Image aided dynamic reconfiguration of SPV array under non uniform illumination

Abstract

In this paper, we studied Solar Photovoltaic (SPV) array output power as a function of dynamic illumination intensity pattern. A fine resolution pixel level analysis of 2D dynamic illumination intensity pattern is mapped to the module architecture. Images taken by high resolution cameras at regular intervals will help track the iso-insolation contours by mapping pixel pattern and estimate the illumination intensity state of the module. The aggregate intensity pattern on such an SPV module will be a mixed pattern classified as non-uniform illumination (NUI) states (BRIGHT, GREY, DARK). Optimization of a SPV array under such NUI conditions is a considerable challenge. We propose an image aided dynamic reconfiguration technique based on the classification of modules after pixelating the image of the array. The nature of the BRIGHT state and DARK state coverage on the SPV module has a significant implication on output power depending on the inter-connectivity of the PV cells (module architecture) and if bypass diodes are attached to individual cells or a packet of cells. The promise of this approach is to reduce disruption of the SPV array operation. Another significant point is the ability to predict the power. It has been demonstrated that a small change in the NUI pattern can cause a significant change in the power output. While certain individual hardware blocks have been developed, ongoing activity is focused on realization of a fully online implementation.