This paper presents an approach for adaptive speed control of large induction motors by means of electric frequency variations. In the proposed method, the equations and the parameters of the motor and load are considered unknown. An adaptive scheme, using feedforward neural networks, is used to estimate a function that is applied as feedback. The neural network is trained in two sessions:off-line and on-line phases. Initially, the system is put into operation and some measurements are taken from the inputoutput response. These measurements are used to train a feedforward neural network with the system inverse dynamics. After training, the neural network is connected to the system, as feedback control, to calculate the actuating signal that produces the desired load velocity. During normal operation, the input-output response is sampled and some training sessions are applied to the neural network to accommodate possible variations in the parameters and the model of the set composed by the induction motor and the load.Some preliminary simulation results have shownthat feedforward neural networks are able to assimilate the inherent non-linearities of the motor-load set, allowing the application of a nonlinear control procedure.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.