Using NSGA II multiobjective genetic algorithm for EKF-based estimation of speed and electrical torque in AC induction machines

Abstract

High-performance AC drives require accurate speed and torque estimations to provide a proper system operation. The selection and quality of extended Kalman fitter (EKF) covariance elements have a considerable bearing on the effectiveness of motor drive. Many EKF-based optimization techniques involve only a single objective for the optimal estimation of speed without giving concern to the torque. This paper presents a new methodology for the selection of EKF filters that uses non-dominated sorting genetic algorithm-II (NSGA-II) developed for filter element selection in order to investigate the concurrent optimization of speed and torque. The proposed optimizing technique for EKF-based estimation scheme is used in the combination with the sensorless direct torque control of induction motor. The multi-optimal based-EKF is tested with three regions of Pareto front curve.

Improved EKF-based direct torque control at the start-up using constant switching frequency

“…The gains for the PI controller and the approximated crossover frequency (f c ) are given in the following. For EKF filter initial values with assumption of white noise, the state estimation error matrix P is initiated with the diagonal matrix one, whereas the initial values of the R and Q filters in the EKF algorithm are found by using the genetic algorithm [11] to achieve a rapid initial convergence as well as the desired transient-and steady-state performance. Thus, the initial values for EKF scheme are defined by (29)-(31).…”

Sensorless Speed Estimation for the Diagnosis of Induction Motors via MCSA. Review and Commercial Devices Analysis

“…In the early 2000s, AI techniques emerged as complementary tools for MRAS and EKF closed-loop schemes aimed to overcome three of their major problems: complexity of non-linear mathematical models, instability and parameter compensation [ 80 ]. Among them, the most used for these purposes are: Artificial Neural Networks [ 9 , 10 , 12 ] (ANNs), Fuzzy Logic [ 13 , 15 ] (FL) and Genetic Algorithms [ 16 , 17 ] (GAs).…”

“…Regarding these two needs, in the last decade, an effort has been made to increase accuracy and stability using modern techniques such as neural networks, genetic algorithms, etc. [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. Nevertheless, since the possibility of working with large data records is restricted in the field of controlled AC drives, the maximum degree of accuracy is yet to be fully exploited.…”

“…SSE techniques, regardless of whether they are used for diagnostic or sensorless control, can be classified into two large groups of methods: Fundamental Model-Based (FMB) [ 1 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 ] and Magnetic Anisotropy-Based (MAB). The latter can be subdivided into two more groups of methods: Signal Injection-Based (SIB) [ 59 , 60 , 61 , 62 , 63 , 64 ] and Slotting and Eccentricity Harmonics Based (SaEHB) [ 4 , 65 , 66 , 67 , 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 ,…”

Energy Modeling and Parameter Identification of Dual-Motor-Driven Belt Conveyors without Speed Sensors

“…The basic idea is as follows: firstly, the linearized state space model of the system will be established. Then, the EKF algorithm is used to estimate the speed and total load torque of the belt conveyors [26][27][28]. Finally, the RLS algorithm is adopted to identify the energy model parameters.…”