Improving the efficiency of grid connected PV system for real operating conditions

Di Vincenzo, Maria Carla (2012) Improving the efficiency of grid connected PV system for real operating conditions. PhD thesis, University Of Strathclyde.

[thumbnail of T13175]
Preview
PDF. Filename: T13175.pdf
Final Published Version

Download (22MB)| Preview

Abstract

This PhD thesis is focused on modelling and development of an improved Maximum Power Point Tracking (MPPT) designed for real operating conditions. Real operating conditions involve changing irradiance and temperature and also often partial shading of the array. It is also common for there to be temperature variation across the array, and also some dierences in the intrinsic quality and eciency of individual cells and modules. These eects combine to give a degree of mismatch between the cells and modules within the array that is time varying. Commercial inverters are not designed to deal with the resulting nonideal system IV curves, and thus can deliver poor MPPT performance that can degrade signicantly the overall eciency of power conversion. The novelty of this research is the development of a Maximum Power Point Tracking algorithm able to indentify accurately and rapidly the MPP under real operating conditions, and thus improve the system performance especially when the mismatch issues outlined above lead to multiple local maxima in the power output of the array (as a function of array voltage). To underpin the development of the new MPPT algorithm, a detailed model of the PV system was developed. This is built up from models of individual cells and modules so as to properly represent cell mismatch. This model has been tested and validated using real measured data from a test rig installed on the roof of James Weir Building of Strathclyde University. The test rig was equipped with comprehensive and appropriate instrumentation to measure both the ambient conditions and the PV performance. Over an extended period of monitoring a substantial amount of high quality detailed data was collected from the roof test rig, and this has been used to develop and rene an algorithm able to track the MPP highly eectively under time varying real outdoor operating conditions. The algorithm uses an Articial Neural Network (ANN) to predict the MPP in the case of partial shading and also any other operating conditions likely to be experienced; the algorithm includes additional code to assist the ANN in tracking the true maximum within a variable time step. It has been implemented on a modelled DC/DC converter to test dierent power conditions and also dierent types of modules with dierent Fill Factors. Finally, the control technique developed has been implemented in a real DC/DC converter but using an electronic PV array simulator rather than the outdoor system to provide more controlled operational conditions