Constrained neural adaptive predictive control of SMES for dynamic performance improvement of power systems

Abstract

For dynamic performance improvement of modern power systems, the use of fast acting energy systems like superconducting magnetic energy storage (SMES) is imperative. In this paper, incorporation of a small rating SMES in a solar and wind power penetrated multi-area power system is proposed. A non-linear neural adaptive predictive controller is used to generate an optimal power command taking into account the converter rating and energy level constraints of SMES unit. The SMES is represented by a control relevant model comprising of a first order lag compensator cascaded by an integrator to translate the hardware constraints pertaining to its coil current into modified power constraints. Moreover for avoiding the sudden SMES outage, the power thresholds are forcibly varied as the SMES current reaches near its maximum and minimum values. The uprightness of the designed scheme is illustrated by simulation studies performed on a three area power system.