CN105160166A - Photovoltaic array state determination method - Google Patents

Abstract

The present invention relates to a photovoltaic array state determination method. The method comprises: collecting feature parameters that affect each photovoltaic array state; introducing a time sequence node and a photovoltaic array monitoring point, and establishing a parametric value model from latitudinal and longitudinal dimensions; constructing a photovoltaic array running state assessment model based on a positive and negative ideal point assessment scheme; and establishing a time function relationship between a state assessment result set and each longitudinal and latitudinal parametric value by using a Gaussian curve fitting method, and in combination with an array I-V curve and an irradiance-temperature curve, accurately assessing a current photovoltaic array state from longitudinal and latitudinal angles, analyzing an abnormality cause, and proposing an operation and maintenance measure. By performing state assessment according to a plurality aspects of feature parameters, accuracy of state assessment can be improved, real-time operation and maintenance of a photovoltaic array is facilitated, and the method lays a good foundation for improving power generation efficiency of a photovoltaic array group and searching for a better MPPT algorithm and has high application value.

Description

A kind of photovoltaic array state judging method

Technical field

The present invention relates to a kind of photovoltaic array state judging method, belong to photovoltaic array safe operation technical field.

Background technology

Along with photovoltaic power generation technology is applied more and more widely, existing a large amount of photovoltaic system puts into operation, more to the research of photovoltaic system supervisory system both at home and abroad, and less to the operation maintenance research of photovoltaic plant.Photovoltaic array is by the integral module that multi-path light photovoltaic assembly is connected or connection in series-parallel forms in order to reach certain direct current energy output, photovoltaic array is as the important component part of photovoltaic generating system, and assembly quality and photovoltaic array state directly affect generating efficiency and the serviceable life of whole photovoltaic generating system.Therefore, in unattended solar photovoltaic power plant, in real time effective status monitoring and assessment are carried out to photovoltaic array and seem very important being in for a long time.

Along with the promotion and application of large-scale photovoltaic technology, how carrying out online evaluation to the state of photovoltaic array in photovoltaic system, Timeliness coverage photovoltaic array failure cause, will be a very significant job.But up to the present, overall research is also just laid particular emphasis on to the monitoring of photovoltaic generating system both at home and abroad, namely only pay close attention to the quantitative target such as output power, electric current, voltage of whole photovoltaic plant, and often pay close attention to seldom on external factor on the impact of photovoltaic plant.In real work, variation of ambient temperature, illumination condition changes, the factors such as irregular irregular shade blocks make each photovoltaic module output characteristics in photovoltaic array occur difference, namely each assembly is in different duties, cause whole photovoltaic array output characteristics various, complicated, duty also may correspondingly change, ignore polynary environmental variable, assembly self deterioration and damage, the rule of combination of photovoltaic array and installation quality, the external factor such as the degree of assembly front cleaning can cause the impact of photovoltaic array state and occur deviation to photovoltaic array state estimation.

Summary of the invention

The object of this invention is to provide a kind of photovoltaic array state judging method, in order to solve in prior art assess photovoltaic array time, only pay close attention to the quantitative targets such as the output power of whole photovoltaic plant, electric current, voltage, and the inaccurate problem of the assessment caused is not paid close attention to external factor.

For achieving the above object, the solution of the present invention comprises a kind of photovoltaic array state judging method, comprises the following steps:

(1) characteristic parameter of each photovoltaic array state of impact, is gathered;

(2), introduce time series node, photovoltaic array monitoring point, set up parametric model from horizontal, longitudinal two dimensions, by characteristic parameter matrix representation be:

Wherein: r _{m × n}for dimensional characteristics parameter matrix in length and breadth, represent m × n scheme to be assessed; M: the time series node number of selection; N: the photovoltaic array monitoring point number of selection; I: the i-th time series node; J: a jth photovoltaic array monitoring point; R _ij: the characteristic parameter collection of monitoring point j when timing node i, represents a scheme to be assessed; R _ijk: kth the characteristic parameter of a jth photovoltaic array at i timing node place; K: the number of characteristic parameter;

(3) the photovoltaic array running status assessment models based on positive and negative ideal point evaluation scheme, is constructed;

(4), adopt gaussian curve approximation method, set up the time function relation of condition evaluation results and each parameter in length and breadth, then in conjunction with array I-V curve and irradiance-temperature curve, from two angle estimator photovoltaic array present states in length and breadth.

Described characteristic parameter comprises: irradiance, environment temperature, assembly backboard temperature, component mismatch rate, eclipse factor, laying dust rate, ageing rate, failure rate.

Described step (3) is specially:

1), positive and negative ideal point state estimation matrix is constructed;

2), calculate contribution margin, and construct positive and negative difference matrix according to contribution margin;

3), calculate projection coefficient, and construct positive and negative difference projection matrix according to projection coefficient;

4), the optimum distance degree of each scheme to be estimated is calculated, then the respectively running status assessment of photovoltaic array in scheme to be estimated.

Described step 1) be specially:

First, the state matrix X to be assessed of selected a certain moment, different photovoltaic array _{n × k}, for:

Wherein, n represents photovoltaic array number, k representation feature parameter number;

Selected a certain photovoltaic array, not state matrix Y to be assessed in the same time _{m × k}, for:

Wherein, m is the moment point number of image data on time shaft, k representation feature parameter number;

Then, for matrix X _{n × k}construct positive and negative ideal point state estimation matrix X ⁺, X ^-: described X ⁺in, the first row data are and X _{n × k}corresponding positive desirable evaluation scheme, all the other are described matrix X _{n × k}; Described X ^-in, the first row data are and X _{n × k}corresponding negative desirable evaluation scheme, all the other are described matrix X _{n × k};

For matrix Y _{n × k}construct positive and negative ideal point state estimation matrix Y ⁺, Y ^-: described Y ⁺in, the first row data are and Y _{m × k}corresponding positive desirable evaluation scheme, all the other are described matrix Y _{m × k}; Described Y ^-in, the first row data are and Y _{m × k}corresponding negative desirable evaluation scheme, all the other are described matrix Y _{m × k};

Wherein, described positive desirable evaluation scheme is the evaluation scheme that each characteristic parameter gets optimum value simultaneously, and negative desirable evaluation scheme is the evaluation scheme that each characteristic parameter gets worst-case value simultaneously.

Described step 2) be specially:

A) when carrying out state estimation to synchronization different photovoltaic array:

First, the contribution margin w of k characteristic parameter is calculated _j, computing formula is:

$w_j=\frac{diff(X_{i_1,j}^{+},X_{i_2,j}^{+})}{\underset{j=1}{\overset{k}{\Σ}}diff(X_{i_1,j}^{+},X_{i_2,j}^{+})}$ I ₁≠ i ₂and i ₁, i ₂=2 ..., n+1,

Wherein, i ₁, i ₂represent different schemes to be estimated, $diff(X_{i_1,j}^{+},X_{i_2,j}^{+})=\frac{1}{2}\underset{i_2=2}{\overset{n+1}{\Σ}}\underset{i_1=2}{\overset{n+1}{\Σ}}|X_{i_1,j}^{+}-X_{i_2,j}^{+}|;$

Then, positive variance matrix E is calculated ⁺=X ⁺w, and negative variance matrix E ^-=X ^-w, wherein, the difference vector of normalization characteristic parameter is w=(w ₁, w ₂..., w _k), and positive and negative difference matrix is expressed as follows:

B) when to same photovoltaic array not in the same time under state estimation time:

First, the contribution margin w of k characteristic parameter is calculated _o, computing formula is:

$w_o=\frac{diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})}{\underset{o=1}{\overset{k}{\Σ}}diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})}$ L ₁≠ l ₂and l ₁, l ₂=2 ..., m+1,

Wherein, l ₁, l ₂represent different schemes to be estimated, $diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})=\frac{1}{2}\underset{l_2=2}{\overset{n+1}{\Σ}}\underset{l_1=2}{\overset{n+1}{\Σ}}|Y_{l_1,o}^{+}-Y_{l_2,o}^{+}|;$

Then, positive variance matrix E is calculated ⁺=Y ⁺w, and negative variance matrix E ^-=Y ^-w, wherein, the difference vector of normalization characteristic parameter is w=(w ₁, w ₂..., w _k), and positive and negative difference matrix is expressed as follows:

Described step 3) be specially:

A) described when carrying out state estimation to synchronization different photovoltaic array:

First, the negative projection coefficient that t characteristic parameter of t characteristic parameter of s scheme the to be estimated orthogonal projection coefficient and the individual scheme to be estimated of s that project to t characteristic parameter of positive desirable evaluation scheme projects to t characteristic parameter of negative desirable evaluation scheme is respectively:

$\begin{array}{cc}{\mathrm{\μ}}_{st}^{+}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}}{d_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}},& {\mathrm{\μ}}_{st}^{-}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}{d_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}\end{array},$

Wherein, to make a comment or criticism t characteristic parameter value of desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., n, n+1;

refer to t characteristic parameter value of negative desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., n, n+1;

Then, positive and negative difference projection matrix P is calculated ⁺, P ^-, computing formula is respectively:

B) described when to same photovoltaic array not in the same time under state estimation time:

First, the negative projection coefficient that t characteristic parameter of t characteristic parameter of s scheme the to be estimated orthogonal projection coefficient and the individual scheme to be estimated of s that project to t characteristic parameter of positive desirable evaluation scheme projects to t characteristic parameter of negative desirable evaluation scheme is respectively:

$\begin{array}{cc}{\mathrm{\μ}}_{st}^{+}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}}{d_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}},& {\mathrm{\μ}}_{st}^{-}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}{d_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}\end{array},$

Wherein, to make a comment or criticism t characteristic parameter value of desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., m, m+1;

refer to t characteristic parameter value of negative desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., m, m+1;

Then, positive and negative difference projection matrix Q is calculated ⁺, Q ^-, computing formula is respectively:

Described step 4) be specially:

A) described when carrying out state estimation to synchronization different photovoltaic array:

First, the respectively scheme to be estimated that calculates relative to the Euclidean distance of positive and negative desirable evaluation scheme is:

$\begin{array}{cc}{D}_s^{+}=\sqrt{\underset{j=1}{\overset{k}{\Σ}}{(P_{sj}^{+}-P_{1j}^{+})}^2}& s=2,3,...,n+1,\end{array}$

$\begin{array}{cc}{D}_s^{-}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\Σ}}}{(P_{sj}^{-}-P_{1j}^{-})}^2}& s=2,3,...,n+1;\end{array}$

Then, optimum distance degree D is calculated _s, wherein S is scheme to be assessed, and computing formula is as follows:

$D_s=\frac{{\left(D_s^{+}\right)}^2}{{\left(D_s^{+}\right)}^2+{\left(D_s^{-}\right)}^2},$

Finally, according to optimum distance degree D _scarry out the state estimation of synchronization different photovoltaic array;

B) described when to same photovoltaic array not in the same time under state estimation time:

First, the respectively scheme to be estimated that calculates relative to the Euclidean distance of positive and negative desirable evaluation scheme is:

$\begin{array}{cc}{D}_s^{+}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\Σ}}}{(Q_{sj}^{+}-Q_{1j}^{+})}^2}& s=2,3,...,n+1,\end{array}$

$\begin{array}{cc}{D}_s^{-}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\Σ}}}{(Q_{sj}^{-}-Q_{1j}^{-})}^2}& s=2,3,...,n+1;\end{array}$

Then, optimum distance degree D is calculated _s, wherein S is scheme to be assessed, and computing formula is as follows:

$D_s=\frac{{\left(D_s^{+}\right)}^2}{{\left(D_s^{+}\right)}^2+{\left(D_s^{-}\right)}^2},$

Finally, according to optimum distance degree D _scarry out same photovoltaic array not in the same time under state estimation.

In described step (4), described time function relation comprises irradiance-state estimation value curve, assembly backboard temperature-state estimation value curve, component mismatch rate-state estimation value curve, laying dust rate-state estimation value curve.

In described step (4), describedly to be specially from two angle estimator photovoltaic array present states in length and breadth: at regulation of longitudinal angle: when photovoltaic array is in uniform illumination, if its state estimation value is lower than a setting value, the funtcional relationship that photovoltaic array respectively organizes string I-V curve satisfied is identical, and under time series, powertrace is the curve continuing to reduce, then judge that photovoltaic array laying dust increases; In lateral angles: when the optimum distance degree of photovoltaic array A and B is respectively 0.832,0.576, and photovoltaic array A with B is identical with irradiance curve at the I-V curve of same period, if the assembly backboard temperature of B is higher than the assembly backboard temperature of A, and if time of putting into operation the A of A _tbe greater than time of putting into operation the B of B _t, then there is local hot spot phenomenon in B; If A _t<B _t, then there is component aging situation in B.

In photovoltaic array state judging method provided by the invention, gather the characteristic parameter affecting photovoltaic array state, then a series of matrix disposal is carried out to characteristic parameter, finally according to the comprehensive assessment coefficient of each evaluation scheme of calculation of characteristic parameters, the condition met according to metewand carries out the assessment of state, the method will affect the external factor of photovoltaic array state: such as variation of ambient temperature, illumination condition changes, the factors such as irregular irregular shade blocks also incorporate in the condition of assessment photovoltaic array state, by carrying out state estimation according to many-sided characteristic parameter, the accuracy of state estimation can be improved, be conducive to the real-time O&M of photovoltaic array.Further, also for improving the generating efficiency of photovoltaic array colony and seeking more excellent MPPT algorithm and establish good basis, there is stronger using value.

Accompanying drawing explanation

Fig. 1 is photovoltaic array state judging method process flow diagram.

Embodiment

Below in conjunction with accompanying drawing, the present invention will be further described in detail.

As shown in Figure 1, first photovoltaic array state judging method of the present invention will analyse in depth duty, the load condition of photovoltaic array, and affect the characteristic parameter of its state, then parametric model is set up based on dimension in length and breadth, photovoltaic array is equivalent to photovoltaic module by recycling Analogy, build photovoltaic array colony analogy assessment models, realize the state estimation to the photovoltaic array under Real-Time Monitoring.Make a concrete analysis of as follows:

First, when a comprehensive assessment object, all influence factors are conceived to, therefore in conjunction with the characteristic parameter of the practical operation situation analyzing influence photovoltaic array state of multiple, multiple photovoltaic plant.State estimation is in the past more take state parameter as data bases, and state parameter refers to parameter attribute at a time, is namely instantaneous value, cannot embodies the dynamic perfromance of parameter, also namely cannot accurately judgment device or system running state; And process variable sign changes in time and occurs continuous print physics wave process, embody the dynamic of parameter, continuity, contain the tendency information of parameter, the parameter form of expression is generally the curve of cyclical fluctuations, is namely function type data.Such as, electric current, voltage are the most direct major parameters of reflection photovoltaic array running status, and the process of recording parameters change continuously directly reflects the time of day of photovoltaic array.Known, utilize procedure parameter to detect assessment apparatus or system running state more comprehensive than state parameter more reliable.For photovoltaic array state estimation, its influence factor is more, from transverse dimensions, comprise array configurations mode, environmental factor, assembly oneself factor, shade, dust etc., from longitudinal dimension, comprise the consecutive variations of solar irradiance, environment temperature, battery back-sheet temperature, wind speed etc., and the service time etc. of photovoltaic cell.These factor actings in conjunction, affect the current-power output characteristic curve of photovoltaic array, current-voltage output characteristic curve.When photovoltaic array breaks down or other states change, its major parameter also can change, namely can according to the status flag of the data assessment photovoltaic array of these mutations.In a photovoltaic field, the area coverage of photovoltaic array is comparatively large, in order to simplify the work of photovoltaic array state estimation, adopts analogy Evaluation Method to carry out population evaluation to the photovoltaic array of whole photovoltaic plant.Namely be that photovoltaic array is analogized to a photovoltaic module, choose from photovoltaic array colony there is characteristic feature photovoltaic array as sample, carry out colony's analogy state estimation by the combination of vertical, horizontal parameter.

In concrete enforcement, choose eight characteristic parameters characterizing photovoltaic array state, namely irradiance, environment temperature, assembly backboard temperature, component mismatch rate, eclipse factor (containing cloud block), laying dust rate, ageing rate (i.e. power attenuation rate), failure rate, use P={P ₁, P ₂, P ₃, P ₄, P ₅, P ₆, P ₇, P ₈represent, this characteristic parameter system can more comprehensively for photovoltaic array colony state estimation provides result accurately and reliably.

Wherein, by the collecting devices such as environment monitor can continuous acquisition irradiance, environment temperature, assembly backboard temperature value, in conjunction with geographical environment residing for photovoltaic plant, state estimation sampling time section, and physical significance, as selected a certain large-scale ground photovoltaic plant in Xinjiang to be example, its longitude and latitude known is east longitude 79 °, north latitude 36 °, in conjunction with power station historical data, the span of irradiance, environment temperature, assembly backboard temperature is respectively [600,900], [32,38], [46,52]; For component mismatch rate, be namely inter-module unbalance factor, can the variance yields approximate representation of obtaining current, its span is [0,1]; Eclipse factor, laying dust rate, ageing rate, failure rate directly can not be monitored and be obtained, value can be carried out in conjunction with Practical Project historical experience and expertise, wherein the span of eclipse factor is [0,1], consider laying dust, aging, fault to a certain extent, assembly, by unavailable, arranges the span of these three characteristic parameters for [0,0.8].

Set up parametric model from horizontal, longitudinal two dimensions, use characteristic parameter matrix representation.Lateral angles analogy can have the state of the different photovoltaic array of different characteristic parameter at synchronization, can also assess the influence degree of certain characteristic parameter to photovoltaic array state; Regulation of longitudinal angle can change along the state of time shaft by the same photovoltaic array of analogy, realizes the Real-Time Monitoring to photovoltaic array state and assessment.Introduce time series node, photovoltaic array monitoring point, construction feature parameter matrix is:

Wherein:

In above-mentioned matrix, each parameter declaration is as follows:

R: dimensional characteristics parameter matrix in length and breadth, represents m × n scheme to be assessed;

M: the time series node number of selection;

N: the photovoltaic array monitoring point number of selection;

I: the i-th time series node;

J: a jth photovoltaic array monitoring point;

R _ij: the characteristic parameter collection of monitoring point j when timing node i, represents a scheme to be assessed;

R _ijk: kth the characteristic parameter of a jth photovoltaic array at i timing node place;

K: the number of characteristic parameter.

The selected a certain moment, then the state matrix to be assessed of different photovoltaic array is:

Wherein, n represents the photovoltaic array number participating in contrast, k representation feature parameter number.

Selected a certain photovoltaic array, then the state matrix to be assessed that this photovoltaic array runs along time shaft is:

Wherein, m is the moment point number of image data on time shaft, k representation feature parameter number.

Then, positive and negative desirable evaluation scheme is increased for matrix X, Y respectively, simultaneously, consider that characteristic parameter belongs to profit evaluation model index, or cost type index, or moderate type index, only have " irradiance " to belong to profit evaluation model index in the present invention, other indexs all belong to cost type index.Positive desirable evaluation scheme refers to that each characteristic parameter gets the evaluation scheme of optimum value simultaneously, and negative desirable evaluation scheme refers to that each characteristic parameter gets the evaluation scheme of worst-case value simultaneously.For matrix X, positive desirable proper vector when moment point is 11:00 is: P={900,32,46,0,0,0,0,0}, and negative desired characteristics vector is: P={600,38,52,1,1,0.8,0.8,0.8}.

Then, matrix X, Y are converted into augmented matrix, for matrix X (follow-up all for X, realize to the different photovoltaic array under synchronization carry out running status assessment contrast; If select Y, be then to a photovoltaic array carry out not in the same time under state assessment analysis, identical with the estimation flow of X), carry out standardization processing, be designated as X' _{n × k}, then construct Positive ideal point state estimation matrix, be designated as X ⁺=(X _ij ⁺) _{(n+1) × k}, structure Negative ideal point state estimation matrix, is designated as X ^-=(X _ij ^-) _{(n+1) × k}, wherein, i=1,2 ..., n, n+1; J=1,2 ..., k.In Positive ideal point matrix, the first row data X _1j ⁺=(X ₁₁ ⁺, X ₁₂ ⁺..., X _1k ⁺) be positive desirable evaluation scheme, set it to reference scheme, and other row data X _2j ⁺, X _3j ⁺..., X _{(n+1) j} ⁺be the scheme X to be estimated of participant status evaluation _{n × k}in data.In like manner, for matrix Y _{n × k}construct positive and negative ideal point state estimation matrix Y ⁺, Y ^-: Y ⁺in, the first row data are and Y _{m × k}corresponding positive desirable evaluation scheme, all the other are matrix Y _{m × k}in data; Y ^-in, the first row data are and Y _{m × k}corresponding negative desirable evaluation scheme, all the other are matrix Y _{m × k}in data.

Usually, the assessment difference of each scheme under certain characteristic parameter is larger, and the contribution of this characteristic parameter to assessment result is larger, otherwise, as the same.

Based on this, when carrying out state estimation to synchronization different photovoltaic array,

First, by the contribution margin of k characteristic parameter, w is used _jrepresent, computing formula has:

$w_j=\frac{diff(X_{i_1,j}^{+},X_{i_2,j}^{+})}{\underset{j=1}{\overset{k}{\&Sigma;}}diff(X_{i_1,j}^{+},X_{i_2,j}^{+})}$ I ₁≠ i ₂and i ₁, i ₂=2 ..., n+1

Wherein, i ₁, i ₂represent different schemes to be estimated.For assessment difference value definition have:

$diff(X_{i_1,j}^{+},X_{i_2,j}^{+})=\frac{1}{2}\underset{i_2=2}{\overset{n+1}{\&Sigma;}}\underset{i_1=2}{\overset{n+1}{\&Sigma;}}|X_{i_1,j}^{+}-X_{i_2,j}^{+}|$

The difference vector of normalization characteristic parameter is w=(w ₁, w ₂..., w _k), and then, positive variance matrix E is drawn ⁺=X ⁺w, in like manner, can obtain negative variance matrix E ^-=X ^-w.

Positive and negative difference matrix is expressed as follows:

Then, at normalized ideal matrix X ⁺in, order wherein, to make a comment or criticism t characteristic parameter value of desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., n, n+1; Order refer to t characteristic parameter value of negative desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., n, n+1.

T characteristic parameter of orthogonal projection coefficient and Ke get s scheme to be estimated that t the characteristic parameter then, defining s scheme to be estimated projects to t characteristic parameter of positive desirable evaluation scheme projects to the negative projection coefficient of t characteristic parameter of negative desirable evaluation scheme be respectively:

$\begin{array}{cc}{\mathrm{\&mu;}}_{st}^{+}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}}{d_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}},& {\mathrm{\&mu;}}_{st}^{-}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}{d_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}\end{array}$

Calculate positive and negative difference projection matrix P ⁺, P ^-, computing formula is respectively:

Then, the respectively scheme to be estimated that calculates relative to the Euclidean distance of positive and negative reference scheme is:

$\begin{array}{cc}{D}_s^{+}=\sqrt{\underset{j=1}{\overset{k}{\&Sigma;}}{(P_{sj}^{+}-P_{1j}^{+})}^2}& s=2,3,...,n+1\end{array}$

$\begin{array}{cc}{D}_s^{-}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\&Sigma;}}}{(P_{sj}^{-}-P_{1j}^{-})}^2}& s=2,3,...,n+1\end{array}$

From formula, less, scheme S to be estimated is nearer apart from positive reference scheme, less, the negative reference scheme of scheme S distance to be estimated is nearer.For scheme S to be estimated, it is the smaller the better, be the bigger the better, preferred plan is closest to positive reference scheme, simultaneously farthest away from negative reference scheme, and definition D _sfor the optimum distance degree of scheme S to be estimated and positive reference sequences, be then 1-D with the optimum distance degree of negative reference sequences _s.

$D_s=\frac{{\left(D_s^{+}\right)}^2}{{\left(D_s^{+}\right)}^2+{\left(D_s^{-}\right)}^2},$

By optimum distance degree D _sas the state estimation value of photovoltaic array, each scheme trap queuing can be realized, the characteristic parameter setting each scheme only has a difference, all the other characteristic ginseng values are consistent, then optimum distance degree can judge the influence degree of this characteristic parameter to photovoltaic array state, in like manner, the Degree of interaction of multiple characteristic parameter to photovoltaic array state can be assessed.

In like manner, when to same photovoltaic array not in the same time under state estimation time,

First, the contribution margin w of k characteristic parameter is calculated _o, computing formula is:

$w_o=\frac{diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})}{\underset{o=1}{\overset{k}{\&Sigma;}}diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})}$ L ₁≠ l ₂and l ₁, l ₂=2 ..., m+1,

Wherein, l ₁, l ₂represent different schemes to be estimated, $diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})=\frac{1}{2}\underset{l_2=2}{\overset{n+1}{\&Sigma;}}\underset{l_1=2}{\overset{n+1}{\&Sigma;}}|Y_{l_1,o}^{+}-Y_{l_2,o}^{+}|;$

Then, positive variance matrix E is calculated ⁺=Y ⁺w, and negative variance matrix E ^-=Y ^-w, wherein, the difference vector of normalization characteristic parameter is w=(w ₁, w ₂..., w _k), and

Positive and negative difference matrix is expressed as follows:

Then, the negative projection coefficient that t characteristic parameter of t characteristic parameter of s scheme the to be estimated orthogonal projection coefficient and the individual scheme to be estimated of s that project to t characteristic parameter of positive desirable evaluation scheme projects to t characteristic parameter of negative desirable evaluation scheme is respectively:

$\begin{array}{cc}{\mathrm{\&mu;}}_{st}^{+}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}}{d_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}},& {\mathrm{\&mu;}}_{st}^{-}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}{d_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}\end{array},$

Wherein, to make a comment or criticism t characteristic parameter value of desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., m, m+1;

refer to t characteristic parameter value of negative desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., m, m+1;

Calculate positive and negative difference projection matrix Q ⁺, Q ^-, computing formula is respectively:

Then, the respectively scheme to be estimated that calculates relative to the Euclidean distance of positive and negative desirable evaluation scheme is:

$\begin{array}{cc}{D}_s^{+}=\sqrt{\underset{j=1}{\overset{k}{\&Sigma;}}{(Q_{sj}^{+}-Q_{1j}^{+})}^2}& s=2,3,...,n+1,\end{array}$

$\begin{array}{cc}{D}_s^{-}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\&Sigma;}}}{(Q_{sj}^{-}-Q_{1j}^{-})}^2}& s=2,3,...,n+1;\end{array}$

Calculate optimum distance degree D _s, wherein S is scheme to be assessed, and computing formula is as follows:

$D_s=\frac{{\left(D_s^{+}\right)}^2}{{\left(D_s^{+}\right)}^2+{\left(D_s^{-}\right)}^2},$

In like manner, according to optimum distance degree D _sjudge same photovoltaic array not in the same time under state.

Finally, set up the time function relation of condition evaluation results and each parameter in length and breadth, comprise irradiance-state estimation value curve, assembly backboard temperature-state estimation value curve, component mismatch rate-state estimation value curve, laying dust rate-state estimation value curve etc.By funtcional relationship Time Created, longitudinally can contrast the running status of same an array different time sections, the running status of across comparison same time period different array.

Curve selects suitable curve type to realize the curve of monitoring target, and analyze the relation between two variablees with the curvilinear equation of matching.In enforcement, according to principle of least square method, adopt the Gauss curve fitting method that degree of fitting is higher, its expression formula is:

$G_i\left(x\right)=y_{max}*e^{(-\frac{{(x-x_{\max })}^2}{S^2})}---\left(7\right)$

In formula, solve for parameter y _max, x _max, physical significance representated by S is peak value, peak, the half width information of Gaussian curve respectively.

As shown in table 1, photovoltaic array Status Type comprises normally, shade, laying dust, fault, aging amount to five large classes.Pass through said method, regulation of longitudinal angle, in conjunction with I-V curve (relation curve between the electric current of the generation of photovoltaic array and voltage) and irradiance curve, analyze the running status curve (with optimum distance degree represent) of certain photovoltaic array at one time in section, analyze abnormal cause, O&M decision-making is provided.In enforcement, under photovoltaic array is in uniform illumination, its state estimation value lower (namely lower than a setting value), photovoltaic array respectively organizes string I-V curve tendency consistent (funtcional relationship that namely each group of string I-V curve is satisfied is identical), and powertrace is the tendency slowly continuing to reduce under time series, then can judge that photovoltaic array laying dust increases, provide the O&M suggestion of " this photovoltaic array needs cleaning badly " simultaneously.

Lateral angles, can the running status of analogy same time period different array.In enforcement, for photovoltaic array A and B.When the optimum distance degree of photovoltaic array A and B is respectively 0.832,0.576, and photovoltaic array A with B is substantially identical with irradiance curve at the I-V curve of same period, contrast its characteristic parameter, if the assembly backboard temperature of B, higher than the assembly backboard temperature of A, is then divided into two kinds of situations: if time of putting into operation the A of A _tbe greater than time of putting into operation the B of B _t, then there is local hot spot phenomenon in B, need investigate in time; If A _t<B _t, then judge that B exists component aging situation.

Table 1

The part described in detail is not had to belong to the common practise of prior art or this area in instructions.

Be presented above concrete embodiment, but the present invention is not limited to described embodiment.Basic ideas of the present invention are above-mentioned basic scheme, and for those of ordinary skill in the art, according to instruction of the present invention, designing the model of various distortion, formula, parameter does not need to spend creative work.The change carried out embodiment without departing from the principles and spirit of the present invention, amendment, replacement and modification still fall within the scope of protection of the present invention.

Claims (9)

1. a photovoltaic array state judging method, is characterized in that, comprises the following steps:

(1) characteristic parameter of each photovoltaic array state of impact, is gathered;

(2), introduce time series node, photovoltaic array monitoring point, set up parametric model from horizontal, longitudinal two dimensions, by characteristic parameter matrix representation be:

Wherein: r _{m × n}for dimensional characteristics parameter matrix in length and breadth, represent m × n scheme to be assessed; M: the time series node number of selection; N: the photovoltaic array monitoring point number of selection; I: the i-th time series node; J: a jth photovoltaic array monitoring point; R _ij: the characteristic parameter collection of monitoring point j when timing node i, represents a scheme to be assessed; R _ijk: kth the characteristic parameter of a jth photovoltaic array at i timing node place; K: the number of characteristic parameter;

(3) the photovoltaic array running status assessment models based on positive and negative ideal point evaluation scheme, is constructed;

(4), adopt gaussian curve approximation method, set up the time function relation of condition evaluation results and each parameter in length and breadth, then in conjunction with array I-V curve and irradiance-temperature curve, from two angle estimator photovoltaic array present states in length and breadth.

2. photovoltaic array state judging method according to claim 1, is characterized in that, described characteristic parameter comprises: irradiance, environment temperature, assembly backboard temperature, component mismatch rate, eclipse factor, laying dust rate, ageing rate, failure rate.

3. photovoltaic array state judging method according to claim 1, is characterized in that, described step (3) is specially:

1), positive and negative ideal point state estimation matrix is constructed;

2), calculate contribution margin, and construct positive and negative difference matrix according to contribution margin;

3), calculate projection coefficient, and construct positive and negative difference projection matrix according to projection coefficient;

4), the optimum distance degree of each scheme to be estimated is calculated, then the respectively running status assessment of photovoltaic array in scheme to be estimated.

4. photovoltaic array state judging method according to claim 3, is characterized in that, described step 1) be specially:

First, the state matrix X to be assessed of selected a certain moment, different photovoltaic array _{n × k}, for:

Wherein, n represents photovoltaic array number, k representation feature parameter number;

Selected a certain photovoltaic array, not state matrix Y to be assessed in the same time _{m × k}, for:

Wherein, m is the moment point number of image data on time shaft, k representation feature parameter number;

Then, for matrix X _{n × k}construct positive and negative ideal point state estimation matrix X ⁺, X ^-: described X ⁺in, the first row data are and X _{n × k}corresponding positive desirable evaluation scheme, all the other are described matrix X _{n × k}; Described X ^-in, the first row data are and X _{n × k}corresponding negative desirable evaluation scheme, all the other are described matrix X _{n × k};

For matrix Y _{n × k}construct positive and negative ideal point state estimation matrix Y ⁺, Y ^-: described Y ⁺in, the first row data are and Y _{m × k}corresponding positive desirable evaluation scheme, all the other are described matrix Y _{m × k}; Described Y ^-in, the first row data are and Y _{m × k}corresponding negative desirable evaluation scheme, all the other are described matrix Y _{m × k};

Wherein, described positive desirable evaluation scheme is the evaluation scheme that each characteristic parameter gets optimum value simultaneously, and negative desirable evaluation scheme is the evaluation scheme that each characteristic parameter gets worst-case value simultaneously.

5. photovoltaic array state judging method according to claim 4, is characterized in that, described step 2) be specially:

A) when carrying out state estimation to synchronization different photovoltaic array:

First, the contribution margin w of k characteristic parameter is calculated _j, computing formula is:

$w_j=\frac{diff(X_{i_1,j}^{+},X_{i_2,j}^{+})}{\underset{j=1}{\overset{k}{\&Sigma;}}diff(X_{i_1,j}^{+},X_{i_2,j}^{+})}$ I ₁≠ i ₂and i ₁, i ₂=2 ..., n+1,

Wherein, i ₁, i ₂represent different schemes to be estimated, $diff(X_{i_1,j}^{+},X_{i_2,j}^{+})=\frac{1}{2}\underset{i_2=2}{\overset{n+1}{\&Sigma;}}\underset{i_1=2}{\overset{n+1}{\&Sigma;}}|X_{i_1,j}^{+}-X_{i_2,j}^{+}|;$

Then, positive variance matrix E is calculated ⁺=X ⁺w, and negative variance matrix E ^-=X ^-w, wherein, the difference vector of normalization characteristic parameter is w=(w ₁, w ₂..., w _k), and positive and negative difference matrix is expressed as follows:

B) when to same photovoltaic array not in the same time under state estimation time:

First, the contribution margin w of k characteristic parameter is calculated _o, computing formula is:

$w_o=\frac{diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})}{\underset{o=1}{\overset{k}{\&Sigma;}}diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})}$ L ₁≠ l ₂and l ₁, l ₂=2 ..., m+1,

Wherein, l ₁, l ₂represent different schemes to be estimated, $diff(Y_{l_1,o}^{+},Y_{l_2,o}^{+})=\frac{1}{2}\underset{l_2=2}{\overset{n+1}{\&Sigma;}}\underset{l_1=2}{\overset{n+1}{\&Sigma;}}|Y_{l_1,o}^{+}-Y_{l_2,o}^{+}|;$

Then, positive variance matrix E is calculated ⁺=Y ⁺w, and negative variance matrix E ^-=Y ^-w, wherein, the difference vector of normalization characteristic parameter is w=(w ₁, w ₂..., w _k), and positive and negative difference matrix is expressed as follows:

6. photovoltaic array state judging method according to claim 5, is characterized in that, described step 3) be specially:

A) described when carrying out state estimation to synchronization different photovoltaic array:

First, the negative projection coefficient that t characteristic parameter of t characteristic parameter of s scheme the to be estimated orthogonal projection coefficient and the individual scheme to be estimated of s that project to t characteristic parameter of positive desirable evaluation scheme projects to t characteristic parameter of negative desirable evaluation scheme is respectively:

${\mathrm{\&mu;}}_{st}^{+}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}}{d_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}},$ ${\mathrm{\&mu;}}_{st}^{-}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}{d_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}},$

Wherein, to make a comment or criticism t characteristic parameter value of desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., n, n+1;

refer to t characteristic parameter value of negative desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., n, n+1;

Then, positive and negative difference projection matrix P is calculated ⁺, P ^-, computing formula is respectively:

B) described when to same photovoltaic array not in the same time under state estimation time:

First, the negative projection coefficient that t characteristic parameter of t characteristic parameter of s scheme the to be estimated orthogonal projection coefficient and the individual scheme to be estimated of s that project to t characteristic parameter of positive desirable evaluation scheme projects to t characteristic parameter of negative desirable evaluation scheme is respectively:

${\mathrm{\&mu;}}_{st}^{+}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}}{d_{st}^{+}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{+}},$ ${\mathrm{\&mu;}}_{st}^{-}=\frac{\underset{s}{\min }\underset{t}{\min }{d}_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}}{d_{st}^{-}+\frac{1}{2}\underset{s}{\max }\underset{t}{\max }{d}_{st}^{-}},$

Wherein, to make a comment or criticism t characteristic parameter value of desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., m, m+1;

refer to t characteristic parameter value of negative desirable evaluation scheme, wherein t=1,2 ..., k; refer to t characteristic parameter value of s scheme to be estimated, wherein s=2,3 ..., m, m+1;

Then, positive and negative difference projection matrix Q is calculated ⁺, Q ^-, computing formula is respectively:

7. photovoltaic array state judging method according to claim 6, is characterized in that, described step 4) be specially:

A) described when carrying out state estimation to synchronization different photovoltaic array:

First, the respectively scheme to be estimated that calculates relative to the Euclidean distance of positive and negative desirable evaluation scheme is:

$D_s^{+}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\&Sigma;}}}{(P_{sj}^{+}-P_{1j}^{+})}^2}$ s＝2,3,…,n+1，

$D_s^{-}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\&Sigma;}}}{(P_{sj}^{-}-P_{1j}^{-})}^2}$ s＝2,3,…,n+1；

Then, optimum distance degree D is calculated _s, wherein S is scheme to be assessed, and computing formula is as follows:

$D_s=\frac{{\left(D_s^{+}\right)}^2}{{\left(D_s^{+}\right)}^2+{\left(D_s^{-}\right)}^2},$

Finally, according to optimum distance degree D _scarry out the state estimation of synchronization different photovoltaic array;

B) described when to same photovoltaic array not in the same time under state estimation time:

First, the respectively scheme to be estimated that calculates relative to the Euclidean distance of positive and negative desirable evaluation scheme is:

$D_s^{+}=\sqrt{\underset{j=1}{\overset{k}{\&Sigma;}}{(Q_{sj}^{+}-Q_{1j}^{+})}^2}$ s＝2,3,…,n+1，

$D_s^{-}=\sqrt{\underset{j=1}{\overset{k}{\mathrm{\&Sigma;}}}{(Q_{sj}^{-}-Q_{1j}^{-})}^2}$ s＝2,3,…,n+1；

Then, optimum distance degree D is calculated _s, wherein S is scheme to be assessed, and computing formula is as follows:

$D_s=\frac{{\left(D_s^{+}\right)}^2}{{\left(D_s^{+}\right)}^2+{\left(D_s^{-}\right)}^2},$

Finally, according to optimum distance degree D _scarry out same photovoltaic array not in the same time under state estimation.

8. photovoltaic array state judging method according to claim 1, it is characterized in that, in described step (4), described time function relation comprises irradiance-state estimation value curve, assembly backboard temperature-state estimation value curve, component mismatch rate-state estimation value curve, laying dust rate-state estimation value curve.

9. photovoltaic array state judging method according to claim 7, it is characterized in that, in described step (4), describedly to be specially from two angle estimator photovoltaic array present states in length and breadth: at regulation of longitudinal angle: when photovoltaic array is in uniform illumination, if its state estimation value is lower than a setting value, the funtcional relationship that photovoltaic array respectively organizes string I-V curve satisfied is identical, and under time series, powertrace is the curve continuing to reduce, then judge that photovoltaic array laying dust increases; In lateral angles: when the optimum distance degree of photovoltaic array A and B is respectively 0.832,0.576, and photovoltaic array A with B is identical with irradiance curve at the I-V curve of same period, if the assembly backboard temperature of B is higher than the assembly backboard temperature of A, and if time of putting into operation the A of A _tbe greater than time of putting into operation the B of B _t, then there is local hot spot phenomenon in B; If A _t<B _t, then there is component aging situation in B.