V.P. Lukic and A. Trehan (USA)
Area power system control, Adaptive control, Neural Networks, Radial Basis Functions Network.
This paper presents an approach to a real time adaptive control, an integral adaptive control, in a multi-area power system configuration. The method is based on the artificial neural network theory and specifically using a Radial Basis Functions Neural (RBFN) network. Nonlinear synchronizing tie line power function is monitored and an adaptive controller is tuned in real time to maintain a desired system response, and over a wide range of operating conditions. Integral channel gains in the areas are expressed in terms of system structure that includes tie line power, a function of synchronizing coefficients , area angles and voltages and their variations. The gains are calculated to keep constant the system eigen values, at the desired level as all of the variable change due to change of operating state. An integral controller is then designed as a RBFN network that calculates the gains continuously with the inputs into the RBFN calculated in real time. The outputs of the RBFN network are the integral controller gains. The neural network, once trained by an appropriate set of input output patterns, training set, will continuously , in real time , adjust the gain and keep the system dynamics as prescribed under wide range of operating states, practically, at any disturbance scenario. The subject of an adaptive real time controller for two area power system is presented in two parts. Part I considers general problem formulation and solution. A companion paper, Part II considers RBFN network regulator design and a numerical example.
Important Links:
Go Back
