Multivariate Time Series Models Applied to the Assessment of Energy Storage in Power Systems

This paper introduces the infeed modeling requirements for investigations on energy storage in power systems with complex spatial dependence structures between stochastic infeeds. It is shown that synthesizing multivariate time series, including correct modeling of persistence and spatial dependence, is crucial for obtaining statistically relevant results. A suited new multivariate time series model is introduced, which is based on the inversion method and an appropriately identified vector autoregressive process. Application of the method can be found in planning and economic evaluation of distributed energy storage in interconnected power systems with stochastic generation. A case study shows considerable potential for stochastic generation, supported by energy storage, and illustrates the necessity for correct infeed modeling.