Measurement sensors, such as current sensors (CSs), are key components of any advanced AC motor drive. To ensure the highest quality and safety of the drive, CS faults, as well as significant measurement disturbances, must be detected and compensated for. This research presents the current sensor fault-tolerant control (CS-FTC) system using innovative extended Kalman filter (d-EKF) for induction motor (IM) drive. The parametric uncertainty correction factor estimated by d-EKF allows us to improve the quality of stator current estimation. The presented d-EKF not only performs CS fault compensation, but also provides reference signals to the residuum-based fault detector. The proposed CS-FTC enables the proper and stable speed regulation in the vector-controlled drive, also in a post-fault mode. An experimental validation of this solution, carried out over a wide range of speed and load changes in post-fault operation, confirmed the high quality of the stator current reconstruction and demonstrated the advantages and limitations of using the proposed d-EKF in the CS fault detection algorithm for the most frequently occurring types of CS faults.
The presented dataset contains the results of simulation and experimental studies of a current sensor fault-tolerant control system (CS-FTC) for an induction motor drive. The system is based on an Extended Kalman Filter, which enables the estimation of stator current, rotor flux, and a coefficient representing overall changes in motor resistance (d).
File description: The files are organized into two folders.
In the folder "KF_comparison", the results of simulation studies (Matlab/Simulink 2024b) are collected, focusing on the performance of stator current estimators: the Kalman Filter (KF), the Kalman Filter extended with the general resistance variation coefficient d (dEKF), and the Kalman Filter extended with both stator and rotor resistances (ddEKF).
The studies were carried out for varying stator and rotor resistances, with a sampling period of 1/8000 s (folder "dynamic"). Resistance changes were applied exclusively in the motor model.
File naming convention:
m100_w100_R_H_dr1.25_ds1.25.xlsx means:
m100 – load equal to 100% of the rated value,
w100 – speed equal to 100% of the rated value,
R – generator operation mode (M denotes motor operation),
H – all current sensors healthy (A denotes a fault in phase A sensor, B denotes a fault in phase B sensor),
dr1.25 – rotor resistance equal to 125% of the rated value,
ds1.25 – stator resistance equal to 125% of the rated value.
Variable naming convention in files:
iA_est_KF, iB_est_KF – phase A/B current estimated by KF
iA_mea_KF, iB_mea_KF – phase A/B current measured in simulations with the KF estimator
iA_est_dEKF, iB_est_dEKF – phase A/B current estimated by dEKF
iA_mea_dEKF, iB_mea_dEKF – phase A/B current measured in simulations with the dEKF estimator
iA_est_ddEKF, iB_est_ddEKF – phase A/B current estimated by ddEKF
iA_mea_ddEKF, iB_mea_ddEKF – phase A/B current measured in simulations with the ddEKF estimator
In the folder "dynamic", experimental data of the CS-FTC system operation under dynamically changing speed and load torque conditions are collected. The stator current is estimated using dEKF, and current sensor faults of the complete signal loss type are considered.
The file "dyn_load_w100.xlsx" contains waveforms for a constant speed equal to 100% of the rated value and a variable load in the range of 30–100% of the rated load. A fault in the phase B sensor occurs at t = 9 s, and a fault in the phase A sensor occurs at t = 20 s.
The file "dyn_speed_m30.xlsx" contains waveforms for a constant load equal to 30% of the rated value and a variable speed in the range of 1–100% of the rated speed. A fault in the phase A sensor occurs at t = 5.5 s, and a fault in the phase B sensor occurs at t = 15 s.
Variable naming convention in dynamic files:
time – time
iA_EKF, iB_EKF – phase A and B currents estimated by dEKF
iA_mea, iB_mea – phase A and B currents measured by faulty sensors
w_ref, w_mea – reference and measured speed
tL_ref, tem – reference load torque and torque estimated based on dEKF
d_EKF – resistance variation coefficient estimated by dEKF
iA_mea_KF, iB_mea_KF – phase A/B current measured in simulations with the KF estimator.
(2026-02-11)
