CN110175760A - A kind of paper oil insulation method for diagnosing status of return voltage and depolarization current composite character amount - Google Patents

Abstract

The present invention relates to the paper oil insulation method for diagnosing status of a kind of return voltage and depolarization current composite character amount, the Time Domain Dielectric Spectroscopy data obtained according to more paper oil insulation transformer on-site tests, according to the different characteristic amount for being able to reflect state of insulation, transformer matter-element model is established；Using the weight coefficient that can open up Hierarchy Analysis Method and determine each layer index of transformer；Then, the result obtained based on a large amount of measured data and calculating analysis, it is determined that the Classical field of four kinds of state of insulations of transformer and section domain；Finally, correlation function value of the follow-up transformer about four kinds of state of insulations is determined by correlation function, so that it is determined that the class of insulation of follow-up transformer.The present invention can more effectively realize the quantitative Diagnosis of paper oil insulation state, further investigation insulation ag(e)ing etc..

Description

Oil paper insulation state diagnosis method based on recovery voltage and depolarization current mixed characteristic quantity

Technical Field

The invention relates to the field of oil paper insulation aging diagnosis research, in particular to an oil paper insulation state diagnosis method based on a return voltage and depolarization current mixed characteristic quantity.

Background

In the field of oil paper insulation aging diagnosis research, certain research is generally carried out on the oil paper insulation state analysis by means of dielectric response characteristic parameters. Has been paid attention by experts and is widely used. However, most researchers only stay in qualitative analysis when evaluating the insulation state of the oil paper by applying dielectric characteristic parameter analysis, cannot make powerful diagnosis and analysis on a single transformer, and can obtain a diagnosis result by comparing and analyzing actual detection data of two or more power transformers. The aging evaluation method not only needs a complicated detection process, but also is difficult to ensure the accuracy of a qualitative analysis result.

Disclosure of Invention

In view of the above, the present invention is directed to a method for diagnosing an insulation state of an oiled paper based on a characteristic quantity of a mixture of a recovery voltage and a depolarization current, which can effectively diagnose the grade of the insulation state of the oiled paper.

The invention is realized by adopting the following scheme: a method for diagnosing the insulation state of oiled paper by using the characteristic quantity of mixed recovery voltage and depolarization current comprises the following steps:

step S1: measuring time domain dielectric spectrum by using a plurality of supplied oil paper transformers, and acquiring a plurality of characteristic values for evaluating the insulation state to establish an attribute matrix V'_IJKAs shown in formula (1);

moment of torsionMatrix X_IJKIn the method, the insulation condition of the transformer is divided into N₁、N₂、…、N_mM insulation states, the number of transformers in each insulation state is t₁、t₂、…、t_m(ii) a The total number of the transformers is t ═ t₁+t₂+...+t_m； x_ijkRepresenting the state of insulation N attached to the transformer_jAn ith insulation evaluation characteristic value i of the kth transformer is 1, 2.. times, n; j is 1,2,. said, m; k 1,2,. t;

step S2: the attribute matrix X in step S1_IJKCarrying out dimensionless transformation on the characteristic values of the middle rows according to the benefit measure evaluation coefficient in the formula (2) or the cost measure evaluation coefficient in the formula (3); after dimensionless transformation, the attribute matrix X is transformed_IJKConversion into a dimensionless matrix V_IJKAs shown in formula (4);

benefit measure evaluation coefficient:

cost measure evaluation coefficient:

in the formulae (2) and (3), x_i-minIs the minimum value of the characteristic value of each row; x is the number of_i-maxThe maximum value of the characteristic value of each row; v. of_ijkThe characteristic value is the characteristic value after dimensionless change;

step S3: firstly, providing a transformer with diagnosis, carrying out time domain dielectric spectrum test on the transformer to be diagnosed, acquiring a plurality of characteristic values of the estimated insulation condition, and determining an object model of the transformer to be diagnosed as follows:

secondly, determining the classical domain R of each state class matter element of the transformer_pComprises the following steps:

in the formula (6), v_pij1,2, n; j is 1,2,. said, m; representing historical statistical data and characteristic quantity c obtained by time domain dielectric spectrum test of t transformers_iClassical field v in the j-th insulation condition_pij＝(μ_ij-3σ_ij,μ_ij+3σ_ij)；μ_ij、σ_ijRespectively represent the characteristic values c_iIn j-th state, t is counted_jThe mean and standard deviation of the transformer; t is t_jFor the historical statistical data of the time domain dielectric spectrum, the insulation state of the transformer is N_jThe number of the cells;

finally, the section region R of each state grade object element of the transformer is determined_qComprises the following steps:

in the formula (7), v_qij1,2, n; j is 1,2,. said, m; representing historical data and characteristic quantity c obtained by testing return voltage of t transformers_iSection area v under j insulation condition_qij＝(v_ijmin,v_ijmax)；v_ijminRepresents a characteristic value c_iMinimum value of corresponding characteristic value in j-th statev_ijmaxDenotes c_iMaximum value of characteristic value corresponding to j-th state

Step S4: determining the weight coefficient of each characteristic value in the step S2 by using an extension analytic hierarchy process;

step S5: calculating the relevance of each characteristic value in the step S2 on the insulation state grade;

step S6: determining the relevance of the object to be evaluated with respect to each state grade by combining the weight coefficients obtained in the step S4, as shown in formula (8);

in the formula (8), K_j(N₀) The evaluation value is a comprehensive value and represents the relevance of the object element to be evaluated with respect to the insulation state grade of each transformer;

step S7: determining the state grade of the transformer to be diagnosed according to the formula (9), and determining the maximum correlation degree K_j′Corresponding transformer insulation state grade N_jJ is 1,2,3,4, which is the state grade of the transformer to be diagnosed;

K_j′＝max{K_j(N₀)}(j＝1,2,3,4) (9)。

further, the step S4 specifically includes the following steps:

step S41: constructing an extension judgment matrix E ═ E_ij)_n×nWherein e is_ijThe following condition needs to be satisfied, for all i, j ═ 1, 2.And ise_ii＝1，

Step S42: solving a left decision matrixAnd right decision matrixRespectively corresponding to the eigenvectors x^-And x⁺：

For x^-And x⁺Carrying out normalization treatment to obtain X^-And X⁺：

Wherein,

step S43: obtaining a comparison weight coefficient λ ═ λ₁,λ₂,...,λ_n)^T＝(rX^-,sX⁺) Wherein

Step S44: solving the single-layer sequencing weight of each index by comparing the weight coefficient lambda;

let two interval numbers e ═ e^-,e⁺)，f＝(f^-,f⁺)，V(e≥f)＝2(e⁺-f^-)/[(e⁺-e^-)+(f⁺-f^-)]；

And (3) calculating:

V(λ_i＞λ_j)(i＝1,2,...,n；i≠j) (10)

if V (lambda)_i＞λ_j) More than or equal to 0(i ═ 1, 2.., n; i ≠ j), then ω_j＝1，ω_i＝V(λ_i＞λ_j)，(i＝1,2,...,n；i≠j)；

Wherein ω is_iRepresenting the single sequence of the ith element of the lower layer in the transformer hierarchical structure framework to a certain factor of the upper layer, and obtaining omega (omega) after normalization₁,ω₂,…,ω_n) And the single-rank-order weight vector of each element in the lower layer to a certain element in the upper layer is represented.

Further, the specific content of step S5 is:

correlation function K (v)_i) Comprises the following steps:

in the formula (11), ρ (v)_i,v_pij)、ρ(v_i,v_qij) Respectively represent the characteristic quantities v_iAnd the classical domain v_pijAnd section area v_qijThe distance of (a) to (b),

the specific calculation formula is shown as formula (12) and formula (13);

in the formula (11), the correlation function K (v)_i) Reflecting the characteristic quantity v of the transformer to be diagnosed_iAnd an insulation state rating of N_jDegree of transformer correlation, correlation function K (v)_i) The larger the characteristic value v of the transformer to be diagnosed is, the larger the characteristic value v_iSubordinate to transformer state class N_jThe more attributes there are.

Compared with the prior art, the invention has the following beneficial effects:

at present, the oil paper insulation state of the power transformer is evaluated only by qualitative analysis, and the quality of the oil paper insulation state can be evaluated only by comparing and analyzing two or more actual detection data values of the transformer. The method can realize quantitative evaluation of the oil-paper insulation state of the power transformer and can effectively diagnose the grade of the oil-paper insulation state.

Detailed Description

The present invention will be further described with reference to the following examples.

Firstly, the insulation state of the transformer is divided into N_jAnd (j) 1,2, a.m), establishing classic domains and section domains of various insulation state grades of the transformer according to the grasped insulation conditions and insulation classification conditions of the t transformers, wherein the classic domains and the section domains can be used for representing the condition of the state grade of the transformer, calculating the association degree of the transformer to be diagnosed about each state grade of the transformer through an association function, and performing one-to-one reference association on the transformer to be diagnosed and each state grade of the established transformer so as to determine the state grade of the transformer.

Specifically, the present embodiment provides a method for diagnosing an insulation state of an oiled paper based on a mixed characteristic quantity of a recovery voltage and a depolarization current, including the following steps:

step S1: measuring time domain dielectric spectrum by using a plurality of supplied oil paper transformers, and acquiring a plurality of characteristic values for evaluating the insulation state to establish an attribute matrix V'_IJKAs shown in formula (1);

in matrix X_IJKIn the method, the insulation condition of the transformer is divided into N₁、N₂、…、N_mM insulation states, the number of transformers in each insulation state is t₁、t₂、…、t_m(ii) a The total number of the transformers is t ═ t₁+t₂+...+t_m；x_ijkRepresenting the state of insulation N attached to the transformer_jAn ith insulation evaluation characteristic value i of the kth transformer is 1, 2.. times, n; j is 1,2,. said, m; k 1,2,. t;

step S2: the attribute matrix X in step S1_IJKCarrying out dimensionless transformation on the characteristic values of the middle rows according to the benefit measure evaluation coefficient in the formula (2) or the cost measure evaluation coefficient in the formula (3); after dimensionless transformation, the attribute matrix X is transformed_IJKConversion into a dimensionless matrix V_IJKAs shown in formula (4);

benefit measure evaluation coefficient:

cost measure evaluation coefficient:

in the formulae (2) and (3), x_i-minIs the minimum value of the characteristic value of each row; x is the number of_i-maxThe maximum value of the characteristic value of each row; v. of_ijkThe characteristic value is the characteristic value after dimensionless change;

step S3: firstly, providing a transformer to be diagnosed, performing time domain dielectric spectrum test on the transformer to be diagnosed, acquiring a plurality of characteristic values of the estimated insulation condition, and determining an object model of the transformer to be diagnosed as follows:

secondly, determining the classical domain R of each state class matter element of the transformer_pComprises the following steps:

in the formula (6), v_pij1,2, n; j is 1,2,. said, m; representing historical statistical data and characteristic quantity c obtained by time domain dielectric spectrum test of t transformers_iClassical field v in the j-th insulation condition_pij＝(μ_ij-3σ_ij,μ_ij+3σ_ij)；μ_ij、σ_ijRespectively represent the characteristic values c_iIn j-th state, t is counted_jThe mean and standard deviation of the transformer; t is t_jFor the historical statistical data of the time domain dielectric spectrum, the insulation state of the transformer is N_jThe number of the cells;

finally, the section region R of each state grade object element of the transformer is determined_qComprises the following steps:

in the formula (7), v_qij1,2, n; j is 1,2,. said, m; representing historical data and characteristic quantity c obtained by testing return voltage of t transformers_iSection area v under j insulation condition_qij＝(v_ijmin,v_ijmax)；v_ijminRepresents a characteristic value c_iOf corresponding characteristic values in the j-th stateMinimum valuev_ijmaxDenotes c_iMaximum value of characteristic value corresponding to j-th state

Step S4: determining the weight coefficient of each characteristic value in the step S2 by using an extension analytic hierarchy process;

step S5: calculating the relevance of each characteristic value in the step S2 on the insulation state grade;

step S6: determining the relevance of the object to be evaluated with respect to each state grade by combining the weight coefficients obtained in the step S4, as shown in formula (8);

in the formula (8), K_j(N₀) The method is a comprehensive value under the condition of considering the weight of the characteristic value, and represents the relevance of an object element to be evaluated with respect to the insulation state grade of each transformer;

step S7: determining the state grade of the transformer to be diagnosed according to the formula (9), and determining the maximum correlation degree K_j′Corresponding transformer insulation state grade N_jJ is 1,2,3,4, which is the state grade of the transformer to be diagnosed;

K_j′＝max{K_j(N₀)}(j＝1,2,3,4) (9)。

further, the step S4 specifically includes the following steps:

step S41: constructing an extension judgment matrix E ═ E_ij)_n×nWherein e_ijThe following condition needs to be satisfied, for all i, j ═ 1, 2.And ise_ii＝1，

Step S42: solving a left decision matrixAnd right decision matrixRespectively corresponding to the eigenvectors x^-And x⁺：

For x^-And x⁺Carrying out normalization treatment to obtain X^-And X⁺：

Wherein,

step S43: obtaining a comparison weight coefficient λ ═ λ₁,λ₂,...,λ_n)^T＝(rX^-,sX⁺) Wherein

Step S44: solving the single-layer sequencing weight of each index by comparing the weight coefficient lambda;

let two interval numbers e ═ e^-,e⁺)，f＝(f^-,f⁺)，V(e≥f)＝2(e⁺-f^-)/[(e⁺-e^-)+(f⁺-f^-)]；

And (3) calculating:

V(λ_i＞λ_j)(i＝1,2,...,n；i≠j) (10)

if V (lambda)_i＞λ_j) More than or equal to 0(i ═ 1, 2.., n; i ≠ j), then ω_j＝1，ω_i＝V(λ_i＞λ_j)，(i＝1,2,...,n；i≠j)；

Wherein ω is_iRepresenting the single sequence of the ith element of the lower layer in the transformer hierarchical structure framework to a certain factor of the upper layer, and obtaining omega (omega) after normalization₁,ω₂,…,ω_n) And the single-rank-order weight vector of each element in the lower layer to a certain element in the upper layer is represented.

Further, the specific content of step S5 is:

correlation function K (v)_i) Comprises the following steps:

in the formula (11), ρ (v)_i,v_pij)、ρ(v_i,v_qij) Respectively represent the characteristic quantities v_iAnd the classical domain v_pijAnd section area v_qijThe specific calculation formula of (a) is shown in formula (12) and formula (13);

in the formula (11), the correlation function K (v)_i) Reflecting the characteristic quantity v of the transformer to be diagnosed_iAnd an insulation state rating of N_jDegree of transformer correlation, correlation function K (v)_i) The larger the characteristic value v of the transformer to be diagnosed is, the larger the characteristic value v_iSubordinate to transformer state class N_jThe more attributes there are.

Preferably, in this embodiment, the RVM5461 tester measures the recovery voltage of nearly 60 transformers, and the maximum value c of the recovery voltage of eight characteristic quantities, which can better reflect the insulation state of the transformer, is extracted from the statistical data of the transformers obtained through actual measurement and analysis₁Main time constant c₂Initial slope c₃Peak time c₄And an insulation resistance c₅Number of polarization branches c₆Maximum depolarization energy c₇Depolarization energy maximum time c₈. According to the regulations of various indexes of preventive test regulations of power equipment, the insulation state of the transformer oil paper can be classified as good (N)₁) Insulating state, etc. (N)₂) Insulation slight aging (N)₃) Severe aging of insulation (N)₄)4 states.

Establishing an attribute matrix V 'of 10 transformers to be diagnosed according to the step 1'_IJK：

Obtaining attribute matrix V'_IJKThe maximum and minimum values of the characteristic values of each row. X_-min＝(162.58，849.57，30.57 ，246.35，5.68，4，0.054，42.2)^T，X_-max＝(598.36，2578.68，189.57，651.25，20.85，7，0.198 ，204.3)^T。

According to the step 2, an attribute matrix V'_IJKEach characteristic value is transformed into a dimensionless matrix V according to the evaluation indexes of cost measure and benefit measure of formula (2) or formula (3)_IJK：

According to the step 3, the characteristic value c of the transformer in the historical database is solved_iWith respect to the insulating state N_jThe mean and variance of (a) are shown in the following matrix:

according to the step 3, establishing the classical domains R of the eight characteristic quantities respectively related to the four matter element state grades according to the formulas (6) and (7)_pAnd the node region R_qAs follows:

according to the step 4, inviting multiple experts to obtain a comprehensive extension judgment matrix as follows:

obtaining the weight of each grade information index of the transformer according to the extension analytic hierarchy process in the step 4 as follows: the first-order index has a weight of omega₁＝(0.57720.0372, 0.3856); weight omega of each index of return voltage₂₁(0.2599,0.4471, 0.2779, 0.0152); each index weight omega of Debye model₂₂(0.0829, 0.9171); each index weight omega of depolarized current₂₃＝ (0.0415,0.9585)。

According to step 5, the relevance of each feature with respect to each state grade is first obtained according to the formula of the relevance function.

According to the step 6, combining the weight of the characteristic value index obtained in the step 4, finally obtaining the association degree of the transformer to be diagnosed about each state equivalent pole as shown in the following table:

TABLE 1 correlation of Ten transformers with respect to each state class

According to step 7, the insulation state class of the transformer can finally be determined.

As can be seen from table 1: the diagnostic results of 10 transformers are substantially consistent with the actual insulation conditions. Now with T₁The transformer is taken as an example to illustrate the diagnosis result: four calculated degrees of relevance N₁,N₂,N₃,N₄Respectively-6.3841, -11.946, -0.168, 0.9489, wherein the maximum value N₄0.9489, the corresponding state is "insulation severe aging", which is the same as the actual state. Furthermore, at T₂The maximum value of the correlation values calculated by the transformer is N₃I.e., the corresponding insulation state is "slightly aged", it is also substantially consistent with the actual condition. Due to space limitations, the diagnostic results of other transformers are not described.

The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.

Claims (3)

1. A recovery voltage and depolarization current mixed characteristic quantity oiled paper insulation state diagnosis method is characterized by comprising the following steps: the method comprises the following steps:

step S1: measuring time domain dielectric spectrum by using a plurality of supplied oil paper transformers, and acquiring a plurality of characteristic values for evaluating the insulation state to establish an attribute matrix V'_IJKAs shown in formula (1);

in matrix X_IJKIn the method, the insulation condition of the transformer is divided into N₁、N₂、…、N_mM insulation states, the number of transformers in each insulation state is t₁、t₂、…、t_m(ii) a The total number of the transformers is t ═ t₁+t₂+...+t_m；x_ijkRepresenting the state of insulation N attached to the transformer_jAn ith insulation evaluation characteristic value i of the kth transformer is 1, 2.. times, n; j is 1,2,. said, m; k 1,2,. t;

step S2: the attribute matrix X in step S1_IJKCarrying out dimensionless transformation on the characteristic values of the middle rows according to the benefit measure evaluation coefficient in the formula (2) or the cost measure evaluation coefficient in the formula (3); after dimensionless transformation, the attribute matrix X is transformed_IJKConversion into a dimensionless matrix V_IJKAs shown in formula (4);

benefit measure evaluation coefficient:

cost measure evaluation coefficient:

in the formulae (2) and (3), x_i-minIs the minimum value of the characteristic value of each row; x is the number of_i-maxThe maximum value of the characteristic value of each row; v. of_ijkThe characteristic value is the characteristic value after dimensionless change;

step S3: firstly, providing a transformer to be diagnosed, performing time domain dielectric spectrum test on the transformer to be diagnosed, acquiring a plurality of characteristic values of the estimated insulation condition, and determining an object model of the transformer to be diagnosed as follows:

secondly, determining the classical domain R of each state class matter element of the transformer_pComprises the following steps:

in the formula (6), v_pij1,2, n; j is 1,2,. said, m; representing historical statistical data and characteristic quantity c obtained by time domain dielectric spectrum test of t transformers_iClassical field v in the j-th insulation condition_pij＝(μ_ij-3σ_ij,μ_ij+3σ_ij)；μ_ij、σ_ijRespectively represent the characteristic values c_iIn j-th state, t is counted_jThe mean and standard deviation of the transformer; t is t_jFor the transformer insulation state N in the historical statistical data of the time domain dielectric spectrum_jThe number of the cells;

finally, the section region R of each state grade object element of the transformer is determined_qComprises the following steps:

in the formula (7), v_qij1,2, n; j is 1,2,. said, m; representing historical data and characteristic quantity c obtained by testing return voltage of t transformers_iSection area v under j insulation condition_qij＝(v_ijmin,v_ijmax)；v_ijminRepresents a characteristic value c_iMinimum value of corresponding characteristic value in j-th statev_ijmaxDenotes c_iMaximum value of characteristic value corresponding to j-th state

Step S4: determining the weight coefficient of each characteristic value in the step S2 by using an extension analytic hierarchy process;

step S5: calculating the relevance of each characteristic value in the step S2 on the insulation state grade;

step S6: determining the relevance of the object to be evaluated with respect to each state grade by combining the weight coefficients obtained in the step S4, as shown in formula (8);

in the formula (8), K_j(N₀) The evaluation value is a comprehensive value and represents the relevance of the object element to be evaluated with respect to the insulation state grade of each transformer;

step S7: determining the state grade of the transformer to be diagnosed according to the formula (9), and determining the maximum correlation degree K_j'Corresponding transformer insulation state grade N_jJ is 1,2,3,4, which is the state grade of the transformer to be diagnosed;

K_j'＝max{K_j(N₀)}(j＝1,2,3,4) (9)。

2. the method for diagnosing the insulation state of the oiled paper with the recovery voltage and depolarization current mixed characteristic quantity according to claim 1, wherein the method comprises the following steps: the step S4 specifically includes the following steps:

step S41: constructing an extension judgment matrix E ═ E_ij)_n×nWherein e is_ijThe following condition needs to be satisfied, for all i, j ═ 1, 2.And ise_ii＝1，

Step S42: solving a left decision matrixAnd right decision matrixRespectively corresponding to the eigenvectors x^-And x⁺：

For x^-And x⁺Carrying out normalization treatment to obtain X^-And X⁺：

Wherein,

step S43: obtaining a comparison weight coefficient λ ═ λ₁,λ₂,...,λ_n)^T＝(rX^-,sX⁺) Wherein

Step S44: solving the single-layer sequencing weight of each index by comparing the weight coefficient lambda;

let two interval numbers e ═ e^-,e⁺)，f＝(f^-,f⁺)，V(e≥f)＝2(e⁺-f^-)/[(e⁺-e^-)+(f⁺-f^-)]；

And (3) calculating:

V(λ_i＞λ_j)(i＝1,2,...,n；i≠j) (10)

if V (lambda)_i＞λ_j) More than or equal to 0(i ═ 1, 2.., n; i ≠ j), then ω_j＝1，ω_i＝V(λ_i＞λ_j)，(i＝1,2,...,n；i≠j)；

Wherein ω is_iRepresenting the single sequence of the ith element of the lower layer in the transformer hierarchical structure framework to a certain factor of the upper layer, and obtaining omega (omega) after normalization₁,ω₂,…,ω_n) And the single-rank-order weight vector of each element in the lower layer to a certain element in the upper layer is represented.

3. The method for diagnosing the insulation state of the oiled paper with the recovery voltage and depolarization current mixed characteristic quantity according to claim 1, wherein the method comprises the following steps: the specific content of step S5 is:

correlation function K (v)_i) Comprises the following steps:

in the formula (11), ρ (v)_i,v_pij)、ρ(v_i,v_qij) Respectively represent the characteristic quantities v_iAnd the classical domain v_pijAnd section area v_qijThe specific calculation formula of (a) is shown in formula (12) and formula (13);

in the formula (11), the correlation function K (v)_i) Reflecting the characteristic quantity v of the transformer to be diagnosed_iAnd an insulation state rating of N_jDegree of transformer correlation, correlation function K (v)_i) The larger the characteristic value v of the transformer to be diagnosed is, the larger the characteristic value v_iSubordinate to transformer state class N_jThe more attributes there are.