An Enhanced Fault Tolerant Control Against Current Sensor Failures in Induction Motor Drive by Applying Space Vector

In this paper, an enhanced active fault-tolerant control (FTC) is proposed to solve a current sensor failure in the induction motor drive (IMD) using two current sensors. The proposed FTC method applies only one observer to diagnose the faults and recon gure the control signals by the space stator current. The diagnosis function is made up of a comparison algorithm between the measured current space vector and the estimated space vector. Then, incorrect feedback stator currents are replaced by the estimated values in the recon guration function. The amplitude of a healthy measured current is applied to adjusted the accuracy of estimated current signals. The IMD uses the eld-oriented control (FOC) technique to control the speed and torque. The e ectiveness in stabilizing the IMD system when a current sensor error occurs is veri ed by various simulations in the Matlab-Simulink environment.

Abstract. In this paper, an enhanced active fault-tolerant control (FTC) is proposed to solve a current sensor failure in the induction motor drive (IMD) using two current sensors. The proposed FTC method applies only one observer to diagnose the faults and recongure the control signals by the space stator current. The diagnosis function is made up of a comparison algorithm between the measured current space vector and the estimated space vector. Then, incorrect feedback stator currents are replaced by the estimated values in the reconguration function. The amplitude of a healthy measured current is applied to adjusted the accuracy of estimated current signals. The IMD uses the eld-oriented control (FOC) technique to control the speed and torque. The eectiveness in stabilizing the IMD system when a current sensor error occurs is veried by various simulations in the Matlab-Simulink environment.  Among many control methods, the eld-oriented control (FOC) algorithm is considered as a typical representation by its robustness and precision in controlling speed and torque of the IMD [1]. In this paper, we proposed an upgrading fault tolerance control against the total failure of current sensors in the IMD system, which applies two current sensors. In this method, the only current-observer is applied to detect the failure of current-sensors. Besides, the healthy-current also is used to adjust the correction of the es- Based on IM's model, the FOC method is applied to control the IMD. The primary principle in the FOC technique is a separate control between torque and rotor ux at the same time. In the rotating coordinate system [x, y], the stator current space vector is separated into two perpendicular elements, as shown in Fig. 1. By the way, the rotor ux is controlled by the i Sx , and the electrical torque is controlled by the i Sy [9,10].
The stator current signals in a three-phase coordinate system [a, b, c] can be transformed into a two-phase coordinate system [α, β] by Clarke's formulas as [10]: Here, i c = −i a − i c .
Here, γ is the angle of rotor ux.
By applying the FOC technique, we can control the electrical torque and the rotor speed by the relationship with two [x, y] current components, as the below formulas: From the above dynamic model, the control structure of IMD based on the FOC technique is described, as in Fig. 2. Here: "R x " is the block of the integrator, "Ri x " is the block of PI controller, "BVN" is the block of vector rotation, and "BZV" is the block of Decoupling.
The feedback signals from two current sensors and the speed sensor are used to estimate the modulus of magnetizing current and the angle of rotor ux by the Current-model of IM [11,12], as in (7)- (9).
Then, the angle of rotor ux is used to transform  measured-current, the failure of each current sensor can be located, as shown in (14).
Finally, the wrong current signal of the sensor is replaced by estimated-currents of the suitable observer.

2.3.
Fault-tolerant control against current sensor faults by the proposed space vector method As mentioned above, the feedback current signals play a decisive role in controlling the IMD.
In normal operation mode, the feedback signals are provided by two current sensors. When one failure of the sensor occurs, these signals are incorrect, which leads the FOC controller can work wrongly. Therefore, it is necessary to have an FTC system integrated into the control unit to ensure the system operates stably and reliably.
The estimated current can be derived from the voltage signals, the rotors speed, and the IM's parameter [13,14] as follows: Here The estimated algorithms are improved the accuracy by addition of two correction-factors, as follows: The deviation between measured and estimated current is calculated, as below: Here: trol structure of the IMD is the same in both two FTC methods.
The IMD has been operated according to the reference rotor speed in two ranges: the nominal speed and low-speed, as shown in Fig. 6(a) and Fig. 7(a) also shown in Fig. 6(b). Similarly, the reference speed is set at 10% and 5% of the rated speed in the low-speed area as Fig. 7(b). The IMD applying the FOC controller operates stably during the steady-state and transient state in both cases above.
When a total failure of the current sensor occurs at t = 2.0 sec, the value of feedback A phase current back to zero that leads the FOC controller operates wrong. As a result, the IMD becomes unstable and collapses, as shown in Fig.   8 and Fig. 9. Therefore, the purpose of FTC methods is to maintain the stability of the drive system, even under fault operating conditions.
In the rst case, the FTC function of the axes transformation method against a current sensor fault is simulated by Matlab/Simulink. Fig.   10(a) and Fig. 11(b) depicted a failure of Aphase current occurrence, and feedback current signal equals to zero. The fault-location function set to a high level to detect the false as in Fig. 10(b) and Fig. 11(b). The rotor speed slightly uctuates in a short time, and after the incorrect current signals are replaced by the es- timated value, the IMD works stably again, as in Fig. 10(c) and Fig. 11(c).
Next step, the simulations have been implemented to demonstrate the eectiveness of the proposed FTC method. Similar to the above case, Fig. 12(a) and Fig. 13(a) shows a current sensor fault with A-phase. After the current fault occurrence, the fault diagnosis of the proposed method works immediately, as in Fig.   12(b) and Fig. 13(b). The wrong signals are isolated and replaced by the estimated current, which is corrected by the healthy current sensor. Fig. 12(c) and Fig. 13  "This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium provided the original work is properly cited (CC BY 4.0)."