Probabilistic analysis of hybrid energy systems using synthetic renewable and load data

Abstract

This paper focuses on probabilistic analysis of hybrid energy systems (HES), which integrate multiple energy inputs and multiple energy outputs for effective management of variability in renewable energy and grid demand. To characterize the volatility, a statistical model combining Fourier series and autoregressive moving average (ARMA) is used to generate synthetic weather condition (e.g., wind speed) and grid demand data. Specifically, Fourier series is used to model the seasonal trends in historical data, while ARMA is applied to characterize the autocorrelation in residue time series (e.g., measurements with seasonal trends subtracted). The synthetic data is shown to have same statistic characteristics with historical measurements, but possesses different temporal profile. The probabilistic analysis of a particular HES configuration is then performed, which consists of nuclear power plant, wind farm, battery storage, and desalination plant. Requirements on component ramping rate, and the effects of deploying different sizes of batteries in smoothing renewable variability, are all investigated.