|Publication number||US20090171597 A1|
|Application number||US 12/316,354|
|Publication date||Jul 2, 2009|
|Filing date||Dec 11, 2008|
|Priority date||Jan 1, 2008|
|Also published as||EP2077453A1, US8874392, US20120101751|
|Publication number||12316354, 316354, US 2009/0171597 A1, US 2009/171597 A1, US 20090171597 A1, US 20090171597A1, US 2009171597 A1, US 2009171597A1, US-A1-20090171597, US-A1-2009171597, US2009/0171597A1, US2009/171597A1, US20090171597 A1, US20090171597A1, US2009171597 A1, US2009171597A1|
|Original Assignee||Sma Solar Technology Ag|
|Export Citation||BiBTeX, EndNote, RefMan|
|Referenced by (3), Classifications (13), Legal Events (1)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The invention relates to an evaluation method for determining a power reduction due to aging of at least one photovoltaic module at constant radiation intensity by measuring an electrical variable that may vary after a while such as cell current, cell voltage and/or cell power without additional sensors for measuring the radiation intensity.
The power of solar cells or of photovoltaic modules is subject to aging. The loss of efficiency ranges from 10% to 20% over a period of time of 20 years.
For mounting, solar cells are combined into modules, so-called solar or photovoltaic modules or solar panels. After 20 years, a solar module only has 90% to 80% of the power indicated.
From a study “Pratt R. G. et al: “Power of a 4 kW amorphous-silicon alloy photovoltaic array at Oakland Community College, Auburn Hills, Mich.” XP010750513” it is known to record the efficiency and the energy output of the plant over a short period of time. It appears from the study that the mere comparison of e.g., the energy output in the same month of e.g., two consecutive years does not allow inference on the aging of the PV modules since the values characterizing the energy output such as radiation intensity and temperature differ too much.
Methods are known, which are based on artificially accelerated aging tests. These tests are only performed for certain conditions such as temperature and irradiation and are thus usually allocated safety factors. Hence, the efficiency a manufacturer guarantees after a certain time is usually less than the actual efficiency drop. For such tests, additional sensors are moreover utilized for sensing for example the temperature or radiation intensity.
In practice, the power drop of the solar module is very difficult to follow over time. It is difficult to locate whether the power drop is within the limits indicated by the manufacturer since certain general conditions are needed for this purpose such as a certain outside temperature, precise sensors or calibratable radiation sensors and the like.
It is the object of the invention to find a method of the type mentioned herein above by means of which a long-term power reduction of solar modules in an installed solar plant can be readily determined with high accuracy.
This object is solved by the following method steps:
Further advantageous implementations of the invention are characterized in the dependent claims.
Thanks to the measurement method of the invention, it is possible to determine very precisely a long-term power reduction without the need of additional sensors.
The invention relies on the observation that comparable days in terms of radiation intensity and temperature of the solar plant can be found over several years, these days allowing a reliable statement with regards to power reduction thanks to their comparability. In order to ensure comparability in terms of solar radiation, it is possible to compare over the years days in which a power curve P(t) is almost identical for example. Since the radiation intensity and the radiation time fluctuate in the course of a year, the year is divided in several periods, i.e., in several classes k. k may for example be equal to 52 so that the comparison may be made weekly. The actual comparison then only occurs within one class k over the years n.
Accordingly, the invention relies on the idea consisting in observing the energy output of the solar plant over several years. Actual data are compared with the values in previous years. This allows locating in which way the energy output decreases over the years.
In every year, the days must be found at which the energy output of the solar plant is comparable with values of previous years in order to be capable of making a reliable statement in terms of power reduction.
Accordingly, for each year n, there may be a maximum of k energy values. These energy values can be compared to the energy values of a previous year, for example to those of the previous year or of the first year. The difference between the energy values is for example a measure of the power reduction. This difference can be normalized. The energy amount of the previous year may serve as a standard.
Accordingly, the invention allows observing over the years power reduction within one class. Thanks to this accurate measurement, a solar plant can thus be connected for a longer time to a grid without maintenance works. It is also possible to locate early increased power reduction so that photovoltaic modules can be exchanged in time.
To measure the energy output, a variable of the generator such as the generator current, the generator voltage or the generator power can be measured.
In an advantageous embodiment of the measurement method of the invention, there are provided at least two classes k, which are distributed over the year and within which the daily power curves or the portion of the daily power curve are comparable on the basis of the radiation intensity and the outside temperature as well as of the radiation time. This allows for taking into consideration the seasonal fluctuations of the radiation intensity and of the outside temperature as well as of the irradiation time to be expected.
Sensing or filing a daily curve of the power P of the generator as well as of the energy output E of the photovoltaic generator of the day is particularly advantageous, a day power curve P(t) being determined, the course of which is comparable to at least one day curve Pk,n(t) from previous years. Accordingly, a measurement of the day curve of the power of the solar generator P(t) as well as of the daily energy output Etag, i.e., of the two variables is advantageous. These values can be recorded and filed in a data bank, for example by means of a data logger. Within each class k, one day power curve Pk,n(t) the course of which is comparable to the day power curves Pk,i(t) from previous years can be determined for each year n. This also means that at most k energy output values Ek,n are determined for each year, said energy output values being compared to the energy output values Ek,I of the previous years i, for example the first year values Ek, 1 so that at the most k values are determined for the energy differences ΔEk,n. At least one of these energy differences ΔEk,n related to an energy output value Ek,i of a previous year, for example of the first year Ek,1 is used to indicate the power reduction of the solar generator.
Accordingly, within each class, one day or one measurement period is determined the power curve P(t) and the energy output E of which is comparable to measurements performed in earlier days.
It is advantageous that a number ktmax of day classes kj is formed. The day classes correspond to time periods that are associated with fixed time intervals and that are distributed over the daily sunshine period to be expected. This comparison then occurs between measurement intervals comprising both the same k and kj classes. One class may also be a class having a same pair of coefficients (k, kj) or a same tuple. In the variant widened to include the day classes, there is a maximum of k*kj energy values E((k,kj),n), which are compared to the energy values of the previous years. Differences and averages can be calculated here.
In an advantageous embodiment of the invention, comparable days are determined within one class in two steps at most, namely by calculating and evaluating the first derivation Pk′(t) of the function Pk(t).
The first derivation Pk′(t) of the data host Pk(t) is hereby calculated and it is checked whether certain limit values have been respected. If these limit values are exceeded or not reached, it can be assumed that this day was cloudless.
Alternatively, the first derivation Pk′(t) is calculated and evaluated by testing a curve area included therein for fixed limit values. In a first step, the first derivation Pk′(t) then enters an evaluation method which yields values that are not allowed to exceed or fall short of imposed limit values.
Using a plausibility criterion, it can be clearly determined in a second step whether the day is really cloudless. One plausibility criterion may for example be measurement data from the data host Pk(t) lying within tolerance bands such as a tolerance band for a certain region or a tolerance band for summer days and one for winter days. This means that this plausibility criterion is applied to the data host P(t) for example.
An alternative plausibility criterion is applied to the energy output Ek,tag. The energy output Ek,tag is not allowed to fall short of an imposed limit value either.
As an alternative or in addition thereto zero crossings of the first derivation Pk′(t) are advantageously evaluated in another implementation. A day is clearly cloudless if there is only one zero crossing. This evaluation method can be performed in only one step.
It is further sensible to impose limit values in the first year and to verify these limit values in the following years.
In another advantageous developed implementation of the method of the invention there is provided that at least one average value ΔEmittel is calculated from the k energy difference values ΔEk,n for the year n as a measure for power reduction. By calculating the average, it is possible to make a very precise statement about the aging condition of the solar cells. The advantage of this way of proceeding is that quotients or percentage values obtained thereby, meaning daily and weekly values, can be averaged both for the k and for the kj classes. One thus obtains a very wide statistical basis, which ensures good accuracy of the values.
A particular benefit is obtained if a daily power curve is determined, which substantially comprises a daily power maximum. This case occurs if the day is cloudless. In this variant of the invention, only energy values from measurement periods are compared in which there was no shadow. Shadowing can be recognized with the methods mentioned herein above.
In this variant, time periods of shadow are recognized by evaluating the change in luminosity occasioned by passing clouds. In principle, one locates the change in luminosity by calculating the derivation of the power generated by the solar plant with respect to time. High values of such a derivation are evaluated by an evaluation algorithm which calculates whether the time range of the measurement was influenced by clouds. An effect of benefit is obtained by implementing an evaluation standard through fix or adaptive threshold values. An evaluation standard with amount averages, quadratic averages or other values can be utilized. An advantage is obtained if a time range with shadowing is recognized by comparing the amount of power or energy generated in the time range with a comparative value. If there are significant negative differences, there were shadows. Such a comparative value can be generated from a model of radiation on a cloudless day. A comparative value can also be calculated from the radiation values of previous days, in particular if these corresponding time intervals were recognized to be cloudless. An effect of benefit is obtained if a comparative value is calculated from values of previous years that have been stored.
An exemplary embodiment will be described in closer detail with reference to the drawings, additional developed implementations of the invention and advantages thereof being described herein.
In said drawings:
In accordance with the invention, the method is based on the measurement or on the acquisition of the daily curve of the power of the solar generator P(t) as well as on a daily energy output Etag. These data can be stored in a data bank.
Distributed over the year, there are several classes k. Two classes k are required though. This is beneficial in order to take into account a seasonal radiation intensity as well as a seasonal dependent outside temperature.
Within each class k, a power curve Pk,n(t) is determined for each year n, said curve having a course that is comparable with the daily power curves Pk,I of previous years i.
I.e., for each year, a maximum number k of energy output values Ek,n are ascertained, which are compared to the energy output values Ek,I of the previous years, for example with the energy output values Ek,l, so that a maximum number k of values is determined for the energy differences ΔEk,n.
At least one of these energy differences ΔEk,n related to an energy output value Ek,l of a previous year i, for example of the first year value Ek,l, is used to indicate the power reduction of the photovoltaic generator.
Preferably, the energy outputs are related to one day. I.e., within each class one ascertains a day the power curve P(t) and the energy output E of which are comparable with measurements performed in earlier days. For each year n, one then has k energy values Ek,n, which relate to one day. These daily energy values Ek,n can then be compared with the daily energy values Ek,n−1 of the year before or with the energy values Ek,l of the first year or of any year.
From these variables, mean values from all the k energy differences Δ(Ek,n) can be acquired for one year n, namely with the formula:
For normalized energy differences, the above mentioned formula takes the following form:
Then, the values ΔEmittel,n or ΔE*mittel,n indicate a yearly average value for power reduction of the photovoltaic generator in the respective year n.
It is also possible to consider quadratic averages or an average to the power of p, i.e., Δ(Ek,n)p or Δ(E(k,kj),n)p. As a result, a significant change in the values is possible through which the ageing process of the cells can be normalized. An effect of benefit is obtained if p is a constant comprised between 2 and 6. Herein after, location is described for comparable days.
As shown in
If a day generates a curve as shown in
P(t) for days with different weather conditions, these being in discrete form, i.e., they constitute effective measured variables. For simplicity's sake, these functions are shown schematically or as continuous functions in the
Preferably, in a first step, the first derivation P′(t) of the power curve P(t) is formed and evaluated for each day, as shown in the
These results are shown in the curve through measurement points on the curve. The curve has a day maximum peak, which is typically about noon. The area below the curves corresponds to the curve integral or to the energy.
In order to be capable of determining a power reduction of the solar module with very high accuracy, data having a comparable power curve P(t) of the photovoltaic module or of the solar plant are preferably observed over several years. As they are comparable, it is possible to make a reliable statement with respect to power reduction of the photovoltaic module. In principle, completely cloudy days as illustrated in
The measurement method preferably uses the power curve P(t) as well as the energy output E that corresponds to the area enclosed by the curve. This area is hatched in the
In a first step, the function P′(t) is formed.
The first derivation P′(t) can be evaluated in different ways. The evaluation clearly indicates whether the day is cloudless or not, as shown in
In a second step, one then analyzes and makes certain whether a cloudless day has indeed been found. For this purpose, the power curve P(t) or its first derivation P′(t) is evaluated. Preferably, two evaluation steps are utilized in order to reliably acquire a comparable cloudless day.
In one of the steps, the first derivation P′(t) is evaluated. For this purpose, there are two possibilities of evaluating the first derivation P′(t). A first possibility is based on the fact that the evaluation method is based on analyzing a maximum for P′(t). If, as shown in
As shown in
As shown in
A second possibility of evaluating the first derivation P′(t) is described herein after. In this variant of the evaluation method, P′(t) is also formed from the power curve P(t) for each day.
For each day, the integral is for example determined
If the value I acquired lies below a maximum allowable limit I_max so that I<I_max, it can be assumed that the day is cloudless and this day can be included in calculating the power reduction. This limit is obtained from typical radiation values and depends for example on the geographical situation. The corresponding day is then fixed according to the same pattern as in the previous example of the evaluation method of P′(t).
The integral I is a measure for the area included in the first derivation P′(t). A comparison between the
In another possibility of evaluating the first derivation P′(t), only the zero crossings of the P′(t) are taken into consideration. If the day is for example cloudless, the number of zero crossings of the curve P(t) is equal to 1. This zero crossing takes place at the time of power maximum, as shown in
Herein after, the second step of evaluating the power curve P(t) is described in closer detail.
Since in the first step it is at first only supposed that the day is cloudless or not, this must be confirmed in a second step.
There are different variants to achieve this. The first possibility is to evaluate the daily power curve P(t) in the second step.
The first method for evaluating the daily power curve P(t) consists in determining the daily energy output Etag and in comparing it with an imposed minimum value. If the daily energy output Etag exceeds this minimum value, it is certain that the day is cloudless.
The second evaluation method in the second step consists in evaluating the extreme values of the power curve P(t). For this purpose, an absolute value of the power Pabs is acquired from the host of data P(t) measured within one day. It may for example be the maximum value Pmax of the power P(t) for the day observed or also an average of several power maxima. If this value Pabs lies within a tolerance band ranging from Pabs_min to Pabs_max, then it may well be a relatively cloudless day. Indirectly one also considers the radiation intensity and the duration without the need for an additional sensor.
This method can be even further improved by using the measurement or the measurement values of the temperature of the modules and/or of the outside temperature.
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7835832||Jan 5, 2007||Nov 16, 2010||Hemisphere Gps Llc||Vehicle control system|
|US9112078 *||Jun 8, 2011||Aug 18, 2015||Sma Solar Technology Ag||Method of recognizing and assessing shadowing events|
|US20110307199 *||Dec 15, 2011||Sma Solar Technology Ag||Method of recognizing and assessing shadowing events|
|U.S. Classification||702/60, 702/64|
|International Classification||H02S50/00, G01R21/06, G01R21/00|
|Cooperative Classification||H02S50/10, H01L31/02, Y02E10/50, H01L31/186, H01L31/042|
|European Classification||H01L31/18G, H01L31/042, H01L31/02|
|Apr 7, 2011||AS||Assignment|
Owner name: SMA SOLAR TECHNOLOGY AG, GERMANY
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:DREWS, PETER;REEL/FRAME:026087/0086
Effective date: 20110407