1. Introduction
Recently, due to the increase in the price of fossil resources, their unsustainability, and the damage that they cause to the environment through carbon emissions, there has been a tendency towards the use of renewable energy sources, which encompass solar, wind, geothermal, and hydro energy. Wind energy stands out as one of the most significant contributors in the renewable energy area [
1,
2,
3,
4].
Wind power is expanding quickly on a global scale. According to the International Renewable Energy Agency (IRENA) Renewable Energy Capacity Statistics 2023, the total installed capacity of globally renewable energy has increased to approximately 3372 GW. As of 2022, the total installed capacity of wind energy among renewable energy technologies is reported to be 26.66% [
5]. Grid-connected wind energy conversion systems (WECS) of a medium to large size are the world’s most significant and rapidly expanding electricity sources. The affordability of wind power technology, the growth of the sector, environmental issues, the rise in energy consumption, the increasing cost of producing fossil fuels, and the accessibility of a strong wind resource in many places globally are projected to support this trend in the near future. These gains include the substantial assistance offered by various governments in the form of investment subsidies and incentives [
6,
7,
8].
As per the 2022 report by the Global Wind Energy Council (GWEC), the global total installed wind power, which was 745 GW in 2020, increased by 93.6 GW and reached 837 GW by the end of 2021. The value of the total wind installed power is obtained as the sum of the total onshore power and total offshore power [
9]. At the end of 2021, the total onshore value had increased by 72,499 MW compared to the previous year and reached 780,275 MW. Moreover, the total offshore value had increased by 21,106 MW compared to the previous year and reached 57,176 MW [
10]. The wind power capacity had increased by approximately 78 GW in 2022 worldwide. Thus, the total installed capacity had reached 906 GW by 2022 [
11]. The historical change in total installations is demonstrated in
Figure 1.
Generators are used to convert the mechanical energy obtained from wind turbines into electrical energy. The generators used in WECS are generally classified into two types, namely induction generators and permanent magnet synchronous generators (PMSGs) [
12,
13,
14].
Originally, the use of PMSGs was limited to small wind power turbines. It was not extended to large-scale electrification because it required large and powerful permanent magnets and these magnets were very expensive. However, due to the gearless design, the complete controllability of the system, and the excellent efficiency of the PMSG, it is widely preferred in WECS [
15,
16,
17]. The PMSG is widely utilized in variable speed operations, particularly at extremely low and medium power levels. PMSGs offer several advantages, including a decreased volume and weight, minimal maintenance, gearbox elimination, brushless operation, and not requiring any energy source for excitation. One of the most important advantages is that they can generate power at any speed. PMSGs are classified as radial, axial, and cross flux according to the direction of the produced flux in the machine. However, radial flux PMSGs are preferred for both motoring and generator actions, because they are simpler, more reliable, and more practical to manufacture than the other two types. PMSGs have surface-mounted and rotor-inserted magnet types, depending on the position of the magnet on the rotor surface. In addition, inner and outer rotor types are available for positioning the rotor inside or outside the stator [
18,
19].
The proposed WECS in this study is composed of a three-phase PMSG, a six-diode bridge rectifier for the AC–DC conversion, a DC–DC boost converter, a Packed U-Cell (PUC) multilevel inverter (MLI) for the DC–AC conversion, and an adjustable RL load. The descriptions of each component are given and their input and output voltages for the simulation and experimental study are analyzed and their changes are observed.
MLIs offer several advantages over conventional two-level inverters in a wide range of industrial applications due to their effectiveness, reduced switching losses, lower dv/dt stress, and diminished total harmonic distortion (THD). Various studies have been conducted to propose different MLI topologies, including cascaded H-bridge (CHB), flying capacitor (FC), and neutral-point clamped (NPC) converters [
18,
19,
20,
21,
22]. Among these traditional MLIs, the PUC-based MLI represents a promising structure, as it combines the advantages of both CHB and FC configurations [
23,
24]. Ounejjar proposed the PUC-based MLI for the first time in 2011 [
25]. This novel MLI is composed of a single primary DC supply voltage for one level and capacitors for the higher levels [
26].
Rita Khawaja et al. designed a novel seven-level single-stage/phase grid-connected photovoltaic (PV) inverter based on a PUC topology with finite set model predictive control (FS-MPC). This innovative design is configured to work with PV panels producing an output voltage of 180 VDC. In this study, the authors converted the DC energy obtained from solar panels into 120 VAC energy using the seven-level PUC MLI. They claimed that the key benefit of the suggested approach is less complexity as an additional conversion stage is not required [
27].
Boutheyna Hadmer et al. introduced a sliding mode controller (SMC) specifically developed for a five-level PUC MLI, utilizing a voltage sensorless switching technique. The purpose of using a control modulation technique without using a voltage sensor is to make the control system simple and less expensive. The suggested switching technique can also stabilize the capacitor voltage of the PUC MLI. In the five-level PUC MLI simulation, the input voltage of the inverter was set to 100 VDC, which was obtained from PV. In addition, the capacitor value was 150 uF, the switching frequency was 1 kHz, and the RL load was 10 Ω, 45 mH [
28].
Mirsajed Pourmirasghariyan et al. proposed and simulated an MPC-based scheme for a PUC-based grid-connected inverter for PV applications. The suggested approach employs a cost function optimization method to select the most suitable switching states for the PUC-based inverter. This enables the PUC inverter to achieve maximum power point tracking in PV arrays. Simulation results reveal that the suggested approach is capable of correctly regulating grid current injection as well as the PV array output voltage [
29].
Mohammad Babaie et al. suggested an optimized SMC system for switching control in PUC7 MLIs. This robust and versatile controller is designed to accomplish both capacitor self-voltage balance and current reference tracking concurrently in both grid-connected and stand-alone PUC7 inverters. The proposed DC–AC converter is designed for an input voltage of 160 VDC, a 900 uF capacitor, and a 2000 Hz switching frequency. Moreover, the SMC’s extensive experimental and simulation results confirm the feasibility of properly controlling the PUC7 inverter even under dynamic modes, non-linear loads, and parameter mismatch [
30].
In [
31], Sanjay Upreti et al. focused on reducing the harmonic content of the output voltage of a grid-connected seven-level PUC MLI. They conducted a power quality analysis of a seven-level three-phase PUC employing multi-carrier switching strategies. Additionally, this proposed system was three-phase. There were two solar PV arrays operating at different voltages on each phase. It was noted that no auxiliary capacitor was used to increase the level.
Researchers have already carried out several studies on MLI topologies for renewable energy applications: a CHB MLI with a three-phase triple voltage source inverter [
32], a single-phase five-level T-type topology in a renewable energy system [
33], a direct-driven PMSG-based WECS based on different average value modeling techniques [
34], a buck–boost/Cuk converter and a five-level inverter with new selected harmonic elimination pulse width modulation (PWM) [
35], a PMSG-based WECS with a back-to-back voltage source converter [
36], an NPC MLI and phase disposition sinusoidal PWM [
37], and a switched capacitor MLI [
38].
In the literature, PV systems have been used as the input sources of PUC MLIs. Consequently, it is seen that the most of the PUC MLI-based studies have been conducted for PV-based systems [
27,
28,
29,
30,
31]. It is important to note that there are not many studies on the application of PUC MLIs for wind energy applications. Furthermore, compared to other conventional MLI circuit topologies, PUC MLIs provide a number of benefits, including the use of fewer semiconductor switches and a smaller filter to achieve the same voltage level, the less complicated control method, and the use of circuit parts with smaller ratings and sizes [
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44]. Taking these aspects into consideration, PUC MLI topologies seem to be innovative for PMSG-based wind energy applications. Moreover, the dual-mode PI-PI control technique is adapted from the literature [
45]. To the best of the authors’ knowledge, this control method is used for the first time for PUC MLI capacitor voltage control. The dual-mode PI-PI control technique is based on the sequential use of two PI controllers. In this approach, the load current measurement has been eliminated and the load voltage reference has been taken into account. In this respect, both simulation and experimental studies are carried out in a stand-alone WECS prototype in this study at two different voltage levels, five and seven. The comparison of the proposed topology with the recently introduced topologies is tabulated in
Table 1. It is observed that the proposed system exhibits the lowest THD value among the MLI topologies provided.
The novelty and contributions of this work can be highlighted as follows.
According to the extensive literature review, it is observed that the PUC MLI topology is widely utilized in PV system applications. In this respect, this topology is used for wind energy conversion systems, and the performance of the system is verified through both simulation and experimental studies.
The dual-mode PI-PI control technique is adopted from the reference [
45] and applied to the voltage control of the auxiliary capacitor in the PUC MLI topology. To the best of authors’ knowledge, this control method is being used for the first time for PUC MLI capacitor voltage control.
The remainder of this paper is organized as follows. In
Section 2, the components of the PMSG-based WECS are examined and analyzed.
Section 2.1 focuses on the PMSG, three-phase diode rectifier, and DC–DC boost converter blocks, while
Section 2.2 deals with the working principle of the PUC MLI topology, its advantages, etc. Simulation results and experimental results are presented and discussed in
Section 3 and
Section 4, respectively. Finally, the conclusions are drawn in
Section 5.
3. Simulation Studies
In this study, the simulation model of the PMSG-based WECS is as shown in
Figure 8. The PUC MLI is able to generate five or seven voltage levels at the output. As can be seen, the output of the DC–DC converter
is used to feed the PUC MLI as
. Obviously, by using only a single DC source, five or seven levels can be produced to feed the load. The design parameters of the proposed PMSG-based WECS are tabulated in
Table 4.
The PMSG is provided to generate voltages at different speeds. As a result of the change in the speed of the wind, the voltage output of the PMSG ranges from about 170 V to 240 V depending on this speed.
The wind speed and the three-phase voltage form at the PMSG output are demonstrated in
Figure 9 and
Figure 10, respectively.
The conversion of the AC voltage at the output of the three-phase PMSG into DC voltage is carried out by an uncontrolled full rectifier with six diodes. The shape of the voltage at the rectifier output is indicated in
Figure 11. The voltage values at the rectifier output adjust accordingly as the output voltage of the PMSG varies.
The AC voltage produced as a result of the changing wind speed of the PMSG is rectified with a diode rectifier. This obtained DC voltage is connected to the input of the DC–DC boost converter. As the wind speed changes, the input voltage of the boost converter also changes. In this study, the main purpose of the DC–DC converter is to keep the output voltage constant even though the input voltage is variable. A PI controller is used to obtain a constant output voltage. The switching trigger of the DC–DC boost converter is generated from the PWM generator block, as shown in
Figure 12. The PWM generator’s switching frequency is selected as 20 kHz.
Figure 13 shows the input and output voltages of the DC–DC boost converter.
It was previously mentioned that the circuit topology in
Figure 5a can be designed as an MLI with five or seven levels depending on the capacitor voltage. In this section, the results of the simulation studies for five and seven levels are given in Case 1 and Case 2, respectively.
3.1. Case 1: Seven-Level PUC MLI
The seven-level Packed U-Cell is obtained by setting the capacitor voltage to one third of the voltage source voltage. To equalize the voltage of the capacitor to one third of the input voltage of the MLI, the dual-mode PI-PI control method is employed, which is utilized here for the first time in PUC MLI topologies. The dual-mode control structure is formed by the sequential use of two PI controllers. In the control approach, the current effect is eliminated and only the voltage reference is taken into account. The control schematic of the PUC7 MLI is shown in
Figure 14.
The input voltage and the auxiliary capacitor voltage of the PUC MLI are given in
Figure 15a. One third of the input voltage of the PUC MLI and the auxiliary capacitor voltage are shown in
Figure 15b. One of the main components of the system is the load. The voltage falling on this load produces different levels according to the switching situation.
Figure 16 demonstrates the load voltage of the system designed by combining all the blocks.
The voltage and current THD for this simulated load value are 10.52% and 3.74%, respectively. The active power absorbed by the load is approximately 425 W.
Figure 17 represents the load current obtained in the simulation study.
3.2. Case 2: Five-Level PUC MLI
PUC5 is obtained by setting the capacitor voltage to one half of the voltage source voltage. To equalize the voltage of the capacitor to one half of the input voltage of the MLI, the dual-mode PI-PI control method is employed, which is utilized here for the first time in PUC MLI topologies. The dual-mode control structure is formed by the sequential use of two PI controllers. In the control approach, the current effect is eliminated and only the voltage reference is taken into account. The control schematic of the PUC5 MLI is shown in
Figure 18.
The input voltage and the auxiliary capacitor voltage of the PUC MLI are indicated in
Figure 19a. One half of the input voltage of the PUC MLI and the auxiliary capacitor voltage are shown in
Figure 19b. Different levels are produced according to the voltage switching situation falling on the load, which is one of the main components of the system. The load voltage of the system designed by combining all the blocks is shown in
Figure 20.
The THD of the voltage and current of the PUC5 MLI is 17.18% and 6.21%, respectively. The load current of PUC5 is shown in
Figure 21. The active power consumed by PUC5 on the load is 480 W.
4. Experimental Studies
The proposed PUC MLI has been designed and validated experimentally as well as through simulation studies. The experimental realization is performed for both seven and five levels.
Figure 22 shows the experimental setup of the proposed PMSG-based WECS. The electrical parameters of the experimental setup are tabulated in
Table 5.
A 1.1 kW asynchronous motor is used in the laboratory environment to model the changing wind speed. The rotor of the PMSG and the rotor of the asynchronous motor are coupled to each other. The speed of the asynchronous motor is changed using a motor driver. Thus, different values of voltage are generated at the output of the PMSG.
Figure 23 demonstrates the coupling of a three-phase asynchronous motor and PMSG. As shown in
Figure 24, different voltage values ranging from approximately 25 V to 35 V are generated at the output of the PMSG. The three-phase AC voltage generated at the output of the PMSG is converted into DC voltage using a diode rectifier.
The state of the selected PMSG operating without a load is expressed as the situation where no load is connected to its output. The result of the test carried out to determine the relationship between the speed and the output voltage of the generator at no load is given in
Figure 25.
The voltage at the output of the diode rectifier varies as the wind speed changes. The input voltage of the PUC MLI must be a constant DC voltage. To obtain this voltage, a DC–DC boost converter is used.
Figure 26 shows the output voltage of the DC–DC boost converter corresponding to the change in the voltage of the diode rectifier. The voltage at the output of the diode rectifier varies between approximately 25 V and 35 V, while the output voltage of the DC–DC boost converter is set to 45 V.
While designing the PUC MLI, six IRFZ44N power MOSFETs are included, and a TLP250 MOSFET gate driver output optocoupler integrator is used to drive these MOSFETs in the circuit. In addition, the voltage of the auxiliary capacitor is realized with closed loop PI control. The PUC MLI’s input, output, and auxiliary capacitor voltage values and output current values are given to the closed loop control system using sensors. The TMS320F2837xD microcontroller is used to perform the control of the circuit and its switching.
4.1. Case 1: Seven-Level PUC MLI
It has already been mentioned that, to obtain seven levels in the PUC MLI, the voltage of the auxiliary capacitor must be one third of the input voltage of the PUC. Experimental verification is carried out using two different loads (RL
1 and RL
2). While the input voltage of the PUC MLI is 45 V, the voltage of the auxiliary capacitor is 15 V with the closed loop system.
Figure 27a,b represent the output voltage, output current, input voltage, and auxiliary capacitor voltage of the PUC7 MLI for RL
1. When the RL
1 load is used, the voltage and current THD of the PUC7 MLI are 14.24% and 4.62%, respectively.
Figure 28a,b represent the output voltage, output current, input voltage, and auxiliary capacitor voltage of the PUC7 MLI for RL
2. When the RL
2 load is used, the voltage and current THD of the PUC7 MLI are 11.65% and 4.10%, respectively.
In the comparison of the RL1 load and RL2 load, the output current is lower for RL2 since the L2 value is larger than the L1 value. Furthermore, in terms of harmonic values, it is seen that the THD values of the RL2 load are lower than the THD values of the RL1 load.
4.2. Case 2: Five-Level PUC MLI
The PUC5 MLI is realized when the voltage value of the auxiliary capacitor is half of the input voltage of the PUC. Two different loads are selected as RL
1 and RL
2 for the experimental studies. While the input voltage of the PUC MLI is 45 V, the voltage of the auxiliary capacitor is 22.5 V with the closed loop system.
Figure 29a,b represent the output voltage, output current, input voltage, and auxiliary capacitor voltage of the PUC5 MLI for RL
1. When the RL
1 load is used, the voltage and current THD of the PUC5 MLI are 28.8% and 5.49%, respectively.
Figure 30a,b represents the output voltage, output current, input voltage, and auxiliary capacitor voltage of the PUC5 MLI for RL
2. When the RL
2 load is used, the voltage and current THD of the PUC5 MLI are 24.96% and 5.09%, respectively.
In the comparison of the RL1 load and RL2 load, the output current is lower for RL2 since the L2 value is larger than the L1 value. Furthermore, in terms of harmonic values, it is seen that the THD values of the RL2 load are lower than the THD values of the RL1 load.
5. Conclusions
As a result of the increase in power electronics applications, such as the development of semiconductor technology, studies on WECSs have increased recently. The energy produced in wind turbines at variable speeds is not smooth and does not reach the desired levels. For this reason, electrical conversion methods are used to make the energy obtained more smooth and more useful. In this study, a novel MLI topology in a PMSG-based WECS is investigated. This inverter topology is called the PUC MLI. Compared to other inverters of the same level, the PUC MLI has advantages in terms of reducing the switching number, switching losses, harmonics, and control complexity. For this reason, it is preferred. The PMSG-based WECS is designed, simulated, and experimentally verified in the laboratory at different speeds and loads. In conclusion, according to the results obtained, some changes are observed as the inverter level changes depending on the level of the generated output voltage and the load values. In the experimental study, the output voltage of the PMSG varies approximately between 25 V and 35 V, while the input voltage of the PUC MLI is determined as 45 V. This value is obtained from the output of the DC–DC boost converter. When the inverter level is increased from 5 to 7, it is observed that the THD of the voltage of the PUC7 MLI decreases from 28.8% to 14.24% and the THD of the current decreases from 5.49% to 4.62% for the RL1 load, while the THD of the voltage decreases from 24.96% to 11.65% and the THD of the current decreases from 5.09% to 4.10% for the RL2 load. When the inverter changes from five to seven levels, it is observed that the THD value decreases, more power is transferred to the load, and it is transformed into a smoothed sine waveform. In future studies, it is planned to adopt advanced controller approaches based on deep learning. It is also planned to design a grid-connected wind energy conversion system with higher power output.