| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 1550877 | Solar Energy | 2012 | 15 Pages |
Correct modelling of solar photovoltaic (PV) system yields is necessary to optimize system design, improve reliability of projected outputs to ensure favourable project financing and to facilitate proper operations and maintenance. An improved methodology for fine resolution modelling of PV systems is presented using module short-circuit current (Isc) at 5-min time-scales, and clearly identifies pertinent error mechanisms that arise when working at this high resolution. This work used a modified version of the Sandia array performance model, and introduces new factors to the calculation of Isc to account for identified error mechanisms, including instrumentation alignment, spectral, and module power tolerance errors. A simple methodology was introduced and verified where specific module parameters can be derived solely from properly filtered performance time series data. In particular, this paper focused on methodologies for determining the predicted Isc for a variety of solar PV module types. These methods of regressive analysis significantly reduced the error of the predicted model, and demonstrate the need for this form of modelling when evaluating long term PV array performance. This methodology has applications for current systems operators, which will enable the extraction of useful module parameters from existing data in addition to more precise continuous monitoring of existing systems, and can also be used to more accurately model and optimize new systems.
► Incorporates data from a large and diverse outdoors test field. ► Presents novel methods of deriving performance parameters from time series data. ► Introduces error mechanisms and corrections for high temporal resolution modeling. ► Presents model validation, showing decreased errors through the use of the proposed modeling technique. ► Demonstrates the use of regressive analysis on system performance monitoring and design.
