Consiglio Nazionale delle Ricerche

Tipo di prodottoArticolo in rivista
TitoloSensorless Direct Torque Control of an Induction Motor by a TLS based MRAS Observer with Adaptive Integration
Anno di pubblicazione2005
FormatoCartaceo
Autore/iM. Cirrincione, M. Pucci
Affiliazioni autoriISSIA - CNR
Autori CNR e affiliazioni
  • MARCELLO PUCCI
Lingua/e
  • inglese
AbstractThis article presents a new speed and flux estimation algorithm for high-performance direct torque control (DTC) induction motor drives based on model reference adaptive systems (MRAS) observers using linear artificial neural networks (ANNs). Two completely new improvements of MRAS speed and flux observers are presented here: the first is a solution to the open-loop integration problem in the reference model, based on the voltage model of the induction machine, by means of a new adaptive neural integrator, the second is the employment of a new adaptation law in the ANN adaptive model, based on the total least-squares (TLS) technique. In particular, the adaptive neural integrator is based on two adaptive noise filters which completely cancel any DC drift present in the voltage or current signals to be integrated. This neural integrator does not need any a priori training of its two only neurons, adapting itself on-line. With regard to the ANN-based adaptive model, since the most suitable least-square technique to be used for training is the TLS technique, here the neuron is trained on-line by means of a TLS EXIN algorithm which is the only neural network able to solve a TLS problem recursively. Also the TLS EXIN algorithm does not require any a priori training, since it adapts itself recursively on-line. Moreover, to improve the dynamical performances of the speed loop of the drive, the adaptive model has been used as predictor, i.e. without any feed-back between its outputs and its inputs. The sensorless algorithm has been verified experimentally both on the classic DTC technique and on the DTC- SVM (space vector modulation), by adopting a proper test set-up. The speed observer has been tested in the most challenging operating conditions. The experimental results show that the dynamical performances of the sensorless drive are comparable or even better than those obtained with the corresponding DTC drives with encoders as for the medium to high-speed ranges. As for low-speed ranges, the presented sensorless DTC algorithm outcomes the performance presented in the literature for MRAS systems, thus permitting to have an accurate estimation equal or better than that obtainable with more complex observers. Finally, experimental results show that the MRAS speed observer is robust to load torque perturbations and permits zero-speed operation at no-load conditions.
Lingua abstractinglese
Altro abstract-
Lingua altro abstract-
Pagine da-
Pagine a-
Pagine totali-
RivistaAutomatica (Oxf.)
Attiva dal 1963
Editore: Pergamon, - Oxford [etc.]
Paese di pubblicazione: Regno Unito
Lingua: multilingue
ISSN: 0005-1098
Titolo chiave: Automatica (Oxf.)
Titolo proprio: Automatica (Oxf.)
Titolo abbreviato: Automatica (Oxf.)
Numero volume della rivista41
Fascicolo della rivista-
DOI10.1016/j.automatica.2005.06.004
Verificato da refereeSì: Internazionale
Stato della pubblicazione-
Indicizzazione (in banche dati controllate)
  • ISI Web of Science (WOS) (Codice:000232674300001)
Parole chiaveElectrical drives; Induction motor; Direct torque control; Sensorless drives; Model adaptive reference systems; Total least-squares; Artificial neural networks
Link (URL, URI)-
Titolo parallelo-
Licenza-
Scadenza embargo-
Data di accettazione-
Note/Altre informazioniArticolo pubblicato su prestigiosa rivista scientifica con impact factor pari a 3.944
Strutture CNR
  • ISSIA — Istituto di studi sui sistemi intelligenti per l'automazione
Moduli/Attività/Sottoprogetti CNR
  • SP.P03.011.001 : Convertitori, attuatori e azionamenti elettrici
Progetti Europei-
Allegati

Dati storici
I dati storici non sono modificabili, sono stati ereditati da altri sistemi (es. Gestione Istituti, PUMA, ...) e hanno solo valore storico.
Area disciplinareElectrical & Electronics Engineering
Area valutazione CIVRIngegneria industriale e informatica
Rivista ISIAUTOMATICA [17974J0]