Performance Enhancement of Induction Motor Based on Three-Level T-Type Inverter Using DTC-SVM

. Induction motors (IM) that are driven using the conventional direct torque control algorithm (DTC) suffer from vital drawbacks which are high ripples in the torque and flux. These ripples result due to the use of a look-up table and hysteresis comparators which causes variable switching frequency. Moreover, the use of a traditional two-level inverter leads to the production of low-quality voltage and current. The work presented in this paper proposed a modified control system to overcome these drawbacks. The proposed control system is based on the use of the space vector modulation-based DTC algorithm (DTC-SVM) for driving three-phase IM via a three-level T-Type inverter. The conventional DTC-SVM algorithm has been modified to match the work of the proposed inverter. The modification process was based on the mapping of the standard DTC-SVM algorithm in the space vector of the three-level inverter. The conventional and the modified control systems are implemented using MATLAB/Simulink package. The comparison of the standard and the modified DTC-SVM has been performed by simulation. Simulation results showed the superiority of the proposed algorithm in terms of reducing ripple in torque and flux and improving the quality of current and voltage supplied to the motor.


INTRODUCTION
Induction motors are the major electrical machin used in manufacturing and further applications [1]. They are popular in comparison to other motors, such as DC motors, because of their low cost, reliability, and flexibility [2,3]. Induction motors utilize more than twothirds of all industrial electrical energy [4]. Important advances in materials, power electronics, and high-performance microcontrollers are being made all the time, even though the basics of IM functioning have been known since the turn of the century. [5,6]. So with high-performance microcontrollers and advanced power electronics switches IM drives offer a competitive advantage over DC machines [7]. To comply with norms such four-quadrant operation, little torque variation, and quick speed recovery during load disturbances, highperformance drivers are used. The field-oriented control (FOC) and the Direct Torque (DTC) control method are the two most prevalent control methods for induction motors. Due to its ease of use, the DTC control algorithm has become the standard for governing induction motors. [3,5,8] Takahashi proposed the DTC method in 1986; the basis of this technique does not necessitate complex computations or conversions; It includes a look-up table and flux and torque estimators. [9,10]. DTC, on the other hand, has some significant drawbacks. It has strong torque ripples, particularly at low switching frequencies, resulting in fluctuating switching frequency and difficulty managing flux and torque at low speeds [11]. To overcome these disadvantages and increase DTC performance, a variety of approaches have been proposed. The most extensively used method for improving DTC performance, known as DTC-SVM, is based on the principle of space vector modulation [12][13][14]. However, with the presence of two-level inverters, improved ways to increase direct torque control make no contribution enough to the process of improvement. The non-sinusoidal output voltage produced by the two-level inverters contains significant harmonics distortion, which reduces drive performance [15,16]. The Multi-Level Inverter (MLI) is used instead of a conventional 2-level inverter to mitigate these drawbacks by permitting the inverter to operate at higher voltage ratings and decreasing the harmonic content of the output voltage and current waveforms. [17]. There are three common types of multi-level inverter topology the cascaded H-bridge (CHB) [18][19][20], Neutral Point Clamped (NPC) [21], and Flying Capacitor (FC) [22][23][24], the cascaded H bridge inverter is suited for photovoltaic (PV) systems. The inverter's restriction is the vast number of DC sources for the upper level. Neutral Point Clamped and Flying Capacitor inverters only utilize a single DC supply, but they still require several semiconductor components and incur significant conduction losses. [25]. Therefore, a threelevel T-type inverter has been proposed Because it has the advantages of using a single DC source, fewer components, and reduced conduction losses [26,27]. The purpose of this paper is to investigate the impact of employing the DTC-SVM on an induction motor. When comparing the use of a DTC-SVM with a 2-level inverter to the use of a DTC-SVM with a 3-level T-type inverter.   Table I  This construction is a 2-level inverter extension using a two-way switch attached to the midpoint n. This inverter's arms each have three switching states, resulting in three voltage levels. As a result, the three-phase three-level inverter will have 27 (3³) switching states, and space vector's voltage vector count being increased, as illustrated in Fig.2. There are three kinds of voltage vectors (small vectors s1-s6, medium vectors m1-m6, and large vectors L1-L6). Furthermore, increasing the number of voltage vectors allows for an increase in the number of sectors that split the space vector, as seen in Fig.2. Table II shows how voltage vectors and their associated switching states are classified. There are 8 redundant switching states and 19 active switching states [28].

DTC-SVM ALGORITHM
A major drawback of traditional DTC systems is the high frequencies and switching frequency. As a result, the augmented system will have high current harmonics, acoustic noise, and poor control performance, especially at low speeds. Because of the hysteresis comparator's discrete nature, high torque, flux ripples, and variable switching frequency remain significant even when the bandwidth is lowered. [29,30]. Due to the comparatively small bandwidth values, the inverter switching frequency and switching losses are also enhanced. To address these flaws, a group of researchers collaborated to combine the benefits DTC and FOC into a single framework, resulting in the novel DTC-SVM approach [31]. Unlike conventional DTC, which uses instantaneous values and direct computations to produce the inverter switching state, the DTC-SVM uses average values and the SVM algorithm to produce the inverter switching state [32]. The DTC-SVM technique is used to estimate the reference voltage vector, which is subsequently adjusted using the SVM technique to build inverter switches. The aim of this design is to regulate the flux and torque of an induction motor. In this method, the hysteresis control and the switching table are deleted, and instead, the proportional-integral (PI) controller and the SVM modulator are employed, as shown in Fig. 3. [8].

A. DTC-SVM Algorithm for 2-Level Inverter
The two zero voltage vectors and six active space vectors formed in the two-level inverter are used to try and switch the space vector to fit the sinusoidal reference voltage. In order to bring the reference voltage vector Vs into phase with its neighboring vectors at the lowest switching frequency, the cycle Ts must be divided into three segments: Ta, Tb, and T0. The following on-time or volt-seconds computations can be conducted using the simplified geometry in Fig. 4. [31,33]: In which Va = (1,0) and Vb = (0.5h) Taking the components of ̅ in quadrature axes yields: From (5) to (7), the following are the on-time durations that can be determined: The rest of the period will be: The resultant leg voltage for producing reference vectors in sector 1 is shown in Fig. 5. A circular path is formed by the sinusoidal reference space vector. The SVPWM produces the greatest significant output voltage at the radius of the largest circle that can fit inside the hexagon.
The DC bus voltage is denoted by Vdc. The following is the formula for calculating switching times (duty cycles) [34,35].

B. DTC-SVM Algorithm for 3-Level T-type Inverter
Utilizing the identical arrangement, the DTC-SVM algorithm for controlling IM with a 3level T-type inverter is built. According to Section II, the inverter contains eight redundant switching states in addition to 19 active switching states. The six sectors in Fig. 4 are divided so that each sector contains 4 triangles as in Fig. 6, the reference vector A0P of immensity | * |is considered to reside in Δ3 and creates an angle with the main axis α, extending the splitting of sector 1 into four triangles. As shown in Fig. 7 (a), (b) defining a little vector in the Δ3, which defines point P. The needed volt-second for the A0P is the same as for the A2P. The vector A2P's on-time calculations are identical to the two-level inverter's on-time calculations at the specified sector. It is necessary to build the triangle into which the reference voltage vector is placed in a 3-level inverter, and the components of the small vector ( 0 and 0 ) must be computed to determine the on-time required for every reference voltage vector. The on-time Ta, Tb, and T0 values are calculated by (5)-(7) [33]. The following is a summary of the technique for finding the on-time switching for a 3-level T-type inverter.
• Using the following equations, we can identify the sector Si and γ within the sector for the desired reference vector: • Define two integers to identify the Δ1 and the tiny vector vs:  Fig. 9 Leg voltage for reference vector at sector 1, Δ1.
A 3-level T-type inverter's switching waveforms are shown in Fig. 9.

SIMULINK IMPLEMENTATION
The DTC-SVM algorithm for a 2-Level inverter has been implemented using the MATLAB/Simulink program, as illustrated in Fig.10.  Fig.11. In this figure, and concerning Fig. 8  and 9, is the calculation of the switching times (output) for each sector, as shown in the equations and Table V. Where To get Y, where Y=Region as shown in Table IV.  In Table VI, to calculate the switching timings (output) for each sector, the Simulink (DTC-SVM) is used to drive the IM, which is based on a three-level T-type inverter.

RESULTS OF SIMULATION
MATLAB/Simulink software is used to compare the performance of the IM utilizing the DTC-SVM algorithm for a 2-level inverter with the DTC-SVM algorithm for a 3-level Ttype inverter. Table VII lists the simulation parameters for the control system.  Fig.12 depicts the behavior of the induction motor's torque and speed in response to their DTC-SVM algorithm reference value for a 2-level inverter. As demonstrated in Fig. 12(a), there is a notable ripple in the torque and as seen in Fig. 12(b), the speed maintains its reference value. Using the modified DTC-SVM algorithm for a 3-level T-type inverter, on the other hand, resulted in a noticeable improvement in IM performance, as shown in Fig. 13 The torque ripples in Fig. 13(a) are decreased, and the IM's speed responsiveness is improved, as illustrated in Fig. 13   Using a DTC-SVM with a 2-level inverter there are stator flux ripples as in Fig. 14 and they are significantly reduced when compared to using a DTC-SVM with a 3-level T-type inverter as in Fig. 15. The quality of the stator currents greatly improves, when the DTC-SVM algorithm for a 3level T-type inverter as shown in Fig. 16 is compared with the DTC-SVM algorithm for a 2level inverter as shown in Fig. 17.

CONCLUSIONS
The influence of utilizing a multi-level inverter on the performance of the DTC-SVM to define the reference vector and switch-on-time is discussed in this paper, and it is compared to the DTC-SVM using a two-level inverter. The realized of employing a two-level inverter of DTC-SVM algorithms and a three-level T-type inverter of DTC-SVM algorithms to drive a three-phase IM using Matlab/Similink. The DTC-SVM generated switching pulses as a function of space vector modulation. Although there are complex calculations as a result of the modification, the simulation results indicate a significant improvement in the performance of the induction motor