judgment module and the fault judgment module, and the daily fault judgment module is communicated with the daily fault judgment module.

A method for using a GIS switch equipment action state monitoring system comprises the following steps:

step 1, detecting the working state of a GIS switching system by using a monitor and forming a detection signal;

Step 2, the signal conditioning circuit conditions the detection signal into an electric analog signal;

step 3, the data acquisition board card receives the electric analog signal and converts the electric analog signal into a digital signal;

step 4, fusing the digital signals into a fault state vector;

step 5, using principal component analysis method to carry out dimension reduction processing on the normal state vector X by using singular value grading method to obtain dimension reduced normal state vector

Step 6, extracting an air leakage signal and a daily power loss signal of SF6 gas from the digital signal to form a daily fault feature library, and calculating a daily fault weight;

step 7, receiving the fault state vector by the fault state fuzzy feature library and calculating to obtain the weight of the fault state;

and 8, performing fault judgment or fault prediction according to the weight of the fault state.

Preferably, step 5 comprises the steps of:

step 5.1, data standardization treatment:

working state X of GIS switching system in n periods_mWriting into a state matrix form:

X_m＝(X₁,X₂...X_n)

wherein, X₁,X₂,...,X_nRespectively representing the working states of the 1 st to the nth period of the GIS switch system(ii) a The working state of each period of the GIS switching system is composed of n-dimensional detection data collected by a monitor;

Mixing X_mCarrying out normalization processing to obtain X:

$X=\frac{X_m-\stackrel{\‾}{X}}{\mathrm{\σ}}$

wherein,represents X_mMean value of (a) represents X_mStandard deviation of (d);

step 5.2, singular value decomposition is carried out by utilizing the covariance matrix:

$S=\frac{1}{n-1}{X}^{T}X={\mathrm{U\ΛU}}^T$

where S represents the covariance matrix between the data elements, σ_i1, 2.. times.n, which respectively represent the 1 st to nth singular values of the matrix S, Λ represents a matrix of singular values, U and U^TIs a representation of singular value decomposition;

step 5.3, taking principal element elements:

taking the first k principal elements of the matrix lambda as analysis elements; and taking the corresponding P ═ u (u)₁,u₂...u_k) (ii) a Wherein u is_iI ═ 1, 2.. times, k, which respectively represent the first 1 st to first k th vectors of the corresponding analysis elements of the matrix U, P represents the vector formed by the first k vectors of the matrix U;

step 5.4, obtaining the dimensionality reduction form of X

$\hat{X}={\mathrm{TP}}^T$

Wherein, T is XP.

Preferably, in step 7, the establishment of the fuzzy feature library of the fault state is based on a TS fuzzy model, and includes the following steps:

step 7.1, establishing a fuzzy rule:

ith fuzzy rule RⁱContribution component y at (k +1) th time to TS fuzzy model outputⁱ(k +1) is:

$\begin{array}{c}{R}^i:If{x}_1\left(k\right)is{A}_1^{i}and{x}_2\left(k\right)is{A}_2^{i}and\mathrm{...}and{x}_n\left(k\right)is{A}_n^i\\ Then{y}^i(k+1)=p_0^i+p_1^i{x}_1+\mathrm{...}+p_n^i{x}_n;\end{array},i=1,2,\mathrm{...}c$

wherein c is the number of fuzzy rules, n is the number of input variables of the TS fuzzy model, and x₁(k),x₂(k),…,x_n(k) Each of the n-dimensional input/output data regression variables of the TS fuzzy model at the k-th time, x (k) ═ x ₁(k),x₂(k),…,x_n(k)]An input vector of the TS fuzzy model at the k moment;a fuzzy set with a linear membership function representing n fuzzy subspaces corresponding to the ith fuzzy rule,n-dimensional back-part parameters of the ith fuzzy rule;

and 7.2, outputting and calculating:

$\begin{array}{c}y(k+1)=\underset{i=1}{\overset{c}{\Σ}}{\mathrm{\β}}_i{y}^i(k+1)=\underset{i=1}{\overset{c}{\Σ}}{\mathrm{\β}}_i(p_0^i+p_1^i{x}_1(k)+\mathrm{...}+p_n^i{x}_n(k\left)\right)\\ =\underset{i=1}{\overset{c}{\Σ}}(p_0^i+p_1^i+\mathrm{...}+p_n^i){({\mathrm{\β}}_i+{\mathrm{\β}}_i{x}_1(k)+\mathrm{...}+{\mathrm{\β}}_i{x}_n(k\left)\right)}^T\end{array}$

wherein, beta_iDefining the fitness of the ith fuzzy rule

$\left\{\begin{array}{c}\mathrm{\Θ}\left(k\right)={\[{\mathrm{\θ}}_1,{\mathrm{\θ}}_2,\mathrm{...},{\mathrm{\θ}}_r\]}^T={\[p_{10},p_{20},\mathrm{...},p_{c0},p_{11},p_{21},\mathrm{...},p_{c1},\mathrm{...},p_{cn}\]}^T;\\ \mathrm{\Φ}\left(k\right)={\[{\mathrm{\β}}_1,\mathrm{...},{\mathrm{\β}}_c,{\mathrm{\β}}_1{x}_1\left(k\right),\mathrm{...},{\mathrm{\β}}_c{x}_1\left(k\right),\mathrm{...},{\mathrm{\β}}_1{x}_n\left(k\right),\mathrm{...},{\mathrm{\β}}_c{x}_n\left(k\right)\]}^T;\end{array}\right.$

Where r ═ c (n +1), gives:

y(k+1)＝Φ(k)^TΘ(k)

where Φ (k) and Θ (k) represent the parameters of the fuzzy rule at the kth instant, θ₁,θ₂,…,θ_rIs a column vector of Θ (k), r denotes the column vector index, r ═ c (n +1), Θ (k) sets each c column vectors as a set of column vectors in ascending order, p_uvIs the u-th column vector in the column vector group with the sequence number v of theta (k), u is 1 to C, v is 0 to n, p_uvThe back-piece parameter at the kth instant is indicated.

Preferably, step 8 comprises the steps of:

step 8.1, establishing a fuzzy rule:

establishing an equation to calculate a clustering center vector v of an input x (k) of the TS fuzzy model at the kth moment and a clustering center vector v corresponding to the ith fuzzy rule at the kth-1 moment_iDistance d of (k-1)_i′(k)：

$d_i^{\′}\left(k\right)=\sqrt{\underset{j=1}{\overset{n}{\Σ}}{\[x_j\left(k\right)-v_i(k-1)\]}^2};i=1,2,...,c$

Wherein x (k) ═ x₁(k),x₂(k),...,x_n(k))，x_j(k) Representing the jth input vector of the TS fuzzy model at the kth moment, wherein c is the fuzzy rule number;

step 8.2, evaluate input x (k) for each cluster center v_i(k-1), i ═ 1,2, …, degree of membership u of c _i' (k) wherein v_i(k-1) represents the cluster center vector corresponding to the ith fuzzy rule at the k-1 time,

$u_i^{\′}\left(k\right)={\[\underset{j=1}{\overset{c}{\Σ}}\left(\frac{d_i^{\′}\left(k\right)}{d_j^{\′}\left(k\right)}\right)\]}^{\frac{(f-1)}{2}},i=1,2,...,c$

wherein f represents a fuzzy factor, and the value of f is more than 1; d'_j(k) Representing the distance between the input x (k) of the TS fuzzy model at the kth moment and the clustering center vector of the jth fuzzy rule corresponding to the kth-1 moment;

step 8.3, correcting the clustering center vector:

v_i(k)＝v_i(k-1)+λu_i′(k)²[x(k)-v_i(k-1)]

wherein, λ represents learning rate, and the range of the value of λ is a number less than 1 and greater than 0; v. of_i(k) A clustering center vector corresponding to the ith fuzzy rule at the kth moment;

step 8.4, update distance d_i' (k) and degree of membership u_i′(k)：

Step 8.5, calculating the fitness beta of the ith fuzzy rule to the TS fuzzy model output_i：

${\mathrm{\β}}_i=\underset{j=1}{\overset{c}{\Σ}}\left(\frac{u_i}{u_j}\right),i=1,2,...,c$

Wherein u is_jThe first jth vector representing the corresponding analysis element of the matrix U;

step 8.6, according to the formula y (k +1) ═ Φ (k)^TΘ (k) is obtained by the least square method, and Θ (k) is (Φ)^TΦ)^-1Φ^TY (k + 1); wherein y (k +1) represents a contribution component of the fuzzy rule to the TS fuzzy model output at the k +1 th moment, and phi (k) and theta (k) represent parameters of the fuzzy rule at the k th moment;

step 8.7, carrying out statistics on phi (k) and theta (k) at different moments, carrying out statistics on y and yf, and establishing a y fuzzy specification library and a yf fuzzy rule library; where y represents the state in the dimension reduced data Probability of failure in case of occurrence, yf represents the state of the data in the reduced dimensionProbability of failure of the next operation under the circumstances;

and 8.8, comparing y and yf with a preset threshold value to judge or forecast the fault.

Preferably, in step 8.8, the step of performing fault judgment or fault prediction by comparing y and yf with the preset threshold value is as follows:

if the y value is larger than the upper limit threshold Au, judging that the fault occurs at this time;

if the y value is smaller than the lower limit threshold AL, judging that the current state is in a normal working state and no fault occurs;

if the y value is less than or equal to the upper threshold Au and more than or equal to the lower threshold AL, the state of the current time cannot be judged, and then yf, the upper threshold Au and the lower threshold AL are started to carry out fault prediction according to the following method:

if the yf value is larger than the upper limit threshold Au, judging that the next operation fails;

if the yf value is smaller than the lower limit threshold AL, judging that the next operation cannot be failed;

if the yf value is less than or equal to the upper threshold Au and greater than or equal to the lower threshold AL, it is considered that no failure will occur for the next operation.

Compared with the prior art, the beneficial effects of the utility model are as follows: the GIS fault diagnosis can be carried out based on the information acquired by various sensor signals, the diagnosis accuracy is improved, and the occurrence of false alarm is reduced. The utility model discloses both can judge this switch whether normally work according to the operating condition of current GIS switch, also can locate next time GIS switch work according to multisource information forecast and whether can break down, play the effect of forecast to the trouble. The utility model discloses also can judge the possibility that GIS will break down according to the normal operating condition's of GIS switch abnormity, eliminate the consequence that probably takes place in advance. The main work is realized in software, the requirement on hardware is low, and the system cost is greatly saved.

Drawings

Other characteristic objects and advantages of the invention will become more apparent from a reading of the detailed description of non-limiting embodiments with reference to the following figures.

Fig. 1 is a schematic structural view of a system for monitoring the operating state of the GIS switch device of the present invention;

fig. 2 is an information fusion schematic diagram of the GIS switch device action state monitoring system of the utility model;

fig. 3 is the utility model discloses GIS switchgear operating condition monitoring system fault characteristic fuzzy library schematic diagram.

Detailed Description

The present invention will be described in detail with reference to the following examples. The following examples will assist those skilled in the art in further understanding the present invention, but are not intended to limit the invention in any way. It should be noted that various changes and modifications can be made by one skilled in the art without departing from the spirit of the invention. These all belong to the protection scope of the present invention.

As shown in FIG. 1, the utility model discloses GIS switchgear motion state monitoring system, multisource information acquisition unit, information fusion unit, fault decision and prediction unit.

The multi-source information acquisition unit detects the working state of the GIS switching system and forms a detection signal, and the multi-source information acquisition unit conditions the detection signal into an electric analog signal; the monitor and the signal conditioning circuit are arranged in the multi-source information acquisition unit, and the monitor comprises a relay control loop state detection module, a connection execution high-voltage switch state detection module, a mechanical mechanism state detection module and an air leakage detection module.

Multisource information acquisition unit utilizes multiple sensor technology, detects GIS switching system's operating condition, includes: the relay control loop state detection circuit comprises a relay control loop state detection circuit, a connection execution high-voltage switch state detection circuit, a mechanical mechanism state detection circuit, other equipment state detection circuit and a signal conditioning circuit.

Wherein, the relay control circuit detects, and what mainly detected is: daily secondary circuit power loss detection, opening and closing coil transient current detection and switching action power detection.

The connection executes the detection of the state of the high-voltage switch, and the main detection is as follows: opening and closing transient magnetic field detection and arc electromagnetic wave detection.

The method mainly comprises the following steps of detecting the state of a mechanical mechanism: and detecting the running distance of the contact and detecting the mechanical vibration.

And detecting the state of other equipment, mainly detecting: detection of leakage of the quenching insulating medium gas SF 6. The signal conditioning circuit conditions the signals detected by the sensor into electric analog signals which can be identified by the signal acquisition board card and transmitted to the data acquisition card.

The information fusion unit is arranged on the PCB and arranged in the computer terminal through a PCI slot, receives the electric analog signal, compares the electric analog signal with a preset value and judges the working state of the GIS switch system; the information fusion unit comprises a data acquisition board card, a fault extraction module, an information fusion module and a daily fault judgment module which are connected with each other; the data acquisition board card is communicated with the number conditioning circuit.

The information fusion unit is a module which extracts the characteristics of the fault and fuses the information by using the information acquired by the multi-source information acquisition unit. The system mainly comprises a data acquisition board card, a fault extraction module, an information fusion unit and a daily fault judgment module.

The data acquisition board card acquires electrical signals of a GIS working state of the multi-source information acquisition unit, converts the electrical signals of the GIS working state into digital signals and inputs the digital signals into the PC.

And the fault characteristic extraction module fuses the digital signals into fault state vectors by using GIS working state data acquired by the data acquisition board card, and transmits the normal state vectors and the fault state vectors to the information fusion module in a classified manner.

The fault decision and prediction unit is used for alarming according to a comparison result of the electric analog signal and a preset value; the fault decision and prediction unit comprises a fault state fuzzy feature library, a fault judgment module and a fault prediction module which are connected with each other; the fault state fuzzy feature library is communicated with the information fusion module, the fault judgment module is communicated with the fault state fuzzy feature library, and the fault prediction module is respectively communicated with the daily fault judgment module and the fault judgment module.

The information fusion module is obtained by performing dimensionality reduction on the working state vector X by using a singular value grading method through a Principal Component Analysis (PCA) methodAnd (4) daily fault characteristics, namely directly recording abnormal characteristics of the GIS system according to the gas leakage condition of SF6 gas and the abnormal condition of daily power loss, and establishing a daily fault characteristic library by using the two pieces of high-level information to calculate the daily fault weight.

A fault decision and forecast unit for filtering the GIS working state according to the information fusion unitA TS model-based fault state fuzzy feature library is established by utilizing historical data, the weight of the current fault state can be calculated, and then the weight is input to a fault judgment module to judge the fault and forecast the future fault.

And for the daily fault judgment, the daily fault weight calculated by the daily fault judgment module in the information fusion module does not need to be combined with other information, and the GIS system can be directly judged to have a fault if air leakage and abnormal daily power occur, and has the highest priority.

As shown in fig. 2, a PCA principal component analysis method is mainly used to perform dimension reduction processing on the GIS operating state X, and the specific algorithm is as follows:

And (6) carrying out data standardization processing. The multi-information acquisition module acquires the working state of the GIS system in the period:

X₁＝(x₁,x₂...x_n)^T

writing the state of n cycles into a state matrix form:

X_m＝(X₁,X₂...X_n) (ii) a Wherein, X₁,X₂,...,X_nRespectively representing the working states of 1 st to nth periods of the GIS switch system; the working state of each period of the GIS switching system is composed of n-dimensional detection data collected by a monitor;

and carrying out normalization processing on Xm to obtain:

wherein: n denotes the dimension of the data, T denotes the matrix transpose,represents X_mMean value of (a) represents X_mStandard deviation of (d);

singular value decomposition, singular value decomposition is carried out on the covariance matrix:

wherein,wherein: s represents the covariance matrix, σ, between the data elements_i1, 2.. times.n, which respectively represent the 1 st to nth singular values of the matrix SΛ represents a matrix of singular values, U and U^TIs a representation of singular value decomposition;

taking principal element elements, and taking the first k principal elements of the lambda as analysis elements; and taking the corresponding P ═ u (u)₁,u₂...u_k) (ii) a Wherein u is_iI ═ 1, 2.. times, k, which respectively represent the first 1 st to first k th vectors of the corresponding analysis elements of the matrix U, P represents the vector formed by the first k vectors of the matrix U;

the dimension-reduced form of X is obtained,wherein T ═ XP.

Fig. 3 is a diagram illustrating a fuzzy library of failure features according to the present invention.

The module mainly fuzzifies input quantity according to the fuzzy rule of the TS fuzzy model by the following main mechanism:

and (1.1) fuzzy rules.

Contribution component y of ith TS fuzzy rule to system outputⁱ(k +1) can be expressed by the statement "If … Then" as follows:

$\begin{array}{c}{R}^i:If{x}_1\left(k\right)is{A}_1^{i}and{x}_2\left(k\right)is{A}_2^{i}and\mathrm{...}and{x}_n\left(k\right)is{A}_n^i\\ Then{y}^i(k+1)=p_0^i+p_1^i{x}_1+\mathrm{...}+p_n^i{x}_n;\end{array},i=1,2,\mathrm{...}c$

wherein c is the number of fuzzy rules, n is the number of input variables of the TS fuzzy model, and x₁(k),x₂(k),…,x_n(k) Each of the n-dimensional input/output data regression variables of the TS fuzzy model at the k-th time, x (k) ═ x₁(k),x₂(k),…,x_n(k)]An input vector of the TS fuzzy model at the k moment;a fuzzy set with a linear membership function representing n fuzzy subspaces corresponding to the ith fuzzy rule,n-dimensional back-part parameters of the ith fuzzy rule;

and (1.2) outputting the calculation.

Definition of beta_iTo blur the fitness of rule i, the output y (k +1) of the model at time (k +1) can be calculated by the following formula:

$\begin{array}{c}y(k+1)=\underset{i=1}{\overset{c}{\Σ}}{\mathrm{\β}}_i{y}^i(k+1)=\underset{i=1}{\overset{c}{\Σ}}{\mathrm{\β}}_i(p_0^i+p_1^i{x}_1(k)+\mathrm{...}+p_n^i{x}_n(k\left)\right)\\ =\underset{i=1}{\overset{c}{\Σ}}(p_0^i+p_1^i+\mathrm{...}+p_n^i){({\mathrm{\β}}_i+{\mathrm{\β}}_i{x}_1(k)+\mathrm{...}+{\mathrm{\β}}_i{x}_n(k\left)\right)}^T\end{array}$

wherein, beta_iDefining the fitness of the ith fuzzy rule

$\left\{\begin{array}{c}\mathrm{\Θ}\left(k\right)={\[{\mathrm{\θ}}_1,{\mathrm{\θ}}_2,\mathrm{...},{\mathrm{\θ}}_r\]}^T={\[p_{10},p_{20},\mathrm{...},p_{c0},p_{11},p_{21},\mathrm{...},p_{c1},\mathrm{...},p_{cn}\]}^T;\\ \mathrm{\Φ}\left(k\right)={\[{\mathrm{\β}}_1,\mathrm{...},{\mathrm{\β}}_c,{\mathrm{\β}}_1{x}_1\left(k\right),\mathrm{...},{\mathrm{\β}}_c{x}_1\left(k\right),\mathrm{...},{\mathrm{\β}}_1{x}_n\left(k\right),\mathrm{...},{\mathrm{\β}}_c{x}_n\left(k\right)\]}^T;\end{array}\right.$

Where r ═ c (n +1), one can obtain:

y(k+1)＝Φ(k)^TΘ(k)

where Φ (k) and Θ (k) represent the parameters of the fuzzy rule at the kth instant, θ₁,θ₂,…,θ_rIs a column vector of Θ (k), r denotes the column vector index, r ═ c (n +1), Θ (k) sets each c column vectors as a set of column vectors in ascending order, p _uvIs the u-th column vector in the column vector group with the sequence number v of theta (k), u is 1 to C, v is 0 to n, p_uvThe back-piece parameter at the kth instant is indicated.

Establishing a GIS fault feature fuzzy rule base according to the fuzzy rule, and identifying two parameters phi (k) and theta (k) of the fuzzy rule at the kth moment according to experimental data by a system identification method, wherein the specific process comprises the following steps:

(2.1) statistics of Fault characteristics

According to a large number of GIS switch experiments, the state of the data in the reduced dimension is countedThe probability y of a failure occurring in the case, and the probability yf of a failure occurring in the next operation. There is no experimentally obtained probability, which is found by multidimensional difference.

(2.2) fuzzy rules

Cluster center vector V (1) ═ V₁(1),v₂(1),…,v_c(1)]The initial value of (a) can be obtained by using a C clustering algorithm through off-line identification of a part of experimental data, and then updated. Obtaining the clustering center vector v of the input x (k) and the ith fuzzy rule corresponding to the k-1 time by the following equation_iDistance d of (k-1)_i′(k)：

Wherein x (k) ═ x₁(k),x₂(k),...,x_n(k))，x_j(k) Representing the jth input vector of the TS fuzzy model at the kth moment, wherein c is the fuzzy rule number;

evaluation input x (k) for each cluster center v_i(k-1), i ═ 1,2, …, degree of membership u of c_i' (k) wherein v _i(k-1) represents the cluster center vector corresponding to the ith fuzzy rule at the k-1 time,wherein f represents a fuzzy factor, and the value of f is more than 1; d'_j(k) Representing the distance between the input x (k) of the TS fuzzy model at the kth moment and the clustering center vector of the jth fuzzy rule corresponding to the kth-1 moment;

and modifying the clustering center vector V (k-1) by using the following formula based on the membership degree and the fuzzy rule learning rate of the (k-1) th time x (k): v. of_i(k)＝v_i(k-1)+λu_i′(k)²[x(k)-v_i(k-1)](ii) a Wherein λ represents a learning rate, and a range of a value of λ is a number smaller than 1 and larger than 0; v. of_i(k) A clustering center vector corresponding to the ith fuzzy rule at the kth moment;

updating the distance d between the input x (k) and the central point according to the newly obtained cluster central vector_i' (k) and degree of membership u_i′(k)：

Calculating the fitness beta of the ith fuzzy rule to the system output_i：u_jIs determined by the degree of membership u_i' (k) obtaining a vector: wherein u is^jThe first j vector representing the corresponding analysis element of the matrix U;

calculate the fitness of the ith rule to the system output when the input is x (k):

${\mathrm{\β}}_i=\underset{j=1}{\overset{c}{\Σ}}\left(\frac{u_i}{u_j}\right),i=1,2,...,c$

the vector can then be found:

Φ(k)＝[β₁,…,β_c,β₁x₁(k),…,β_cx₁(k)，…,β₁x_n(k),…,β_cx_n(k)]^T.

according to the formula y (k +1) ═ phi (k)^TΘ (k), where y (k +1) and Φ (k) are known, Θ (k) ═ Φ (Φ) by the least square method^TΦ)^-1Φ^TY (k + 1); where y (k +1) represents the contribution component of the fuzzy rule to the system output at time k +1, and Φ (k) and Θ (k) represent the parameters of the fuzzy rule at time k.

Through the steps, two parameters phi (k) and theta (k) of the fuzzy rule are obtained through a system identification method, and a y fuzzy specification library is established. By the same method, a fuzzy rule base of yf can be established.

(3.1) failure judgment Overall flow

And according to the state acquired by the multi-source information acquisition unit, after the dimensionality reduction of the precision information fusion unit, inputting the state into the established fuzzy rule base for calculation to obtain the weight of the fault in the state, and judging and predicting the fault according to the weight.

(3.2) failure determination mechanism

As shown in fig. 3, the y value is compared with the upper threshold Au, and if the current fault weight exceeds the threshold, the fault is directly determined to occur this time;

if the y value is smaller than the lower limit threshold AL, judging that the current state is in a normal working state and no fault occurs;

if the y value is between the upper threshold Au and the lower threshold AL, the fault can not be judged by using y; and starting a yf judgment mechanism to forecast the fault according to the same method.

The fault can be judged and forecasted according to the method.

The foregoing description of the specific embodiments of the invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by those skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.

Claims (5)

1. The utility model provides a GIS switchgear motion state monitoring system which characterized in that includes:

the system comprises a multi-source information acquisition unit, a GIS switching system and a control unit, wherein the multi-source information acquisition unit detects the working state of the GIS switching system and forms a detection signal, and the multi-source information acquisition unit conditions the detection signal into an electric analog signal;

the information fusion unit is arranged in the computer terminal and used for receiving the electric analog signal, comparing the electric analog signal with a preset value and judging the working state of the GIS switching system;

and the fault decision and prediction unit is used for alarming or predicting according to the comparison result of the electric analog signal and the preset value.

2. The GIS switchgear operating state monitoring system of claim 1, wherein a monitor and a signal conditioning circuit are provided in the multi-source information collecting unit, which are connected to each other.

3. The GIS switchgear operating state monitoring system of claim 2, wherein the monitor comprises a relay control loop state detection module, a connection execution high-voltage switch state detection module, a mechanical mechanism state detection module, and a gas leakage detection module.

4. The GIS switchgear action state monitoring system of claim 2, wherein the information fusion unit comprises a data acquisition board card, a fault extraction module, an information fusion module and a daily fault judgment module which are connected with each other; the data acquisition board card is communicated with the signal conditioning circuit.

5. The GIS switchgear action state monitoring system of claim 4, wherein the fault decision and prediction unit comprises a fault state fuzzy feature library, a fault judgment module and a fault prediction module which are connected with each other; the fault state fuzzy feature library is communicated with the information fusion module, the fault judgment module is communicated with the fault state fuzzy feature library, the fault forecast module is respectively communicated with the daily fault judgment module and the fault judgment module, and the daily fault judgment module is communicated with the daily fault judgment module.