Synchronous machines parameter estimation using artificial neural networks

Economic factors are constantly pushing the operation of power systems to their optimal capacity. For this to be possible, sophisticated models and accurate parameters of the electric components of the systems are required. This work presents an alternative approach that can be used to deal with the parameter estimation problem.; This thesis presents a new on-line based method to estimate the electric parameters of synchronous generators. The solution has been devised similar to that of solving a pattern recognition problem but in a manner which is an entirely new concept in the field.; The strategy is based on the following concept. A synchronous machine operating under different boundary conditions will have a unique response determined by its physical characteristics, which in mathematical terms are expressed by its electric parameters. Alternatively, given the behaviour of the synchronous generator, expressed in currents, voltages, and rotor position, it is possible to think of the existence of an inverse model that will be able to provide the parameters of the machine under consideration. This is a pattern classification problem.; The inverse model is obtained using a feedforward artificial neural network which has excellent properties as a pattern classifier. Artificial neural networks have been used in many areas of power systems, but this is the first time that they have been applied for parameter estimation purposes.; The proposed method has the capability to estimate all the electric parameters of a synchronous generator that govern its steady state and dynamic behaviour, including the saturation characteristics. The method has been extensively tested in a salient pole microalternator. On-line data from a round rotor generator has been used to successfully estimate the steady state characteristics of the synchronous machine and simulation studies have confirmed that the proposed method can also be applied to estimate the dynamic characteristics of round rotor synchronous machines.