Online Identification of Permanent Magnet Flux Based on Extended Kalman Filter for IPMSM Drive With Position Sensorless Control

Determination of Torque Distribution Ratio for Electric Bicycle with Independently Driven Front and Rear Wheels

“…The Kalman filter is promising technique to identify system parameters (13)- (15) . The linear discrete-time state-space equations of the difference of each wheel's slip ratio are given in (30)-(36).…”

Parameter Estimation for Condition Monitoring of PMSM Stator Winding and Rotor Permanent Magnets

“…However, it is known in system identification theory that it is impossible to ensure estimation accuracy if the estimation is based on a rank deficient reference/variable model, and the steady state dq-axis equation of PMSM shows that it is a rank deficient equation for simultaneously estimating winding dq-axis inductances, resistance and rotor flux linkage. Therefore, as detailed in [3], [22], [30] and [34], it is impossible to simultaneously estimate winding resistance and rotor flux linkage using only one set of PMSM states. Taking the method of [1] for example, it did not take into account the rank of the utilized reference model, and its estimator failed to estimate the stator winding resistance due to ill-convergence.…”

“…Taking the method of [1] for example, it did not take into account the rank of the utilized reference model, and its estimator failed to estimate the stator winding resistance due to ill-convergence. For solving this rank deficient problem, some literature [2]- [6], [32]- [34] proposes to estimate each parameter separately by fixing other parameters to their nominal values to get a full rank reference/variable model. However, what could be worse is that if the winding resistance and rotor flux linkage are not simultaneously estimated but estimated separately, the estimation of the two parameters will converge to the wrong points [22] due to the mismatching of un-estimated parameters.…”

GPU Implementation of DPSO-RE Algorithm for Parameters Identification of Surface PMSM Considering VSI Nonlinearity

“…In practical engineering, parameter optimization using numerical methods is an ideal technology for directly estimating the needed parameters based on regular measurable data instead of using additional measurement instruments [7]. Existing research mainly focused on online estimation algorithms including self-commissioning technique [2] ,extended Kalman filter (EKF) [9], model reference adaptive system (MRAS) [10], recursive least-squares (RLS) [3] [11], adaptive observer [12], and artificial neural networks (ANN) [10] [11]. However, with the increasing complexity of operation conditions, these methods may not always work well.…”

“…For example, in [2] a self-commissioning technique was proposed to estimate PMSM parameters under standstill .However, it cannot estimate the permanent magnet when the machine is at standstill state. In [9], an EKF algorithm was proposed to estimate the rotor speed and position of PMSM, but it may be difficult for real applications as the algorithm is sensitive to noise. The MRAS estimators proposed in [10] cannot simultaneously estimate winding resistance, inductance and rotor flux linkage accurately.…”