Spatio-Temporal Analysis and Forecasting of Distributed PV Systems Diffusion: A Case Study of Shanghai Using a Data-Driven Approach

Abstract

In recent years, distributed photovoltaic (PV) systems have witnessed rapid development worldwide. Nevertheless, the diffusion of distributed PV systems in a specific region is still indefinite and hard to predict, which bring uncertainties to the planning and operation of electricity distribution network. This paper investigates the diffusion tendency and forecasting approach of distributed PV systems from macro- and micro-aspects. Macroscopic analysis includes spatial clustering of PV systems and quantitative analysis of PV adoption drivers in the time-dimension. Shanghai, Pudong in China is studied in this paper to offer some insights. Analysis reveals that the capacity and location of PV systems are clustered. These clusters continuously spread to the surrounding with changes of size and location, under the impact of internal and external factors. This indicates that diffusion of PV systems can be simulated by a cellular automation model. For microscopic analysis, a data-driven forecasting approach of PV diffusion is proposed based on cellular automation. Analysis shows that the developing state of PV cells can be forecasted based on multi-source datasets. Besides, statistical distribution of newly installed PV capacity per cell tends to be stable, so that it can also be considered as a predictor of distributed PV systems diffusion.