Optimal Planning of Grid Scale PHES Through Characteristics-Based Large Scale Data Clustering and Emission Constrained Optimization

In today&rsquo
s modern power system, the proportion of renewable energy generation is increasing. The inherent frequent variability of these energy sources creates a power balance and frequency stability problem within the power system. Planning energy storage technologies for the mitigation of this fluctuation requires an analysis of large datasets whose competition is difficult as it increases the computation burden due to the increased variable size of the dataset. The generation of wind energy scenarios based on two notable wind energy generation characteristics and the use of representative data for the generated scenarios is proposed for the optimal sizing of energy storage tools. The IEEE-30 bus system with a one year hourly average wind data of the Northern Ireland wind resource was considered for the sizing of a pumped hydro energy storage (PHES) system. Fifteen data sets were generated and used in the emission constrained optimal sizing process using code written in MATLAB R2017a and particle swarm optimization (PSO) was used as the searching algorithm. The result proves that data grouping based on the combined average and variation method gives a better optimal storage size.