CN104778345A - Nonlinear parameter calculation method for simulating aging failure of photovoltaic cell models - Google Patents

Abstract

The invention relates to a nonlinear parameter calculation method for simulating the aging failure of photovoltaic cell models, which includes the following steps: (1) a translucent film blocking method is adopted to decrease irradiance received by a photovoltaic panel in order to simulate the aging failure of the photovoltaic panel, and the relation between the number of translucent film blocking layers and irradiance attenuation is measured; (2) a relation curve between blocking film layer numbers (C) and aging degrees (L) is created; (3) the difference of blocking degrees is utilized to simulate the different aging degrees, a load in a photovoltaic panel circuit is regulated and the voltage and current values of the circuit are acquired, and thereby an I-V output curve of the photovoltaic panel under different irradiances and the different aging degrees is obtained; an external characteristic I-V relation of the photovoltaic panel is established; (4) for the I-V output curve of the photovoltaic panel under the different irradiances and the different aging degrees, optical parameter values are obtained by parameter identification. The nonlinear parameter calculation method can obtain the nonlinear parameter change laws of the aging failure of the photovoltaic cell models.

Description

A kind of nonlinear parameter computing method simulating photovoltaic battery model degradation failure

Art

The invention belongs to photovoltaic cell technical field, relate to a kind of photovoltaic cell parameter acquiring method.

Background technology

Owing to facing the pressure of energy crisis and environmental protection, photovoltaic solar is subject to various countries and payes attention to its clean environment firendly, become the trend of future source of energy development.The fault of photovoltaic cell is divided into aging, sliver, hot spot etc.The research of hot spot problem has now developed into the aspect of localization of fault, and the domestic and international correlative study of the problem of aging of photovoltaic cell is less.

Summary of the invention

The object of this invention is to provide a kind of method asking for the model aging fault parameter Changing Pattern of photovoltaic cell.Technical scheme of the present invention is as follows:

Simulate nonlinear parameter computing method for photovoltaic battery model degradation failure, comprise step below:

(1) adopt the mode that semitransparent thin film blocks to reduce the irradiance of photovoltaic battery panel reception to simulate the degradation failure of photovoltaic battery panel, the irradiance blocked during simulation degradation failure is more, represents that degree of aging is darker; Under identical testing conditions, adopt the semitransparent thin film of the different number of plies to block the light receiving surface of photovoltaic battery panel, record the number of plies that semitransparent thin film blocks and the relation that irradiance is decayed;

(2) define degree of aging (L) to be expressed as follows:

Irradiance value (1000W/m under degree of aging (L)=irradiance pad value ÷ standard conditions ²) × 100%

Degree of aging (L) is used for representing the attenuation degree of transmittance, degree of aging more high transmission rate is lower, irradiance attenuation degree is higher, the relation that the number of plies of blocking according to the semitransparent thin film recorded and irradiance are decayed, sets up the relation curve of the film number of plies (C) and the degree of aging (L) blocked;

(3) utilize the difference of coverage extent to simulate different degree of agings, adjust the load in photovoltaic battery panel loop and acquisition circuit voltage and current value, the I-V curve of output of photovoltaic battery panel under different irradiance, under different degree of aging of acquisition;

(4) according to single diode model and the equivalent electrical circuit thereof of photovoltaic battery panel, the external characteristic I-V relational expression of photovoltaic battery panel is drawn:

$I=I_{\mathrm{ph}}-I_0\left\{e\mathrm{xp}\right[\frac{q(V+{\mathrm{IR}}_s)}{\mathrm{AkT}}]-1\}-\frac{V+{\mathrm{IR}}_s}{R_{\mathrm{sh}}}$

Wherein,

V---photovoltaic battery panel two ends measuring voltage

I---electric current in photovoltaic battery panel external circuit

A---diode quality factor

The backboard temperature of T---photovoltaic battery panel

K---Boltzmann constant (1.380 × 10-23J/K)

R _s---cell series resistance

Q---electron charge (1.608 × 10-19C)

I _o---diode reverse saturation current

I _ph---photovoltaic battery panel photogenerated current

R _sh---cell parallel resistance

Quote Lambert W function and simplify I-V equation, obtain the explicit expression of photovoltaic battery panel electric current I:

$I=\frac{R_{\mathrm{sh}}(I_{\mathrm{ph}}+I_o)-V}{R_{\mathrm{sh}}+R_s}-\frac{{\mathrm{AV}}_{\mathrm{th}}}{R_s}\×W\left(\frac{R_s{R}_{\mathrm{sh}}{I}_o}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\exp \left(\frac{R_{\mathrm{sh}}{R}_s(I_{\mathrm{ph}}+I_o)+R_{\mathrm{sh}}V}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\right)\right)---\left(2\right)$

Wherein, V _th=KT/q, W () are LambertW function, by photogenerated current I _ph, reverse saturation current I ₀, parallel resistance R _sh, resistance in series R _s, diode quality factor A five parameters just can determine I-V curve;

(4) for the I-V curve of output of photovoltaic battery panel under different irradiance, different degree of aging, if five dimension particle x _k=(I _ph, I ₀, A, R _s, R _sh), through parameter identification, obtain the optimal value of these 5 parameters.

Preferably, step (4) application self-adapting Chaos-Particle Swarm Optimization parameter identification method carries out parameter identification.

The present invention adopts polyethylene film to reduce the method for irradiance to simulate the ageing state of photovoltaic cell, the relation that film blocks between the number of plies and degree of aging is quantized, parameter identification is carried out by the aging experimental data of simulation, obtain the expression formula of nonlinear parameter in photovoltaic cell model thus, thus the nonlinear parameter Changing Pattern of the model aging fault of photovoltaic cell can be obtained.

Accompanying drawing explanation

Fig. 1 is implementing procedure figure

Fig. 2 is measurement mechanism structural drawing

Fig. 3 film blocks the relation curve of the number of plies and degree of aging

Fig. 4 tests main circuit

Fig. 5 photovoltaic cell model equivalent electrical circuit

The practical application process flow diagram of Fig. 6 photovoltaic cell identification of Model Parameters algorithm

Fig. 7 photogenerated current (Iph) is with the change curve of degree of aging

Fig. 8 diode reverse saturation current (Ios) is with the change curve of degree of aging

Fig. 9 diode quality factor (A) is with the change curve of degree of aging

Figure 10 resistance in series is with the change curve of degree of aging

Embodiment

The encapsulating structure of photovoltaic cell is: ultra-clear glasses---ethylene-vinyl acetate copolymer (EVA)---cell piece---ethylene-vinyl acetate copolymer (EVA)---backboard.Due to photovoltaic cell stable in properties and backboard mainly play a supportive role in photovoltaic module, when analyzing the aging outside of photovoltaic cell and characterizing, the impact (digestion period of semiconductor material relative longer) that cause aging with EVA of main consideration ultra-clear glasses.

Under ultra-clear glasses is exposed to external environment, the extraneous factors such as light oxygen, temperature, humidity all can cause it aging, and what it was aging is characterized primarily by glass board material surface jaundice, embrittlement, physical strength decline, and transmittance declines.The main cause that EVA is aging is the only a few oxygen of photovoltaic module finished product internal residual, EVA and oxygen generation chemical reaction, and the speed of reaction increases with ultraviolet enhancing and temperature and accelerates.The outside that EVA is aging characterizes the change being mainly EVA color, and then affects the change of photovoltaic cell transmittance.

The principle of photovoltaic cell is " photoproduction electricity ", the size of generated energy depends on that photovoltaic cell receives the power of terrestrial solar radiation degree, when photovoltaic cell is aging, the feature reflected is the reduction that light transfers power to, also can wait the reduction (such as cloudy day) being all terrestrial solar radiation degree, first of the present invention is the catabiosis being simulated photovoltaic cell by the transmittance on reduction photovoltaic cell surface accordingly.Utilize blocking of the different number of plies of semitransparent thin film, can the transmittance on reduction photovoltaic cell surface in various degree, simulate state during the different degree of aging of photovoltaic cell.It is the uniform semitransparent thin film of transmittance that the present invention chooses density polyethylene film with high.Density polyethylene film with high has good thermotolerance and physical strength, and tensile elongation is little, and transmittance is slightly low, and tear resistance is strong.The transmittance that density polyethylene film with high is slightly low meets the mentality of designing of simulation photovoltaic cell degradation, and tensile elongation is little, tear resistance forces it to be in stable level to blocking of irradiance.

After photovoltaic cell is aging, the electric energy power of output can decline.By analyzing the mechanism of photovoltaic cell, can by the degradation failure reducing irradiance that photovoltaic battery panel receives and simulate photovoltaic cell, the method that the present invention adopts PVC film to block, the irradiance blocked during simulation degradation failure bright degree of aging of more speaking more is darker.

First composition graphs 1 illustrates the method implementation process of being blocked simulation photovoltaic cell degradation by polyethylene film below:

The number of plies of film and the foundation of photovoltaic cell degree of aging relational expression is blocked in the present invention, identical with the factory calibration test condition (irradiance is 1000w/m^2, and temperature is 25 DEG C) of photovoltaic cell.Under standard conditions, often increase the number of plies of blocking film and often survey an irradiance value, record the relation that the number of plies of semitransparent thin film material blocks and irradiance decay as shown in table 1.

Table 1 blocks the film number of plies and irradiance attenuation relation

Due to the change of the change reflection degree of aging of transmittance, therefore, we define degree of aging (L) and are expressed as follows:

Irradiance value (1000W/m under degree of aging (L)=irradiance pad value ÷ standard conditions ²) × 100% degree of aging (L) is used for representing the attenuation degree of transmittance, and degree of aging is high transmission rate lower (irradiance attenuation degree is higher) more.When the different film numbers of plies in table 1-1 being blocked, corresponding irradiance pad value is brought in above-mentioned formula, the relation curve of the film number of plies (C) that can block and degree of aging (L), as shown in Figure 3.

Under different degree of aging (difference of coverage extent), adjust the load in photovoltaic cell loop and acquisition circuit voltage and current value, the I-V curve of output under the photovoltaic panel different condition of acquisition.

Concrete steps are as follows:

First get out PVC film, measure different film block under irradiance reduction level, and record.

Start to measure the photovoltaic cell output voltage current relationship curve under different simulation degree of aging below.Measure the irradiance of sunlight, at irradiance 995W/m ², under temperature 25 DEG C of conditions, first all 8 tunics are superimposed together, cover on photovoltaic solar cell plate, block, then, experimental facilities is connected as shown in Fig. 4 or Fig. 2, two voltages are surveyed in data acquisition board collection continuously: Rp and r both end voltage, r both end voltage.Namely data collecting card two pairs of measurement points connect Rp and r two ends, r two ends respectively, notice that place is battery cathode altogether.

Slide rheostat resistance is transferred to minimum value 0 Europe, (data acquisition board connects computer to open data measurin system, the data acquisition system (DAS) interface of opening computer starts to gather) start measurement data, the resistance of slide rheostat is increased to maximal value with suitable speed from minimum value, then disconnects slide rheostat; Then remove one deck Protective film fast, slide rheostat resistance is transferred to minimum simultaneously, then connects slide rheostat, carry out the operation same with when 8 tunic, regulate slide rheostat to maximal value, then disconnect slide rheostat; Continue the step repeated above afterwards, reduce film, slide rheostat resistance is transferred to minimum, then connects slide rheostat, rheostat value is transferred to maximal value with suitable speed, then disconnecting circuit, until do not have film to block.So just obtain the simulation senile experiment result under this irradiance.Then temperature 25 DEG C, irradiance is respectively 750W/m next, ², 400W/m ²and 240W/m ²the step that lower repetition is same, completes experiment and just can obtain the simulation senile experiment data under different irradiance.

The R obtained like this _pwith the output voltage that r both end voltage is exactly photovoltaic cell, r both end voltage is exactly the output current of photovoltaic cell divided by the resistance of r.

Due to the point of some apparent errors of generation such as noise, can may filter out by application data average value filtering algorithm in experimental data.Because each curve recorded all is made up of a lot of point, and to carry out to so many data point the calculated amount that parameter asks for be excessive, data extraction must be carried out, choose some to curve shape representational unique point morphogenesis characters curve, be foundation with voltage in I-V curve, when voltage is less than 3.5V, do on average again to suing for peace a little within the scope of 0.1V, obtain an average point, when voltage is greater than 3.5V, because most curve enters the larger decline district of slope from here on, after 3.5V, its average is obtained again to the some summation within the scope of 0.05V, these average point points taken out are linked to be line, exactly to measuring the representational curve of I-V obtained.Remain 35 groups of electric current and voltage values and in like manner carry out data filtering according to the method, data decimation, obtain I-V curve.

Following application self-adapting Chaos-Particle Swarm Optimization parameter identification method carries out parameter identification.The most widely used single diode model of photovoltaic solar cell is used, five parameters comprised: photogenerated current I in this patent _ph, reverse saturation current I ₀, parallel resistance R _sh, resistance in series R _s, diode quality factor A.Physical model as shown in Figure 5.

According to photovoltaic cell model equivalent electrical circuit, the external characteristic I-V relational expression of photovoltaic cell can be drawn:

$I=I_{\mathrm{ph}}-I_0\left\{e\mathrm{xp}\right[\frac{q(V+{\mathrm{IR}}_s)}{\mathrm{AkT}}]-1\}-\frac{V+{\mathrm{IR}}_s}{R_{\mathrm{sh}}}---\left(1\right)$

Wherein,

V---photovoltaic module two ends measuring voltage

I---electric current in component external loop

A---diode quality factor

The backboard temperature of T---battery

K---Boltzmann constant (1.380 × 10-23J/K)

R _s---cell series resistance

Q---electron charge (1.608 × 10-19C)

I _o---diode reverse saturation current

I _ph---photovoltaic cell photogenerated current

R _sh---cell parallel resistance

Electric current I in equation cannot obtain explicit expression formula by elementary function process, quotes Lambert W function for this reason and simplifies I-V equation, obtain the explicit expression of photovoltaic cell electric current I:

$I=\frac{R_{\mathrm{sh}}(I_{\mathrm{ph}}+I_o)-V}{R_{\mathrm{sh}}+R_s}-\frac{{\mathrm{AV}}_{\mathrm{th}}}{R_s}\×W\left(\frac{R_s{R}_{\mathrm{sh}}{I}_o}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\exp \left(\frac{R_{\mathrm{sh}}{R}_s(I_{\mathrm{ph}}+I_o)+R_{\mathrm{sh}}V}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\right)\right)---\left(2\right)$

Wherein, V _th=KT/q, $W\left(\frac{R_s{R}_{\mathrm{sh}}{I}_0}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\exp \left(\frac{R_{\mathrm{sh}}{R}_s(I_{\mathrm{ph}}+I_0)+R_{\mathrm{sh}}V}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\right)\right)$ For LambertW function, not containing unknown quantity in bracket, because algorithm is realized by matlab, directly can call lambertw () statement and just can solve its value in matlab.

Therefore, by photogenerated current I _ph, reverse saturation current I ₀, parallel resistance R _sh, resistance in series R _s, diode quality factor A five parameters just can determine I-V curve.

Next just can application group optimized algorithm (self-adaptation Chaos particle swarm optimization algorithm), carry out parameter estimation, to any curve, if five tie up particle x _k=(I _ph, I ₀, A, R _s, R _sh), the globally optimal solution obtained after utilizing the optimization of self-adaptation Chaos particle swarm optimization algorithm is the optimal value of 5 parameters of wanted identification.Self-adaptation Chaos particle swarm optimization algorithm is realized by computing machine matlab, and concrete steps are as follows:

Step1: set the number of population as n=100, iterations is set to 150 times.The span of each dimensional parameter of particle is set.

Logistic mapping equation (3) is utilized to carry out chaos intialization to the population that group's number is n:

Z _i+1＝μZ _i(1-Z _i),i＝0,1,2,…,μ∈(0,4] (3)

Wherein, 0≤Z ₀≤ 1, Z _ibe i-th variable.μ is controling parameter, and during elected μ=4, system is in Complete Chaos state, and its chaotic space is [0,1].Within the scope of the traversal of restriction, random starting values is arranged to the position of particle and speed.

Step2: the optimal location p of position as himself getting each particle _best, particle position optimum in population is assigned to colony optimal value g _best.

Step3: the position current according to particle and speed by the Position And Velocity upgrading formula (4) below, (5), (6) upgrade oneself, and utilize carrier format to be limited to by Position And Velocity to allow in span.

v _id＝w×v _id+c ₁×rand()×(p _id-x _id)+c ₂×rand()×(p _gd-x _id) (4)

x _id＝a×x _id+v _id(5)

Wherein, the value of c1, c2 is generally the random number between 2, rand () expression [0,1], and w is the coefficient keeping original speed, is called inertia weight, and α is constraint factor.Particle upgrades the traversal scope of location variable and speed variables for being respectively [X _min, X _max], [V _min, V _max].At no point in the update process, if get boundary value when the speed of particle renewal and position exceed traversal scope.

Make use of the strategy of Automatic adjusument inertia weight w, upgrade the weighted value in iterative process, formula is as follows:

$w=(w_{\max }-w_{\min })\×\exp (-{(\mathrm{\τ}\×\frac{\mathrm{cur}}{\mathrm{Loop}})}^2)+w_{\min }---\left(6\right)$

Wherein, w _maxwith w _minit is the maxima and minima of inertia weight w in formula (4).Cur is current iteration number of times, and Loop is maximum iteration time.τ is experience factor, and value is between [20,55].Owing to containing negative exponent part in formula, at the iteration initial stage of algorithm, cur value is less, and inertia weight w value is comparatively large, and the speed of particle and position upgrade within the scope of whole traversal; And in the iteration later stage, cur value is comparatively large, inertia weight w value is less, and the speed of particle and position upgrade among a small circle.Thus this regulation strategy enhances the harmony of Chaos particle swarm optimization algorithm between global search and layout search.

Step4: the ordinate value (I calculating the respective point identical with each point horizontal ordinate in curve _cal), and utilize the ordinate value I of respective point in formula (7) and virgin curve _meacalculate the fitness value of each particle respectively.For each particle, fitness value as current in particle is better than the optimum extreme value of particle individuality, then replace individual optimal value p with the current fitness value of particle _best.Find out the optimal value of population according to the individual optimal value of particle each in population, optimal value as current in population is better than history optimal value, then replace population optimal value g by current optimal value _best.

$f\left(X\right)=\sqrt{\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{(I_{\mathrm{cal}}-I_{\mathrm{mea}})}^2}---\left(7\right)$

Wherein, X=(Iph, Io, A, Rs, Rsh), namely the position vector of each particle represents the parameter value of five battery models.I _caland I _meathe parameter being respectively algorithm identification brings the current value and actual current measurement value that obtain in formula (7) into.Fitness value less expression identified parameters is more accurate.Be brought in colony optimization algorithm by the data of any I-V curve, setting cycle index, is finally obtained very close to the model parameter of actual value by the continuous optimizing of algorithm.

Step5: evaluation algorithm reaches maximum iteration time or meets the condition of convergence (fitness reaches setting value), directly forwards Step7 to if meet, otherwise performs next step.

Step6: utilize formula (8) evaluation algorithm whether Premature Convergence.If so, upgrade population position carry out Chaos Search according to formula (5), formula (9) again, then turn to Step3.Otherwise, perform next step.

${\mathrm{\σ}}^2=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\left(\frac{f_i-f_{\mathrm{avg}}}{f}\right)}^2,f=\max [1,\max |f_i-f_{\mathrm{avg}}\left|\right]---\left(8\right)$

In formula, n is the total number of particles of population.F _ibe the fitness value of i-th particle, f _avgfor the average fitness value that population is current.Colony fitness variance σ ²the convergence state of reflection population, if σ ²less, illustrate that population is more tending towards convergence.The present invention is σ ²set a threshold value, work as σ ²when being less than this threshold value, be then judged to be that algorithm is absorbed in precocity.In order to avoid global optimum is mistaken for Premature Convergence, also need setting optimal-adaptive degree threshold value.

When algorithm occurs precocious, the location variable of particle carries out Chaos Search by following formula, and upgrades:

x _t+1＝X _min+Z _i+1×(X _max-X _min),t＝0,1,2… (9)

Wherein, x _i+1for the location variable of particle, Z _i+1for chaotic particle variable, determined by formula (9).

Step7: globally optimal solution X=(Iph, Io, A, Rs, Rsh) is exported, obtains the optimum solution of model 5 parameters.Algorithm terminates.

The process flow diagram of algorithm is as shown in Figure 3:

In like manner, residue 35 curve datas are brought in algorithm the parameter identification result obtained under different irradiance under different simulation degree of aging.

The present invention is on the basis of existing photovoltaic cell universal model (single diode model), the method of successively blocking in conjunction with PVC film material is to simulate the measurement data of photovoltaic panel ageing process, and application group optimized algorithm obtains photovoltaic cell photovoltaic cell model 5 parameters (photovoltaic cell photogenerated current, diode reverse saturation current, diode quality factor, resistance in series, parallel resistance) of (difference of coverage extent) under different irradiance, temperature, different degree of aging.Graphically parameter is expressed as (at a certain backboard temperature) shown in Fig. 4, Fig. 5, Fig. 6, Fig. 7 with the Changing Pattern of degree of aging, irradiance below.

According to photogenerated current (I in Fig. 4 and table 1 _ph) with the variation relation of degree of aging, known I _phapproximate linearly with L, and be negative correlation.The impact of degree of aging is considered, definition photogenerated current (I in formula _ph) as follows with the relational expression between irradiance (S), battery back-sheet temperature (T), degree of aging (L):

I _ph＝a×I _ph,ref×(1-b×L)×S/S _ref

In formula,

a×I _ph,ref＝c

C is a determined value, brings data in table 1 into, carries out linear fit with related algorithm, obtains the relational expression (under determining temperature) of photovoltaic cell:

I _ph＝5.0226×(1-1.122×L)×S/1000

Backboard temperature in experiment is 41 degree, and its goodness of fit index is as follows:

Goodness of fit:

SSE:1.349

R-square:0.9791

Adjusted R-square:0.9778

RMSE:0.2053

Can find out that fitting result and the raw data goodness of fit of formula are very high, fitness is fine, has carried out extraordinary parsing to raw data.

According to diode reverse saturation current (I in Fig. 5 and table 1 _o) with the variation relation of degree of aging, known diode reverse saturation current and the approximate exponentially relation of degree of aging, and affect by irradiance.Definition diode reverse saturation current (I _o) as follows with the relational expression between irradiance (S), battery temperature (T), degree of aging (L):

$I_o=a\×I_{o,\mathrm{ref}}\×(b+\frac{S}{S_{\mathrm{ref}}})\×e^{c\×(1-L)}$ (T is when specified temp)

Make in formula:

d×I _o,ref＝h

H is a determined value, then data in table 1 is brought in above formula and carry out data fitting, can obtain relational expression:

$I_o=2.775\×{10}^{-5}\×(-0.1342+\frac{S}{1000})\×e^{10.63\×(1-L)}$

In formula, I _ounit be mA.The goodness of fit index of fitting result is as follows:

Goodness of fit:

SSE:0.08914

R-square:0.9423

Adjusted R-square:0.9387

RMSE:0.05278

When photovoltaic battery temperature is certain particular value, diode quality factor (A) with irradiance (S) and degree of aging (L) change curve as shown in Figure 6, can roughly find out from curve, diode quality factor and degree of aging relation object liny, and present negative correlation, become positive correlation with irradiance.

$A=a\×A_{\mathrm{ref}}\×(1+b\×(1-L))\×(1+c\×1n\left(\frac{S}{S_{\mathrm{ref}}}\right))$

Data according to table 1 carry out matching, obtain relational expression:

$A=10.07\×(1+0.8857\×(1-L))\×(1+0.1879\×1n\left(\frac{S}{1000}\right))$

The goodness of fit index of fitting result is as follows:

Goodness of fit:

SSE:7.287

R-square:0.9703

Adjusted R-square:0.9664

RMSE:0.4929

As shown in Figure 7 for photovoltaic battery temperature be particular value time, resistance in series (R _s) with the change curve of irradiance and degree of aging, as seen from the figure, resistance in series is approximate with degree of aging exponentially relation, is subject to the impact of irradiance level.Therefore, estimation formulas form is as follows:

$R_s=a\×R_{s,\mathrm{ref}}\×(1+b\×\frac{S_{\mathrm{ref}}}{S})\×(1+\exp (c\×L+d))$

Experimental data is carried out matching, obtains Relationship of Coefficients as follows:

$R_s=0.02878\×(1+2.042\×\frac{1000}{S})\×(1+\exp (3.846\×L-1.475))$

The goodness of fit level of fitting result is as follows:

Goodness of fit:

SSE:0.2367

R-square:0.9534

Adjusted R-square:0.9489

RMSE:0.08739

Although fitting formula is above very high with the goodness of fit of experimental data, result relative complex, in order to formula of reduction, make it be more suitable for being applied in engineering, the fitting result of application relative ease is as follows:

$R_s=a\×R_{s,\mathrm{ref}}\×\left(\frac{S_{\mathrm{ref}}}{S}\right)\×(1+\exp (b\×L+d))$

Bring data into, obtain fitting result:

$R_s=0.06371\×\left(\frac{S_{\mathrm{ref}}}{S}\right)\×(1+\exp (3.697\×L-1.25))$

Its goodness of fit level is as follows:

Goodness of fit:

SSE:0.3072

R-square:0.9396

Adjusted R-square:0.9358

RMSE:0.09797

Experimental result also finds, if continued to be reduced to following formula by above-mentioned formula:

$R_s=a\×R_{s,\mathrm{ref}}\×\left(\frac{S_{\mathrm{ref}}}{S}\right)\×\exp (b\×L)$

Bring data into, fitting result is as follows:

$R_s=0.05531\×\left(\frac{1000}{S}\right)\×\exp (2.523\×L)$

But its goodness of fit still has following level:

Goodness of fit:

SSE:0.3684

R-square:0.9275

Adjusted R-square:0.9253

RMSE:0.1057

Above formula fitting result, both ensure that the error precision of fitting result and raw data, considered again the achievement in research of Practical Project physical significance and forefathers, had very high engineering practical value.

Claims (2)

1. simulate nonlinear parameter computing method for photovoltaic battery model degradation failure, comprise step below:

(1) adopt the mode that semitransparent thin film blocks to reduce the irradiance of photovoltaic battery panel reception to simulate the degradation failure of photovoltaic battery panel, the irradiance blocked during simulation degradation failure is more, represents that degree of aging is darker; Under identical testing conditions, adopt the semitransparent thin film of the different number of plies to block the light receiving surface of photovoltaic battery panel, record the number of plies that semitransparent thin film blocks and the relation that irradiance is decayed;

(2) define degree of aging (L) to be expressed as follows:

Irradiance value (1000W/m under degree of aging (L)=irradiance pad value ÷ standard conditions ²) × 100%

Degree of aging (L) is used for representing the attenuation degree of transmittance, degree of aging more high transmission rate is lower, irradiance attenuation degree is higher, the relation that the number of plies of blocking according to the semitransparent thin film recorded and irradiance are decayed, sets up the relation curve of the film number of plies (C) and the degree of aging (L) blocked;

(3) utilize the difference of coverage extent to simulate different degree of agings, adjust the load in photovoltaic battery panel loop and acquisition circuit voltage and current value, the I-V curve of output of photovoltaic battery panel under different irradiance, under different degree of aging of acquisition;

(4) according to single diode model and the equivalent electrical circuit thereof of photovoltaic battery panel, the external characteristic I-V relational expression of photovoltaic battery panel is drawn:

$I=I_{\mathrm{ph}}-I_0\left\{\exp \right[\frac{q(V+{\mathrm{IR}}_s)}{\mathrm{AkT}}]-1\}-\frac{V+{\mathrm{IR}}_s}{R_{\mathrm{sh}}}$

Wherein,

V---photovoltaic battery panel two ends measuring voltage

I---electric current in photovoltaic battery panel external circuit

A---diode quality factor

The backboard temperature of T---photovoltaic battery panel

K---Boltzmann constant (1.380 × 10-23J/K)

R _s---cell series resistance

Q---electron charge (1.608 × 10-19C)

I _o---diode reverse saturation current

I _ph---photovoltaic battery panel photogenerated current

R _sh---cell parallel resistance

Quote Lambert W function and simplify I-V equation, obtain the explicit expression of photovoltaic battery panel electric current I:

$I=\frac{R_{\mathrm{sh}}(I_{\mathrm{ph}}+I_o)-V}{R_{\mathrm{sh}}+R_s}-\frac{{\mathrm{AV}}_{\mathrm{th}}}{R_s}\×W\left(\frac{R_s{R}_{\mathrm{sh}}{I}_o}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\exp \left(\frac{R_{\mathrm{sh}}{R}_s(I_{\mathrm{ph}}+I_o)+R_{\mathrm{sh}}V}{{\mathrm{AV}}_{\mathrm{th}}(R_{\mathrm{sh}}+R_s)}\right)\right)---\left(2\right)$

Wherein, V _th=KT/q, W () are LambertW function, by photogenerated current I _ph, reverse saturation current I ₀, parallel resistance R _sh, resistance in series R _s, diode quality factor A five parameters just can determine I-V curve;

(4) for the I-V curve of output of photovoltaic battery panel under different irradiance, different degree of aging, if five dimension particle x _k=(I _ph, I ₀, A, R _s, R _sh), through parameter identification, obtain the optimal value of these 5 parameters.

2. method according to claim 1, is characterized in that, step (4) application self-adapting Chaos-Particle Swarm Optimization parameter identification method carries out parameter identification.