CN104933505A - Decision and evaluation method for intelligent power distribution network group based on fuzzy assessment - Google Patents

Abstract

The invention discloses a decision and evaluation method for an intelligent power distribution network group based on fuzzy assessment. An evaluation index system of a recursion order hierarchy system of the intelligent power distribution network is established, experts evaluates indexes according to index positive and dimensionless results, subjective evaluation is combined with objective evaluation, combined evaluation is carried out by combining the expert weight and the index weight, and evaluation values of comprehensive effectiveness is given to different schemes. The decision and evaluation method is suitable for the condition that construction schemes of intelligent distribution network are compared, and the evaluation result can provide decision basis for programming, design, construction and operation management staff.

Description

A kind of intelligent distribution network group decision evaluation method based on fuzzy evaluation

Technical field

The invention belongs to power distribution network assessment technique field, particularly relate to a kind of intelligent distribution network group decision evaluation method based on fuzzy evaluation.

Background technology

The development of intelligent distribution network is a complicated arduous systems engineering, has that scale is large, uncertain factor is many, relate to the wide feature in field.Along with the propelling of intelligent power distribution network construction, be necessary evaluation index and the evaluation method of exploring applicable intelligent distribution network, from security, technical, the aspect such as reliability, economy, intelligent power distribution network construction level is evaluated, make every effort to the overall picture of complete, accurate, objective reaction intelligent power distribution network construction, to providing guidance and reference for the theoretical research in China's intelligent distribution network field and Construction Practice.

2009, the U.S. issued intelligent power distribution network construction Development Assessment index system, proposed 6 characteristics of intelligent distribution network: (1) participates in based on the user of full information; (2) all generatings and energy storage can be received; (3) introducing of new product, new service etc. is allowed; (4) the different quality of power supply is provided according to user's request; (5) assets utilization efficiency and operation of power networks efficiency is optimized; (6) operation of power networks has more flexibility, can tackle all kinds of disturbance attack and disaster.Domestic experts and scholars have also carried out large quantity research in intelligent grid evaluation.New forms of energy and intelligent grid develop in harmony assessment indicator system from generating electricity, electrical network, electricity consumption, scheduling 4 links, construct the assessment indicator system comprising 3 levels, 15 indexs.Intelligent grid demonstration project System of Comprehensive Evaluation has refined one-level to level Four index from technical, economy, social, engineering management, practical and novelty 6 dimensions.Intelligent grid pilot project assessment indicator system emphasis, for pilot projects such as the intelligent substation carried out, power distribution automation and power information collections, builds corresponding special assessment indicator system.

More be absorbed in the foundation of index system and the determination of index weights about the research of intelligent distribution network evaluation at present, have ignored the setting of metrics evaluation criterion, cause evaluation criteria to set and too simplify.Evaluation criteria needs the importance and the feature that embody index itself, need set in conjunction with expert opinion.Evaluation criteria characterizes the corresponding relation between discrete desired value and mark, for forward index, increases along with desired value increases its mark, for negative sense index, reduce along with desired value increases its mark, for appropriate index, value mark in desired value interval is higher.Compared the quality of different schemes by mark, the standard such as " excellent, good, in, poor " is carried out weighing more directly perceived than simple adopting.In practice, due to the limitation that the ambiguity of many problems itself, complicacy and expert are familiar with problem, expert may take fuzzy language to the evaluation of index, need carry out quantitative description, then assemble expert opinion and multiple index the fuzzy evaluation language of expert.

Intelligent distribution network evaluation belongs to Multiple Attribute Decision Problems, for improving evaluation level and precision, making up the defect of single evaluation, can adopt combination evaluation methods.In theory, combination evaluation methods is more reasonable than single evaluation method, science.There are three class combinatorial problems at present: the first kind is Index Weights combined method; Equations of The Second Kind is that the expert of multidigit expert is clustered to; 3rd class is the combination of multiple single method.

Summary of the invention

The object of the present invention is to provide a kind of intelligent distribution network group decision evaluation method based on fuzzy evaluation, be intended to carry out comprehensive analysis and evaluation to multiple intelligent power distribution network construction scheme, provide the integrated ordered of scheme, thus select the scheme of comprehensive optimum.

The present invention realizes like this, a kind of intelligent distribution network group decision evaluation method based on fuzzy evaluation, expert is according to the result of index forward and nondimensionalization, with the form of Triangular Fuzzy Number, index is evaluated, integrated use index calculate and expert's subjective evaluation result, overall target weight and Weight of Expert carry out combination evaluation, provide the aggreggate utility evaluation of estimate of each scheme, are applicable to the situation compared multiple intelligent power distribution network construction scheme.

Further, should comprise the following steps based on the intelligent distribution network group decision evaluation method of fuzzy evaluation:

Step one, sets up the intelligent distribution network assessment indicator system of recursive hierarchy structure, namely comprises the Hierarchy indicataors system of destination layer, rule layer and indicator layer;

Step 2, index calculate, and forward and nondimensionalization process are carried out to each evaluation index;

Step 3, agriculture products weight and Weight of Expert, and be normalized;

Step 4, expert provides the evaluation to the different index of each scheme, and is represented by Triangular Fuzzy Number by the fuzzy evaluation of expert;

Step 5, comprehensive Weight of Expert and index weights, obtain weighted synthetical evaluation battle array, obtains combination evaluation result.

Further, in step one, intelligent distribution network assessment indicator system comprises destination layer, rule layer and indicator layer;

Destination layer comprises power supply capacity, the quality of power supply, reliability, economy aspect;

Rule layer comprises load capacity, turns for ability, voltage, frequency, user dependability, line loss per unit, power factor, plant factor;

Indicator layer comprises load capacity index, turns for capacity index, voltage indexes, Frequency Index, user dependability index, line loss per unit index, power factor specification, plant factor index.

Further, load capacity index comprises capacity load ratio of network, 10kV line load rate, 10kV distribution transforming load factor;

Turn and comprise 10kV line looped network rate, the 10kV line load rate of transform for capacity index;

Voltage indexes comprises integrated voltage qualification rate, voltage deviation, non-equilibrium among three phase voltages;

Frequency Index refers to frequency departure;

User dependability index comprises power supply reliability, the average power off time of user, the average frequency of power cut of user;

Line loss per unit index comprises comprehensive line loss per unit, hi-line loss rate, low voltage line loss rate;

Power factor specification comprises main transformer power factor, 10kV line power factor, distribution transforming power factor;

Plant factor index comprises the unbalanced degree of line load rate, the unbalanced degree of distribution transforming load factor.

Further, the method that evaluation index carries out forward in step 2 specifically comprises:

For the forward of reverse index, take directly to get negative method, that is: y _i=-x _i;

In formula: x _i, y _irepresent the desired value before and after index forward respectively.

For the forward of appropriate index, calculate the difference with set-point, then get negative to the absolute value of difference, that is: y _i=-| x _i-x _given|, x _givenfor set-point.

In formula: x _givenrepresent index limits.

Further, in step 2, evaluation index adopts standardized method to carry out nondimensionalization, that is:

$y_i=\frac{x_i-\stackrel{\‾}{x}}{s};$

In formula: represent the mean value of the same index series of each scheme, s represents the mean square deviation of the same index series of each scheme, n represents the number of scheme to be evaluated.

After standardization, variable y _iinterval is uncertain; The mean value of index is mean square deviation is s _y=1;

Mean value: $\stackrel{\‾}{y}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{y}_i=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{x_i-\stackrel{\‾}{x}}{s}=\frac{1}{n}\×\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i-n\stackrel{\‾}{x}}{s}=0;$

Mean square deviation: $s_y=\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(y_i-\stackrel{\‾}{y})}^2}=\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\left(\frac{x_i-\stackrel{\‾}{x}}{s}\right)}^2}=\frac{1}{s}\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_i-\stackrel{\‾}{x})}^2}=1.$

In formula: the mean value of index series after expression forward, s _ythe mean square deviation of index series after expression forward;

Further, in step 3, Weight of Expert normalized specifically comprises:

Determine first class index weight two-level index weight the fuzzy semantics of expert to index weights is converted into Triangular Fuzzy Number:

represent that expert is to the fuzzy evaluation of the i-th class first class index weight, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent that expert is to the fuzzy evaluation of a jth two-level index weight, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

Weight of Expert label taking value W _e=(W _e1, W _e2..., W _eK), have

In formula: W _eirepresent the evaluation weight of expert i, K represents expert's number.

Further, in step 3, index weights normalized specifically comprises:

With for benchmark, first adopt linear method pair be normalized, then basis scaling pair with carry out equal proportion convergent-divergent, realize the normalization of fuzzy weighted values;

Fuzzy weighted values after normalization is ${\tilde{W}}^{i^{\′}}=(W_1^{i^{\′}},W_2^{i^{\′}},W_3^{i^{\′}}),$ Then have: $W_2^{i^{\′}}=\frac{W_2^i}{\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{W}_2^i};$

Equal proportion convergent-divergent, then have:

: $W_1^{i^{\′}}=W_1^i\frac{W_2^{i^{\′}}}{W_2^i},W_3^{i^{\′}}=W_3^i\frac{W_2^{i^{\′}}}{W_2^i};$

After normalization, have:

In formula: I represents the number of first class index.

In like manner to two-level index weight be normalized, formed

Further, in step 4, the fuzzy evaluation of expert to index specifically comprises:

K expert provides the evaluation of the two-level index to each scheme respectively form fuzzy evaluation battle array:

In formula: for the expert k represented by Triangular Fuzzy Number is to the evaluation of a jth two-level index of m kind scheme;

represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively.

Further, step 5 specifically comprises:

The first step, comprehensive Weight of Expert and fuzzy evaluation battle array:

Fuzzy evaluation battle array describe K expert to the evaluation of M kind scheme J two-level index, often kind of scheme has K × J fuzzy number, total M × K × J fuzzy number; Triangular Fuzzy Number is utilized to be similar to multiplication and generalized addition computing, comprehensive Weight of Expert W _ewith fuzzy evaluation battle array, form expert's comprehensive evaluation battle array expert's comprehensive evaluation battle array combines K expertise, total M × J fuzzy number:

${\tilde{P}}^{\mathrm{mj}}=(W_{E1}\⊗{\tilde{P}}_1^{\mathrm{mj}})\⊕(W_{E2}\⊗{\tilde{P}}_2^{\mathrm{mj}})\⊕...\⊕(W_{\mathrm{EK}}\⊗{\tilde{P}}_K^{\mathrm{mj}})=(P_1^{\mathrm{mj}},P_2^{\mathrm{mj}},P_3^{\mathrm{mj}});$

In formula: represent a comprehensive K expert opinion, expert's comprehensive evaluation of the jth two-level index to m kind scheme represented by Triangular Fuzzy Number, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

W _eirepresent the evaluation weight of expert i;

for the expert k represented by Triangular Fuzzy Number is to the evaluation of a jth two-level index of m kind scheme; M=1,2 ..., M; J=1,2 ..., J;

⊕, represent addition and the multiplying of Triangular Fuzzy Number respectively.

Second step, overall target weight and expert's comprehensive evaluation battle array:

Integrated two-stage index weights with expert's comprehensive evaluation battle array form secondary weighted synthetical evaluation battle array weighted synthetical evaluation battle array combines index weights and expert opinion, total M × J fuzzy number:

${\tilde{P}}^{{\mathrm{mj}}^{\′}}=({\tilde{w}}^{j^{\′}}\⊗{\tilde{P}}^{\mathrm{mj}})=({P_1^{{\mathrm{mj}}^{\′}}P}_2^{{\mathrm{mj}}^{\′}},P_3^{{\mathrm{mj}}^{\′}});$

In formula: represent and combine K expert opinion and a jth two-level index weight, the secondary weighted synthetical evaluation of a jth two-level index of the m kind scheme represented by Triangular Fuzzy Number, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent the fuzzy weighted values of the jth two-level index after normalization;

represent a comprehensive K expert opinion, expert's comprehensive evaluation of the jth two-level index to m kind scheme represented by Triangular Fuzzy Number;

m＝1,2,...,M；j＝1,2,...,J；

Comprehensive first class index weight with secondary weighted synthetical evaluation battle array form one-level weighted synthetical evaluation battle array

${\tilde{P}}^{{\mathrm{mi}}^{\′\′}}={\tilde{W}}^{i^{\′}}\⊗({\tilde{P}}^{{m1}^{\′}}\⊕...\⊕{\tilde{P}}^{{\mathrm{mj}}^{\′}}\⊕...\⊕{\tilde{P}}^{{\mathrm{mn}}^{\′}})=(P_1^{{\mathrm{mi}}^{\′\′}},P_2^{{\mathrm{mi}}^{\′\′}},P_3^{{\mathrm{mi}}^{\′\′}});$

In formula: represent and combine expert opinion and first class index weight, the weighted synthetical evaluation of i-th first class index to m kind scheme represented by Triangular Fuzzy Number, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent the fuzzy weighted values of the i-th class first class index after normalization,

represent and combine K expert opinion and a jth two-level index weight, the secondary weighted synthetical evaluation of a jth two-level index under the i-th class first class index of the m kind scheme represented by Triangular Fuzzy Number;

N represents the number of two-level index under the i-th class first class index, m=1,2 ..., M; J=1,2 ..., n;

3rd step, calculates aggreggate utility value:

The aggreggate utility value of each scheme is calculated according to one-level weighted synthetical evaluation battle array

${\tilde{P}}^m={\tilde{P}}^{{m1}^{\′\′}}\⊕{\tilde{P}}^{{m2}^{\′\′}}\⊕...\⊕{\tilde{P}}^{{\mathrm{mI}}^{\′\′}}=(P_1^m,P_2^m,P_3^m)$

In formula: represent the aggreggate utility fuzzy number of m kind scheme, represent that the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum, I represent the classification number of first order index respectively, m=1,2 ..., M;

4th step, calculates sequence desired value:

Calculate the sequence desired value F of each scheme respectively _m(m=1,2 ..., M):

$F_m=\mathrm{\βmean}{\tilde{P}}^m+(1-\mathrm{\β})(1-\mathrm{\σ}{\tilde{P}}^m);$

In formula: F _mrepresent the sequence desired value of m kind scheme; represent average and the standard deviation of the evaluation fuzzy number of m kind scheme respectively; β is weights;

$\mathrm{mean}{\tilde{P}}^m=\frac{P_1^m+P_2^m+P_3^m}{3};$

$\mathrm{\σ}{\tilde{P}}^m=\sqrt{\frac{{\left(P_1^m\right)}^2+{\left(P_2^m\right)}^2+{\left(P_3^m\right)}^2-P_1^m{P}_2^m-P_1^m{P}_3^m-P_2^m{P}_3^m}{18}};$

Decision maker attaches equal importance to average and deviation, gets β=0.5, then:

$F_m=0.5\mathrm{mean}{\tilde{P}}^m+0.5(1-\mathrm{\σ}{\tilde{P}}^m);$

According to F _msize, namely judge the quality of each scheme, sequence desired value larger, scheme is more excellent.

Intelligent distribution network group decision evaluation method based on fuzzy evaluation provided by the invention, establish the intelligent distribution network assessment indicator system of recursive hierarchy structure, expert evaluates index according to the result of index forward and nondimensionalization, in conjunction with subjective assessment and objective evaluation, comprehensive Weight of Expert and index weights carry out combination evaluation, provide the aggreggate utility evaluation of estimate of each scheme, be applicable to the situation that multiple intelligent power distribution network construction scheme is compared, the result of evaluation can be planning, design, construction, operational management personnel provide decision-making foundation.

Accompanying drawing explanation

Fig. 1 is the intelligent distribution network group decision evaluation method process flow diagram based on fuzzy evaluation that the embodiment of the present invention provides;

Fig. 2 is the intelligent distribution network assessment indicator system schematic diagram that the embodiment of the present invention provides.

Embodiment

In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with embodiment, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

Below in conjunction with drawings and the specific embodiments, application principle of the present invention is further described.

As shown in Figure 1, the intelligent distribution network group decision evaluation method based on fuzzy evaluation of the embodiment of the present invention comprises the following steps:

S101: the intelligent distribution network assessment indicator system setting up recursive hierarchy structure, namely comprises the Hierarchy indicataors system of destination layer, rule layer and indicator layer;

S102: index calculate, and forward and nondimensionalization process are carried out to each evaluation index;

S103: agriculture products weight and Weight of Expert, and be normalized;

S104: expert provides the evaluation to the different index of each scheme, and is represented by Triangular Fuzzy Number by the fuzzy evaluation of expert;

S105: comprehensive Weight of Expert and index weights, obtain weighted synthetical evaluation battle array, obtains combination evaluation result.

In step S101, intelligent distribution network assessment indicator system:

As Fig. 2, intelligent distribution network assessment indicator system comprises destination layer, rule layer and indicator layer, and wherein destination layer comprises the aspects such as power supply capacity (A), the quality of power supply (B), reliability (C), economy (D); Rule layer relates to load capacity, turns for aspects such as ability, voltage, frequency, user dependability, line loss per unit, power factor, plant factor; Load capacity index comprises capacity load ratio of network (A1), 10kV line load rate (A2), 10kV distribution transforming load factor (A3); Turn and comprise 10kV line looped network rate (A4), the 10kV line load rate of transform (A5) for capacity index; Voltage indexes comprises integrated voltage qualification rate (B1), voltage deviation (B2), non-equilibrium among three phase voltages (B3); Frequency Index mainly refers to frequency departure (B4); User dependability index comprises power supply reliability (C1), the average power off time of user (C2), the average frequency of power cut of user (C3); Line loss per unit index comprises comprehensive line loss per unit (D1), hi-line loss rate (D2), low voltage line loss rate (D3); Power factor specification comprises main transformer power factor (D4), 10kV line power factor (D5), distribution transforming power factor (D6); Plant factor index comprises the unbalanced degree of line load rate (D7), the unbalanced degree of distribution transforming load factor (D8);

Index forward and nondimensionalization in step s 102:

In multiattribute assessment, some is the index that the larger evaluation of desired value is better, is called forward index (also claim profit evaluation model or hope large-scale index); Some is the index that the less evaluation of desired value is better, and be called reverse index (also claim cost type index or hope small-sized index), also some refers to scale value more close to the index that certain value is better, is called appropriate index; In comprehensive evaluation, first need index Communalities, be generally that reverse index and appropriate index are converted into forward index, this process is called the forward of index;

Forward index comprises: { A1, A2, A3, A4, A5, B1, C1, D4, D5, D6}

Reverse index comprises: { B2, B3, B4, C2, C3, D1, D2, D3, D7, D8}

For forward index, without the need to special processing; For reverse index, except C2, C3, all the other indexs all represent with percents, get after bearing and are added with 1.To C2, C3, directly get negative.

This linear transformation can not change the regularity of distribution of desired value, is more practical forward method;

Different evaluation index often has different dimensions and dimensional unit, in order to eliminate the incommensurability brought thus, also needs to carry out nondimensionalization process to each evaluation index; Nondimensionalization also claims standardization, the normalization of data, is eliminated the impact of original variable dimension by mathematic(al) manipulation; Conventional nondimensionalization disposal route has: extreme value method, linear scaling method, normalization method, vectorial laws for criterion, Standardization Act; Said method cuts both ways, and applicable situation is also different; For the Comprehensive Evaluation Problem of being given a mark by subjectivity, should not retain the variation information of index, and the variation information of index should be eliminated, the variation information of index can be eliminated with standardized method;

Standardized method is adopted to carry out nondimensionalization, that is:

$y_i=\frac{x_i-\stackrel{\‾}{x}}{s};$

In formula: mean value $\stackrel{\‾}{x}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i,$ Mean square deviation (standard deviation) $s=\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_i-\stackrel{\‾}{x})}^2};$

After standardization, variable y _iinterval is uncertain; The mean value of index is mean square deviation is s _y=1;

Mean value: $\stackrel{\‾}{y}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{y}_i=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{x_i-\stackrel{\‾}{x}}{s}=\frac{1}{n}\×\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i-n\stackrel{\‾}{x}}{s}=0;$

Mean square deviation: $s_y=\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(y_i-\stackrel{\‾}{y})}^2}=\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\left(\frac{x_i-\stackrel{\‾}{x}}{s}\right)}^2}=\frac{1}{s}\sqrt{\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_i-\stackrel{\‾}{x})}^2}=1;$

In step s 103, weight normalized:

Determine first class index weight two-level index weight the fuzzy semantics of expert to index weights is converted into Triangular Fuzzy Number:

represent that expert is to the fuzzy evaluation of the i-th class first class index weight, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent that expert is to the fuzzy evaluation of a jth two-level index weight, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

Weight of Expert label taking value W _e=(W _e1, W _e2..., W _eK), have

In step s 103, index weights is normalized:

With for benchmark, first adopt linear method pair be normalized, then basis scaling pair with carry out equal proportion convergent-divergent, to realize the normalization of fuzzy weighted values;

If the fuzzy weighted values after normalization is ${\tilde{W}}^{i^{\′}}=(W_1^{i^{\′}},W_2^{i^{\′}},W_3^{i^{\′}}),$ Then have: $W_2^{i^{\′}}=\frac{W_2^i}{\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{W}_2^i}$

Equal proportion convergent-divergent, then have:

Can obtain: $W_1^{i^{\′}}=W_1^i\frac{W_2^{i^{\′}}}{W_2^i},W_3^{i^{\′}}=W_3^i\frac{W_2^{i^{\′}}}{W_2^i}$

After normalization, have:

In like manner can to two-level index weight be normalized, formed

In step S104, expert is to the fuzzy evaluation of index:

K expert provides the evaluation of the two-level index to each scheme respectively form fuzzy evaluation battle array:

In formula: for the expert k represented by Triangular Fuzzy Number is to the evaluation of a jth two-level index of m kind scheme;

represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

In step S105, comprehensive Weight of Expert and index weights

First, comprehensive Weight of Expert and fuzzy evaluation battle array:

Fuzzy evaluation battle array describe K expert to the evaluation of M kind scheme J two-level index, often kind of scheme has K × J fuzzy number, total M × K × J fuzzy number; Triangular Fuzzy Number is utilized to be similar to multiplication and generalized addition computing, comprehensive Weight of Expert W _ewith fuzzy evaluation battle array, form expert's comprehensive evaluation battle array expert's comprehensive evaluation battle array combines K expertise, total M × J fuzzy number:

${\tilde{P}}^{\mathrm{mj}}=(W_{E1}\⊗{\tilde{P}}_1^{\mathrm{mj}})\⊕(W_{E2}\⊗{\tilde{P}}_2^{\mathrm{mj}})\⊕...\⊕(W_{\mathrm{EK}}\⊗{\tilde{P}}_K^{\mathrm{mj}})=(P_1^{\mathrm{mj}},P_2^{\mathrm{mj}},P_3^{\mathrm{mj}})$

In formula: represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively; M=1,2 ..., M; J=1,2 ..., J;

Secondly, overall target weight and expert's comprehensive evaluation battle array:

Integrated two-stage index weights with expert's comprehensive evaluation battle array form secondary weighted synthetical evaluation battle array weighted synthetical evaluation battle array combines index weights and expert opinion, total M × J fuzzy number:

${\tilde{P}}^{{\mathrm{mj}}^{\′}}=({\tilde{w}}^{j^{\′}}\⊗{\tilde{P}}^{\mathrm{mj}})=({P_1^{{\mathrm{mj}}^{\′}}P}_2^{{\mathrm{mj}}^{\′}},P_3^{{\mathrm{mj}}^{\′}})$

In formula: represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively; M=1,2 ..., M; J=1,2 ..., J;

Comprehensive first class index weight with secondary weighted synthetical evaluation battle array form one-level weighted synthetical evaluation battle array

${\tilde{P}}^{{\mathrm{mi}}^{\′\′}}={\tilde{W}}^{i^{\′}}\⊗({\tilde{P}}^{{m1}^{\′}}\⊕...\⊕{\tilde{P}}^{{\mathrm{mj}}^{\′}}\⊕...\⊕{\tilde{P}}^{{\mathrm{mn}}^{\′}})=(P_1^{{\mathrm{mi}}^{\′\′}},P_2^{{\mathrm{mi}}^{\′\′}},P_3^{{\mathrm{mi}}^{\′\′}})$

In formula: represent the fuzzy weighted values of the i-th class first class index, represent the weighted synthetical evaluation fuzzy number of a jth two-level index under the i-th class first class index, n represents the number of two-level index under the i-th class first class index, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum, m=1 respectively, 2 ..., M; J=1,2 ..., n;

Then, aggreggate utility value is calculated:

The aggreggate utility value of each scheme is calculated according to one-level weighted synthetical evaluation battle array

${\tilde{P}}^m={\tilde{P}}^{{m1}^{\′\′}}\⊕{\tilde{P}}^{{m2}^{\′\′}}\⊕...\⊕{\tilde{P}}^{{\mathrm{mI}}^{\′\′}}=(P_1^m,P_2^m,P_3^m)$

In formula: represent that the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum, I represent the classification number of first order index respectively, m=1,2 ..., M;

Finally, sequence desired value is calculated:

Calculate the sequence desired value F of each scheme respectively _m(m=1,2 ..., M):

$F_m=\mathrm{\βmean}{\tilde{P}}^m+(1-\mathrm{\β})(1-\mathrm{\σ}{\tilde{P}}^m)$

In formula: $\mathrm{mean}{\tilde{P}}^m=\frac{P_1^m+P_2^m+P_3^m}{3}$

$\mathrm{\σ}{\tilde{P}}^m=\sqrt{{\mathrm{\σ}}^2{\tilde{P}}^m}=\sqrt{\frac{{\left(P_1^m\right)}^2+{\left(P_2^m\right)}^2+{\left(P_3^m\right)}^2-P_1^m{P}_2^m-P_1^m{P}_3^m-P_2^m{P}_3^m}{18}}$

Suppose that decision maker attaches equal importance to average and deviation, get β=0.5, then:

$F_m=0.5\mathrm{mean}{\tilde{P}}^m+0.5(1-\mathrm{\σ}{\tilde{P}}^m)$

According to F _msize, the quality of each scheme can be judged, sequence desired value larger, scheme is more excellent.

The invention provides a kind of intelligent distribution network group decision evaluation method based on fuzzy evaluation, establish the intelligent distribution network assessment indicator system of recursive hierarchy structure, propose the intelligent distribution network combination evaluation method of comprehensive Weight of Expert and index weights, be applicable to province/ground/Utilities Electric Co. of city and comprehensive evaluation is carried out to intelligent power distribution network construction situation, be particularly useful for the situation that multiple intelligent power distribution network construction scheme is compared, evaluation result can be planning, design, construction, operational management personnel provide decision-making foundation.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.

Claims (9)

1. the intelligent distribution network group decision evaluation method based on fuzzy evaluation, it is characterized in that, should based on the result of the intelligent distribution network group decision evaluation method of fuzzy evaluation according to index forward and nondimensionalization, with the form of Triangular Fuzzy Number, index is evaluated, integrated use index calculate and expert's subjective evaluation result, overall target weight and Weight of Expert carry out combination evaluation, provide the aggreggate utility evaluation of estimate of each scheme, be applicable to the situation that multiple intelligent power distribution network construction scheme is compared;

Specifically comprise the following steps:

Step one, sets up the intelligent distribution network assessment indicator system of recursive hierarchy structure, namely comprises the Hierarchy indicataors system of destination layer, rule layer and indicator layer;

Step 2, index calculate, and forward and nondimensionalization process are carried out to each evaluation index;

Step 3, agriculture products weight and Weight of Expert, and be normalized;

Step 4, expert provides the evaluation to the different index of each scheme, and is represented by Triangular Fuzzy Number by the fuzzy evaluation of expert;

Step 5, comprehensive Weight of Expert and index weights, obtain weighted synthetical evaluation battle array, obtains combination evaluation result.

2., as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, in step one, intelligent distribution network assessment indicator system comprises destination layer, rule layer and indicator layer;

Destination layer comprises power supply capacity, the quality of power supply, reliability, economy aspect;

Rule layer comprises load capacity, turns for ability, voltage, frequency, user dependability, line loss per unit, power factor, plant factor;

Indicator layer comprises load capacity index, turns for capacity index, voltage indexes, Frequency Index, user dependability index, line loss per unit index, power factor specification, plant factor index.

3., as claimed in claim 2 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, load capacity index comprises capacity load ratio of network, 10kV line load rate, 10kV distribution transforming load factor;

Turn and comprise 10kV line looped network rate, the 10kV line load rate of transform for capacity index;

Voltage indexes comprises integrated voltage qualification rate, voltage deviation, non-equilibrium among three phase voltages;

Frequency Index refers to frequency departure;

User dependability index comprises power supply reliability, the average power off time of user, the average frequency of power cut of user;

Line loss per unit index comprises comprehensive line loss per unit, hi-line loss rate, low voltage line loss rate;

Power factor specification comprises main transformer power factor, 10kV line power factor, distribution transforming power factor;

Plant factor index comprises the unbalanced degree of line load rate, the unbalanced degree of distribution transforming load factor.

4. as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, the method that evaluation index carries out forward in step 2 specifically comprises:

For the forward of reverse index, take directly to get negative method, that is: y _i=-x _i;

In formula: x _i, y _irepresent the desired value before and after index forward respectively;

For the forward of appropriate index, calculate the difference with set-point, then get negative to the absolute value of difference, that is: y _i=-| x _i-x _given|, x _givenfor set-point;

In formula: x _givenrepresent index limits.

5. as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, in step 2, evaluation index adopts standardized method to carry out nondimensionalization, that is:

In formula: represent the mean value of the same index series of each scheme, s represents the mean square deviation of the same index series of each scheme, n represents the number of scheme to be evaluated;

After standardization, variable y _iinterval is uncertain; The mean value of index is mean square deviation is s _y=1;

Mean value:

Mean square deviation:

In formula: the mean value of index series after expression forward, s _ythe mean square deviation of index series after expression forward.

6., as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, in step 3, Weight of Expert normalized specifically comprises:

Determine first class index weight two-level index weight the fuzzy semantics of expert to index weights is converted into Triangular Fuzzy Number:

represent that expert is to the fuzzy evaluation of the i-th class first class index weight, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent that expert is to the fuzzy evaluation of a jth two-level index weight, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

Weight of Expert label taking value W _e=(W _e1, W _e2..., W _eK), have

In formula: W _eirepresent the evaluation weight of expert i, K represents expert's number.

7., as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, in step 3, index weights normalized specifically comprises:

With for benchmark, first adopt linear method pair be normalized, then basis scaling pair with carry out equal proportion convergent-divergent, realize the normalization of fuzzy weighted values;

Fuzzy weighted values after normalization is then have:

Equal proportion convergent-divergent, then have:

:

After normalization, have:

In formula: I represents the number of first class index;

In like manner to two-level index weight be normalized, formed

8., as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, in step 4, the fuzzy evaluation of expert to index specifically comprises:

K expert provides the evaluation of the two-level index to each scheme respectively form fuzzy evaluation battle array:

In formula: for the expert k represented by Triangular Fuzzy Number is to the evaluation of a jth two-level index of m kind scheme;

represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively.

9., as claimed in claim 1 based on the intelligent distribution network group decision evaluation method of fuzzy evaluation, it is characterized in that, step 5 specifically comprises:

The first step, comprehensive Weight of Expert and fuzzy evaluation battle array:

Fuzzy evaluation battle array describe K expert to the evaluation of M kind scheme J two-level index, often kind of scheme has K × J fuzzy number, total M × K × J fuzzy number; Triangular Fuzzy Number is utilized to be similar to multiplication and generalized addition computing, comprehensive Weight of Expert W _ewith fuzzy evaluation battle array, form expert's comprehensive evaluation battle array expert's comprehensive evaluation battle array combines K expertise, total M × J fuzzy number:

In formula: represent a comprehensive K expert opinion, expert's comprehensive evaluation of the jth two-level index to m kind scheme represented by Triangular Fuzzy Number, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

W _eirepresent the evaluation weight of expert i;

for the expert k represented by Triangular Fuzzy Number is to the evaluation of a jth two-level index of m kind scheme; M=1,2 ..., M; J=1,2 ..., J;

represent addition and the multiplying of Triangular Fuzzy Number respectively;

Second step, overall target weight and expert's comprehensive evaluation battle array:

Integrated two-stage index weights with expert's comprehensive evaluation battle array form secondary weighted synthetical evaluation battle array weighted synthetical evaluation battle array combines index weights and expert opinion, total M × J fuzzy number:

In formula: represent and combine K expert opinion and a jth two-level index weight, the secondary weighted synthetical evaluation of a jth two-level index of the m kind scheme represented by Triangular Fuzzy Number, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent the fuzzy weighted values of the jth two-level index after normalization;

represent a comprehensive K expert opinion, expert's comprehensive evaluation of the jth two-level index to m kind scheme represented by Triangular Fuzzy Number; M=1,2 ..., M; J=1,2 ..., J;

Comprehensive first class index weight with secondary weighted synthetical evaluation battle array form one-level weighted synthetical evaluation battle array

In formula: represent and combine expert opinion and first class index weight, the weighted synthetical evaluation of i-th first class index to m kind scheme represented by Triangular Fuzzy Number, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively;

represent the fuzzy weighted values of the i-th class first class index after normalization;

represent and combine K expert opinion and a jth two-level index weight, the secondary weighted synthetical evaluation of a jth two-level index under the i-th class first class index of the m kind scheme represented by Triangular Fuzzy Number;

N represents the number of two-level index under the i-th class first class index, m=1,2 ..., M; J=1,2 ..., n;

3rd step, calculates aggreggate utility value:

The aggreggate utility value of each scheme is calculated according to one-level weighted synthetical evaluation battle array

In formula: represent the aggreggate utility fuzzy number of m kind scheme, represent the value that the lower limit of Triangular Fuzzy Number, the upper limit and possibility are maximum respectively,

I represents the classification number of first order index, m=1,2 ..., M;

4th step, calculates sequence desired value:

Calculate the sequence desired value F of each scheme respectively _m(m=1,2 ..., M):

In formula: F _mrepresent the sequence desired value of m kind scheme; represent average and the standard deviation of the evaluation fuzzy number of m kind scheme respectively; β is weights;

Decision maker attaches equal importance to average and deviation, gets β=0.5, then:

According to F _msize, namely judge the quality of each scheme, sequence desired value larger, scheme is more excellent.