CN105405066A - Distribution transformer health index determination method - Google Patents

Abstract

The present invention relates to a distribution transformer health index determination method. The method comprises: establishing a health index model, wherein an establishing process of the health index model comprises: determining a health index grade interval of a distribution transformer; selecting a key characteristic amount of the distribution transformer; performing standardization processing on data of the key characteristic amount of the distribution transformer; establishing an evaluation grade of the distribution transformer by using a trapezoidal membership function; acquiring a weight of the key characteristic amount of the distribution transformer; and determining an evaluation model according to the evaluation grade and the weight, thereby determining a health index of the distribution transformer by determining a modified weight. The technical scheme provided by the present invention can be widely applied to health status evaluation of the distribution transformer.

Description

A kind of substation transformer health index defining method

Technical field

The present invention relates to electrical engineering field, more specifically relate to a kind of substation transformer health index defining method.

Background technology

Traditionally, for ensureing the high reliability of Operation of Electric Systems, electric power enterprise adopt in planning and design improves individual device quality more, or guarantees subsystem at different levels with Redundancy Design, and even the global reliability of electric system, and less consideration rate of return on investment; Overhaul of the equipments is also mainly time maintenance and the repair based on condition of component of target with single devices.Rational planning and suitable maintenance can be electrical network and bring great benefit, improve the efficiency of equipment and the usefulness of electrical network; Otherwise will loss be brought, affect reliability and the economy of operation of power networks.Therefore the focus of Electric Power Network Planning principle and O&M maintenance decision Dou Shi academia and engineering circles concern all the time.

Substation transformer is one of visual plant in transformation and distribution system, its operational reliability to power distribution network general safety, stable, economical operation is significant.The substation transformer of China mainly concentrates on 110KV, 35KV and 10KV, have have a large capacity and a wide range, relatively transmission facility cost is lower, the features such as state facility imperfection, thus standing state assessment technology, mostly for transmission facility, is seldom applied to controller switching equipment at present.For above-mentioned analysis, a kind of substation transformer health index defining method is proposed, very important to solve the problems of the technologies described above.

Summary of the invention

The object of this invention is to provide a kind of substation transformer health index defining method, can be widely used in the middle of the evaluation of substation transformer health status.

For achieving the above object, the present invention by the following technical solutions: a kind of substation transformer health index defining method, comprises and set up health index model; The process of establishing of described health index model comprises:

Determine substation transformer health index grade interval;

Choose described substation transformer key feature amount;

The standardization of the data of described substation transformer key feature amount;

The opinion rating of described substation transformer is set up by trapezoidal membership function;

Obtain the weight of the key feature amount of described substation transformer;

According to described opinion rating and described weight determination assessment models, thus by determining that revising weight determines described substation transformer health index.

Described substation transformer health index HI grade interval is defined between [1,10]; When 1≤HI≤2 are fine state; When 2<HI≤5 are normal condition; When 5≤HI≤7 for poor state need strengthen maintenance; When 7<HI≤10 are severe conditions and existing obvious defect pipelines.

Described choosing of transformer key feature amount is determined according to oil chromatogram analysis, oiling test and electrical test; Described oil chromatogram analysis comprises acetylene content according to data, hydrogen content, carbon dioxide are determined relative to gas production rate, the relative gas production rate of carbon monoxide and total hydrocarbon content; Described oiling test comprises liquid water content, oil breakdown voltage, furfural content, interfacial tension, the paper degree of polymerization, oil loss and flash-point data for determining according to data; Described electric test comprises iron core grounding current, specific insulation, polarization index, absorptance and winding dielectric loss according to data and determines.

The standardization of the data of described substation transformer key feature amount is carried out with following formula:

$C_{ij}=\frac{y_{ij}-y_{ij0}}{y_{ijl}-y_{ij0}},y_{ij0}\&le;y_{ij}\&le;y_{ijl}$

In formula: y _ijfor actual institute measured value; y _ij0for initial value; y _ijlfor limit value; C _ijfor income value after standardization.

Described trapezoidal membership function is determined by following formula:

${\mathrm{\&mu;}}_1\left(f_{ij}\right)=\left\{\begin{array}{cc}1& f_{ij}\&le;a\\ \frac{b-f_{ij}}{b-a}& a<f_{ij}<b\\ 0& other\end{array}\right.$ ${\mathrm{\&mu;}}_2\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;aor{f}_{ij}\&GreaterEqual;d\\ \frac{f_{ij}-a}{b-a}& a<f_{ij}<b\\ 1& b\&le;f_{ij}\&le;c\\ \frac{d-f_{ij}}{d-c}& c<f_{ij}<d\end{array}\right.$

${\mathrm{\&mu;}}_4\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;eor{f}_{ij}\&GreaterEqual;h\\ \frac{f_{ij}-e}{f-e}& e<f_{ij}<f\\ 1& f\&le;f_{ij}\&le;g\\ \frac{h-f_{ij}}{h-g}& g<f_{ij}<h\end{array}\right.$ ${\mathrm{\&mu;}}_3\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;cor{f}_{ij}\&GreaterEqual;f\\ \frac{f_{ij}-c}{d-c}& c<f_{ij}<d\\ 1& d\&le;f_{ij}\&le;e\\ \frac{f-f_{ij}}{f-e}& e<f_{ij}<f\end{array}\right.$

${\mathrm{\&mu;}}_5\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;g\\ \frac{f_{ij}-g}{h-g}& g<f_{ij}<h\\ 1& f_{ij}\&GreaterEqual;h\end{array}\right.$

Wherein, f _ijfor the index degree of membership of data, μ ₁, μ ₂, μ ₃, μ ₄, μ ₅what be respectively five kinds of states is subordinate to grade (very good, good, to note, seriously, damage), and a, b, c, d, e, f, g, h are respectively the spacing value of each grade, as shown in Figure 3.

The process of the weight of the key feature amount of the described substation transformer of described acquisition comprises:

Ambiguity in definition judgment matrix:

Every a line of described judgment matrix is added:

${\mathrm{RS}}_i={\mathrm{\&Sigma;}}_{j=1}^n{c}_{ij}=({\mathrm{\&Sigma;}}_{j=1}^n{l}_{ij},{\mathrm{\&Sigma;}}_{j=1}^n{m}_{ij},{\mathrm{\&Sigma;}}_{j=1}^n{u}_{ij})$

Wherein, c _ijfor the vector of described matrix;

I-th ^ththe degree composite value of individual element is:

$S_i=\frac{{\mathrm{RS}}_i}{{\mathrm{\&Sigma;}}_{j=1}^n{\mathrm{RS}}_j}=(\frac{{\mathrm{\&Sigma;}}_{j=1}^n{l}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{u}_{kj}},\frac{{\mathrm{\&Sigma;}}_{j=1}^n{m}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{m}_{kj}},\frac{{\mathrm{\&Sigma;}}_{j=1}^n{u}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{l}_{kj}}),i=1,\mathrm{...},n$

Two Triangular Fuzzy Number M ₁and M ₂between possibility degree be defined as:

Wherein, M ₁=(l ₁, m ₁, u ₁), M ₂=(l ₂, m ₂, u ₂); for triangle ambiguity function;

And define:

$d\left(A_i\right)=\underset{\&Element;\{1,...,n\},j\&NotEqual;i}{\min }V(S_i\&GreaterEqual;S_j),i=1,...,n.$

Then weight vectors matrix is:

W＝(d(A ₁)，d(A ₂)，…，d(A _n)) ^T

The described assessment models index of index is the product of described opinion rating and described weight; And as original evidence; Described assessment models is:

$M_i\left(H\right)=\underset{j=1}{\overset{n}{\&Sigma;}}{w}_{ij}{U}_i\left(H\right)$

Wherein:

$U_i\left(H\right)=\left[\begin{array}{cccc}{\mathrm{\&mu;}}_1\left(f_{11}\right)& {\mathrm{\&mu;}}_2\left(f_{i1}\right)& ...& {\mathrm{\&mu;}}_5\left(f_{i1}\right)\\ {\mathrm{\&mu;}}_1\left(f_{i2}\right)& {\mathrm{\&mu;}}_2\left(f_{i2}\right)& ...& {\mathrm{\&mu;}}_5\left(f_{i2}\right)\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ {\mathrm{\&mu;}}_1\left(f_{in}\right)& {\mathrm{\&mu;}}_2\left(f_{in}\right)& ...& {\mathrm{\&mu;}}_5\left(f_{in}\right)\end{array}\right]$

W _ijfor weight, U _i(H) opinion rating of j index under i factor is represented; M _i(H) evaluation result of each index is represented.

Described substation transformer health index deterministic process comprises:

Suppose that two Basic Probability As-signment are m ₁and m ₂, the distance between them is:

$d(m_1,m_2)=\sqrt{\frac{1}{2}{(m_1-m_2)}^T\underset{\&OverBar;}{\underset{\&OverBar;}{D}}(m_1-m_2)}$

Wherein, it is one 2 ⁿ× 2 ⁿmatrix, the element in matrix is

$\begin{array}{cc}D(A,B)=\frac{\left|A\&cap;B\right|}{\left|A\&cup;B\right|}& A,B\&Element;R\left(\mathrm{\&Theta;}\right)\end{array}$

Namely the distance between burnt first A and B be about | A|, | B|, | A ∩ B| and | a function of A ∪ B|, || represent the base of burnt unit.

The similar matrix obtaining them is:

$Sim=\left[\begin{array}{cccccc}1& {\mathrm{Sim}}_{12}& ...& {\mathrm{Sim}}_{1j}& ...& {\mathrm{Sim}}_{1n}\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ {\mathrm{Sim}}_{i1}& {\mathrm{Sim}}_{i2}& ...& {\mathrm{Sim}}_{ij}& ...& {\mathrm{Sim}}_{in}\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ {\mathrm{Sim}}_{n1}& {\mathrm{Sim}}_{n2}& ...& {\mathrm{Sim}}_{nj}& ...& 1\end{array}\right]$

The correction weight that we obtain each evidence by similar matrix is:

$w_{ei}=\frac{Sup\left(m_i\right)}{\underset{i=1}{\overset{n}{\&Sigma;}}Sup\left(m_i\right)}$

Now, the evidence of correction is:

$MAE\left(m\right)=\underset{i=1}{\overset{n}{\&Sigma;}}(w_i\&times;m_i)$

Wherein, w _ifor revising weight;

Suppose there be n evidence, so final result passes through average evidence fusion n-1 time; Wherein in income value, maximal value is required substation transformer health index.

With immediate prior art ratio, the invention provides technical scheme and there is following excellent effect

1, technical solution of the present invention have for current substation transformer have a large capacity and a wide range, relatively transmission facility cost is lower, the features such as state facility imperfection propose a kind of defining method adopting fuzzy set theory, Fuzzy AHP and correction weighted average to merge the substation transformer health index combined;

2, technical solution of the present invention can avoid classical layer fractional analysis and classical evidence theory Problems existing;

3, technical solution of the present invention can effectively calculate substation transformer health index, and the repair based on condition of component for substation transformer provides favourable foundation;

4, technical solution of the present invention improves the efficiency of equipment and the usefulness of electrical network, reduces loss;

5, technical solution of the present invention is verified by substation transformer measured data, can draw, the Multiple Attribute Decision Model of this technical scheme is effective, can be widely used in the middle of the evaluation of substation transformer health status.

Accompanying drawing explanation

Fig. 1 is the substation transformer status information schematic diagram of the embodiment of the present invention;

Fig. 2 is the method flow diagram of the embodiment of the present invention;

Fig. 3 is the shape subordinate function pictorial diagram of the embodiment of the present invention.

Embodiment

Below in conjunction with embodiment, the invention will be described in further detail.

Embodiment 1:

The invention of this example provides a kind of substation transformer health index defining method, comprises and sets up health index model; The foundation of described model as shown in Figure 2, comprising:

(1) division of state grade.The health index HI of substation transformer is defined between [1,10].Being fine state in 1≤HI≤2, is normal condition in 2<HI≤5, in 5≤HI≤7 for poor state need strengthen maintenance, is severe conditions in 7<HI≤10, existing obvious defect pipelines.

(2) key feature amount choose the acquisition with data.As shown in Figure 1, the selection of transformer key feature amount with oil chromatogram analysis, oiling test and electrical test for foundation.Oil chromatogram analysis data mainly with acetylene content, hydrogen content, carbon dioxide relative to gas production rate, the relative gas production rate of carbon monoxide and total hydrocarbon content for foundation; Oil test main with liquid water content, oil breakdown voltage, furfural content, interfacial tension, the paper degree of polymerization, oil loss and flash-point data for foundation; Electric test mainly with iron core grounding current, specific insulation, polarization index, absorptance and winding dielectric loss for foundation.The acquisition of data is imported and electric substation's field survey gained by PMS2.0.

(3) standardization of controller switching equipment key feature amount data.The key feature amount characterizing controller switching equipment health status is numerous, and each index magnitude differences is very large, and dimension is different, for avoiding the evaluation grade caused thus not mate, carries out standardization with following formula.

$C_{ij}=\frac{y_{ij}-y_{ij0}}{y_{ijl}-y_{ij0}},y_{ij0}\&le;y_{ij}\&le;y_{ijl}$

Wherein y _ijactual institute measured value, y _ij0initial value, y _ijllimit value, C _ijincome value after being standardization.

(4) being subordinate to of index function

This patent adopts trapezoidal membership function to be subordinate to grade evaluation to each index foundation.Attachedly Figure 3 shows that trapezoidal membership function figure.

${\mathrm{\&mu;}}_1\left(f_{ij}\right)=\left\{\begin{array}{cc}1& f_{ij}\&le;a\\ \frac{b-f_{ij}}{b-a}& a<f_{ij}<b\\ 0& other\end{array}\right.$ ${\mathrm{\&mu;}}_2\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;aor{f}_{ij}\&GreaterEqual;d\\ \frac{f_{ij}-a}{b-a}& a<f_{ij}<b\\ 1& b\&le;f_{ij}\&le;c\\ \frac{d-f_{ij}}{d-c}& c<f_{ij}<d\end{array}\right.$

${\mathrm{\&mu;}}_3\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;cor{f}_{ij}\&GreaterEqual;f\\ \frac{f_{ij}-c}{d-c}& c<f_{ij}<d\\ 1& d\&le;f_{ij}\&le;e\\ \frac{f-f_{ij}}{f-e}& e<f_{ij}<f\end{array}\right.$ ${\mathrm{\&mu;}}_4\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;eor{f}_{ij}\&GreaterEqual;h\\ \frac{f_{ij}-e}{f-e}& e<f_{ij}<f\\ 1& f\&le;f_{ij}\&le;g\\ \frac{h-f_{ij}}{h-g}& g<f_{ij}<h\end{array}\right.$

${\mathrm{\&mu;}}_5\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;g\\ \frac{f_{ij}-g}{h-g}& g<f_{ij}<h\\ 1& f_{ij}\&GreaterEqual;h\end{array}\right.$

(5) acquisition of controller switching equipment key feature amount weight.The acquisition of weight adopts Fuzzy AHP, and concrete steps are as follows: the preference information of 1) assembling S expert, provides fuzzy judgment matrix; 2) normalized Synthetic Judgement Matrix is calculated; 3) try to achieve the corresponding possibility degree of each index, set up Possibility Degree Matrix; 4) according to Possibility Degree Matrix, ask for the relative weighting of each index, get its minimum value.Its step is as follows:

The desired value of Triangular Fuzzy Number is as shown in table 1:

The desired value of table 1 Triangular Fuzzy Number

Fuzzy judgment matrix is defined as follows:

Every a line of judgment matrix is added:

${\mathrm{RS}}_i={\mathrm{\&Sigma;}}_{j=1}^n{c}_{ij}=({\mathrm{\&Sigma;}}_{j=1}^n{l}_{ij},{\mathrm{\&Sigma;}}_{j=1}^n{m}_{ij},{\mathrm{\&Sigma;}}_{j=1}^n{u}_{ij})$

I-th ^ththe degree composite value of individual element is (asking arithmetic mean to row vector):

$S_i=\frac{{\mathrm{RS}}_i}{{\mathrm{\&Sigma;}}_{j=1}^n{\mathrm{RS}}_j}=(\frac{{\mathrm{\&Sigma;}}_{j=1}^n{l}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{u}_{kj}},\frac{{\mathrm{\&Sigma;}}_{j=1}^n{m}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{m}_{kj}},\frac{{\mathrm{\&Sigma;}}_{j=1}^n{u}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{l}_{kj}}),i=1,\mathrm{...},n$

Possibility degree between two Triangular Fuzzy Number is defined as:

Definition:

$d\left(A_i\right)=\underset{\&Element;\{1,...,n\},j\&NotEqual;i}{\min }V(S_i\&GreaterEqual;S_j),i=1,...,n.$

Then weight vectors matrix is:

W＝(d(A ₁)，d(A ₂)，…，d(A _n)) ^T

(6) revise weighted average to merge

Suppose that two Basic Probability As-signment (BPA) are m ₁and m ₂, the distance between them is

$d(m_1,m_2)=\sqrt{\frac{1}{2}{(m_1-m_2)}^T\underset{\&OverBar;}{D}(m_1-m_2)}$

Therefore the similar matrix that we can obtain them is:

$Sim=\left[\begin{array}{cccccc}1& {\mathrm{Sim}}_{12}& ...& {\mathrm{Sim}}_{1j}& ...& {\mathrm{Sim}}_{1n}\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ {\mathrm{Sim}}_{i1}& {\mathrm{Sim}}_{i2}& ...& {\mathrm{Sim}}_{ij}& ...& {\mathrm{Sim}}_{in}\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ .& .& .& .& .& .\\ {\mathrm{Sim}}_{n1}& {\mathrm{Sim}}_{n2}& ...& {\mathrm{Sim}}_{nj}& ...& 1\end{array}\right]$

The correction weight that we obtain each evidence by similar matrix is:

$w_{ei}=\frac{Sup\left(m_i\right)}{\underset{i=1}{\overset{n}{\&Sigma;}}Sup\left(m_i\right)}$

Now, the evidence of correction is:

$MAE\left(m\right)=\stackrel{n}{\mathrm{\&Sigma;}}(w_i\&times;m_i)$

Suppose have n and evidence, so final result passes through average evidence fusion n-1 time.

(7) illustrate below by way of Longjiang, Nanjing transformer station main transformer 110KV transformer actual collection data.

Table 2 transformer measured value

Table 2 is measured data values of this transformer, is drawn the opinion rating of each index by standard of index process and trapezoidal membership function, as shown in table 3.

Table 3 index assessment grade

Then by Fuzzy AHP, the weighted value of factor and corresponding index can be drawn, as shown in table 4:

Table 4 factor and corresponding exponential quantity

By table 3 and table 4, then merge by the weighted average revised, show that final evaluation result is as following table:

Can be found out by table 5, final health index value is: HI=0.82 × 10=8.2, and this transformer of surface is in severe conditions, matches with actual result.

Table 5 evaluation result

By evidence fusion rule, the maximal value merging gained is final health index value, shows that this equipment is in the middle of any state, is beneficial to repair based on condition of component.

Substation transformer is one of visual plant in transformation and distribution system, in order to overcome during transformer health index calculates the ambiguity existed, uncertain, and conflicting information, the technical program proposes a kind of employing fuzzy set theory, and fog-level analytical hierarchy process and correction weighted average fusion method carry out calculating transformer health index.Transformer health index computation model comprises three factors: oil chromatogram analysis, oil test and electric test.Have different index content under each factor, such as, oil chromatogram analysis is because have acetylene content, hydrogen content, carbon dioxide relative to index content such as gas production rate, the relative gas production rate of carbon monoxide and total hydrocarbon contents.The evaluation grade of each index is obtained and is calculated by trapezoidal membership function and obtain; The acquisition of the weight of all factors and its corresponding index adopts fog-level analytical hierarchy process, and the method can avoid the limitation of tradition stratum fractional analysis, fully demonstrates the ambiguity of expert's thinking; Based on the original evidence that above-mentioned steps draws, final result can by the weighted average calculating fusion revised, and the method can effectively solve the problem of conflicting between evidence.

Finally should be noted that: above embodiment is only in order to illustrate that technical scheme of the present invention is not intended to limit; although those of ordinary skill in the field are to be understood that with reference to above-described embodiment: still can modify to the specific embodiment of the present invention or equivalent replacement; these do not depart from any amendment of spirit and scope of the invention or equivalent replacement, are all applying within the claims of the present invention awaited the reply.

Claims (8)

1. a substation transformer health index defining method, is characterized in that: comprise and set up health index model; The process of establishing of described health index model comprises:

Determine substation transformer health index grade interval;

Choose described substation transformer key feature amount;

The standardization of the data of described substation transformer key feature amount;

The opinion rating of described substation transformer is set up by trapezoidal membership function;

Obtain the weight of the key feature amount of described substation transformer;

According to described opinion rating and described weight determination assessment models, thus by determining that revising weight determines described substation transformer health index HI.

2. a kind of substation transformer health index defining method as claimed in claim 1, is characterized in that: described substation transformer health index HI grade interval is defined between [1,10]; When 1≤HI≤2 are fine state; When 2<HI≤5 are normal condition; When 5≤HI≤7 for poor state need strengthen maintenance; When 7<HI≤10 are severe conditions and existing obvious defect pipelines.

3. a kind of substation transformer health index defining method as claimed in claim 1 or 2, is characterized in that: choosing of described transformer key feature amount is determined according to the data of oil chromatogram analysis, oiling test and electrical test; The data of described oil chromatogram analysis comprise acetylene content, hydrogen content, carbon dioxide relative to gas production rate, the relative gas production rate of carbon monoxide and total hydrocarbon content; The data of described oiling test comprise liquid water content, oil breakdown voltage, furfural content, interfacial tension, the paper degree of polymerization, oil loss and flash-point; The data of described electric test comprise iron core grounding current, specific insulation, polarization index, absorptance and winding dielectric loss.

4. a kind of substation transformer health index defining method as claimed in claim 1 or 2, is characterized in that: the standardization carrying out the data of described substation transformer key feature amount with following formula:

$C_{ij}=\frac{y_{ij}-y_{ij0}}{y_{ijl}-y_{ij0}},y_{ij0}\&le;y_{ij}\&le;y_{ijl}$

In formula: y _ijfor actual institute measured value; y _ij0for initial value; y _ijlfor limit value; C _ijfor income value after standardization.

5. a kind of substation transformer health index defining method as claimed in claim 2, is characterized in that: described trapezoidal membership function is determined by following formula:

$\begin{array}{cc}{\mathrm{\&mu;}}_1\left(f_j\right)=\left\{\begin{array}{cc}1& f_{ij}\&le;a\\ \frac{b-f_{ij}}{b-a}& a<f_{ij}<b\\ 0& other\end{array}\right.& {\mathrm{\&mu;}}_2\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;aor{f}_{ij}\&GreaterEqual;d\\ \frac{f_{ij}-a}{b-a}& a<f_{ij}<b\\ 1& b\&le;f_{ij}\&le;c\\ \frac{d-f_{ij}}{d-c}& c<f_{ij}<d\end{array}\right.\end{array}$

$\begin{array}{cc}{\mathrm{\&mu;}}_4\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;eor{f}_{ij}\&GreaterEqual;h\\ \frac{f_{ij}-e}{f-e}& a<f_{ij}<f\\ 1& f\&le;f_{ij}\&le;g\\ \frac{d-f_{ij}}{h-g}& g<f_{ij}<h\end{array}\right.& {\mathrm{\&mu;}}_3\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;cor{f}_{ij}\&GreaterEqual;f\\ \frac{f_{ij}-c}{d-c}& c<f_{ij}<d\\ 1& b\&le;f_{ij}\&le;e\\ \frac{f-f_{ij}}{f-e}& c<f_{ij}<f\end{array}\right.\end{array}$

${\mathrm{\&mu;}}_5\left(f_{ij}\right)=\left\{\begin{array}{cc}0& f_{ij}\&le;g\\ \frac{f_{ij}-g}{h-g}& g<f_{ij}<h\\ 1& f_{ij}\&GreaterEqual;h\end{array}\right.$

Wherein, f _ijfor the index degree of membership of data, μ ₁, μ ₂, μ ₃, μ ₄, μ ₅what be respectively five kinds of states is subordinate to grade, and a, b, c, d, e, f, g, h are respectively the spacing value of each grade.

6. a kind of substation transformer health index defining method as claimed in claim 5, is characterized in that: the process of the weight of the key feature amount of the described substation transformer of described acquisition comprises:

Ambiguity in definition judgment matrix:

Every a line of described judgment matrix is added:

${\mathrm{RS}}_i={\mathrm{\&Sigma;}}_{j=1}^n{c}_{ij}=({\mathrm{\&Sigma;}}_{j=0}^n{l}_{ij},{\mathrm{\&Sigma;}}_{j=1}^n{m}_{ij},{\mathrm{\&Sigma;}}_{j=1}^n{u}_{ij})$

Wherein, c _ijfor the vector of described matrix;

In matrix-vector, the degree composite value of i-th element is:

$S_i=\frac{{\mathrm{RS}}_i}{{\mathrm{\&Sigma;}}_{j=1}^n{\mathrm{RS}}_j}=(\frac{{\mathrm{\&Sigma;}}_{j=1}^n{l}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{u}_{kj}},\frac{{\mathrm{\&Sigma;}}_{j=1}^n{m}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{m}_{kj}},\frac{{\mathrm{\&Sigma;}}_{j=1}^n{u}_{ij}}{{\mathrm{\&Sigma;}}_{k=1}^n{\mathrm{\&Sigma;}}_{j=1}^n{l}_{kj}}),i=1,\mathrm{...},n$

Two Triangular Fuzzy Number M ₁and M ₂between possibility degree be defined as:

Wherein, M ₁=(l ₁, m ₁, u ₁), M ₂=(l ₂, m ₂, u ₂); for triangle ambiguity function;

And define:

$d\left(A_i\right)=\underset{\&Element;\{1,...,n\},j\&NotEqual;i}{\min }V(S_i\&GreaterEqual;S_j),i=1,...,n.$

Then weight vectors matrix is:

W＝(d(A ₁)，d(A ₂)，...，d(A _n)) ^T。

7. a kind of substation transformer health index defining method as claimed in claim 6, is characterized in that: the described assessment models index of index is the product of described opinion rating and described weight; And as original evidence; Described assessment models is:

$M_i\left(H\right)=\underset{j=1}{\overset{n}{\&Sigma;}}{w}_{ij}{U}_i\left(H\right)$

Wherein:

W _ijfor weight, Ui (H) represents the opinion rating of j index under i factor; M _i(H) evaluation result of each index is represented.

8. a kind of substation transformer health index defining method as claimed in claim 7, is characterized in that: described substation transformer health index deterministic process comprises:

Suppose that two Basic Probability As-signment are m ₁and m ₂, the distance between them is:

$d(m_1,m_2)=\sqrt{\frac{1}{2}{(m_1-m_2)}^T\underset{\&OverBar;}{D}(m_1-m_2)}$

Wherein, it is one 2 ⁿ× 2 ⁿmatrix;

The similar matrix obtaining them is:

$Sim=\left[\begin{array}{cccccc}1& {\mathrm{Sim}}_{12}& ...& {\mathrm{Sim}}_{1j}& ...& {\mathrm{Sim}}_{1n}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ {\mathrm{Sim}}_{i1}& {\mathrm{Sim}}_{i2}& ...& {\mathrm{Sim}}_{ij}& ...& {\mathrm{Sim}}_{in}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ {\mathrm{Sim}}_{n1}& {\mathrm{Sim}}_{n2}& ...& {\mathrm{Sim}}_{nj}& ...& 1\end{array}\right]$

The correction weight that we obtain each evidence by similar matrix is:

$w_{ei}=\frac{Sup\left(m_i\right)}{\underset{i=1}{\overset{n}{\&Sigma;}}Sup\left(m_i\right)}$

Now, the evidence of correction is:

$MAE\left(m\right)=\underset{i=1}{\overset{n}{\&Sigma;}}(w_i\&times;m_i)$

Wherein, w _ifor revising weight;

Suppose there be n evidence, so final result passes through average evidence fusion n-1 time; Wherein in income value, maximal value is required substation transformer health index.