A methodology for neural network training for control of drives with nonlinearities

Low, T.S.; Lee, T.-H.; Lim, H.K.

doi:10.1109/41.222646

Cited by 30 publications

(6 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To obtain the data an excitation signal must be chosen. A possible approach would be to use a pseudo-random binary signal [23]. However, this signal is not the best choice for drive systems because it is filtered by mechanical time constants.…”

Section: Modelling An Electrical Drivementioning

confidence: 99%

Language identification of controlled systems: modeling, control, and anomaly detection

Martins

Dente²,

Pires

et al. 2001

IEEE Trans. Syst., Man, Cybern. C

View full text Add to dashboard Cite

Formal language techniques have been used in the past to study autonomous dynamical systems. However, for controlled systems, new features are needed to distinguish between information generated by the system and input control. We show how the modelling framework for controlled dynamical systems leads naturally to a formulation in terms of context-dependent grammars. A learning algorithm is proposed for on-line generation of the grammar productions, this formulation being then used for modelling, control and anomaly detection. Practical applications are described for electromechanical drives. Grammatical interpolation techniques yield accurate results and the pattern detection capabilities of the language-based formulation makes it a promising technique for the early detection of anomalies or faulty behaviour.

show abstract

Section: Modelling An Electrical Drivementioning

confidence: 99%

Language identification of controlled systems: modeling, control, and anomaly detection

Martins

Dente²,

Pires

et al. 2001

IEEE Trans. Syst., Man, Cybern. C

View full text Add to dashboard Cite

show abstract

“…Some slight modifications are inserted in these simulations to accomodate the training set richness and the parameter variations. In particular, following the procedure detailed in [15], random signals uniform in the interval of 10% of the reference voltages, are added to the stator voltages in order to ensure the richness of the training set in the neighborhood of the desired operating conditions. Moreover, at fixed time steps, the motor parameters are varied within a suitable designed region in the parameter space.…”

Section: The Training Set Designmentioning

confidence: 99%

Linear quadratic state feedback and robust neural network estimator for field-oriented-controlled induction motors

Marino¹,

Milano

Vasca

1999

IEEE Trans. Ind. Electron.

View full text Add to dashboard Cite

A field-oriented control scheme for an induction motor with a linear quadratic optimal regulator and a robust neural network estimator is proposed. The state feedback is designed by using the synchronous frame motor model. The number of the states is increased in order to take into account the presence of two integrators on the flux and torque errors. The resulting model is suitably simplified and the corresponding approximations are discussed. The procedure proposed is shown to be suitable also for the design of the state feedback via pole placement technique. A comparison with standard proportional integral regulators is provided. The rotor flux is estimated by using a robust neural network observer. The network training set is suitably designed in order to preserve the drive effectiveness also in the presence of large parameter uncertainties. The robust neural observer is compared with an extended Kalman filter and a standard neural network observer. Using a 250-kW induction motor as a case study, the simulation results show the effectiveness of the proposed solution, both during transient and steady-state operating conditions.

show abstract

“…To obtain the training set, a sinusoidal reference signal is imposed into the system with different amplitudes and frequencies into the speed limits of the electrical motor [8]. The training process is responsible for acquiring a good learning set.…”

Section: A Learning Process With Incomplete Informationmentioning

confidence: 99%

An experiment in automatic modeling an electrical drive system using fuzzy logic

Branco

Dente

1998

IEEE Trans. Syst., Man, Cybern. C

View full text Add to dashboard Cite

Electrical drives are usually modeled using circuit theory, with currents or linking fluxes chosen as state variables for its electrical part and rotor speed or position chosen for its mechanical part. Often, its internal structure contains nonlinear relations difficult to model as dead-time, hysteresis, and saturation effects. On the contrary, if the available model is accurate enough, its parameter values are generally difficult to obtain and/or be estimated in real time. Therefore, this paper investigates the use of fuzzy logic for automatic modeling electrical drive systems. An experimental system composed by a DC motor supplied from a DC-DC converter is used. We underline the unsupervised learning characteristics of the fuzzy algorithm, its memory and generalization capabilities. Some learning situations with critical effects in model performance are presented and discussed, pointing out some results and conclusions concerning the fuzzy modeling process in practice.

show abstract

A methodology for neural network training for control of drives with nonlinearities

Cited by 30 publications

References 8 publications

Language identification of controlled systems: modeling, control, and anomaly detection

Language identification of controlled systems: modeling, control, and anomaly detection

Linear quadratic state feedback and robust neural network estimator for field-oriented-controlled induction motors

An experiment in automatic modeling an electrical drive system using fuzzy logic

Contact Info

Product

Resources

About