CN105447646A - Health index assessment method for power distribution system - Google Patents

Abstract

The invention provides a health index assessment method for a power distribution system. The method comprises: dividing a life cycle of a power distribution device, wherein the life cycle comprises a design stage, a production stage, an experimental stage, an operational stage and a maintenance stage; according to the five stages of the life cycle, constructing a judgment matrix; according to actual operational data of the power distribution device, obtaining ordering vectors of experts; determining participating experts; configuring weights according to key eigenvectors of the power distribution device and correcting the weights by adopting a historical information correction based intelligent algorithm; and performing comprehensive assessment on the power distribution device to obtain a health index optimization result of the power distribution device. The health index assessment is effectively combined with expert experience, so that the accuracy and effectiveness of health assessment of the power distribution device are further improved.

Description

A kind of distribution system health index appraisal procedure

Technical field

The present invention relates to a kind of appraisal procedure, be specifically related to a kind of distribution system health index appraisal procedure.

Background technology

Due to China's expanding economy, the national demand to electric power is also in the growth continued.As the ring be directly connected with user in electric system, distribution system is the most direct on the impact of customer power supply reliability.But compared with developed countries, all there is larger gap in China in distribution net work structure, Standardization Construction, power distribution automation and intelligent level etc.; Compared with power transmission network, controller switching equipment has a large capacity and a wide range, kind is numerous and diverse, it is low and manufacturing firm is numerous to manufacture threshold, and existence foundation data accumulation is not enough, and the level of informatization and lean manage that degree is not high and the problem such as automaticity is low.In order to manage the controller switching equipment that these have a large capacity and a wide range better, understand the health status of distribution network, in the urgent need to carrying out health index evaluation to the running status of power distribution network.

And the present method evaluated for health index a lot, but be mostly Direct Modeling computing, do not fully take into account the experience of expert.And in the process for power distribution network health assessment, these experts can provide more information for the evaluation of power distribution network health index, expertise add the accuracy also improving health assessment.

In traditional investigation method, expert can be subject to the impact of extraneous thinking on the evaluation procedure of equipment, and this impact is favourable also has fraud, and wherein drawback is that expert's thinking is interfered in this process, lacks the space of thinking independently.

Summary of the invention

In order to make up above-mentioned defect, the application proposes a kind of distribution system health index appraisal procedure, the comprehensive integration investigation method used can make full use of the heuristic thinking of expert and computing machine to the processing power of complex calculation, and it is auxiliary with information, the network system, and the evaluation of health index is effectively combined with expertise, the health index of equipment thus improve accuracy and the validity of controller switching equipment health assessment further.

In order to realize foregoing invention object, the present invention takes following technical scheme:

A kind of distribution system health index appraisal procedure, described method comprises:

(1) life cycle of controller switching equipment is divided; Described life cycle, comprises design phase, production phase, experimental phase, operational phase and maintenance phase;

(2) according to the double teacher of life cycle, development of judgment matrix;

(3) according to controller switching equipment actual operating data, the ordering vector of expert is obtained;

(4) participant expert is determined;

(5) according to controller switching equipment key feature vector configure weights, and the intelligent algorithm correction weight based on historical information correction is adopted;

(6) comprehensive evaluation is carried out to controller switching equipment, obtain controller switching equipment health index optimum results.

Preferably, described step (2) specifically comprises: with the essential information of controller switching equipment for foundation, compares the accuracy in life cycle each stage between two, development of judgment matrix A:

$A=\left(\begin{array}{c}{a}_{11}{a}_{12}\mathrm{...}{a}_{15}\\ a_{21}{a}_{22}\mathrm{...}{a}_{25}\\ \mathrm{.........}\\ a_{51}{a}_{52}\mathrm{...}{a}_{55}\end{array}\right)={\left(a_{ij}\right)}_{5\×5}$

$a_{ij}=\frac{w_i}{w_i+w_j}(i=1,2...5;j=1,2...5)---\left(1\right)$

Wherein, w _iand w _jbe respectively life cycle i-th stage and weight score value corresponding to jth stage; The essential information of described controller switching equipment, comprises Time To Market, maintenance cost and device parameter.

Preferably, described step (3) specifically comprises:

3-1 obtains characteristic vector W by power process of iteration _c,

3-2 builds decision matrix A _d:

3-3 adopts revised simplex algorithm to obtain the ordering vector b of expert;

b＝A _dW _c(2)

$A_d=\left(\begin{array}{c}{y}_{11}{y}_{21}\mathrm{...}{y}_{51}\\ y_{12}{y}_{22}\mathrm{...}{y}_{52}\\ \mathrm{.........}\\ y_{1m}{y}_{2m}\mathrm{...}{y}_{5m}\end{array}\right)={\left(y_{ij}\right)}_{5\×m}$

Wherein, y _ijfor i stage expert is according to w in a model _twith value after normalized.

Preferably, described step (4) determines that participant expert comprises, and is sorted from big to small by element in b, and its expression formula is:

$A_d{W}_c=\left(\begin{array}{cccc}{y}_{11}& y_{21}& ...& y_{51}\\ y_{12}& y_{22}& ...& y_{52}\\ ...& ...& ...& ...\\ y_{1m}& y_{2m}& ...& y_{5m}\end{array}\right)\·\left(\begin{array}{c}{W}_1\\ W_2\\ \mathrm{...}\\ W_5\end{array}\right)=\left(\begin{array}{c}{b}_1\\ b_2\\ \mathrm{...}\\ b_m\end{array}\right)---\left(3\right)$

Wherein, m is expert's sum.

Preferably, described step (5) configure weights, and adopt the intelligent algorithm correction weight based on historical information correction to comprise:

A. determine the key feature vector of the health index model that each expert proposes, obtain the influence degree of key feature vector, and be that controller switching equipment adds mark according to influence degree;

Described key feature vector is chosen by Data Dimensionality Reduction method, comprises electrical specification, mechanical property, insulation characterisitic, oil chromatogram analysis and natural influence factor;

B. quantification matching is carried out to expert t, obtain the weight after expert t quantification;

C. according to the accuracy of the described key feature vector of each selection of specialists, modified weight is carried out to the health assessment index that expert t provides.

Further, in described step a, the influence degree obtaining key feature vector comprises:

e _t＝Σx _i(4)

Wherein, key feature vector classification value X _ibe expressed as controller switching equipment and add tagged summation.

Further, the weights W after described expert t quantification _texpression formula be:

W _t＝a _t+b _t+c _t+d _t(5)

Wherein, a _t, b _t, c _tand d _trepresent the historical information weights that the academic title of expert t, field power, qualifications and record of service and problem familiarity are corresponding respectively.

Further, in described step b, the health assessment index provided expert t carries out modified weight and comprises: according to the accuracy variance of expert's history investigation record, revise the historical information weights of expert

If the mark value of absolute difference is x _i, described accuracy variance is:

${\stackrel{\‾}{S}}^2=\frac{\underset{i=1}{\overset{n}{\Σ}}{(x_i-1)}^2}{n-1}---\left(6\right)$

Wherein, x _ifor proper vector classification value X _icorresponding component, n is sample size;

With the evaluating data in history investigation record for Comparative indices, calculate the accuracy of Comparative indices respectively:

$\stackrel{\‾}{\mathrm{\α}}=\frac{\underset{i=0}{\overset{n}{\Σ}}{\mathrm{\α}}_i}{n}---\left(7\right)$

In formula (7), α _ifor discussing the standard value of the absolute difference of acquisition at every turn, for the component about absolute difference in historical information weights, then

$\stackrel{\‾}{\mathrm{\β}}=\frac{\underset{i=0}{\overset{n}{\Σ}}{\mathrm{\β}}_i}{n}---\left(8\right)$

In formula (8), β _ifor discussing the variance criterion value of acquisition at every turn, for in historical information weights about component of variance; Then

$\stackrel{\‾}{\mathrm{\γ}}=\frac{\underset{i=0}{\overset{n}{\Σ}}{\mathrm{\γ}}_i}{n}---\left(9\right)$

In formula (9), γ _ifor discussing the accuracy variance criterion value of acquisition at every turn, for in historical information weights about accuracy component of variance.

Further, the historical information weights of expert t are obtained by power multiplication its expression formula is:

$\stackrel{\‾}{{\mathrm{\λ}}_t}={\stackrel{\‾}{\mathrm{\α}}}^3{\stackrel{\‾}{\mathrm{\β}}}^2\stackrel{\‾}{\mathrm{\γ}}---\left(10\right)$

The weight of each expert of normalized, its expression formula is:

${\mathrm{\λ}}_t=\frac{(W_t+e_t)\×\stackrel{\‾}{{\mathrm{\λ}}_t}}{\underset{t=1}{\overset{m}{\Σ}}(W_t+e_t)\×\stackrel{\‾}{{\mathrm{\λ}}_t}}---\left(11\right)$

Wherein, m is expert's sum, e _tfor weights influence degree.

Preferably, described step (6) specifically comprises: establish the health assessment Index A that expert t proposes _t, complete controller switching equipment health index optimum results by following formula and obtain:

$\underset{t=1}{\overset{m}{\Σ}}{A}_t\×{\mathrm{\λ}}_t---\left(12\right)$

Compared with immediate prior art, the beneficial effect that the present invention reaches is:

(1) method of this patent proposition, discusses comprehensive integration in the middle of approach application to the example of " software development of controller switching equipment health status comprehensive evaluation ".Compensate for the effective integration means having lacked qualitative and quantitative knowledge in power status health assessment technology, and the defect that the evaluation of health index and expertise effectively cannot be combined, the evaluation of health index is effectively combined with expertise, the accuracy of further raising controller switching equipment health assessment, validity.

(2) expert can be met in the process using comprehensive integration investigation method, provide the space of thinking independently, information interaction is carried out in the process of comprehensive integration, both the common advantage inquired in traditional investigation method had been remained, also eliminate the interference of inconsistent thinking, ensure that the independence of thinking.

Accompanying drawing explanation

Fig. 1 is a kind of distribution system health index appraisal procedure process flow diagram; Fig. 2 is distribution system health index assessment Organization Chart.

Embodiment

Below with reference to accompanying drawing, the specific embodiment of the present invention is described in further detail.

As shown in Figure 1, a kind of distribution system health index appraisal procedure, described method comprises:

(1) life cycle of controller switching equipment is divided; Described life cycle, comprises design phase, production phase, experimental phase, operational phase and maintenance phase;

(2) according to the double teacher of life cycle, development of judgment matrix; Concrete basis is discussed the basic condition of equipment each time and determines.Such as, equipment outbalance in design pilot production of just listing, and service time longer equipment in use, maintenance aspect outbalance.Temporally concrete sequence is as shown in table 1, thus calculates the order of priority of each expert.

Table 1 marks life cycle by tenure of use

Service time	Life cycle sorts	Corresponding score value w
			0-1	Design > (producing, experiment) > (use, safeguard)	(5，3，1)
1-3	(design, produces, experiment) > (use, safeguard)	(2，1)
			3-10	(use, safeguard) > (design, produces, experiment)	(2，1)
More than 10 years	Safeguard use (design, produces, experiment)	(5，3，1)

With the essential information of controller switching equipment for foundation, compare the accuracy in life cycle each stage between two, development of judgment matrix A:

$A=\left(\begin{array}{c}{a}_{11}{a}_{12}\mathrm{...}{a}_{15}\\ a_{21}{a}_{22}\mathrm{...}{a}_{25}\\ \mathrm{.........}\\ a_{51}{a}_{52}\mathrm{...}{a}_{55}\end{array}\right)={\left(a_{ij}\right)}_{5\×5}$

$a_{ij}=\frac{w_i}{w_i+w_j}(i=1,2...5;j=1,2...5)---\left(1\right)$

Wherein, w _iand w _jbe respectively life cycle i-th stage and weight score value corresponding to jth stage; The essential information of described controller switching equipment, comprises Time To Market, maintenance cost and device parameter.

(3) according to controller switching equipment actual operating data, the ordering vector of expert is obtained; Comprise:

3-1 obtains characteristic vector W by power process of iteration _c,

3-2 builds decision matrix A _d:

3-3 adopts revised simplex algorithm to obtain the ordering vector b of expert;

b＝A _dW _c(2)

$A_d=\left(\begin{array}{c}{y}_{11}{y}_{21}\mathrm{...}{y}_{51}\\ y_{12}{y}_{22}\mathrm{...}{y}_{52}\\ \mathrm{.........}\\ y_{1m}{y}_{2m}\mathrm{...}{y}_{5m}\end{array}\right)={\left(y_{ij}\right)}_{5\×m}$

Wherein, y _ijfor i stage expert is according to w in a model _twith value after normalized.Form decision matrix A _das shown in table 2:

Table 2 expert decision-making vector table

Expert numbers	Design phase	Production phase	Experimental phase	Operational phase	Maintenance phase
						1	y 11	y 21	y 31	y 41	y 51
2	y 12	y 22	y 32	y 42	y 52
						3	y 13	y 23	y 33	y 43	y 53
…	…	…	…	…	…

(4) participant expert is determined;

Sorted from big to small by element in b, its expression formula is:

$A_d{W}_c=\left(\begin{array}{cccc}{y}_{11}& y_{21}& \mathrm{...}& y_{52}\\ y_{12}& y_{22}& \mathrm{...}& y_{52}\\ \mathrm{...}& \mathrm{...}& \mathrm{...}& \mathrm{...}\\ y_{1m}& y_{2m}& \mathrm{...}& y_{5m}\end{array}\right)\·\left(\begin{array}{c}{w}_1\\ w_2\\ \mathrm{...}\\ w_5\end{array}\right)=\left(\begin{array}{c}{b}_1\\ b_2\\ \mathrm{...}\\ b_m\end{array}\right)---\left(3\right)$

Wherein, m is expert's sum.

(5) according to controller switching equipment key feature vector configure weights, and the intelligent algorithm based on historical information correction is adopted

Revise weight; After complete for the investigation of life double teacher, the model about health index that each expert proposes by we carries out comprehensively according to the weight of configuration, the health index evaluation that the equipment that draws is final.The importance of each key feature amount related to according to controller switching equipment after expert proposes evaluation model is integrated model.By expert info such as popularity, academic title, qualifications and record of service, field contribution degree, the quantification such as problem familiarity.And form the modified weight method based on the intelligence of expert info according to the accuracy of expert's history investigation.

Described step (5) configure weights, and adopt the intelligent algorithm correction weight based on historical information correction to comprise:

A. determine the key feature vector of the health index model that each expert proposes, obtain the influence degree of key feature vector, and be that controller switching equipment adds mark according to influence degree;

Described key feature vector is chosen by Data Dimensionality Reduction method, comprises electrical specification, mechanical property, insulation characterisitic, oil chromatogram analysis and natural influence factor; Choose normative reference as table 3.

Table 3 key feature amount affects marker values table

Characteristic quantity class	Key feature amount	Mark value
			Electrical specification X 1	Voltage, electric current, resistance	(1,1,1)
Mechanical property X 2	Working time	(2)
			Insulation characterisitic X 3	Insulation resistance, contact temperature	(1,1)
Oil chromatogram analysis X 4	Hydrogen, total alkynes, carbon monoxide, carbon dioxide	(1,1,1,1)
			Natural cause affects X 5	Lightning monitoring, bad environments situation (humiture)	(1,2)

Different mark value is given in this process according to the difference of characteristic quantity to the influence degree that last health index is evaluated.

B. quantification matching is carried out to expert t, obtain the weight after expert t quantification;

C. according to the accuracy of the described key feature vector of each selection of specialists, modified weight is carried out to the health assessment index that expert t provides.

In step a, the influence degree obtaining key feature vector comprises:

e _t＝Σx _i(4)

Wherein, key feature vector classification value X _ibe expressed as controller switching equipment and add tagged summation.According to academic title, field power, qualifications and record of service, problem familiarity four expert infos mark expert, and mark outcome record is as shown in table 4.

Table 4 expert personal information affects marker values table

Judging basis	Expert info	Weight score value
			Academic title a	Senior engineer, slip-stick artist, assistant engineer, technician	[4.5,3.5，2.5，1.5]
Field power b	Very large, greatly, more greatly, less	[4,3,2,1]
			Qualifications and record of service (length of service) c	0-3,3-10,10-20, more than 20 years	[4,3,2,1]
Problem familiarity d	Very familiar, more familiar, be generally familiar with, be slightly familiar with	[5,4,3,2]

In these five expert infos, particularly important to the familiarity of problem, so its score value is divided into slightly higher than other several.And academic title is relative also more convincing, so score value divides also higher.

Weights W after expert t quantification _texpression formula be:

W _t＝a _t+b _t+c _t+d _t(5)

Wherein, a _t, b _t, c _tand d _trepresent the historical information weights that the academic title of expert t, field power, qualifications and record of service and problem familiarity are corresponding respectively.

In step b, the health assessment index provided expert t carries out modified weight and comprises: according to the accuracy variance of expert's history investigation record, revise the historical information weights of expert

If the mark value of absolute difference is x _i, described accuracy variance is:

${\stackrel{\‾}{S}}^2=\frac{\underset{i=1}{\overset{n}{\Σ}}{(x_i-1)}^2}{n-1}---\left(6\right)$

Wherein, x _ifor proper vector classification value X _icorresponding component, n is sample size;

With the evaluating data in history investigation record for Comparative indices, calculate the accuracy of Comparative indices respectively:

$\stackrel{\‾}{\mathrm{\α}}=\frac{\underset{i=0}{\overset{n}{\Σ}}{\mathrm{\α}}_i}{n}---\left(7\right)$

In formula (7), α _ifor discussing the standard value of the absolute difference of acquisition at every turn, for the component about absolute difference in historical information weights, then

$\stackrel{\‾}{\mathrm{\β}}=\frac{\underset{i=0}{\overset{n}{\Σ}}{\mathrm{\β}}_i}{n}---\left(8\right)$

In formula (8), β _ifor discussing the variance criterion value of acquisition at every turn, for in historical information weights about component of variance; Then

$\stackrel{\‾}{\mathrm{\γ}}=\frac{\underset{i=0}{\overset{n}{\Σ}}{\mathrm{\γ}}_i}{n}---\left(9\right)$

In formula (9), γ _ifor discussing the accuracy variance criterion value of acquisition at every turn, for in historical information weights about accuracy component of variance.

The historical information weights of expert t are obtained by power multiplication its expression formula is:

$\stackrel{\‾}{{\mathrm{\λ}}_t}={\stackrel{\‾}{\mathrm{\α}}}^3{\stackrel{\‾}{\mathrm{\β}}}^2\stackrel{\‾}{\mathrm{\γ}}---\left(10\right)$

Due to all be less than 1, thus descending according to falling power arrangement according to significance level.

And after investigation each time, we carry out the feedback of information, retain the result of Experts ', for the reference as history weights during investigation next time.

The weight of each expert of normalized, its expression formula is:

${\mathrm{\λ}}_t=\frac{(w_t+e_t)\×\stackrel{\‾}{{\mathrm{\λ}}_t}}{\underset{t=1}{\overset{m}{\Σ}}(W_t+e_t)\×\stackrel{\‾}{{\mathrm{\λ}}_t}}---\left(11\right)$

Wherein, m is expert's sum, e _tfor weights influence degree.

First the health index HI of substation transformer is defined between [1,10] by we.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.The exact value of mark value is 1.The health index of the health index provided with expert and the final gained of equipment compares, the result from following consideration is compared:

Table 5 accuracy historical information affects marker values table

(6) comprehensive evaluation is carried out to controller switching equipment, obtain controller switching equipment health index optimum results.

Described step (6) specifically comprises: establish the health assessment Index A that expert t proposes _t, complete controller switching equipment health index optimum results by following formula and obtain:

$\underset{t=1}{\overset{m}{\&Sigma;}}{A}_t\&times;{\mathrm{\&lambda;}}_t---\left(12\right).$

Said method can be utilized to build software platform, for expert builds friendly operation interface.After in comprehensive integration investigation, process terminates, we carry out the feedback of information, analyze and stored in database to the accuracy of expert opinion result, for the investigation correction as historical information confession expert during investigation next time.

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; those of ordinary skill in the field still can modify to the specific embodiment of the present invention with reference to above-described embodiment 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 (10)

1. a distribution system health index appraisal procedure, is characterized in that, described method comprises:

(1) life cycle of controller switching equipment is divided; Described life cycle, comprises design phase, production phase, experimental phase, operational phase and maintenance phase;

(2) according to the double teacher of life cycle, development of judgment matrix;

(3) according to controller switching equipment actual operating data, the ordering vector of expert is obtained;

(4) participant expert is determined;

(5) according to controller switching equipment key feature vector configure weights, and the intelligent algorithm correction weight based on historical information correction is adopted;

(6) comprehensive evaluation is carried out to controller switching equipment, obtain controller switching equipment health index optimum results.

2. the method for claim 1, is characterized in that, described step (2) specifically comprises: with the essential information of controller switching equipment for foundation, compares the accuracy in life cycle each stage between two, development of judgment matrix A:

$A=\left(\begin{array}{cccc}{a}_{11}& a_{12}& ...& a_{15}\\ a_{21}& a_{22}& ...& a_{25}\\ ...& ...& ...& ...\\ a_{51}& a_{52}& ...& a_{55}\end{array}\right)={\left(a_{ij}\right)}_{5\&times;5}$

$a_{ij}=\frac{w_i}{w_i+w_j}(i=1,2...5;j=1,2...5)---\left(1\right)$

Wherein, w _iand w _jbe respectively life cycle i-th stage and weight score value corresponding to jth stage; The essential information of described controller switching equipment, comprises Time To Market, maintenance cost and device parameter.

3. the method for claim 1, is characterized in that, described step (3) specifically comprises:

3-1 obtains characteristic vector W by power process of iteration _c,

3-2 builds decision matrix A _d:

3-3 adopts revised simplex algorithm to obtain the ordering vector b of expert;

b＝A _dW _c(2)

$A_d=\left(\begin{array}{cccc}{y}_{11}& y_{21}& ...& y_{51}\\ y_{12}& y_{22}& ...& y_{52}\\ ...& ...& ...& ...\\ y_{1m}& y_{2m}& ...& y_{5m}\end{array}\right)={\left(y_{ij}\right)}_{5\&times;m}$

Wherein, y _ijfor i stage expert is according to w in a model _twith value after normalized.

4. the method as described in claim 1 or 3, is characterized in that, described step (4) determines that participant expert comprises, and is sorted from big to small by element in b, and its expression formula is:

$A_d{W}_c=\left(\begin{array}{cccc}{y}_{11}& y_{21}& ...& y_{51}\\ y_{12}& y_{22}& ...& y_{52}\\ ...& ...& ...& ...\\ y_{1m}& y_{2m}& ...& y_{5m}\end{array}\right)\&CenterDot;\left(\begin{array}{c}{w}_1\\ w_2\\ ...\\ w_5\end{array}\right)=\left(\begin{array}{c}{b}_1\\ b_2\\ ...\\ b_m\end{array}\right)---\left(3\right)$

Wherein, m is expert's sum.

5. the method for claim 1, is characterized in that, described step (5) configure weights, and adopts the intelligent algorithm correction weight based on historical information correction to comprise:

A. determine the key feature vector of the health index model that each expert proposes, obtain the influence degree of key feature vector, and be that controller switching equipment adds mark according to influence degree;

Described key feature vector is chosen by Data Dimensionality Reduction method, comprises electrical specification, mechanical property, insulation characterisitic, oil chromatogram analysis and natural influence factor;

B. quantification matching is carried out to expert t, obtain the weight after expert t quantification;

C. according to the accuracy of the described key feature vector of each selection of specialists, modified weight is carried out to the health assessment index that expert t provides.

6. method as claimed in claim 5, is characterized in that, in described step a, the influence degree obtaining key feature vector comprises:

e _t＝Σx _i(4)

Wherein, key feature vector classification value X _ibe expressed as controller switching equipment and add tagged summation.

7. method as claimed in claim 5, is characterized in that, the weights W after described expert t quantification _texpression formula be:

W _t＝a _t+b _t+c _t+d _t(5)

Wherein, a _t, b _t, c _tand d _trepresent the historical information weights that the academic title of expert t, field power, qualifications and record of service and problem familiarity are corresponding respectively.

8. method as claimed in claim 7, is characterized in that, in described step b, the health assessment index provided expert t carries out modified weight and comprises: according to the accuracy variance of expert's history investigation record, revise the historical information weights of expert

If the mark value of absolute difference is x _i, described accuracy variance is:

${\stackrel{\&OverBar;}{S}}^2=\frac{\underset{i=1}{\overset{n}{\&Sigma;}}{(x_i-1)}^2}{n-1}---\left(6\right)$

Wherein, x _ifor proper vector classification value X _icorresponding component, n is sample size;

With the evaluating data in history investigation record for Comparative indices, calculate the accuracy of Comparative indices respectively:

$\stackrel{\&OverBar;}{\mathrm{\&alpha;}}=\frac{\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&alpha;}}_i}{n}---\left(7\right)$

In formula (7), α _ifor discussing the standard value of the absolute difference of acquisition at every turn, for the component about absolute difference in historical information weights, then

$\stackrel{\&OverBar;}{\mathrm{\&beta;}}=\frac{\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&beta;}}_i}{n}---\left(8\right)$

In formula (8), β _ifor discussing the variance criterion value of acquisition at every turn, for in historical information weights about component of variance; Then

$\stackrel{\&OverBar;}{\mathrm{\&gamma;}}=\frac{\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&gamma;}}_i}{n}---\left(9\right)$

In formula (9), γ _ifor discussing the accuracy variance criterion value of acquisition at every turn, for in historical information weights about accuracy component of variance.

9. method as claimed in claim 8, is characterized in that, is obtained the historical information weights of expert t by power multiplication its expression formula is:

$\stackrel{\&OverBar;}{{\mathrm{\&lambda;}}_t}={\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}^3{\stackrel{\&OverBar;}{\mathrm{\&beta;}}}^2\stackrel{\&OverBar;}{\mathrm{\&gamma;}}---\left(10\right)$

The weight of each expert of normalized, its expression formula is:

${\mathrm{\&lambda;}}_t=\frac{(w_t+e_t)\&times;\stackrel{\&OverBar;}{{\mathrm{\&lambda;}}_t}}{\underset{t=1}{\overset{m}{\&Sigma;}}(W_t+e_t)\&times;\stackrel{\&OverBar;}{{\mathrm{\&lambda;}}_t}}---\left(11\right)$

Wherein, m is expert's sum, e _tfor weights influence degree.

10. the method for claim 1, is characterized in that, described step (6) specifically comprises: establish the health assessment Index A that expert t proposes _t, complete controller switching equipment health index optimum results by following formula and obtain:

$\underset{t=1}{\overset{m}{\&Sigma;}}{A}_t\&times;{\mathrm{\&lambda;}}_t---\left(12\right).$