CN103235991A - Condition evaluation method of distribution network transformer based on fuzzy theory - Google Patents

Abstract

The invention discloses a condition evaluation method of a distribution network transformer based on a fuzzy theory. The method comprises the following steps that a three-level index system of condition evaluation of a distribution transformer is established in accordance with data of daily check, routine tests and diagnostic tests of the distribution network transformer, an evaluation index set and a comment set are determined, wherein the evaluation index set contains all indexes in the three-level index system, and the evaluation set contains four conditions of normality, caution, abnormity and severity; an improved analytic hierarchy process is adopted to determine the weight of each index in the evaluation index set; a relatively deteriorative concept is adopted to determine an evaluation matrix of the distribution network transformer; and a grey relation theory is adopted to determine the condition level of the transformer. The condition evaluation method has the following beneficial effects that the condition of the distribution network transformer can be evaluated in batches, losses of manpower and material resources are reduced, the working efficiency and the condition evaluation accuracy are greatly increased, and the condition evaluation method has the advantage of strong practicability.

Description

Distribution Transformer State Assessment method based on fuzzy theory

Technical field

The present invention relates to distribution Transformer State Assessment technical field.

Background technology

Along with the continuous development of modern economy, the scale of power distribution network is also in continuous expansion, and it is more important that the safe operation of distribution network systems just seems, the reliability of power supply is directly connected to the economic interests of user and enterprise itself.Distribution system is positioned at the end of electric system, is the tie that connects between electric system and the user, and people are more and more higher for the requirement of power distribution network power supply reliability, and this just presses for carries out state estimation in power distribution network.Transformer carries out reliable state estimation to it and has important practical significance as the visual plant of distribution.

The state estimation of substation transformer relates to the state estimation of and each parts whole to it, carry out analyzing and processing to equipment a large amount of tests and Monitoring Data over the years, thereby draw correct result.The analytical calculation of mass data is manually handled iff relying on, not only labour intensity big, waste time and energy, and easily make mistakes, need lot of manpower and material resources, this method of carrying out Transformer State Assessment by tactics of human sea be with the current power industry in the policy of cutting payroll to improve efficiency contrary, be impracticable.Therefore press for multiple functional, convenient, the practical substation transformer state estimation decision system of exploitation one cover, thereby can realize the robotization of data analysis, processing, and then realization finally is reached for the target that the operation power personnel provide the overhaul of the equipments strategy with directive significance to the accurate assurance of equipment time of day.

Summary of the invention

The technical problem to be solved in the present invention provides a kind of can increasing work efficiency greatly and the distribution Transformer State Assessment method based on fuzzy theory of state estimation accuracy rate.

For solving the problems of the technologies described above, the technical solution used in the present invention is:

A kind of distribution Transformer State Assessment method based on fuzzy theory, its method step is as follows:

(1) according to distribution transformer daily patrol and examine, routine test and diagnostic test data, set up three grades of index systems of substation transformer state estimation, determine judging quota collection and comment collection, wherein the judging quota collection comprises all indexs in described three grades of index systems, passes judgment on collection and is normal, attention, unusual, serious four kinds of states;

(2) adopt improved analytical hierarchy process to determine the weight of each index in the described judging quota collection;

(3) concept of the relative impairment grade of employing is determined the judge matrix of distribution transformer;

(4) adopt the theoretical state grade of determining transformer of grey correlation.

The concrete grammar of described step (2) is as follows:

1) each state index at the same level in three grades of index systems of distribution transformer is compared in twos, obtain fuzzy judgment matrix, and fuzzy judgment matrix is represented with the trigonometric function form;

2) comprehensive evaluation value of calculating fuzzy evaluation;

3) each index weight in the judging quota collection is carried out consistency check, when assay is unreasonable, fuzzy judgment matrix is adjusted, calculate again then and check, reasonable until assay.

The specific implementation method of described step (3) is as follows:

1) qualitative quantity of state is carried out state grade and divide, determine degree of membership;

2) to quantitative quantity of state, adopt triangle and trapezoidal combined subordinate function to determine degree of membership.

The degree of association in step (4) the grey correlation theory is a kind of method of correlation degree in the analytic system, and the basic thought of the degree of association is to judge the degree of association according to the similarity degree in the curve, is a kind of quantitative test.The present invention finally determines the state grade of transformer by the correlation coefficient of each index evaluation result of calculating distribution transformer with normal, attention, unusual, serious four kinds of state reference sequences.

Beneficial effect of the present invention is as follows:

The correlation theories knowledge of fuzzy mathematics is applied in the state estimation of distribution transformer, can assess by state whole to transformer and each parts, final appraisal results provide with the form that may spend, have overcome the one-sidedness of traditional assessment result to determine that state provides.

In addition, the present invention can carry out the batch assessment to the state of distribution transformer, reduces the loss of manpower and materials, increases work efficiency greatly and the state estimation accuracy rate, has very strong practicality.

Description of drawings

The present invention is further detailed explanation below in conjunction with the drawings and specific embodiments.

Fig. 1 is three grades of index system figure of distribution Transformer State Assessment of the present invention.

Embodiment

Distribution Transformer State Assessment method based on fuzzy theory of the present invention, its method step is as follows:

(1) according to distribution transformer daily patrol and examine, routine test and diagnostic test data, set up three grades of index systems of substation transformer state estimation, as shown in Figure 1, determine judging quota collection and comment collection, wherein the judging quota collection comprises all indexs in described three grades of index systems, passes judgment on collection and is normal, attention, unusual, serious four kinds of states.

(2) adopt improved analytical hierarchy process to determine the weight of each index in the described judging quota collection, the specific implementation method is as follows:

1) each state index at the same level in three grades of index systems of distribution transformer is compared in twos, obtain fuzzy judgment matrix

Be designated as

Wherein n represents the number of state index in the indicator layer at the same level,

Represent in the indicator layer at the same level n state index and the 1st significance level that state index is compared, and a is arranged _Ij=1/a _Ji

Above-mentioned fuzzy judgment matrix is expressed as follows with the trigonometric function form:

2) comprehensive evaluation value of calculating fuzzy evaluation, concrete grammar is as follows:

At first, calculate the fuzzy evaluation value ${\stackrel{\‾}{u}}_i=(\frac{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{l}_{\mathrm{ij}}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{l}_{\mathrm{ij}}},\frac{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{m}_{\mathrm{ij}}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{m}_{\mathrm{ij}}},\frac{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{u}_{\mathrm{ij}}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{u}_{\mathrm{ij}}})$ Expectation value, The expectation on the left side

Value is $I_L\left({\stackrel{\‾}{u}}_i\right)=(l_i+m_i)/2,$

The expectation value on the right is $I_R\left({\stackrel{\‾}{u}}_i\right)=(u_i+m_i)/2;$

Then, introduce sense organ coefficient η, 0＜η＜1 wherein, then Expectation value be

Wherein the η value can be adjusted according to actual conditions;

3) weight of each index at first definite judging quota collection, the formula of parameter weight is as follows:

After determining the weight of each index, it is carried out consistency check, with judge the weight of being asked whether reasonable, the consistency check condition formula is as follows:

$l_{\mathrm{ij}}\stackrel{\‾}{\≤}\frac{w_i}{w_j}\stackrel{\‾}{\≤}{m}_{\mathrm{ij}},$

Wherein, w _i＞0, w _j＞0, i ≠ j, symbol

Expression is fuzzy to be less than or equal to; When assay does not satisfy above-mentioned condition, show that institute's weighted value of asking is unreasonable, need adjust fuzzy judgment matrix, till satisfying consistency check.

(3) concept of the relative impairment grade of employing is determined the judge matrix of distribution transformer, and the specific implementation method is as follows:

1) qualitative quantity of state is carried out state grade and divide, determine degree of membership, concrete grammar is as follows:

Inviting some experts that qualitative quantity of state is carried out state grade in the mode of survey divides.For example, total m name expert participates in assessment, thinks that a certain quantity of state is in normally, notes, unusually, seriously expert's number of state is respectively a, b, c, d, then use the number of selecting every kind of state divided by total number of persons, through normalized, then can obtain the degree of membership of single index;

2) to quantitative quantity of state, adopt triangle and trapezoidal combined subordinate function to determine degree of membership, concrete grammar is as follows:

Subordinate function corresponding to four kinds of states of comment collection is as follows:

Subordinate function under the normal condition is:

$\mathrm{\μ}\left(x\right)=\left\{\begin{array}{cc}1& x\≥0.8\\ 5x-3& 0.6\≤x\≤0.8\\ 0& x\≤0.6\end{array}\right.,$

Subordinate function under the attention state is:

Subordinate function under the abnormality is:

Subordinate function under the serious state is:

$\mathrm{\μ}\left(x\right)=\left\{\begin{array}{cc}-5x+2& 0.2\≤x\≤0.4\\ 1& x\≤0.2\\ 0& x\≥0.4\end{array}\right..$

(4) adopt the theoretical state grade of determining transformer of grey correlation.

The degree of association in the grey correlation theory is a kind of method of correlation degree in the analytic system, and the basic thought of the degree of association is to judge the degree of association according to the similarity degree in the curve, is a kind of quantitative test.The present invention finally determines the state grade of transformer by the correlation coefficient of each index evaluation result of calculating distribution transformer with normal, attention, unusual, serious four kinds of state reference sequences.

In sum, advantage of the present invention is as follows:

(1) calculates the fuzzy evaluation value

Expectation value the time, consider decision maker's psychological factor, the present invention introduces sense organ coefficient η, significantly reduces the influence of subjective factor.

(2) concept of having introduced relative impairment grade is determined the judge matrix of distribution transformer, not only the span of quantity of state can be defined between [0～1], can also come the size of clear judgement transformer normal degree according to the size of degradation.

(3) transformer end-state grade determines, traditional grade determines that method is the maximum membership degree method, what the present invention adopted is the grey correlation theory, has overcome maximum membership grade principle and has judged the defective that the end-state of substation transformer can be lost many important informations;

What (4) the present invention finally obtained is that substation transformer is in may spending of a certain state, meets objective reality more.

(5) can develop the practical operating system of a cover according to the method for the invention, this system can be according to the accurate assessment that realizes and each parts time of day whole to substation transformer automatically with dynamic two kinds of approach.

Embodiment

Be example with Xingtai City 10kV substation transformer, the data of patrolling and examining, the running test information collected are handled and its running status is assessed.Data information after the section processes is as follows: the above substation transformer of 1.6MVA is tested through winding D.C. resistance, alternate direct current resistance maximal value is 0.64% of three-phase mean value, insulation resistance is compared with initial value and is descended approximately 8%, and the alternate temperature difference of joint is less than 9K, and connector temperature is 78C.Maximum running load rate reaches more than 80% for a long time, and the transformer body temperature surpasses 5% of producer's permissible value approximately, and the three-phase imbalance rate is 13% under the Yyn0 mode of connection, and neutral line current is 120% of ratings, and the insulating oil withstand voltage test is 31.2kV.Patrolling and examining data is: outdoor transformer winding and sleeve pipe filth are not obvious, slightly damaged, disappearance.Shunting switch, temperature control, blower fan apparatus move normal, the slight oil impregnate of fuel tank, slight corrosion.The no abnormal demonstration of oil level, down conductor does not have obvious ground connection, perhaps connects problems such as loosening, imperfect earth.

Evaluation process is as follows:

According to above-mentioned data, for quantitative target, become the value between 0～1 after treatment, be brought in the subordinate function, can obtain the degree of membership matrix.For qualitative index, then can determine its degree of membership by the method for fuzzy statistics.After trying to achieve the degree of membership of each state index for four comments, can constitute and be subordinate to matrix, as follows: the degree of membership of winding and sleeve pipe is R _P1,

$R_{\mathrm{pl}}=\left[\begin{array}{cccc}0.768& 0.232& 0& 0\\ 0.668& 0.212& 0.120& 0\\ 0.871& 0.129& 0& 0\\ 1& 0& 0& 0\\ 0.2& 0.6& 0.2& 0\\ 0.1& 0.7& 0.2& 0\\ 0.137& 0.621& 0.242& 0\\ 0.202& 0.449& 0.349& 0\end{array}\right],$

Try to achieve weight allocation, can obtain

The degree of membership of winding and sleeve pipe then:

$B_{p1}=\left[\begin{array}{ccccccccc}0.192& 0.168& 0.109& 0.088& 0.073& 0.106& 0.105& |& 0.159\end{array}\right]$

$\×\left[\begin{array}{cccc}0.768& 0.232& 0& 0\\ 0.668& 0.212& 0.120& 0\\ 0.871& 0.129& 0& 0\\ 1& 0& 0& 0\\ 0.2& 0.6& 0.2& 0\\ 0.1& 0.7& 0.2& 0\\ 0.137& 0.621& 0.242& 0\\ 0.202& 0.449& 0.349& 0\end{array}\right]=\left[\begin{array}{cccc}0.5143& 0.3488& 0.1369& 0\end{array}\right],$

In like manner can get:

The degree of membership B of shunting switch _P2=[0.9 0.1 0 0],

The degree of membership B of cooling system _P3=[0.8355 0.1645 0 0],

The degree of membership B of fuel tank _P4=[0.4936 0.4242 0.0822 0],

The degree of membership B of non electrical quantity protective device _P5=[0.563 0.211 0.226 0],

The degree of membership B of earthing device _P5=[0.8983 0.1017 0 0],

The degree of membership B of insulating oil _P7=[0.8222 0.1524 0.0254 0],

The degree of membership B of sign _P8=[0 0.7 0.3 0],

Then the multifactorial evaluation result of whole substation transformer is:

Net result is as follows:

If two reference sequence B ^I=[1 00 0] and B ^II=[0 10 0] are respectively the clear judge of normal condition and attention state, calculate B respectively _pWith B ^I, B ^IICorrelation coefficient: r ^I=0.854, r ^II=0.318.This shows that the comprehensive state of this transformer and normal condition are more approaching, we think that it is normal condition.

Can see from actual result: the part test data have departed from factory-said value, and certain degradation trend is arranged, but do not reach demand value as yet.Insulation resistance obviously descends, and near demand value, should give attention, needs to strengthen trace analysis.

Claims (3)

1. distribution Transformer State Assessment method based on fuzzy theory is characterized in that its method step is as follows:

(1) according to distribution transformer daily patrol and examine, routine test and diagnostic test data, set up three grades of index systems of substation transformer state estimation, determine judging quota collection and comment collection, wherein the judging quota collection comprises all indexs in described three grades of index systems, passes judgment on collection and is normal, attention, unusual, serious four kinds of states;

(2) adopt improved analytical hierarchy process to determine the weight of each index in the described judging quota collection;

(3) concept of the relative impairment grade of employing is determined the judge matrix of distribution transformer;

(4) adopt the theoretical state grade of determining transformer of grey correlation.

2. the distribution Transformer State Assessment method based on fuzzy theory according to claim 1 is characterized in that the concrete grammar of described step (2) is as follows:

1) each state index at the same level in three grades of index systems of distribution transformer is compared in twos, obtain fuzzy judgment matrix, and fuzzy judgment matrix is represented with the trigonometric function form;

2) comprehensive evaluation value of calculating fuzzy evaluation;

3) each index weight in the judging quota collection is carried out consistency check, when assay is unreasonable, fuzzy judgment matrix is adjusted, calculate again then and check, reasonable until assay.

3. the distribution Transformer State Assessment method based on fuzzy theory according to claim 1 is characterized in that the specific implementation method of described step (3) is as follows:

1) qualitative quantity of state is carried out state grade and divide, determine degree of membership;

2) to quantitative quantity of state, adopt triangle and trapezoidal combined subordinate function to determine degree of membership.

Images