Comparison of Uncertainties in Membership Function of Adaptive Lyapunov NeuroFuzzy-2 for Damping Power Oscillations

Introduction

In a large interconnected power system, the machine angles of a group of machines oscillate against the others due to the imbalance of energy conversion resulting from a three phase fault, load outage and/or any other discrete event. These oscillations either decay with time restoring the system to its original or a new steady-state, depending upon the nature of the fault, or they may grow indefinitely resulting in system collapse. In the former case, the system has sufficient natural damping to diminish the effect of these oscillations with time, whereas, in later case the system cannot damp these oscillations and additional damping control is needed for secure and reliable operation (Mohagheghi, Venayagamoorthy and Harley, 2006; Mohapatra, Panda and Satpathy, 2013). Flexible AC Transmission System (FACTS) is a relatively new technology which optimizes the performance of a power system through one of the many ways like changing the effective reactance of transmission line, exchange of reactive power with transmission network and phase control of injected voltage (Obulesu, et al. 2009; Varma, 2011) etc.

STATCOM is a shunt connected controller, belonging to the second generation of FACTS, which maintains the bus voltage at a constant level by injecting or absorbing reactive power. Due to its efficient performance, STATCOM found extensive applications to optimize power system operation (Abido, 2005; Panda and Padhay 2008; Ganesh, Dahiya and Singh, 2014). Since, the primary objective of the STATCOM is voltage regulation, it may not perform well or even degrade the system performance in terms of damping inter-area oscillations (Gharaveisi, 2009). In order to damp the low frequency electromechanical oscillations, STATCOM is facilitated with some Supplementary Damping Control (SDC) in addition to the voltage control. The supplementary control must provide the STATCOM with the complete information of the system to perform efficiently. Due to highly nonlinear, multivariable and time varying nature of power system, the exact modeling of each component is almost impossible especially those which are not well-defined like different loads and distributed generations (Hiskens and Alseddiqui, 2006). This gives rise to the uncertainties in power system. Therefore, SDC must exploit some technique which could encounter the effect of the uncertainties in the system (Badar and Khan, 2012, 2013a).