DE10233855A1 - Process and arrangement to optimize the production of photovoltaic units uses comparison with test data to sort the individual cells - Google Patents

Abstract

The invention relates to a method and an arrangement for optimizing the production of photovoltaic products, in which, in addition to the characteristic values relating to the standard conditions, additional characteristic values are determined and the scatter of the mean power values of a type and the ratio of cell power to module power used by interconnecting cells and strings adopt minimum values according to a laying plan. DOLLAR A The new process should make it possible to optimize the process parameters during contacting and to remove unsuitable cells. DOLLAR A The task is solved with a classification and sorting process in which branches of the same open circuit voltage are interconnected, whose short-circuit current and impedance have a small class width. The mean values and the class widths of the voltages, currents and impedances of the cells and strings of these branches are determined in a training process and used as a database for the automatic comparison device of the real-time detection system. DOLLAR A The results of the comparison serve to classify and store the cells and strings according to voltage classes, from which parts are taken in the laying processes for the products according to the laying plan determined with the process computer. DOLLAR A The field of application of the invention lies primarily in the process control of photovoltaic products.

Description

The invention relates to a method and an arrangement for optimizing the production of photovoltaic Products. In particular, the invention relates to a method and an arrangement with which cells and strings are classified inexpensively and be sorted.

For the production of photovoltaic products cells are used that are sorted into performance classes.

It is known that the scatter the performance of photovoltaic products of the same type the use of cells with a narrow span in the performance classes can be reduced.

The manufacture of photovoltaic products generally includes laying, contacting, laminating, assembly as well as testing processes.

The yield of products one Design, the average power distribution of which has a minimal standard deviation can, in essence, only by taking measures prior to lamination elevated become. In the known processes, an inexpensive increase in yield is hardly possible. The process-related changes the impedances as well as the ranges of the characteristic values of impedance, The current and voltage of the components of a product are checked at Laying and interconnection not taken into account.

A simple procedure is advantageous to reduce the standard deviation of the power mean distribution for products of one type.

The invention was based on the object to create a method and an arrangement with which from cells different performance classes with photovoltaic products the required nominal power mean with minimal standard deviation as well as minimal ratio generated from cell power to module power can.

Another object of the invention is in the manufacturing process cells or strings with structural as well Performance deficiencies and low reliability to capture and discard.

Another object of the invention is to optimize the process parameters of the contacting devices.

The above tasks are accomplished by the features of the method in claims 1 and 2 and the features the arrangement solved in claim 3.

The solution to the problem is the Process control and reduction of costs in the production of photovoltaic Products as well as improving product reliability.

The specified in the claims solution the problem relates to a method and an arrangement for the same Execution, with which the dispersion of the average power values of a type of photovoltaic products as well as the ratio Cell power used to module power through the interconnection of Cells and strings according to a layout plan based on voltage classes accept minimum values.

The task is solved by one Classification and sorting process for different cells and strings process-related impedances, open-circuit voltages and short-circuit currents.

The comparison device of the process computer contains a real-time detection, analysis, training, classification and sorting system.

The real-time process conditions with the associated Cell and Product Specifications are stored in the system database of the training system.

Be in a training process the test results are saved and saved by the automatic Comparison device of the real-time detection system as an analysis basis used.

The test results are used for process control and increase in yield.

The photovoltaic products are manufactured in a manufacturing process that involves a series of steps includes.

The performance of the manufacturing process is measured by the yield of products that meet the performance specification the design with minimal standard deviation.

The classification and sorting the cells and strings take place after the measurement of impedance, open circuit voltage and short-circuit current, comparing the characteristic values, learning using examples as well as the optimization of class mean values, spans and upper ones n or lower class limits in the process computer.

The permissible ranges, mean values and class widths of open circuit voltage U _Z , impedance Z _Z and short circuit current I _Z for the cells of each class are determined in a training process.

Depending on the number of cells connected in series, the ranges, mean values and class widths of the open circuit voltages U _S and the short-circuit currents I _{S of} the strings are dimensioned in such a way that power losses are avoided when modulating the non-linear current-voltage characteristic curves and scattering of the mean power values within one Design of photovoltaic products can be minimized.

The mean value of the open circuit voltage U _B results from the mean value of the open circuit voltage of the branches U _A , which are formed by strings (U _S ) with the sum of the cell voltages U _z . U B ± ΔU B = U al ± (I · ΔU A 2 ) 1.2 U A ± ΔU A = ΣU S m ± (m · ΔU S 2 ) 1.2 U S ± ΔU S = ΣU Zn ± (n · ΔU Z 2 ) 1.2 ,

The branches of a product have same mean values in the open circuit voltages. The branches become parallel interconnected to the product.

By comparing the characteristic values the final test and the required characteristic values become the characteristic mean values and class sizes in the classification procedures during the Manufacturing processes optimized with the training system.

The characteristic values of the active components the impedances of cells and strings before and after contacting - measured at a frequency of one kHz - are compared and their process-related scattering to correct the process parameters of the Contacting devices used.

You can also use this comparison the inductors or the damping at a frequency of 1 kHz, for example, or the resonance frequencies be used.

The effective part of the impedance of penetrated, broken and similar Unsuitable cells are identified in the training process so that they are faulty Cells in the real-time recognition process from further processing can be excluded. The active components of the impedances and the inductivities of these defective cells are much lower than the more standard ones Cells.

The resonance curves of defective cells have - according to higher Damping - one flatter course than the standard cells, so these cells also measured by measuring the effective series resistance can be.

The characteristic values for the classification and sorting cells and strings are done during the manufacturing process in a training process by comparison with the characteristic values of the products after the final inspection corrected until the spread of benefits from the mean one Design as well as the ratio a minimum of the cell services used for module performance to have.

Practical application of the invention is carried out during the process control by a method with which Classification and sorting processes of photovoltaic products with minimal standard deviations from the mean power of the design can be manufactured.

The invention is intended in an embodiment be explained in more detail with reference to schematic drawings.

Show it

1 The frequency distributions of the performance of products of a type with unsorted and with classified sorted strings.

2 The arrangement for checking, contacting, classifying and sorting the cells

3 The arrangement for checking and classifying the strings and for laying in branches

4 The flow chart for the implementation of the process for the optimization of the production of photovoltaic products

5 The flow of data in the classification and sorting process

6 The scheme of the measurement arrangement

7 The scheme of the probe

1 shows the effect of optimization on the standard deviation of the mean distribution of performance. The frequency distribution of the mean power values of products of a type with unsorted and with classified sorted strings is shown.

2 shows the arrangement for performing the method for checking, contacting, classifying and sorting the cells.

The cell 1 is or will have a cell-specific indicator 2 provided and on the pad 3 stored. With the help of the probe 4 the impedance of the unirradiated cell is determined at a frequency of 1 kHz. Then the open circuit voltage and the short circuit current of the irradiated cell 1 with the help of the probe 4 measured. The characteristic values determined in this way are sent to the analog-digital converters of the process computer 6 fed.

The irradiance and spectral radiation distribution of the radiation source 5 have a reproducible constant amount on the measurement object in the measurement period. The measured values for impedance, current and voltage of the cells 1 are in the process computer 6 related to the standard conditions and with the indicator 2 saved.

The cell 1 is in the contacting device 7 with the contact strips 8th and 9 Mistake.

With the probe 10 becomes the impedance of the unirradiated contacted cell 11 measured at a frequency of 1 kHz. Then the open circuit voltage and the short circuit current of the irradiated cell 11 measured.

The irradiance and spectral radiation distribution of the radiation source 12 have the same amount on the measurement object as the radiation source 5 ,

The measured values for the impedance, the current and the voltage of the cells 11 are in the process computer 6 related to the standard conditions and with the indicator 2 saved.

In the process computer 6 become the evaluated characteristic values of the cells 1 in front of and the cells 11 compared after contacting. When contacting, no or only minimal process-related changes in the active components of the cell impedance or the inductance should occur.

The process parameters time, temperature and pressure of the contacting device 7 are recalculated and corrected if the cell impedance deviates from the permitted values.

The cells 11 are in the process computer 6 classified into voltage classes with the class _{mean values} U _z1 to U _zn and the class _width ΔU _z and then sorted in the cell magazines 13 stored.

For a product - knowing the range and the measured value distribution of voltage, current and impedance within a batch of products of one type - each average cell voltage U _{Zn is} an impedance Z _ZK and a short-circuit current 1z assigned to: U Z1 ± ΔU Z with Z ZK ± ΔZ ZK and I Z ± ΔI Z U Z2 ± ΔU Z with Z ZK ± ΔZ ZK and I Z ± ΔI Z U Zn ± ΔU Z with Z ZK ± ΔZ ZK and I Z ± ΔI Z

3 shows the arrangement for checking and classifying the strings.

The cells 11 are in the process computer 6 compiled into strings of the voltage classes U _S1 to U _{S m} and the class width ΔU _S. U S m ± ΔU S = Σ U Zn ± (n · ΔU Z 2 ) 1.2 Z S ± ΔZ S = Σ Z ZK ± (N · .DELTA.Z ZK 2 ) 1.2 I S ± ΔI S = I Z ± (n · ΔI Z 2 ) 1.2 An impedance ZS and a short-circuit current Ι _{S are} assigned to the mean voltage values U _Sm : U S1 ± ΔU S with Z S ± ΔZ S and IS ± ΔI S U S2 ± ΔU S with Z S ± ΔZ S and IS ± ΔI S U sm ± ΔU S with Z S ± ΔZ S and I S ± ΔI S The contacted cells 11 of the cell magazines 13 after one from the process computer 6 created layout plan for strings 14 on the string pad 15 filed and then contacted that with minimal use of cell power the specifications of the photovoltaic module 16 be met exactly.

The product voltage U _B ± ΔU _B is equal to the branch voltage U _A ± (I · ΔU _A ² ) ^1/2 . The product current I _B ± ΔI _B results from the sum of the branch currents Σ IA ± (I · ΔI _A ² ) ^1/2 .

The branches 17 of a module 16 have the same open circuit voltages U _A ± (I · ΔU _A ² ) ^1/2 after laying and contacting. The sum of the mean voltage values of the cells U _Zn must therefore lie within the permissible range of the mean string voltage values U _Sn so that their sum can give the branch voltage U _A.

The class sizes of the considered Characteristic values of current, impedance and voltage determine the standard deviations the mean power distributions one type.

The mean values of the string currents I _S and their class widths ± ΔI _S are determined by modulating the nonlinear I _Z -U _{Z characteristic} curves of the cells according to radiation of the irradiance E, its spectral radiation distribution, the resulting string voltage U _S , the effective series resistance R _{s, eff} , the load resistance R _L and the temperature: I S ± ΔI S = f (E, I Z ± (n · ΔI Z 2 ) 1.2 , U S , R s, eff , R L , T) The following correlations result for a product: U B ± ΔU B = U A , ± (I · ΔU A 2 ) 1.2 U A ± ΔU A = Σ U S m ± (m · ΔU S 2 ) 1.2 U S ± ΔU S = Σ U Zn ± (n · ΔU Z 2 ) 1.2 Z S ± ΔZ S = Σ Z ZK ± (n · ΔZ ZK 2 ) 1.2 I A ± ΔI A = I S ± (n · ΔI S 2 ) 1.2 I S ± ΔI S = I Z ± (n · ΔI Z 2 ) 1.2

As an example, the combination of the branches with the mean voltage value U _A from two strings of different voltage classes U _{Sm is} considered.

With the help of the probes 18 . 19 and 20 are the impedances of the unirradiated contacted strings 14 and of their parts (e.g. half a string) measured at a frequency of 1 kHz.

Then the open circuit voltages and short-circuit currents of the irradiated strings 14 and measured from their parts.

The irradiance and spectral radiation distribution of the radiation sources 21 have the same amount on the measurement object as the radiation sources 5 and 12 ,

The measured values for the impedances, currents and voltages of the strings 14 or branches 17 are in the process computer 6 related to the standard conditions, evaluated and with the markings 2 saved.

In the process computer 6 become the characteristic values of the cells 11 and the strings made from it 14 evaluated, saved and compared after contact.

The process parameters time, temperature and pressure of the stringer 22 If the string impedance deviates from the permitted values, they are recalculated and corrected.

When contacting, the deviation of the active component of the string impedance R _s from the sum of the active components of the cell impedances R _{zk should assume} minimum values. R s ± Δ RS ≈ Σ R ZK ± (n · ΔRZ ZK 2 ) 1.2

The strings 14 are in the process computer 6 classified according to voltage classes U _Sm and sorted in magazines for strings 14 or on documents for modules 16 stored.

4 shows the process flow diagram for optimizing the production of photovoltaic products. The products and their cells must be distinguishable for this.

Before the exams begin, the cells 1 with a machine-readable mark 2 Mistake. The strings made from it 14 have for example the combination of the indicators 2 the first and last cell 1 of the string 14 ,

The optimization of the production of photovoltaic products is based on the parameterization of the automatic comparison device of the real-time detection system 23 of the process computer 6 to solve the classification tasks after learning from examples with a training system 24 ,

In the classification phase of the training process each class of a characteristic value must have a sufficiently large number of characteristic values in order to achieve a sufficient classification quality.

• – die Erzeugnisspezifikation – insbesondere die Standardabweichung der Mittelwertverteilung der Leistung von Zelle und Erzeugnis, die Schaltungsanordnung, der Zellentyp und die Zellenspezifikation,
• – die daraus für den ersten Lernschritt errechneten Spannweiten, Mittelwerte und Klassenweiten für die Impedanzen, Leistungen, Leerlaufspannungen und Kurzschlussströme der Zweige; Strings und Zellen sowie
• – die bezogenen und bewerteten Messwerte für die Impedanzen, Leistungen, Leerlaufspannungen und Kurzschlussströme.

The input data of the object characteristics for the classification tasks are

• - the product specification - in particular the standard deviation of the mean value distribution of the performance of cell and product, the circuit arrangement, the cell type and the cell specification,
• - The ranges, mean values and class widths calculated from this for the first learning step for the impedances, powers, open circuit voltages and short-circuit currents of the branches; Strings and cells as well
• - the related and evaluated measured values for the impedances, powers, open-circuit voltages and short-circuit currents.

The input data are with variable Weighting factors multiplied and after the transformation in one Mapped the matrix of the feature space.

The measured values for the impedances, powers, currents and voltages are stored in the process computer 6 related to the standard conditions, evaluated and with the markings 2 saved.

After further processing of these results in the process computer 6 The class assignment takes place in the real-time recognition phase based on the training results.

The objects become the voltage class assigned that is closest to the evaluated characteristics.

The developed training results are included in a series of classification processes compared the results of the product test and corrected if necessary.

The training results serve in the ongoing production of photovoltaic products as a basis for comparison for current Characteristic values and as a tool to increase reliability and quality.

In the real-time detection system 23 optical and electrical characteristics are processed. The imaging system 25 is connected to the real-time detection system 23 connected.

Before processing into strings 14 become the cells 11 the optical control of an image analysis system 27 with the camera 26 of the imaging system 25 subjected.

The illustration of the camera 26 is converted into a digitized image by an image-relevant signal generation, which is stored in the image analysis system 27 is processed.

To the image analysis system 27 include image preprocessing, image recognition, feature comparison and the output unit of the analysis results.

During image preprocessing reduced the amount of data of the digitized image and thus the Cognitive performance increased by reducing the effort.

There will be suitable local property features extracted.

Image preprocessing is used for error suppression and highlighting relevant image elements as a preliminary stage automatic image interpretation.

The input data of the object characteristics are multiplied by variable weighting factors and in one Matrix depicted. After further processing of the characteristic values The class assignment takes place in the real-time recognition phase Basis of the training results.

With the image analysis system 27 the presence of optical error features is checked. The optical error characteristics include, for example, deviations in the dimensions, in the color, in the contact structure, as well as cutouts, breaks, holes, gaps and cracks. The errors are assigned to error classes based on their characteristics.

The training system 24 and the real-time detection system 23 deliver to every incoming image of the image acquisition system 25 a classification result. After the feature comparison in the real-time recognition system 23 become faulty cells 11 excluded from further processing.

In the real-time detection system 23 the object characteristics of the photovoltaic products impedance, power, open circuit voltage, and short circuit current are processed.

The input data of these object characteristics are multiplied by variable weighting factors and in one Matrix depicted.

The class assignment of the objects takes place in the real-time detection phase after evaluation and comparison with characteristic values of the training phase. In the training phase optimization through learning.

The cells 11 and strings 14 are put together so that the branches 17 of a module 16 - taking into account the irradiance, temperature, power, impedances and short-circuit currents as well as their class widths - have the same open circuit voltages [U _A ± ΔU _A ].

The impedances of the cells 1 , the contacted cells 11 and the strings 14 are determined without radiation. They can also be determined as the effective series resistance from the low-frequency flank of the filter curve.

The open circuit voltages and short circuit currents are measured with the help of the radiation sources 5 . 12 and 21 in the test arrangements for cells 28 , for contacted cells 29 , for strings 30 and modules 31 determined.

The radiation sources 5 . 12 and 21 cause identical irradiance levels and spectral radiation distributions on the measurement object.

The one in the test setup 29 determined active components of the impedances R _{ZK of} the contacted cells 11 after contacting with those in the test setup 28 determined active components of the impedances R _{z of} the cells 1 in the process computer 6 compared.

The process parameters time, temperature and pressure of the contacting device 7 In the event of impermissible deviations of the active components of the impedances R _zK ± ΔR _zK from the values of the non-contacted cell R _z ± ΔR _z , they are recalculated and corrected.

When contacting, the process-related change in cell impedance should assume minimal values. cell 11 with too low active component of the impedance R zK <R zKmin have a significant defect in the crystal structure, in the doping profile or in the contact structure. You will be excluded from further processing.

Preferably the cells 11 before storage in a cell magazine 13 and processing into strings 14 subjected to an optical inspection.

The cells with optical error characteristics are excluded from further processing.

The contacted cells 11 that have no fault characteristics are classified according to the evaluated measured values of open circuit voltage, impedance and short-circuit current and in cell magazines 13 stored.

The cells classified as suitable 11 are in the process computer 6 in laying plans for strings 14 put together, the stringer 22 and then the test arrangement for strings 30 fed.

In the test set for strings 30 become partial strings and the entire strings 14 checked.

The impedances of the partial strings and strings 14 are determined without radiation. The open circuit voltages and short circuit currents of the strings and partial strings are when irradiated with the radiation sources 21 in the test set for strings 30 measured. 'The in the test setup 30 determined impedances of the strings 14 after contacting with those in the test setup 29 determined impedances of the cells of the relevant string compared. Z S ± ΔZ S = Σ Z ZK ± (n · ΔZ ZK 2 ) 1.2 The process parameters time, temperature and pressure of the stringer 22 If the permissible process-related deviations of the active components of the string impedance from the sum of the active components of the cell impedances are exceeded, they are recalculated and corrected.

The strings 14 are classified according to the characteristic values of open circuit voltage, impedance and short-circuit current and in branch layout plans 17 for modules 16 compiled.

The modules laid out in this way 16 are in the test set for modules 31 checked.

The modules 16 are in the laminator 32 and finishing 33 completed and then the test arrangement for the products 34 fed.

The characteristic values of the product inspection are compared in the process computer with the expected characteristic values.

With larger deviations of the characteristic values each other and the approved characteristic values of the products the mean values, class limits and laying plans are corrected.

5 shows schematically the data flow in the classification and sorting process for cells and strings. In the process computer 6 the required characteristic values of the branches for the characteristic values of the product specification 17 , the strings 14 and the cells 11 calculated.

For each of these characteristic values are the mean values, the standard deviations as well as the upper and lower limit values of the classes in the training process determined.

In the process computer 6 are the cell measured values of impedance, voltage and current of the test arrangement for contacted cells, which are based on and evaluated on standard conditions 29 processed and the cells 11 classified. The cells 11 are classified according to voltage classes in the cell magazines 13 stored. The permissible upper and lower limit values of the impedance and the current of the cells 11 each stress class was determined in the training process.

Based on the stock of cells 11 in the cell magazines 13 and the requirements for the product are determined by the process computer 6 Laying plans for modules 16 as well as their branches 17 and strings 14 created.

In the process computer 6 are the string measurement values of impedance, voltage and current of the test arrangement for strings, which are related to standard conditions and evaluated 30 processed and the strings 14 classified. The strings 14 are put together according to voltage classes to branches of the same voltage, stored or modules 16 placed. The permissible upper and lower limit values of the impedance and the current of the strings 14 each stress class was determined in the training process.

By measuring the characteristic values of String parts, for example half strings, and their comparison with Defects can be reduced in the relevant string parameters.

In the process computer 6 are those with the test arrangement for products 34 The characteristic values determined are compared with the characteristic values used in the classification process and the mean values and limit values of the classes are corrected in the event of deviations.

6 shows the diagram of the measuring arrangement. The measurement object 35 can be a cell 1 . 11 , a string 14 or a module 16 his.

The probes 4 . 10 . 18 . 19 and 20 are identical. They have a coaxial or stripline structure. A low overall height is achieved with a stripline structure.

The contact pin 36 of the inner conductor 37 of the probe 19 forms a contact for current measurement, voltage measurement and impedance measurement. The ground contact 38 is connected to the ground contacts of the other arrangements. It is also in the immediate vicinity of the probe 19 a temperature sensor 39 attached, the component of the probe 19 can be.

The probe is over the analog / digital converter 40 with the process computer 6 connected. The characteristic values of short-circuit current and open circuit voltage may have to be switched 41 and ballasts 42 to the input limit values of the A / D converter 40 be adjusted.

The process computer 6 controls the operation of the imaging system 25 with the camera 26 that over the measurement object 35 is arranged. It also controls the operation of the radiation sources 5 . 12 or 21 , the function generator 43 and the measurement process.

The probe 19 is via a circulator 44 (or a directional line or a T-piece) and DC isolation 45 with the sweepable function generator 43 , the A / D converter 40 and the process computer 6 connected.

The results of the process computer 6 are from data output to process control 46 directed to the appropriate control arrangements.

7 shows the schematic of the probe.

The top view 7a shows a section of a measurement object 35 , Over the contact ribbon 9 is the arrangement of the probe 19 with the temperature sensor 39 and the contact pin 36 of the inner conductor 37 on the insulator 51 ,

The side view 7b shows a stripline-shaped probe 19 with the temperature sensor 39 and its connection 47 , the ground contact 38 , the spring-loaded contact pin 36 of the inner conductor 37 on the insulator 51 , the probe guide 48 , the cable connection 49 and the cable 50 , The height of the probe is to avoid shadows 19 smaller than the width of the contact flags 9 ,

1

cell

2

Mark

3

document

4

probe

5

radiation source

6

process computer

7

contacting

8th

contact strips

9

contact strips

10

probe

11

contacted cell

12

radiation source

13

cells Magazine

14

string

15

String support

16

module

17

branch

18

probe

19

probe

20

probe

21

radiation source

22

Stringer

23

Real-time detection system

24

training system

25

Image acquisition system

26

camera

27

Image Analysis System

28

test set for cells

29

test set for contacted cell

30

test set for strings

31

test set for modules

32

laminator

33

finishing

34

test set for products

35

measurement object

36

pin

37

inner conductor

38

mass contact

39

temperature sensor

40

Analog / digital converter

41

switch

42

Vorschaltelement

43

function generator

44

circulator

45

DC separation

46

data output for process control

47

connection of the temperature sensor

48

Tastkopfführung

49

cable

50

electric wire

51

insulator

Claims (3)

Process for optimizing the production of photovoltaic products, in which, in addition to the characteristic values related to the standard conditions, additional characteristic values are determined and the scatter of the mean power values of a design as well as the ratio of cell power to module power used by interconnecting cells and strings according to a layout plan assume minimal values characterized in that - the additional characteristic values are the impedances of unirradiated components measured with high-frequency signals, - the cells and strings are sorted into voltage classes and the display of a product is placed so that the voltages have a minimum class width and the short-circuit currents and impedances are one have a small class width, - the voltages are the open circuit voltages, - the layout plan is compiled by comparing the measured values of voltage, current and impedance in the real-time detection system of the automatic comparison device of the process computer with the test results of a training system, - the probes of the test facilities for cells and strings are connected to the real-time detection system of the process computer, - the characteristic value of the impedance of the unirradiated cells and strings and the characteristic values of the open circuit voltage and short-circuit current of the irradiated cells and strings are measured, and d the comparison system of the process computer, - the difference between the product specification values and the final test values is used to correct the class mean values and class widths so as not to exceed a certain range of the characteristic values, - the ratios of the characteristic values of the strings and their partial strings as well as of cells and contacted cells are used to correct the contacting parameters, - the permissible mean values and class widths of open circuit voltage, impedance and short-circuit current for the cells and strings of each class are determined in a training process, - depending on the number of cells connected in series with different processes -conditional impedances and open circuit voltages, the class widths of the short-circuit currents are dimensioned in such a way that power losses are avoided when modulating the non-linear current-voltage characteristics and scattering of the mean power values be reduced within a type of product, A method according to claim 4, characterized in that - the active component the impedance of cells and strings without radiation as more effective Series resistance at high frequency from the steepness of the low frequency Astes the filter curve is determined - to compare the characteristic values the inductance of the unirradiated component is used, - for comparison the characteristic values the lower resonance frequency of the unirradiated Component is used, - The effective part of the impedance the cells and strings without radiation with a high-frequency impedance measuring device is measured - to the Comparison of the characteristic values of the equivalents Noise resistance is used, - the characteristic values of the active components the impedances of cells and strings compared before and after contacting and their process-related spread to correct the process parameters the contacting devices are used, - the active component the impedance of broken through, broken and unsuitable ones Cells in the training process is identified so that errors Cells in the real-time recognition process from further processing can be excluded. Arrangement for implementation of the method according to claim 1, characterized in that - the probe is built in stripline technology, a temperature sensor and a strip conductor with contact pin for current measurement, voltage measurement as well as for impedance measurement.