Use of Time Series Load Data to Size Energy Storage Systems

Vast amounts of newly-available time series measurements of electrical loads now enable data-driven optimization of energy storage systems. This paper presents an algorithm for sizing the installed capacity of an Energy Storage System (ESS) with the objective of minimizing the total energy cost for a load having a specific time series profile. This algorithm considers not only energy costs, but also demand charges, which often represent a significant portion of electric cost for commercial users. The sizing process-implemented using linear programming-takes into account ESS characteristics and costs as well as time-of-day electricity rates. This paper discusses optimization results for two case studies-a commercial property and a railroad traction substation-using the time series of hourly load profiles for one year. The algorithm successfully identified the optimal ESS capacities for both cases, yielding cost savings of 10% and 28% respectively. The algorithm also shows that a single ESS serving the aggregated loads (assuming they are co-located) would further increase the economic benefit of the ESS installation, while also being 10% smaller in capacity compared to the total sum of optimal capacities for the two individual cases. This algorithm can be used by practitioners to analyze scenarios in which two applications with different load profiles can create synergies in ESS installations, particularly for commercial businesses.