CN105096037A - Failure risk determination method of photovoltaic assembly - Google Patents

Abstract

The invention relates to a failure risk determination method of a photovoltaic assembly. The method comprises the following steps that: fault modes of a photovoltaic assembly are determined; an expert panel evaluates grades of correlated evaluation index factors of the fault modes and a fuzzy evaluation matrix of the fault modes is constructed according to evaluation results; a weight set of the correlated evaluation index factors is constructed based on an analytic hierarchy process; integrated risk evaluation factors of all fault modes are calculated according to the fuzzy determination matrix and the weight set; and failure risks of the fault modes are determined based on the integrated risk evaluation factors. According to the invention, objectification and meaningfulness of index evaluation are realized; with the fuzzy integrated evaluation, weight assignment is carried out by using the analytic hierarchy process, thereby realizing quantification continuity and weighting of the risk evaluation indexes; with the failure risk analysis of the photovoltaic assembly, a phenomenon that the quantification ordering is unreasonable and is repeated according to the failure mode effect and criticality analysis (FMECA) method can be effectively avoided and thus the analysis is closer to the reality, thereby improving the credibility of the failure risk analysis result of the photovoltaic assembly.

Description

A kind of photovoltaic module failure risk method of discrimination

Technical field

The present invention relates to photovoltaic module technical field, particularly relate to a kind of photovoltaic module failure risk method of discrimination.

Background technology

Photovoltaic module is the critical component of solar power system, and photovoltaic module reliability is to the normal operation important of whole electricity generation system.In photovoltaic plant actual motion, because the extraneous factor effects such as temperature, humidity or UV radiation can cause photovoltaic module conversion efficiency to reduce even complete failure, thus affect the generating efficiency in whole power station, therefore be necessary the evaluation studies of carrying out photovoltaic module reliability, wherein failure risk analysis is prerequisite and the basis of carrying out photovoltaic module reliability assessment.

Current dominant failure analytical approach has FMEA/FMECA (FailureModeEffectandCriticalityAnalysis), FTA (FailureTreeAnalysis), BFA (BouncingFailureAnalysis) etc., and wherein failure mode, impact and HAZAN (FMECA) are a kind of common system failure analysis methods.The impact that in FMECA analytic product, all issuable disabled status cause product, system, and classified (RPN method) by each failure mode order of severity S, Frequency O and detection degree D.Although this method is easy to operation, is applied in photovoltaic module failure analysis and there is some problems:

(1) failure risk index uncontinuity: FMECA method utilizes risk priority number method (RPN) to cross three key elements affecting failure mode Severity level: { fault seriousness S, incidence O, detection degree D} carry out risk identification, obtain quantitative RPN value, i.e. ([SOD]={ (1 ... 10), (1 ... 10), (1 ... 10) }).Though method to a certain degree compensate for the deficiency of qualitative analysis, there is the discontinuous problem of risk indicator value in risk priority number method (RPN).As calculated, RPN only has 120 values, also namely in 1 ~ 1000 theoretical coverage, has the risk quantum of 88% to get.Such as, suppose [SOD]={ 10,10,10}, then RPN=1000; Suppose [SOD]={ 9,9,9}, then RPN=729, the risk quantum between 1000 and 729 cannot be got, and larger spacer section appears in value.

(2) failure risk sequence is credible: first 3 indexs that in FMECA, RPN function is chosen self do not have actual physics meaning, and the subjectivity of index easily causes the normal and objective sensation of people to degree of risk of assessment result not to be inconsistent; Secondly 3 index value many employings expert point rating method, comparatively large by the subjective impact such as expertise, experience, RPN assesses ranking results consistance, credibility is theoretically unsound.

With certain 10KW experimental power station photovoltaic module history data and actual conditions for foundation, utilize traditional F MECA method to carry out photovoltaic module failure risk analysis, result is as shown in table 1:

Table 1 is based on the photovoltaic module failure risk analysis of FMECA

According to table 1 traditional RPN value ranking results, fault mode 1 comes risk assessment minimum position, and fault mode 2,5 sorts and repeats, ranking results obviously and actual conditions be not inconsistent.

Summary of the invention

For the deficiency that prior art exists, the object of the present invention is to provide a kind of raising failure risk analysis credible result, make the photovoltaic module failure risk method of discrimination that analysis result and reality conform to more.

For achieving the above object, the present invention can be achieved by the following technical programs:

A kind of photovoltaic module failure risk method of discrimination, comprises the following steps:

Determine the fault mode of photovoltaic module;

Expert group, respectively to fault mode relevant evaluation index factor opinion rating, builds the fuzzy matrix for assessment of fault mode according to evaluation result;

The weight sets of relevant evaluation index factor is built according to analytical hierarchy process;

The Rate of aggregative risk factor of each fault mode is calculated according to fuzzy matrix for assessment and weight sets;

According to the failure risk of Rate of aggregative risk factor determination fault mode.

Further, described relevant evaluation index factor comprises output power loss, breakdown maintenance cost and probability of happening.Output power loss, breakdown maintenance cost substitute the order of severity in traditional RPN, detection degree, and reason is as follows:

(1) output power loss: be one of Key Performance Indicator weighing photovoltaic module and power station, consequence that what most fault caused affect finally can be lost index from output power and be embodied, and has clear and definite physical significance;

(2) breakdown maintenance cost: breakdown maintenance cost can not only weigh fault severity level with clear and definite quantification value, also reflects fault finding degree simultaneously; For the fault of easily detection, discovery is easy and in time, maintenance cost is low, and the fault not easily detected, influence time is long, and detection difficulty is large, causes maintenance cost also high, and therefore breakdown maintenance cost more can reflect that photovoltaic module inefficacy is essential;

(3) output power loss and breakdown maintenance cost from the historical data of photovoltaic module use and analysis of Production Technology value, can improve the value mode relying on expert's subjective determination completely, significantly reduce traditional RPN value sequence subjectivity.

Further, the detailed process building the fuzzy matrix for assessment of fault mode comprises:

Suppose that expert's total number of persons is S, fault mode is k, and evaluation index factor is i, and to be expert's number of grade m be evaluation evaluation index factor i then the evaluation collection of the evaluation index factor i of fault mode k is:

$R_i^k=\left\{\frac{s_{i1}^k}{s},\frac{s_{i2}^k}{s},...,\frac{s_{im}^k}{s}\right\}=\left\{r_{i1}^k,r_{i2}^k,...,r_{im}^k\right\};$

The evaluation collection of all evaluation index factor i of described fault mode k is built into the fuzzy matrix for assessment R of fault mode k ^k, wherein i=1 ..., n,

$R^k={\[R_1^k,R_2^k,...,R_n^k\]}^T=\left[\begin{array}{cccc}{r}_{11}^k& r_{12}^k& ...& r_{1m}^k\\ r_{21}^k& r_{22}^k& ...& r_{2m}^k\\ ...& ...& ...& ...\\ r_{n1}^k& r_{n2}^k& ...& r_{nm}^k\end{array}\right].$

Further, the step building the weight sets of relevant evaluation index factor comprises:

The judgment matrix of evaluation index factor is built according to 1-9 scaling law;

Obtain the proper vector corresponding to judgment matrix Maximum characteristic root;

Carry out consistency check to judgment matrix, if test value is less than 0.1, then this proper vector is weight sets, otherwise rebuilds judgment matrix.

Further, to the formula that judgment matrix carries out consistency check be: wherein, I _cfor consistency check index, I _rfor Aver-age Random Consistency Index.

Further, wherein, λ _maxfor the Maximum characteristic root of judgment matrix, n is the exponent number of judgment matrix.

Further, the step calculating the Rate of aggregative risk factor comprises:

Fuzzy matrix for assessment is multiplied with weight sets and show that fuzzy synthesis judges collection;

Fuzzy synthesis is judged that collection and opinion rating are weighted average treatment, draws the Rate of aggregative risk factor.

Further, the described Rate of aggregative risk factor is larger, represents that failure risk is higher.

Further, the principle followed when evaluating output power loss is: lose larger, opinion rating is higher; The principle followed when evaluating breakdown maintenance cost is: expense is higher, and opinion rating is larger; The principle followed when evaluating probability of happening is: probability is larger, and opinion rating is higher.

The present invention proposes output power loss, breakdown maintenance cost and probability of happening these three brand-new evaluation index factors, achieve the objectifying of metrics evaluation, meaning, improve the evaluation value mode relying on expert's subjective determination completely, significantly reduce the subjectivity of traditional RPN value sequence.Utilize fuzzy comprehensive evoluation, introduce analytical hierarchy process and carry out weight assignment, realize quantification continuity, the weight of evaluation index factor, be applied to photovoltaic module failure risk analysis, solve that traditional analysis exists preferably subjective, physical significance is not obvious, quantize that sequence is unreasonable, sequence such as to repeat at the problem, evaluation result rationalizes, and analysis result and actual conditions is fitted more, substantially increases the credibility of photovoltaic module failure risk analysis result.

Accompanying drawing explanation

Fig. 1 is schematic flow sheet of the present invention.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the present invention is further illustrated:

As shown in Figure 1, photovoltaic module failure risk method of discrimination of the present invention, mainly comprises the following steps:

Step 1: the fault mode (square frame 1) determining photovoltaic module, concrete analysis defining method can adopt existing ripe scheme, such as FMECA analytic approach in prior art, and therefore not to repeat here.

Step 2: expert group, respectively to fault mode relevant evaluation index factor opinion rating, builds the fuzzy matrix for assessment (square frame 2) of fault mode according to evaluation result.

Wherein, relevant evaluation index factor comprises output power loss, breakdown maintenance cost and probability of happening, represents respectively with P, W, O.

Above-mentioned steps 2 specifically comprises:

Step 2.1: suppose that expert's total number of persons is S, fault mode is k, and evaluation index factor is i, to be expert's number of grade m be evaluation evaluation index factor i then the evaluation collection of the evaluation index factor i of fault mode k is:

$R_i^k=\left\{\frac{s_{i1}^k}{s},\frac{s_{i2}^k}{s},...,\frac{s_{im}^k}{s}\right\}=\left\{r_{i1}^k,r_{i2}^k,...,r_{im}^k\right\}---\left(1\right);$

Step 2.2: the evaluation collection of all evaluation index factor i of fault mode k is built into the fuzzy matrix for assessment R of fault mode k ^k, wherein i=1 ..., n,

$R^k={\[R_1^k,R_2^k,...,R_n^k\]}^T=\left[\begin{array}{cccc}{r}_{11}^k& r_{12}^k& ...& r_{1m}^k\\ r_{21}^k& r_{22}^k& ...& r_{2m}^k\\ ...& ...& ...& ...\\ r_{n1}^k& r_{n2}^k& ...& r_{nm}^k\end{array}\right]---\left(2\right).$

Wherein, opinion rating V is divided into 5 grades, i.e. V={1,2,3,4,5}, and the principle followed when evaluating output power loss is: lose larger, opinion rating is higher; The principle followed when evaluating breakdown maintenance cost is: expense is higher, and opinion rating is larger; The principle followed when evaluating probability of happening is: probability is larger, and opinion rating is higher.Evaluating to carry out unifying, according to the service condition of photovoltaic module reality, historical data and analysis of Production Technology, setting up the opinion rating system (table 2) of evaluation index factor:

The opinion rating system of table 2 evaluation index factor

Evaluation index factor	Grade 1	Grade 2	Grade 3	Class 4	Class 5
						Output power loss P	0～10％	10％～30％	30％～70％	70％～90％	90％～100％
Breakdown maintenance cost W	Extremely low	Low	Medium	High	High
						Probability of happening O	Rare	Less	Generally	Higher	High

Step 3: the weight sets (square frame 3) building relevant evaluation index factor according to analytical hierarchy process.Weight sets gives to embody each evaluation index factor significance level of photovoltaic module the set that the respective weight factor forms.Determine that each factor weight is one of link the most key in Comprehensive Evaluation, weight factor whether appropriate, will directly affect Comprehensive Evaluation result.Qualitative and quantitative analysis combines by analytical hierarchy process, structure trip current, and utilizes consistency check, can eliminate the human factor in weight analysis method as far as possible, ensure validity, the practicality of weight.

Above-mentioned steps 3 specifically comprises:

Step 3.1: the judgment matrix building evaluation index factor according to 1-9 scaling law, with representing that influence factor is to factor comparative result, compare value and utilize 1-9 scaling law (see table 3), Judgement Matricies is as follows:

$P=\left[\begin{array}{cccc}{p}_{11}& p_{12}& ...& p_{ln}\\ p_{21}& p_{22}& ...& p_{2n}\\ ...& ...& ...& ...\\ p_{n1}& p_{n2}& ...& p_{nn}\end{array}\right]---\left(3\right)$

In formula (3), n is evaluation index factor;

Table 31-9 scaling law judges value table of grading

Yardstick P	Implication
		1	With " of equal importance "
3	With " important a little "
		5	With " obviously important "
7	With " strongly important "
		9	With " definitely important "
One of 2,4,6,8	Each adjacent rank 1-3, intermediate value between 3-5,5-7 and 7-9

Step 3.2: obtain the proper vector corresponding to judgment matrix Maximum characteristic root;

Step 3.3: in order to reduce the human factor impact of qualitative analysis, weight accuracy is obtained in assessment, and carry out consistency check to judgment matrix, if test value is less than 0.1, then this proper vector is weight sets, otherwise rebuilds judgment matrix.Formula judgment matrix being carried out to consistency check is: wherein, I _cfor consistency check index, I _rfor Aver-age Random Consistency Index, for 1-9 rank judgment matrix, its value is in table 4. wherein, λ _maxfor the Maximum characteristic root of judgment matrix, n is the exponent number of judgment matrix.

Table 41-9 rank judgment matrix I _rvalue table

n	I R
		1	0.00
2	0.00
		3	0.58
4	0.90
		5	1.12
6	1.24
		7	1.32
8	1.41
		9	1.45

Step 4: the Rate of aggregative risk factor (square frame 4) calculating each fault mode according to fuzzy matrix for assessment and weight sets.

Above-mentioned steps 4 specifically comprises:

Step 4.1: fuzzy matrix for assessment is multiplied with weight sets and show that fuzzy synthesis judges collection; If the weight sets of the relevant evaluation index factor of required fault mode k is in formula, n is evaluation index factor, then fuzzy comprehensive evoluation integrates as W ^kbe multiplied with formula (2):

$B^k=W^k{R}^k=\[w_1^k{w}_2^k...w_n^k\]\·\left[\begin{array}{cccc}{r}_{11}^k& r_{12}^k& ...& r_{1m}^k\\ r_{21}^k& r_{22}^k& ...& r_{2m}^k\\ ...& ...& ...& ...\\ r_{n1}^k& r_{n2}^k& ...& r_{nm}^k\end{array}\right]---\left(6\right);$

By fuzzy synthesis, step 4.2: in order to more intuitively find out evaluation result, judges that collection and opinion rating are weighted average treatment, draws Rate of aggregative risk factor C ^k:

C ^k＝B ^k·V ^T(7)。

Step 5: according to the failure risk of Rate of aggregative risk factor determination fault mode, the Rate of aggregative risk factor is larger, represents that failure risk is higher.

Flow and method of the present invention is illustrated below by data:

Step 1: the fault mode determining photovoltaic module, for table 1.

Step 2: the fuzzy matrix for assessment of fault mode, for the fault mode 1 of table 1.

Step 2.1: according to Delphi method evaluation, suppose that expert group's total number of persons is 100, evaluate by table 2,0 is had respectively, 0,10 to 5 opinion ratings of output power loss P, 20,70 expert evaluations, have 0 respectively to 5 opinion ratings of breakdown maintenance cost W, 10,50,30,10 expert evaluations, have 80,20 respectively to 5 opinion ratings of probability of happening O, 0,0,0 expert evaluation.According to formula (1), then evaluating collection is accordingly $R_1^1=\{0.0,0.0,0.1,0.2,0.7\},R_2^1=\{0.0,0.1,0.5,0.3,0.1\},$ $R_3^1=\{0.8,0.2,0.0,0.0,0.0\}.$

Step 2.2: according to formula (2), the fuzzy matrix for assessment of fault mode 1 is:

$R^1={\[R_1^1,R_2^1,R_3^k\]}^T=\left[\begin{array}{ccccc}0.0& 0.0& 0.1& 0.2& 0.7\\ 0.0& 0.1& 0.5& 0.3& 0.1\\ 0.8& 0.2& 0.0& 0.0& 0.0\end{array}\right].$

Step 3: the weight sets building evaluation index factor.

Step 3.1,3.2: according to formula (3) and table 3, construct the judgment matrix of fault mode 1 evaluation index factor and obtain respective weights coefficient, table 5 is judgment matrix and the weights value of the evaluation index factor of fault mode 1.

The evaluation index constructing matrix of table 5 fault mode 1 and weight

Step 3.3: carry out consistency check to judgment matrix, according to formula (4), formula (5) and table 4, calculates R _c=0.0559<0.1, illustrate that the judgment matrix approach of structure meets the demands, therefore the evaluation index factorial power sets of fault mode 1 is W ¹=[0.7306,0.1884,0.0810].

Step 4: the Rate of aggregative risk factor calculating fault mode 1.

Step 4.1: according to formula (6), the synthetic determination collection of fault mode 1 is:

$\begin{array}{c}{B}^1=W^1{R}^1=\left[\begin{array}{ccc}0.7306& 0.1884& 0.0810\end{array}\right]\&CenterDot;\left[\begin{array}{ccccc}0.0& 0.0& 0.1& 0.2& 0.7\\ 0.0& 0.1& 0.5& 0.3& 0.1\\ 0.8& 0.2& 0.0& 0.0& 0.0\end{array}\right]\\ =\left[\begin{array}{ccccc}0.0648& 0.0350& 0.1673& 0.2026& 0.5303\end{array}\right]\end{array}$

By fuzzy synthesis, step 4.2: according to formula (7), judges that collection and opinion rating (i.e. V={1,2,3,4,5}) are weighted average treatment, draws Rate of aggregative risk factor C ¹=4.0986.

Same method can draw the fuzzy matrix for assessment of fault mode 2 ~ 5:

$R^2={\&lsqb;R_1^2,R_2^2,R_3^2\&rsqb;}^T=\left[\begin{array}{ccccc}0.0& 0.2& 0.6& 0.2& 0.0\\ 0.0& 0.2& 0.3& 0.5& 0.0\\ 0.0& 0.0& 0.3& 0.7& 0.0\end{array}\right]$

$R^3={\&lsqb;R_1^3,R_2^3,R_3^3\&rsqb;}^T=\left[\begin{array}{ccccc}0.6& 0.4& 0.0& 0.0& 0.0\\ 0.0& 0.0& 0.2& 0.7& 0.1\\ 0.0& 0.0& 0.4& 0.6& 0.0\end{array}\right]$

$R^4={\&lsqb;R_1^4,R_2^4,R_3^4\&rsqb;}^T=\left[\begin{array}{ccccc}0.8& 0.2& 0.0& 0.0& 0.0\\ 0.0& 0.2& 0.6& 0.2& 0.0\\ 0.0& 0.0& 0.0& 0.1& 0.9\end{array}\right]$

$R^5={\&lsqb;R_1^5,R_2^5,R_3^5\&rsqb;}^T=\left[\begin{array}{ccccc}0.8& 0.0& 0.0& 0.0& 0.2\\ 0.0& 0.1& 0.5& 0.3& 0.1\\ 0.0& 0.0& 0.0& 0.2& 0.8\end{array}\right]$

For easy calculating, the evaluation index factor of each fault mode adopts same weight sets, namely

W ²＝W ³＝W ⁴＝W ⁵＝W ¹＝[0.7306,0.1884,0.0810]

Thus can obtain each fault mode fuzzy synthesis judge collection as:

B ²＝W ²R ²＝[00.18380.51920.29700]

B ³＝W ³R ³＝[0.43840.29220.07010.18050.0188]

B ⁴＝W ⁴R ⁴＝[0.58450.18380.11300.04580.0729]

B ⁵＝W ⁵R ⁵＝[0.58450.01880.09420.07270.2298]

In like manner, the Rate of aggregative risk factor set of each fault mode is drawn:

C＝{C ¹,C ²,C ³,C ⁴,C ⁵}＝{4.0986,3.1132,2.0491,1.8388,2.3445}

Step 5: the Rate of aggregative risk factor pair each fault mode failure risk ranking results utilizing photovoltaic module failure risk method of discrimination of the present invention to obtain is: fault mode 1> fault mode 2> fault mode 5> fault mode 3> fault mode 4, fault mode 1 assembly surface glass fragmentation faced causes risk ranking maximum, sequence replication problem is also resolved, and this is consistent with actual conditions.

For a person skilled in the art, according to above technical scheme and design, other various corresponding change and distortion can be made, and all these change and distortion all should belong within the protection domain of the claims in the present invention.

Claims (10)

1. a photovoltaic module failure risk method of discrimination, is characterized in that, comprises the following steps:

Determine the fault mode of photovoltaic module;

Expert group, respectively to fault mode relevant evaluation index factor opinion rating, builds the fuzzy matrix for assessment of fault mode according to evaluation result;

The weight sets of relevant evaluation index factor is built according to analytical hierarchy process;

The Rate of aggregative risk factor of each fault mode is calculated according to fuzzy matrix for assessment and weight sets;

According to the failure risk of Rate of aggregative risk factor determination fault mode.

2. photovoltaic module failure risk method of discrimination according to claim 1, is characterized in that: described relevant evaluation index factor comprises output power loss, breakdown maintenance cost and probability of happening.

3. photovoltaic module failure risk method of discrimination according to claim 1, is characterized in that, the detailed process building the fuzzy matrix for assessment of fault mode comprises:

Suppose that expert's total number of persons is S, fault mode is k, and evaluation index factor is i, and to be expert's number of grade m be evaluation evaluation index factor i then the evaluation collection of the evaluation index factor i of fault mode k is:

$R_i^k=\left\{\frac{s_{i1}^k}{s},\frac{s_{i2}^k}{s},...,\frac{s_{im}^k}{s}\right\}=\left\{r_{i1}^k,r_{i2}^k,...,r_{im}^k\right\};$

The evaluation collection of all evaluation index factor i of described fault mode k is built into the fuzzy matrix for assessment R of fault mode k ^k, wherein i=1 ..., n,

$R^k=\&lsqb;R_1^k,R_2^k,...,R_n^k\&rsqb;=\left[\begin{array}{cccc}{r}_{11}^k& r_{12}^k& ...& r_{1m}^k\\ r_{21}^k& r_{22}^k& ...& r_{2m}^k\\ ...& ...& ...& ...\\ r_{n1}^k& r_{n2}^k& ...& r_{nm}^k\end{array}\right].$

4. photovoltaic module failure risk method of discrimination according to claim 1, is characterized in that, the step building the weight sets of relevant evaluation index factor comprises:

The judgment matrix of evaluation index factor is built according to 1-9 scaling law;

Obtain the proper vector corresponding to judgment matrix Maximum characteristic root;

Carry out consistency check to judgment matrix, if test value is less than 0.1, then this proper vector is weight sets, otherwise rebuilds judgment matrix.

5. photovoltaic module failure risk method of discrimination according to claim 4, is characterized in that, formula judgment matrix being carried out to consistency check is: wherein, I _cfor consistency check index, I _rfor Aver-age Random Consistency Index.

6. photovoltaic module failure risk method of discrimination according to claim 5, is characterized in that: wherein, λ _maxfor the Maximum characteristic root of judgment matrix, n is the exponent number of judgment matrix.

7. photovoltaic module failure risk method of discrimination according to claim 1, is characterized in that, the step calculating the Rate of aggregative risk factor comprises:

Fuzzy matrix for assessment is multiplied with weight sets and show that fuzzy synthesis judges collection;

Fuzzy synthesis is judged that collection and opinion rating are weighted average treatment, draws the Rate of aggregative risk factor.

8. the photovoltaic module failure risk method of discrimination according to claim 1 or 7, is characterized in that: the described Rate of aggregative risk factor is larger, represents that failure risk is higher.

9. photovoltaic module failure risk method of discrimination according to claim 1, is characterized in that: described opinion rating is divided into 5 grades.

10. photovoltaic module failure risk method of discrimination according to claim 9, is characterized in that: the principle followed when evaluating output power loss is: lose larger, opinion rating is higher; The principle followed when evaluating breakdown maintenance cost is: expense is higher, and opinion rating is larger; The principle followed when evaluating probability of happening is: probability is larger, and opinion rating is higher.