A Modular Multilevel Converter with an Advanced PWM Control Technique for Grid-Tied Photovoltaic System

: Due to global warming and shortage of fossil fuels, the grid-connected solar photovoltaic (PV) system has gained signiﬁcant popularity all over the world. The modular multilevel cascaded (MMC) inverter is the natural choice for step-up transformer and line ﬁlter less direct medium voltage grid integration of solar PV systems. However, power quality and loss are the important issues while connecting the PV system to the medium voltage grid through MMC inverter. Modulation technique is the key to maintain output power quality, e.g., total harmonic distortion (THD) and to ensure low switching and conduction losses. In this paper, an advanced modulation technique named “triangle saturated common mode pulse width modulation (TSCMPWM)” control is proposed for a 3-phase 5-level MMC inverter-based grid-tied PV system. Compared to traditional modulation techniques, the proposed TSCMPWM control offers the lowest voltage THD as well as lower inverter power losses. Performance of the proposed modulation technique is evaluated in MATLAB/Simulink environment and tested with a reduced scale prototype test platform. Both simulation and experimental results show the effectiveness of the proposed modulation technique. voltage direct axis component ( V ∗ d ) and the inverter reference voltage quadrature axis component ( V ∗ q ) are determined. After that, the direct and quadrature axis components of the reference inverter voltage are converted into three-phase reference voltages, which are then added to the common mode signal to form the proposed modulating signal. Finally, the modulating signals are compared to the high-frequency carrier signals to produce the gate pulses for the MMC inverter. respectively. From Figure 13, it can be concluded that the proposed TSCMPWM technique shows better THD performance for three modulation regions compared to the existing SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM techniques. Performance of the proposed modulation technique is evaluated in MATLAB/Simulink environment and tested with a reduced scale prototype test platform. Both simulation and experimental results show the effectiveness of the proposed modulation technique. The proposed PWM technique with the enlisted advantages could be a promising technology for the development of medium voltage power converters used in different emerging applications.


Introduction
To mitigate the increasing energy demand without affecting the environment, the renewable energy sources such as solar, oceanic wave, and wind are being considered as the alternative source of energy all over the world. Among different renewable power generation systems, solar photovoltaic (PV) system has received significant interest due to a number of features such as cost-effectiveness and technological advancement. The number of installed solar PV power plants in the world are increasing rapidly to replace the traditional fossil fuel-based electrical power plants [1,2]. In grid-connected solar PV plants, inverters play an important role in conversion of electrical energy from DC to AC before feeding the solar power to grid. Due to some special features such as capability to handle high voltage and generate staircase sinusoidal output voltage, the multilevel inverters have gained considerable popularity for their use in step-up transformer and line filter less medium voltage grid integration of solar PV power plants. [3,4]. A wide research has been done by the research community on different types of multilevel inverter topologies. The neutral point clamped (NPC), flying-capacitor (FC), cascaded H-bridge (CHB) and modular multilevel cascaded (MMC) are the basic multilevel inverter topologies used in different industrial applications [5].
Among these basic topologies, MMC has drawn significant attention from the research community for its numerous advantages such as capability to control the internal power flow, ease of implementation, low harmonic distortion in the output voltage, and less maintenance cost [6][7][8][9][10][11]. Due to these attractive features, MMC inverters are increasingly used in various applications including renewable power generation and industrial medium voltage motor drives [12,13].
However, the switching and conduction losses are the limitations of multilevel inverters, which depends mainly on the modulation techniques used to control the solar PV inverter's semiconductor switching devices [14][15][16]. To improve the power quality of multilevel inverters, a number of modulation techniques have been proposed by the research community in recent decades, as the amount of conduction and switching losses of a multilevel inverter circuit greatly depends upon the switching technique used in the multilevel inverters.
The pulse width modulation (PWM) technique is the most fundamental switching strategy for multilevel inverters, where the gate pulses are mainly generated by comparing the reference signal with a carrier signal of higher frequency. The PWM technique has some advantages such as easy to implement, better performance [17], and limited amount of power loss in the switching devices [18]. Several PWM-based advanced switching techniques have been presented in the literature including sine PWM (SPWM), third harmonic-injected PWM (THPWM), conventional space vector PWM (CSVPWM), sixtydegree PWM (SDPWM), third harmonic sixty-degree PWM (THSDPWM), trapezoidal PWM (TRPWM), third harmonic trapezoidal PWM (THTRPWM), and fifth harmonic trapezoidal PWM (FHTRPWM) [19]. A comparative analysis of the performance of all these traditional switching techniques has been presented in [17]. The THPWM technique has been the most adopted modulation technique among all the traditional PWM modulation techniques based on THD [20]. On the other hand, the CSVPWM technique has been widely applied for better DC bus voltage utilization and flexible control [21][22][23][24]. Although these two techniques provide a good harmonic spectrum, the THPWM technique causes high power losses in the MMC inverters and the practical implementation of the CSVPWM technique is quite complex in case of the MMC inverters [4]. The thirty-degree bus clamping PWM (TDBCPWM) and sixty-degree bus clamping PWM (SDBCPWM) techniques have also gained popularity due to their use in the MMC inverters with low power losses. However, these techniques show poor spectra property [5].
In this paper, an advanced switching strategy is proposed to minimize the voltage and current THDs at the output and reduce the switching and conduction losses of the MMC inverters. The modulating signal of the proposed switching technique is generated by adding a triangular saturated common-mode signal to a pure sinusoidal reference signal. To test the performance, the proposed switching technique is applied to a solar PV systembased grid-connected 3-phase 5-level MMC inverter. The performance (voltage and current THDs and the power losses) of the proposed switching technique is compared with that of the existing modulation techniques such as SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM. The key features of the proposed modulation technique are: • low output voltage THD, • low injected grid current THD, • low inverter power loss, and • high DC bus voltage utilization.
The paper is organized as follows: the MMC inverter-based grid-tied PV system is described in Section 2, existing PWM techniques are described in Section 3, inverter control for grid synchronization is demonstrated in Section 4, the performance evaluation is described in Section 5, experimental validation is described in Section 6, and finally, the conclusion of the paper is in Section 7.

The MMC Inverter-Based Grid-Tied PV System
Step-up transformer-less direct medium voltage grid integration of PV array using a 5-level MMC inverter is shown in Figure 1. PV array feeds power to the DC link of the high-frequency inverter through a DC-DC converter. The DC-DC converter ensures maximum power extraction from the PV array and maintain a constant output voltage according to the voltage level of the DC link. The inverter is linked to the high-frequency magnetic-link's primary winding, and each secondary winding operates via a bridge rectifier as an isolated source connected to the H-bridge cell. The magnetic-link (no voltage step-up) is used to generate isolated and balanced multiple power sources from a single source and to ensure galvanic isolation. THD decreases as the number of MMC inverter levels increases. The rectifier unit rectifies the high-frequency AC voltage to DC voltage, which creates isolated DC source for the H-bridge module. The THD and the switching loss can also be minimized using proper modulation methods for a specific inverter stage. The performance of the proposed modulation technique is analyzed in terms of the THD and the switching loss in the subsequent sections. described in Section 5, experimental validation is described in Section 6, and finally, the conclusion of the paper is in Section 7.

The MMC Inverter-Based Grid-Tied PV System
Step-up transformer-less direct medium voltage grid integration of PV array using a 5-level MMC inverter is shown in Figure 1. PV array feeds power to the DC link of the high-frequency inverter through a DC-DC converter. The DC-DC converter ensures maximum power extraction from the PV array and maintain a constant output voltage according to the voltage level of the DC link. The inverter is linked to the high-frequency magnetic-link's primary winding, and each secondary winding operates via a bridge rectifier as an isolated source connected to the H-bridge cell. The magnetic-link (no voltage stepup) is used to generate isolated and balanced multiple power sources from a single source and to ensure galvanic isolation. THD decreases as the number of MMC inverter levels increases. The rectifier unit rectifies the high-frequency AC voltage to DC voltage, which creates isolated DC source for the H-bridge module. The THD and the switching loss can also be minimized using proper modulation methods for a specific inverter stage. The performance of the proposed modulation technique is analyzed in terms of the THD and the switching loss in the subsequent sections.

Existing PWM Techniques
The performance of the proposed switching technique is compared with the conventional PWM switching techniques, such as SPWM, THPWM, CSVPWM, THSDPWM TDBCPWM, and SDBCPWM. The modulating signals of these conventional switching techniques are depicted in Figure 2.

Existing PWM Techniques
The performance of the proposed switching technique is compared with the conventional PWM switching techniques, such as SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM. The modulating signals of these conventional switching techniques are depicted in Figure 2.    The proposed switching technique along with other traditional PWM switching techniques are applied to a 3-phase 5-level MMC inverter modeled in MATLAB/Simulink en-

Switching Techniques
Mathematical Expressions The proposed switching technique along with other traditional PWM switching techniques are applied to a 3-phase 5-level MMC inverter modeled in MATLAB/Simulink  Figure 3 shows the output line voltage waveforms for the traditional modulation techniques. For each modulation technique, the frequency of the reference signal is kept 50 Hz and as a carrier signal, a 2 kHz phase-shifted triangular signal is used. In the simulations, the modulation index is considered 1. Figure 4 shows the frequency spectrum of the output line voltages for SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM techniques.
It can be observed from Figure 4 that the THPWM presents the lowest and the SPWM presents the highest THD among all of these traditional switching strategies.        It can be observed from Figure 4 that the THPWM presents the lowest and the SPWM presents the highest THD among all of these traditional switching strategies.

Proposed PWM Technique
The modulating signal for the proposed triangle saturated common mode-injected PWM technique is depicted in Figure 5. The signal is formed by adding a triangle saturated common mode signal to the three-phase sinusoidal reference signals.
Energies 2021, 14, x FOR PEER REVIEW The modulating signal for the proposed triangle saturated common mod PWM technique is depicted in Figure 5. The signal is formed by adding a tria rated common mode signal to the three-phase sinusoidal reference signals.   s a = A sin ωt (1) where A represents the amplitude of the reference signals.
In the next step, a triangular signal, V 1 of lower amplitude is generated, which is expressed by Equation (4).
where C = 0.77A. As depicted in Figure 6a, the common-mode signal, represented by V 2 , is generated by binding V 1 in a region, where the maximum amplitude of the signal is 0.11 C, and the minimum amplitude is −0.11 C. The following equations explains the relation mathematically.
Energies 2021, 14, 331 where A represents the amplitude of the reference signals. In the next step, a triangular signal, V1 of lower amplitude is generated, which is e pressed by Equation (4). Finally, the common-mode signal is added to the reference three-phase sinusoidal signals, as expressed in Equations (8)- (10).
where S a , S b , and S c are the modulating signals of the proposed switching technique for a three-phase MMC inverter. Figure 6b shows the flowchart for the generation of the proposed TSCMPWM technique. Two phase-shifted carriers (where two carriers are out of phase by 90 • ) are compared with the modulating signals in the switching schemes and the corresponding gate pulses are produced for the insulated gate bipolar transistors (IGBTs). Figure 7 depicts the procedure of gate pulse generation using the proposed TSCMPWM technique, where a high-frequency triangular carrier signal (2 kHz used in the simulation) is compared with the proposed modulating signal of 50 Hz to generate the gate pulses. pared with the modulating signals in the switching schemes and the corre pulses are produced for the insulated gate bipolar transistors (IGBTs). Figur procedure of gate pulse generation using the proposed TSCMPWM techn high-frequency triangular carrier signal (2 kHz used in the simulation) is c the proposed modulating signal of 50 Hz to generate the gate pulses. This proposed switching technique is applied to the 3-phase 5-level M for PV system, where the values taken for A and C are 1.09 and 0.84, r modulation index 1. Figure 8 shows the inverter output line voltage and spectra. The THD of the inverter output line voltage for the proposed tech which is lower than the existing six modulation techniques. This proposed switching technique is applied to the 3-phase 5-level MMC inverter for PV system, where the values taken for A and C are 1.09 and 0.84, respectively, at modulation index 1. Figure 8 shows the inverter output line voltage and its frequency spectra. The THD of the inverter output line voltage for the proposed technique is 13%, which is lower than the existing six modulation techniques.       The filtered line voltage waveforms and the THD for the proposed TSCMPWM scheme are shown in Figure 9. The filtered line voltage THD of 5-level MMC inverter for TSCMPWM scheme is 0.64%.

Inverter Control for Grid Synchronization
For grid-tied operation, the amplitude and frequency of the inverter output voltage need to be synchronized with the grid. The output voltage and frequency of the generator are the leading parameters to synchronize with the grid in the conventional synchronous generator system. The generator synchronizes with the grid at the common coupling point (PCC) after proper matching of amplitude and frequency. Whether or not the synchronization is consistently maintained, it needs to be continuously monitored online. The synchronization condition determines the inverter's gating signals. Figure 11 depicts the overall control scheme for the MMC inverter-based PV system utilizing the proposed PWM technique. The reference direct axis component (I * d ) is initially obtained from the dc-link voltage regulator and the reference quadrature axis component (I * q ) is set to zero because of the unity power factor operation of the inverter. The error between the real injected grid currents (I d and Iq) and the reference currents (I * d and I * q ) is then passed through the proportional-integral (PI) controllers, and the inverter reference voltage direct axis component (V * d ) and the inverter reference voltage quadrature axis component (V * q ) are determined. After that, the direct and quadrature axis components of the reference inverter voltage are converted into three-phase reference voltages, which are then added to the common mode signal to form the proposed modulating signal. Finally, the modulating signals are compared to the high-frequency carrier signals to produce the gate pulses for the MMC inverter. Figure 11 depicts the overall control scheme for the MMC inverter-based PV system utilizing the proposed PWM technique. The reference direct axis component ( * ) is initially obtained from the dc-link voltage regulator and the reference quadrature axis component ( * ) is set to zero because of the unity power factor operation of the inverter. The error between the real injected grid currents (Id and Iq) and the reference currents ( * and * ) is then passed through the proportional-integral (PI) controllers, and the inverter reference voltage direct axis component ( * ) and the inverter reference voltage quadrature axis component ( * ) are determined. After that, the direct and quadrature axis components of the reference inverter voltage are converted into three-phase reference voltages, which are then added to the common mode signal to form the proposed modulating signal. Finally, the modulating signals are compared to the high-frequency carrier signals to produce the gate pulses for the MMC inverter.

Loss Evaluation of the MMC Inverter
Inverter losses include switching loss (switch-off, switch-on, and diode turn-off losses) and conduction loss (losses for the switch and the diode conduction). In practice, diode turn-on loss is negligible due to first conducting nature of diode. IGBT module 5SNA 1500E250300 is taken as an example to perform the loss calculation. The equation of the total inverter loss (TIL) can be written as TIL=Conduction loss (Pcond) + Switching loss (PSL).

Loss Evaluation of the MMC Inverter
Inverter losses include switching loss (switch-off, switch-on, and diode turn-off losses) and conduction loss (losses for the switch and the diode conduction). In practice, diode turn-on loss is negligible due to first conducting nature of diode. IGBT module 5SNA 1500E250300 is taken as an example to perform the loss calculation. The equation of the total inverter loss (TIL) can be written as TIL=Conduction loss (P cond ) + Switching loss (P SL ).
where v F , i c , and v ce are the voltage drop of antiparallel diode, collector current, and collector-emitter voltage drop, respectively. For switching loss, switch turn-on loss (E on ) and switch turn-off loss (E off ) are evaluated using the collector current characteristics curve taken from the datasheet of the IGBT module. The total switching loss (P SL ) is calculated using Equation (15). In addition, using diode switch current and reverse recovery characteristics curve of the IGBT, diode switching loss (E rec ) is calculated [9]. If N be the number of IGBTs, total diode switching loss (P rrL ) can be calculated using Equation (16).
where i F is the forward diode current and N = 24 for 3-phase 5-level MMC inverter. Figure 12 depicts various loss components with the proposed TSCMPWM technique.
where vF, ic, and vce are the voltage drop of antiparallel diode, collector curre tor-emitter voltage drop, respectively. For switching loss, switch turn-on loss (Eon) and switch turn-off loss ( ated using the collector current characteristics curve taken from the datashe module. The total switching loss (PSL) is calculated using Equation (15). In a diode switch current and reverse recovery characteristics curve of the IGBT ing loss (Erec) is calculated [9]. If N be the number of IGBTs, total diode switc can be calculated using Equation (16).
where iF is the forward diode current and N = 24 for 3-phase 5-level MMC in 12 depicts various loss components with the proposed TSCMPWM techniq

THD Profile of the MMC Inverter for Modulation Region Variation
Modulation index has significant impact on the THD of the inverter output voltage. Based on the grid conditions, controller varies the modulation index of the PWM technique employed in the inverter switching according to the requirement. There are three modulation regions of the PWM technique. These are under modulation region, linear modulation region, and over modulation region. Modulation region also affects the switching and conduction power losses of the inverter. As shown in Figure 13, detailed comparisons are made between the converter output line voltages with distinct modulation techniques in under, linear, and over modulation regions. It should be noted that the line voltage THD gets lower in over modulation region than that of linear and under modulation and the width of the gate pulses gets wider, which increases the conduction losses of the inverter. On the contrary, in under modulation region, lower order harmonics increase, which puts additional burden on the filtering circuitry. Therefore, for the desired output, the controller should return as soon as possible to the linear modulation mode. The proposed TSCMPWM offers 13.00%, 23.77%, and 12.03% THDs for linear, under, and over modulation regions, respectively. From Figure 13, it can be concluded that the proposed TSCMPWM technique shows better THD performance for three modulation regions compared to the existing SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM techniques.
ulation and the width of the gate pulses gets wider, which increases the conduction losses of the inverter. On the contrary, in under modulation region, lower order harmonics increase, which puts additional burden on the filtering circuitry. Therefore, for the desired output, the controller should return as soon as possible to the linear modulation mode. The proposed TSCMPWM offers 13.00%, 23.77%, and 12.03% THDs for linear, under, and over modulation regions, respectively. From Figure 13, it can be concluded that the proposed TSCMPWM technique shows better THD performance for three modulation regions compared to the existing SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM techniques.    Figure 14 depicts the dynamic load change response of the inverter voltage and current. Basically, the grid is connected to several local loads. At any instant of time, the total load on the inverter may change. So, it is mandatory to analyze the inverter current profile during load perturbations. At t = 0.05 s, 5 kW resistive type load is applied on the inverter suddenly and the load is withdrawn at t = 0.10 s. In this period, the inverter current has increased for the load changing effect, but it also shows the sinusoidal nature in the time  Basically, the grid is connected to several local loads. At any instant of time, the total load on the inverter may change. So, it is mandatory to analyze the inverter current profile during load perturbations. At t = 0.05 s, 5 kW resistive type load is applied on the inverter suddenly and the load is withdrawn at t = 0.10 s. In this period, the inverter current has increased for the load changing effect, but it also shows the sinusoidal nature in the time of load changing period. Similarly, at t = 0.15 s, a 4 kW + j 0.5 VAR resistive-inductive type load is applied on the inverter and the load is withdrawn at t = 0.10 s. Same as before, the inverter current has increased during the load perturbation interval, which is also of load changing period. Similarly, at t = 0.15 s, a 4 kW + j 0.5 VAR resistive-inductive type load is applied on the inverter and the load is withdrawn at t = 0.10 s. Same as before, the inverter current has increased during the load perturbation interval, which is also sinusoidal. Finally, from Figure 14, it can be concluded that the MMC inverter tied with medium voltage grid utilizing the proposed TSCMPWM technique shows satisfactory dynamic response during the load perturbation.  Figure 15a shows the inverter responses, such as the grid voltage sag and swell phenomenon, when there are variations in the grid. The voltage is sensed by the inverter when there is a voltage sag or swell on the grid side. The value of Vd (direct axis component of grid voltage) and Vq (quadrature axis component of grid voltage) are also modified due to the change in voltage amplitude, and the inverter modifies the modulation index accordingly. The inverter decreases its modulation index when voltage sags occur, and when voltage swells occur, the inverter raises its modulation index to cope with the grid voltage. Figure 15b depicts the zoomed inverter output phase voltage during sag and swell with the variation of the modulation index.  Figure 15a shows the inverter responses, such as the grid voltage sag and swell phenomenon, when there are variations in the grid. The voltage is sensed by the inverter when there is a voltage sag or swell on the grid side. The value of V d (direct axis component of grid voltage) and V q (quadrature axis component of grid voltage) are also modified due to the change in voltage amplitude, and the inverter modifies the modulation index accordingly. The inverter decreases its modulation index when voltage sags occur, and when voltage swells occur, the inverter raises its modulation index to cope with the grid voltage. Figure 15b depicts the zoomed inverter output phase voltage during sag and swell with the variation of the modulation index. of load changing period. Similarly, at t = 0.15 s, a 4 kW + j 0.5 VAR resistive-inductive type load is applied on the inverter and the load is withdrawn at t = 0.10 s. Same as before, the inverter current has increased during the load perturbation interval, which is also sinusoidal. Finally, from Figure 14, it can be concluded that the MMC inverter tied with medium voltage grid utilizing the proposed TSCMPWM technique shows satisfactory dynamic response during the load perturbation.  Figure 15a shows the inverter responses, such as the grid voltage sag and swell phenomenon, when there are variations in the grid. The voltage is sensed by the inverter when there is a voltage sag or swell on the grid side. The value of Vd (direct axis component of grid voltage) and Vq (quadrature axis component of grid voltage) are also modified due to the change in voltage amplitude, and the inverter modifies the modulation index accordingly. The inverter decreases its modulation index when voltage sags occur, and when voltage swells occur, the inverter raises its modulation index to cope with the grid voltage. Figure 15b depicts the zoomed inverter output phase voltage during sag and swell with the variation of the modulation index.

Response of Grid Current during Fault
The grid current is observed in this section under different fault conditions. Figure 16 describes the grid current profile under different fault conditions. The fault is applied in-between phases "a" and "b," phases "b" and "c," phases "a" and "c," phases "a," "b," and "c" in Figure 16a-d, respectively. The amplitude of the current increases during fault period. After removing the fault, the current recovers its previous balanced condition very quickly, which demonstrates fast, robust, and dynamic characteristics of the system with the MMC inverter utilizing the proposed TSCMPWM technique for switching.

Response of Grid Current during Fault
The grid current is observed in this section under different fault conditions. Figure  16 describes the grid current profile under different fault conditions. The fault is applied in-between phases "a" and "b," phases "b" and "c," phases "a" and "c," phases "a," "b," and "c" in Figure 16a-d, respectively. The amplitude of the current increases during fault period. After removing the fault, the current recovers its previous balanced condition very quickly, which demonstrates fast, robust, and dynamic characteristics of the system with the MMC inverter utilizing the proposed TSCMPWM technique for switching. Fault response in inverter current: (a) fault is applied in phase "a" and "b" (b) fault is applied in phase "b" and "c" (c) fault is applied in phase "c" and "a," and (d) fault is applied in phase "a," "b," and "c."

Comparative Analysis
In this section, the superiority of the proposed PWM technique with existing PWM techniques is investigated in terms of THD with the variation of modulation index and carrier frequency, inverter power loss, and THD variation with the change of inverter level. Table 3 shows the results of line voltage THDs by varying carrier frequency where modulation index is taken as 1. The proposed TSCMPWM offers low line voltage THD that of existing six PWM techniques.  Table 4 shows the results of line voltage THDs by varying modulation index where carrier frequency is taken as 2 kHz. For modulation index variation, the proposed TSCMPWM offers low line voltage THDs compared to existing six PWM techniques. Fault response in inverter current: (a) fault is applied in phase "a" and "b" (b) fault is applied in phase "b" and "c" (c) fault is applied in phase "c" and "a," and (d) fault is applied in phase "a," "b," and "c."

Comparative Analysis
In this section, the superiority of the proposed PWM technique with existing PWM techniques is investigated in terms of THD with the variation of modulation index and carrier frequency, inverter power loss, and THD variation with the change of inverter level. Table 3 shows the results of line voltage THDs by varying carrier frequency where modulation index is taken as 1. The proposed TSCMPWM offers low line voltage THD that of existing six PWM techniques.  Table 4 shows the results of line voltage THDs by varying modulation index where carrier frequency is taken as 2 kHz. For modulation index variation, the proposed TSCMPWM offers low line voltage THDs compared to existing six PWM techniques.  Figure 17 shows the inverter power losses for different PWM techniques with the variation of modulation index. Most of the practical system utilizes 0.7 to 1 modulation index variation. However, modulation index below 0.7 and over 1 have some detrimental effects on inverter performance. From Figure 17, it should be noted that the proposed TSCMPWM offers lower inverter power loss compared to existing six PWM techniques for a wide range of variation of modulation index.   Figure 17 shows the inverter power losses for different PWM techniques with the variation of modulation index. Most of the practical system utilizes 0.7 to 1 modulation index variation. However, modulation index below 0.7 and over 1 have some detrimental effects on inverter performance. From Figure 17, it should be noted that the proposed TSCMPWM offers lower inverter power loss compared to existing six PWM techniques for a wide range of variation of modulation index.     Table 5.  Figure 18 shows a down-scale laboratory setup of 5 kVA, 400 V, and 5-level MMC inverter. Necessary experiments are carried out on the MMC inverter to substantiate the improved performance of the proposed modulation technique. Three-phase electronic load module with a second-order LC filtering unit is connected at the output of the MMC inverter. To supply each H-bridge with 200 V dc, a battery bank is utilized, as shown in Figure 18. A Xilinx Spartan-3E FPGA is programmed to implement the proposed TSCMPWM scheme. Each H-bridge unit is made of compact SK30GH123 IGBT module from SEMIKRON, and all the switches are driven by a gate driver unit made of isolated gate driver SKHI 20opA.  Figure 18 shows a down-scale laboratory setup of 5 kVA, inverter. Necessary experiments are carried out on the MMC inv improved performance of the proposed modulation technique load module with a second-order LC filtering unit is connected a inverter. To supply each H-bridge with 200 V dc, a battery bank Figure 18. A Xilinx Spartan-3E FPGA is programmed to im TSCMPWM scheme. Each H-bridge unit is made of compact SK from SEMIKRON, and all the switches are driven by a gate driv gate driver SKHI 20opA. Figure 19a shows the TSCMPWM reference modulating wa ules are operated with 1.2 kHz switching frequency. Figure 19b line-line PWM voltages before filter and three-phase line current The voltage signals are collected with the KEYSIGHT N2791A vo rent signals are collected with the Agilent N2779A current probes Figure 19b that the THD of the line-line voltages before fil TSCMPWM is 13.2%, which is highly similar to the one found fr ver, the line current THD is 1.2%, which is below the IEEE 519-2 THD limit (5%).   Figure 19a shows the TSCMPWM reference modulating waveforms. The IGBT modules are operated with 1.2 kHz switching frequency. Figure 19b,c show the three-phase line-line PWM voltages before filter and three-phase line currents after filter, respectively. The voltage signals are collected with the KEYSIGHT N2791A voltage probes and the current signals are collected with the Agilent N2779A current probes. It can be observed from Figure 19b that the THD of the line-line voltages before filter under the proposed TSCMPWM is 13.2%, which is highly similar to the one found from simulations. Moreover, the line current THD is 1.2%, which is below the IEEE 519-2014 standard maximum THD limit (5%).  Figure 20 shows the load current performance under sudden load change conditions. First, a three-phase resistive load of 1.78 kW is connected in parallel, which causes highcurrent injection from the converter. The signals show steady and pollution free performance. Later, a load with pf of 0.9 (lagging) is connected in parallel. It can be observed from Figure 20 that the system offers distortion less signals even with the inclusion of the lagging load, however, the currents will be phase lagged.

Conclusions
In this paper, an advanced switching control technique for a 5 level grid-tied MMC inverter is proposed. The proposed technique offers 13% output line voltage THD (without line filter), which is lower than that of the existing most common PWM techniques. Moreover, the proposed technique ensures lower switching and conduction power losses of the inverter compared with the existing SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM techniques. The proposed technique also provides better responses in terms of dynamic load change, fault, modulation index variation, and voltage sag/swell. Performance of the proposed modulation technique is evaluated in MATLAB/Simulink environment and tested with a reduced scale prototype test platform.  Figure 20 shows the load current performance under sudden load change conditions. First, a three-phase resistive load of 1.78 kW is connected in parallel, which causes high-current injection from the converter. The signals show steady and pollution free performance. Later, a load with pf of 0.9 (lagging) is connected in parallel. It can be observed from Figure 20 that the system offers distortion less signals even with the inclusion of the lagging load, however, the currents will be phase lagged.  Figure 20 shows the load current performance under sudden load change condition First, a three-phase resistive load of 1.78 kW is connected in parallel, which causes high current injection from the converter. The signals show steady and pollution free perfo mance. Later, a load with pf of 0.9 (lagging) is connected in parallel. It can be observe from Figure 20 that the system offers distortion less signals even with the inclusion of th lagging load, however, the currents will be phase lagged.

Conclusions
In this paper, an advanced switching control technique for a 5 level grid-tied MM inverter is proposed. The proposed technique offers 13% output line voltage THD (with out line filter), which is lower than that of the existing most common PWM technique Moreover, the proposed technique ensures lower switching and conduction power losse of the inverter compared with the existing SPWM, THPWM, CSVPWM, THSDPWM TDBCPWM, and SDBCPWM techniques. The proposed technique also provides better re sponses in terms of dynamic load change, fault, modulation index variation, and voltag sag/swell. Performance of the proposed modulation technique is evaluated i MATLAB/Simulink environment and tested with a reduced scale prototype test platform Both simulation and experimental results show the effectiveness of the proposed modu

Conclusions
In this paper, an advanced switching control technique for a 5 level grid-tied MMC inverter is proposed. The proposed technique offers 13% output line voltage THD (without line filter), which is lower than that of the existing most common PWM techniques. Moreover, the proposed technique ensures lower switching and conduction power losses of the inverter compared with the existing SPWM, THPWM, CSVPWM, THSDPWM, TDBCPWM, and SDBCPWM techniques. The proposed technique also provides better responses in terms of dynamic load change, fault, modulation index variation, and voltage sag/swell.