A novel battery network modelling using constraint differential evolution algorithm optimisation

Abstract

The amount of battery storage into the power system network has been increasing in the recent years. The use of battery storage devices has been advocated as one of the main ways of improving the power quality and reliability of the power system, including minimization of energy imbalance and reduction of peak demand. Higher peaks in demand will increase the electricity price and could cause blackouts and infrastructure damage. Lowering peak demand to reduce the use of carbon-intensive fuels and the number of expensive peaking plant generators is thus of major importance. Self-adaptive control methods for individual batteries have been developed to reduce the peak demand. However, these self-adaptive control algorithms of are not very efficient without sharing the energy among different batteries. This paper proposes a novel battery network system with optimal management of energy between batteries. An optimal management strategy has been implemented using a population-based constraint differential evolution algorithm. Taking advantage of this strategy the battery network model can remove more peak areas of forecasted demand data compared to the self-adaptive control algorithm developed for the New York City study case.

The energy demand estimation for Turkey using differential evolution algorithm

“…It does however have certain disadvantages as well, such as unsteady convergences and imperfect local optimum determination [29,30]. DE has been applied in many areas, such as scheduling [31], network modelling [32], image thresholding [33], fan knife designing [34], function optimization [35,36], feature selection [37] and yield curve estimation [38].…”