This paper proposes a snubber circuit for a flying capacitor multilevel inverter and converter. It also explains the concept of constructing a snubber circuit for a multilevel inverter and converter. The proposed snubber circuit makes use of an Undeland snubber as a basic snubber unit, and thus can be regarded as a generalized Undeland snubber for a flying capacitor multilevel inverter and converter. It has such an advantage of Undeland snubber used in the two-level inverter. Compared with a conventional RLD/RCD snubber for multilevel inverter and converter, the proposed snubber keeps such good features as fewer number of components, reduction of voltage stress of main switching devices due to low overvoltage, and improved efficiency of system due to low snubber loss. In this paper, the proposed snubber is applied to a three-level flying capacitor inverter, and its features are in detail demonstrated by computer simulation and experimental result.
4 November 2005 in IEE Proceedings - Electric Power Applications
A deadbeat load torque observer with a system parameter compensator that can be used to improve the performance of a permanent-magnet synchronous motor in a precision position control system is presented. The compensator makes real systems work as it in a nominal parameter system. Therefore, the observer has a performance level equivalent to that obtained if there is no parameter variation. Noise effects are reduced by the use of past-filter implimented using a moving-average process. A comparison study on the system response for a convertional deadbeat gain observer and the proposed parameter-compensated deadbeat observer is performed. The proposed system can be used to cancel out any steady-state or transient position errors created by external disturbances such as friction, load torque or chattering effects.
15 April 2021 in IEEE Transactions on Vehicular Technology
When an electric vehicle (EV) is driving at a very low or high speed, there are cases where energy recovery is not possible because the back electromotive force generated in the motor is too small or large. In this paper, we investigate the application of the regenerative electromagnetic retarder's energy recovery method to the permanent magnet synchronous motor (PMSM) of an EV. When the EV was driven downhill or coasts, L-C resonance via switched capacitor was used to control the voltage generated by the PMSM as a voltage that can recharge the battery. During regeneration, the voltage generated in the PMSM varied along with the degree of resonance through switching the duty of the proposed circuit. Through simulation, DC-Link voltage according to switching duty ratio and motor speed is plotted in a 3D map. Based on this, voltage control algorithm using PI and reverse PI duty control has been proposed. To prove the proposed algorithm, the DC-link stage of the inverter is simply equivalent to the R-C model. And, if the PMSM operates at a constant speed, the generated voltage is controlled so as to charge the battery. In addition, the discharging and charging energies of the battery are simulated when the EV with certain load torque is driving uphill and downhill. Considering the dynamics of the car, the energy recovery rate was simulated when the EV is driving uphill and downhill in a 50:50 ratio at the same slope and distance. The simulations were conducted using MATLAB Simulink, and the experiments were carried out using a motor-generator set and a small electric truck equipped with a 13 kW PMSM, respectively.
1 January 2003 in International Journal of Electronics
A very simple control approach using a neural network for the robust position control of a permanent magnet synchronous motor (PMSM) is presented. The linear quadratic controller plus feedforward neural network compensator is employed to obtain the robust PMSM system approximately linearized using a field-orientation method for an AC servo. The neural network is trained in on-line phases and a feedforward recall and error back-propagation training compose this neural network. Since the total number of nodes is only eight, this system is realized easily by the general microprocessor. During normal operation, the input–output response is sampled and the weighting value is trained multi-times by an error back-propagation method at each sample period to accommodate the possible variations in the parameters or load torque. The state space analysis is also performed to obtain the state feedback gains systematically. In addition, the robustness is also obtained without affecting overall system response. A floating-point digital signal processor DS1102 board (TMS320C31) realizes this method. The basic DSP software is used to write C-program, which is compiled by using ANSI-C style function prototypes.
20 October 2016 in THE TRANSACTIONS OF KOREAN INSTITUTE OF POWER ELECTRONICS