A synchronous generator stabilizer design using neuro inverse controller and error reduction network

A neuro power system stabilizer (PSS) is developed for multimachine power systems. Each machine is identified in its inverse relation by an artificial neural network (inverse dynamics neural network (IDNN)) offline, which is used as a local inverse controller. The control error due to the interactions between generators is predicted and compensated through another network called the error reduction network (ERN). The ERN consists of several IDNNs in a linear combination form. In most neurocontrollers, two neural nets are required, one for system emulation, the other for control. In the proposed controller, the only network requiring training is the IDNN. Simulations are performed on two typical cases: an unstable single-machine power system of nonminimum phase, and a multimachine power system