Search

A unified unsaturated-reluctance-based wide speed range sensorless control method for switched reluctance machines (SRMs) is proposed in this article. First, the sensitivity analysis of the estimated position to the possible reluctance calculation error is analyzed. This analysis reveals the feasible regions with low sensitivity and low risk of entering saturation. Then, the unsaturated reluctance in these feasible regions is obtained and modeled. A simple second-order rational expression is proposed to fit the highly nonlinear unsaturated reluctance curve. With this simple model, the rotor position can be analytically calculated only by solving a simple quadratic problem. Based on the abovepresented analysis, a wide-speed range sensorless control is developed. The rotor position is universally estimated from the real-time reluctance value and the proposed analytical model. The high-frequency voltage injection and normal conduction phase are combined to calculate the phase reluctance over the wide-speed range. The concerns of calculating reluctance at the low-speed range are also discussed. A 5.1 kW four-phase 8/6 SRM is used to verify the proposed method. Various steady state and transient state experimental results prove the effectiveness of the proposed sensorless control method.

This paper presents as a contribution a fast and systematic approach for the estimation of mechanical parameters of switched reluctance motor drives based on the particle swarm optimization (PSO) algorithm. The proposal is straightforward, successfully avoiding the use of traditional cumbersome and time-consuming tests, sophisticated observers and even the need to design additional controllers. Considering the data from only a current controlled step response, the procedure enables fast offline parameter estimation, based on a minimum squared error metric. Experimental results are provided to demonstrate the effectiveness of the proposed approach, showing that the estimated parameters can be satisfactorily used for control design and lead to suitable simulation responses that match the measured experimental data.