Entity

Time filter

Source Type


Gao W.,Naval Bengbu Petty Officer Academy | Mao W.,Naval Bengbu Petty Officer Academy
Proceedings - IEEE 2011 10th International Conference on Electronic Measurement and Instruments, ICEMI 2011 | Year: 2011

Because sing traditional static neural network coping with continuous-time dynamic time may produce unsatisfactory control effect, a dynamic neural network (D-FNN) was adopted to design the speed controller to control PMSM vector control system. The D-FNN input and output are sliding mode switch function, sliding mode control function, respectively. The single input and single output neural network sliding mode control was achieved using D-FNN learning capability, which is not only can fully exert the characteristics of sliding mode control (SMC) which are insensitive to parameters change and disturbance, but also has the ability of fuzzy neural self-adjusting. The simulation results show that the proposed control scheme has stronger robustness. © 2011 IEEE. Source


Gao W.,Naval Bengbu Petty Officer Academy | Guo Z.,Naval Bengbu Petty Officer Academy
Journal of Networks | Year: 2013

In the speed sensorless vector control system, the amended method of estimating the rotor speed about model reference adaptive system (MRAS) based on radial basis function neural network (RBFN) for PMSM sensorless vector control system was presented. Based on the PI regulator, the radial basis function neural network which is more prominent learning efficiency and performance is combined with MRAS. The reference model and the adjust model are the PMSM itself and the PMSM current, respectively. The proposed scheme only needs the error signal between q axis estimated current and q axis actual current. Then estimated speed is gained by using RBFN regulator which adjusted error signal. Comparing study of simulation and experimental results between this novel sensorless scheme and the scheme in reference literature, the results show that this novel method is capable of precise estimating the rotor position and speed under the condition of high or low speed. It also possesses good performance of static and dynamic. © 2013 ACADEMY PUBLISHER. Source


Gao W.,Naval Bengbu Petty Officer Academy | Hua X.,Naval Bengbu Petty Officer Academy | Guo Z.,Naval Bengbu Petty Officer Academy
Communications in Computer and Information Science | Year: 2012

To study the speed estimation of permanent magnet synchronous motor (PMSM) sensorless vector control system, model reference adaptive system (MARS) and radial basis function neural network (RBFN) were adopted. The method was the organic integration of the RBFN and MRAS, selecting the PMSM as the reference model, while selecting current model of PMSM as an adjustable model. Only using the difference of the q-axis estimated current and q-axis actual current as the error signal, the error signal was transported to the regulation of RBFN to gain the estimated speed. With comparative study of simulation between this novel sensorless scheme and the reference scheme, the results show that this method presented in this paper is capable of precise estimating the rotor speed under the condition of high or low speed and achieve better static and dynamic performances. © 2012 Springer-Verlag. Source

Discover hidden collaborations