Intelligent performance check of pv operation using satellite data (PVSAT-2)
A low cost hardware user interface for communication between PV systems and the PVSAT-2 server based on Internet communication structures will be provided. The added hardware- monitoring device will allow for a reliable and precise measurement of the actual photovoltaic energy production. The hardware device automatically records the energy production and transfers the data to the central server once a day. The fully automatic procedure ensures a high reliability of the measurements on the one hand, because typical human errors due to meter readings not done in due time or done incorrectly are avoided. On the other hand it makes the PVSAT-2 application easy to use. There is also an added value for the set-up of the communication infrastructure. The gathering of performance information at the server and the two-way communication can later be used for new applications in grid management.
The irradiance software package will provide algorithms to derive high quality irradiance information from satellite data on an hourly base. The new model to calculate direct and diffuse irradiance will be based on a new irradiance calculation scheme for the new satellite generation MSG developed within the Heliosat3 project. Improved calculation of the diffuse irradiance will allow for a more accurate determination of the plane of array irradiance. The quality of the satellite derived irradiance values is further improved by combination with data from ground based meteorological stations using the kriging of differences method. Evaluations showed, that for monthly values the kriging of differences method reduced the RMSE to 50% of the error of satellite derive values. For daily values the RMSE to 75% of the error satellite derived values respectively. In the PVSAT project (EU JOULE-III program, JOR3-CT98-0230) it was shown that the accuracy of the overall procedure is mainly determined by the accuracy of the irradiance calculation scheme. Improvement of the irradiance accuracy will therefore directly result in a higher accuracy of the overall scheme and thus in earlier fault detection. As an additional feature, information on the accuracy of the irradiance values is required as input for the automated performance check provided by the PVSAT-2. Quality information is be obtained by distinguishing meteorological conditions that correspond with different levels of accuracy of the Heliosat method.
The footprint method developed in this project supports the error detection and error type classification of the occurred error in grid connected PV-systems. On base of a statistical analysis of simulated radiation and yield data and of monitored yield data, a comparison between the actual yield pattern and typical error pattern tables will allow to calculate probability values for a set of system errors. The identification of an error is connected to the probability values of the detected faults. Within the PVSAT2 service the footprint algorithm is integrated to the decision support system for automated error detection.