Title:	Validating the Model for a 250 kW Size Grid Connected PV-System in Rwanda Based on Sparse Operational Data
Author(s):	F. Habyarimana, H.G. Beyer
Keywords:	PV System, Performance Ratio, Grid Connected, Global Solar Irradiance, GeoModel, Automatic Weather Station
Topic:	OPERATIONS, PERFORMANCE AND RELIABILITY OF PHOTOVOLTAICS (FROM CELLS TO SYSTEMS)
Subtopic:	Operation of PV Systems and Plants
Event:	31st European Photovoltaic Solar Energy Conference and Exhibition
Session:	5BV.2.41
Pages:	2294 - 2299
ISBN:	3-936338-39-6
Paper DOI:	10.4229/EUPVSEC20152015-5BV.2.41
Price:	0,00 EUR
Document(s):	paper

Abstract/Summary:

---

For the assessment of the performance of this system, the available data base is comparatively sparse. Especially on site measured irradiance data and power data are available in few timeslots only. In general only monthly energy data from the system is available from June 2007 to March 2012 and the monitored energy within 30 minutes from December 2013 to April 2014 and September to December 2014. Thus a first performance analysis was based on the use of remotely sensed meteorological data (supplied by GeoModel and Automatic Weather Station). Result is that a considerable under performance ratio of the system (compared PR modeled of 79% from AWS data versus PR actual calculated of 73% from 30minutes power monitored and 71% for PR calculated from monthly energy output to 83% for PR modeled within GeoModel data) has to be stated. To get a better understanding of this under performance, the current work is focused on using both, the long-term offsite meteorological data (in addition data from an automated weather located in Rutongo, 1.88o South and 30.05o East, distance to the system at 4 km can be used) and the “spot-like” on site data to drive a simulation model for the system performance. A model structured as described in [1] is used and parameterized according to nameplates data. From this, it is tested whether an adapted simulation model (i.e. a variation of the parameter set for e.g. the rated Standard Test Conditions power) can reflect the actual system performance. This is analyzed the correlation of the measured and simulated monthly energy data and measured and simulated power data for the long term and short term data, respectively. The validity of the adaption is tested by using subsets of the data for parameter identification and other “unseen” subsets to prove that the respective model can be used to reflect the actual system performance. The parameter adoptions necessary to achieve the improved correlation of simulation and measurement are discussed in view of the deviations of the actual characteristics of the system components from their ex-factory status.