Patent Application
20100332167
Photovoltaic-type solar power generation is rapidly being deployed as both isolated residential systems and as utility-grade power generators. In both, an area is covered with solar panels wherein the area is large enough to have significant sun illumination or shading differences between panels in a system. Similarly, there may be significant panel to panel differences in soiling, dust coverage, mechanical or electrical degradation, and even aging in that not all panels in a system may have been installed at the same time nor have been manufactured by the same maker or even have the same capacity or other characteristics. Due to exposure to extreme conditions of heat, cold, moisture, wind, salt spray, sand, and other environmental factors, problems with one or more panels from time to time is to be expected.
System problems must be detected, then appropriately responded to. Typical installations today provide little information regarding the operating conditions of individual panels. In response, some power providers periodically send trained technicians to evaluate performance, visually look for damage or dirt, and measure certain operational parameters; an expensive procedure. If the technician is sent too often with no problems found, the cost is wasted. If a problem develops and the technician does not visit soon or the problem is not one appropriate for a trained technician to be required, there is an opportunity cost to under-delivering power that could have been delivered. For example, if performance is low but the panels are simply dirty, a less expensive cleaning person could be sent to clean the panels rather than having a trained technician do so or the technician make a visit just to request a cleaning crew be sent instead.
As the solar power generation industry has matured and competition become more keen, power providers are being asked to base quotes and charges upon actual power delivered, with a minimum guarantee, rather than on simply installed capacity. Thus maintaining optimum performance is critical. In addition, the aging of a system and its panels must be recognized so that an economic decision can be made regarding when and if a panel or panels should be replaced.
What is needed is a remote means for periodically determining the performance of a system as a whole as well as the individual components of the system. Once less than optimum performance is detected, what is needed is the means to then determine the nature of the problem to enable a decision to be made regarding what, if any, steps should be taken in response.
Fill factor of a photovoltaic cell or module is defined as the ratio of the maximum output power to the product of the open-circuit voltage value times the short-circuit current value of the module. It is a measure of rectilinearity of the I-V curve of the array. The long term ageing effects of crystalline and thin-film photovoltaic cells and modules manifest themselves as a fill-factor reduction on the measured I-V curve
Compared to the theoretical, the IV-curve of a real solar cell with series and shunt resistances is shifted closer to the origin of the coordinate system. This is because the panel series resistance, Rs 110, causes the current values to become smaller, while the shunt resistance, Rsh 102, reduces the voltage.
According to the present invention, a solar panel electronic controller, for example an array converter, periodically measures its associated photovoltaic panel's temperature, determines series and shunt resistances, and computes the fill factor. I some embodiments the system also maintains a long term trend of these parameters. The computations are based on periodically sampled I-V curve and module temperature measurements. These measurements are correlated with a model of the photovoltaic module. The model is calibrated with I-V curve data points measured by an I-V tester by the panel manufacturer at the time of manufacturing.
The data taken makes it possible to identify and diagnose problems within the photovoltaic module. Anomalies in the I-V curve are used to differentiate between performance degradation caused by ageing effects of the photovoltaic module, damage, and external factors causing degradation, such as soiling and shading.
The ability to remotely monitor and appropriately respond to problems found opens the possibility for performance yield guarantees as yield reductions caused by the solar plant can be detected and liabilities assigned to the solar plant owner vs. the photovoltaic module manufacturer.
Long term data harvesting of the fill factor data of thousands of field deployed panels can be used to improve long term predictability of photovoltaic module degradation. This information can be used to fine tune warranty reserves. It may also be used as a marketing tool to substantiate the long term energy yield of a specific photovoltaic module.
As disclosed in U.S. patent application Ser. No. 12/061,025, a direct current to pulse amplitude modulated (“PAM”) current converter, denominated a “PAMCC”, is connected to an individual source of direct current. The PAMCC receives direct current and provides pulse amplitude modulated current at its output. When the PAMCC's output is connected in parallel with the outputs of similar PAMCCs an array of PAMCCs is formed, wherein the output pulses of the PAMCCs are out of phase with respect to each other. An array of PAMCCs form a distributed multiphase inverter whose combined output is the demodulated sum of the current pulse amplitude modulated by each PAMCC. A PAMCC is also termed an “array converter.”
By way of discussion, an array converter will be assumed to be employed to control a solar panel, though the method of the present invention may be practiced with any electronic module that includes the ability to measure voltage and current, to determine the temperature of the panel, to communicate with other similar modules in a common system, and to control a photovoltaic panel. Some embodiments further include the ability to communicate to a central office, for example via an internet connection or a radio transmission.
According to the present invention, array converter modules are mechanically attached and electrically connected to individual photovoltaic modules. The combined assembly is subjected to I/V flash testing. The values of Voc, Isc, Vmp, Rs, and Rsh parameters are recorded during the flash test of the module. These module-specific parameters as well as temperature coefficients and maximum tolerable reverse bias voltage for the specific combination of cell and module technology are programmed into the array converter's non-volatile memory before the unit is shipped.
A maximum power point tracking algorithm (“MPPT”) periodically adjusts the current drawn from the photovoltaic module such that the module voltage times current product (VI) is maximized. The MPPT algorithm also makes an initial estimate of where the voltage for maximum power, “Vmp”, should be based on temperature, current, and the manufacturing parameters for the instant module and other factors. The accuracy of the Vmp estimate is primarily determined by the accuracy of the temperature sensing circuit.
The value of Vmp changes over time due to long-term ageing effects of the photovoltaic cell and deterioration of the module laminates and interconnect. In one embodiment, to calibrate the corresponding changes to Vmp over time, the array converter maintains historical data correlating Vmp to module temperature. This correlation is done based on local observations within the photovoltaic module associated with its attached array converter as well as comparative correlation with neighboring photovoltaic modules in the same system. In some embodiments the data is communicated to a remote facility, wherein the remote facility has means to store the historical data and perform analysis using the stored data.
In a useful solar cell, Rs 110 must be small enough that useful voltage is available at Out 104 when the voltage across diode Voc 106 is near the forward conduction voltage of diode Voc 106. Typically it is desirable that no more than 1% of the current be lost to Rs 110 when operating close to the forward conduction voltage of Voc 106. For example, if current source Isc 108 were 10 Amps when illuminated to 1 sun and the forward conduction voltage of Voc 106 were 0.5V then it would be typical to design the cell so that Rs 110 is <((0.5V*1%)/10 A) or 0.005V/10 A. Rs 110 should be less than 500 micro ohms. Cells optimized for high efficiency in bright light often have lower Rs 110 values lower than 500 micro ohms.
For the reasons above, useful solar cells have Rsh 102 values that are many orders of magnitude higher than Rs 110. In the example given, Rsh 102 is a minimum of 1,000 times higher than Rs 110, given a design requirement that no more than 1% be lost due to each.
Increasing Rs 110 and decreasing Rsh 102 will also result in the knee of the I-V curve around the Vmp point to become flattened. As a result, in some embodiments relative changes in Rsh 102 and Rs 110 are determined by measuring impedance changes close to the maximum power point. Looking to
Impedance=dV/dI=(Vs−Vmp)/(Is−Imp)
where dI is determined by increasing the instantaneously drawn current in a controlled fashion and subtracting the Imp current from the new current. Since Rsh 102 is so much larger then Rs 110, Rs 110 may be neglected when a load is placed at Out 104 such that the voltage at Voc 106 is well below the forward conduction voltage of diode Voc 106. In this case changes in the load will result in changes in voltage proportional to Rsh 102 with the current through Rs 110 nearly constant and therefore the voltage drop across Rs 110 nearly constant. If one divides the change in voltage by the change in current with the application of two different loads such that the voltage at Voc 106 is well below the forward conduction point of Voc 106 then the result of the division will be close to Rsh 102.
A similar procedure is followed to compute Rsh 102, where the difference being the instantenously drawn current is decreased in a controlled fashion and impedance found using
Impedance=(dV/dI)=(Vsh−Vmp)/(Ish−Imp).
Since Rs 110 is so much smaller then Rsh 102, Rsh 102 may be neglected when a load is placed at Out 104 such that the voltage at Voc 106 is well above the forward conduction voltage of diode Voc 106. In this case changes in the load will result in changes in voltage proportional to Rs 110 with the voltage across Rsh 102 nearly constant and therefore the current through Rsh 102 nearly constant. If one divides the change in voltage by the change in current with the application of two different loads such that the voltage at Voc 106 is well above the forward conduction point of Voc 106 then the result of the division will be close to Rs 110.
The maximum power point tracking algorithm operates by the principle of “perturb and observe”. The algorithm periodically alters the current draws from the photovoltaic module and observes the voltage at the new operating point through direct measurement. This process of perturb and observe can typically be operated several hundred times each second. A control algorithm uses the periodic data measurements to find the operating point corresponding to the maximum power point. Once the algorithm has found the maximum power point, it continuously searches for changes in close proximity to the maximum power point. The voltage and current measurements required to compute impedances proportional to Rs 110 and Rsh 102 are a direct byproduct of the maximum power point tracking algorithm. The proportional impedance is averaged over a longer period of time. Average photovoltaic module performance degradation due to ageing is typically on the order of 0.5% to 1% per year. In some embodiments short term averaging of the proportional Rs 110 and Rsh 102 impedance measurements are performed within the array converter. A short term average is typically on the order of 15 minutes. Longer term averages on the order of months, quarters, or years can also be performed within the array converter. Alternatively, the short term averages can be communicated by individual array converters to a central data processing and storage facility, where the long term averages can be computed and saved.
In some embodiments the continuous energy output, temperature data, and proportional Rs 110 and Rsh 102 impedance measurements from a photovoltaic module equipped with an array converter are communicated to and analyzed by a central data processing facility. The central data processing facility maintains historical performance data for each individual photovoltaic module equipped with an array converter. If a specific module experiences an energy output reduction relative to its neighboring modules or other modules showing previous long term correlated performance, the relative degradation can be caused by a) external obstruction such as shading or soiling of the photovoltaic module or b) an ageing related degradation resulting in increasing Rs 110 or decreasing Rsh 102. Continuous measurements of proportional Rs 110 and Rsh 102 impedances provides the means to differentiate whether the differentially reduced energy output was caused by inherently degraded photovoltaic module performance or by external factors.
The values of voltage and current at the maximum power point are sensitive to Rs 110, Rsh 102 and the forward conduction characteristic of Voc 106. At this point the fact that the forward current in the junction of diode Voc 106 is really an exponential in the change of voltage across the diode and strongly influenced by temperature is most evident. In some technologies, several diode characteristics superimposed may be observed. This means that after the maximum power point has been established, its voltage and current at a particular temperature is a sensitive indicator of changes within the cell.
A degradation in photovoltaic module performance due to external obstructions caused by shading or soiling never manifests itself uniformly across all the cells in the photovoltaic module. As a result the cells in the photovoltaic module subjected to shading or soiling will produce less current than the other cells in the module. This will result in the unaffected cells trying to force the degraded cell to operate in reverse bias conditions. Since the array converter continuously measures the module temperature and is aware of the voltage related temperature coefficient for the specific photovoltaic module it can determine if some cells in the module show degraded performance, described in more detail in previously referenced U.S. patent application Ser. No. 12/335,357. The presence of partial shading or soiling can therefore be deterministically detected. An array converter equipped photovoltaic module can communicate that is has detected the presence of partial shading or soiling and report this information to the central data processing facility. If the partially degraded performance is temporal in nature it is caused by shading. If the partially degraded performance is continuous it is caused by soiling. This can be determined by the central data processing facility by analyzing the performance trend of each individual photovoltaic module.
If the maximum power point is seen to shift away from its historical norms for a given panel, the regions above and below the voltage recorded at the maximum power point may be measured by changing the load above and below the maximum power point to measure Rsh 102 and Rs 110 as previously discussed. In one embodiment, if a photovoltaic module shows signs of accelerated degradation of Rs 110 or Rsh 102 the central data processing facility issues a request for diagnostic measurements from the suspected panel. While in diagnostic measurement mode the array converter samples the complete I-V curve of the photovoltaic module. These diagnostic I-V measurements allow for the computation of the absolute Rs 110 and Rsh 102 values to confirm the exact magnitude of the Rs 110 and Rsh 102 degradation. It is also possible to perform these diagnostic measurements on a periodic basis to further improve the detection accuracy of ageing degradation relative to alarm thresholds. The drawback with diagnostic I-V measurements compared with relative impedance measurements is that diagnostic I-V measurements result in energy yield reduction during the diagnostic time period.
If Rsh 102 is found to be smaller than it once was, we conclude that the cell's leakage has increased, a definitive sign of damage within the cell.
If Rs is found to be higher than it once was, we conclude that the cell's resistance has increased or the interconnect to the cell has been degraded. Either way, this is also a definitive sign of damage.
If Rsh 102 and Rs 110 are found to be at their nominal values then we conclude that the voltage characteristic of the diode junction itself has been degraded. This is also a definitive sign of damage.
It should be noted that merely reducing the light available to a cell does not change Rs 110, Rsh 102 or the diode's conduction characteristic. The only significant effect is to reduce the current from Isc 108. This would result in a shift in the maximum power point that was on the conduction characteristic as seen through an unchanged diode conduction characteristic, a definitive sign that the cell was not damaged.
Once we have established that the cell is not damaged, we move from device defect concerns to maintenance or metrological issues. If a panel including the instant cell is producing differently than other nearby panels we conclude that this particular panel is either shaded or dirty. If power production is lower but follows the standard profile of solation of the other panels then we conclude it is dirty. If the pattern is different then we conclude the panel is shaded. If every panel has a reduced production that follows the standard profile of solation for a particular site and the production is seen to be poor independent of changes in cloud cover over time as observed by publicly available weather satellite imagery for the site, we conclude that the entire site needs to be cleaned.
From the above discussion, it is clear that the technique of measuring and comparing Rs 110, Rsh 102 and the maximum power point allows one to separate internal damage from external effects. Further, by examining the differential performance of panels in an array of panels and utilizing publicly available metrological data, one can deduce the nature of the external factor. By being able to distinguish between damage which would typically require replacement of a panel to correct and dirt or shading which would require cleaning, pruning or other maintenance actions, the correct service solution may be provided without need of additional truck rolls, expense and delays.
If any disclosures are incorporated herein by reference and such incorporated disclosures conflict in part or whole with the present disclosure, then to the extent of conflict, and/or broader disclosure, and/or broader definition of terms, the present disclosure controls. If such incorporated disclosures conflict in part or whole with one another, then to the extent of conflict, the later-dated disclosure controls.
This patent application is related to the commonly-owned U.S. utility patent application Ser. No. 12/061,025 titled “DISTRIBUTED MULTIPHASE CONVERTERS”, submitted 2 Apr. 2008 by Kent Kernahan and Sorin Spanoche and to commonly-owned U.S. patent application Ser. No. 12/335,357 titled “DETECTION AND PREVENTION OF HOT SPOTS IN A SOLAR PANEL” submitted 15 Dec. 2008 by Kent Kernahan, both incorporated herein by reference in their entirety.
