A statistical approach to electrical storage sizing with application to the recovery of braking energy

In the context of efficient energy use, electrical energy in electric drives plays a fundamental role. High efficiency energy storage systems permit energy recovery, peak shaving and power quality functions. Due to their cost and the importance of system integration, there is a need for a correct design based on technical-economical optimization. In this paper, a method to design a centralized storage system for the recovery of the power regenerated by a number of electric drives is presented. It is assumed that the drives follow deterministic power cycles, but shifted by an uncertain amount. Therefore the recoverable energy and, consequently, the storage size requires the optimization of a random cost function, embedding both the plant total cost and the saving due to the reduced energy consumption during the useful life of the storage. The underlying stochastic model for the power profile of the drives as a whole is built from a general Markov chain framework. A numerical example, based on Monte Carlo simulations, concerns the maximization of the recoverable potential energy of multiple bridge cranes, supplied by a unique grid connection point and a centralized supercapacitor storage system.

► Recovery of braking power produced by multiple electric drives.
► Temporal power profile modeled through the multinomial distribution and Markov chains.
► Storage sizing via random cost function optimization.
► The search region for the optimization is given explicitly.
► The value of energy recovered during the useful life of the storage outweighs its cost.