Multilevel Inverters

Multilevel inverters (MLI) are power electronics converters that transform electrical energy from a direct current (DC) source into alternating current (AC) at multiple output voltage levels. These structures have been developed to reduce harmonic content in the output waveform, limit voltage stress on switching components, and improve energy efficiency in high-voltage and high-power applications. They are widely used in renewable energy systems, HVDC transmission, motor drives, and electric vehicle technologies.

^{Yedi Seviye Çok Seviyeli Evirici Çıkış Gerilimi Gösterimi}

The first multilevel inverter topology was introduced in 1975 and gained broader application in the 1980s with the adoption of diode-clamped topologies. Subsequent years saw the development of flying capacitor and cascaded H-bridge topologies. Starting in the 2000s, increasing demands for modularity, harmonic reduction, and efficiency led to the integration of advanced control techniques into these structures. Today, multilevel inverters serve as fundamental power conversion units in applications ranging from grid-connected solar energy systems to electric vehicle propulsion systems.

The Neutral Point Clamped (NPC) topology connects each phase leg to different voltage levels through clamp diodes placed between capacitors connected to the DC link. The three-level NPC structure is widely used in industry. Achieving four or more levels is limited by the increased number of diodes and the challenge of balancing capacitor voltages. This topology is typically preferred in medium-voltage motor drives.

^{Tek-Faz Üç Seviye Diyot-Kenetli Evirici Topolojisi}

In the Flying Capacitor (FC) topology, each phase leg generates multilevel output using flying capacitors. This structure operates without clamp diodes and allows balancing of capacitor voltages through phase-internal switching redundancy. However, the initial charging process and regulation of capacitor voltages increase application complexity.

The CHB topology consists of H-bridge cells connected in series, each powered by an independent DC source. Its modular structure enables individual control of each cell and provides strong harmonic suppression capability, making it particularly suitable for photovoltaic (PV) applications. Symmetric configurations use equal DC voltages across all cells, while asymmetric configurations achieve higher output levels with fewer cells by using different voltage levels.

In hybrid topologies, different multilevel structures are combined. For example, the high-voltage stage of a phase leg may use an NPC configuration while the low-voltage stage employs a CHB structure. This approach reduces cost and system complexity by decreasing the number of switching components. Reduced Device Count (RDC) topologies offer simplified configurations for the same purpose.

In carrier-based PWM techniques, a sinusoidal reference signal is compared with N-1 triangular carrier signals. This method offers advantages such as fixed switching frequency and easy integration with digital controllers. The most common variants are:

• Phase Disposition (PD): All carriers are in phase, enabling simple implementation.
• Phase Opposition Disposition (POD): Carriers are divided into two groups, upper and lower, and placed 180° out of phase.
• Alternate Phase Disposition (APOD): Each carrier is phase-shifted by 180° relative to the previous one.

SHE optimizes switching angles to eliminate specific harmonics from the output. Using Fourier series analysis, this method sets the amplitude of targeted harmonics to zero and is especially effective in high-power, low-switching-frequency systems. However, real-time control is challenging.

SVM involves transforming three-phase systems into a two-dimensional vector plane and selecting optimal switching vectors. It offers high harmonic suppression performance and more efficient DC voltage utilization. However, its application to multilevel systems is complex.

PI (Proportional-Integral) controllers generate system response using the integral of the error signal and are the fundamental controllers for voltage and current regulation. PR (Proportional-Resonant) controllers provide zero steady-state error for AC reference signals; their resonant structure effectively suppresses errors at specific frequencies.

This method consists of an inner current loop and an outer voltage loop. The inner loop regulates the output current while the outer loop monitors the DC link voltage or grid voltage. This structure enhances system stability and improves adaptation to dynamic conditions.

• Hysteresis Control: Switching occurs when the current exceeds a predefined tolerance band. It is fast but does not operate at a fixed frequency.
• Model Predictive Control (MPC): Optimal switching is selected by modeling the system’s future behavior.
• Capacitor Voltage Balancing: For stable operation of FC and NPC topologies, capacitor voltages are regulated using redundant switching states or zero-sequence voltage injection.

In grid-connected photovoltaic systems, control of multilevel inverters is critical for both power quality and energy management. The following control structures are commonly used in such applications:

To maximize the output power of the Maximum Power Point Tracking (MPPT) algorithm is applied (e.g., Perturb and Observe). The reference voltage determined by MPPT is fed to the DC link voltage loop. The output current is compared with a grid-synchronized reference and tracked using a PR-type inner current controller.

The output current must be phase-aligned with the grid voltage. For this purpose, a Phase Locked Loop (PLL) is used to track grid frequency and phase. This enables active power delivery to the grid and allows reactive power control.

Harmonics originating from the multilevel structure are suppressed via PR controllers. Simultaneously, the phase of the current injected into the grid is adjusted to maintain the power factor close to unity or to provide reactive power as required by the user.

• Low modularity; system expansion is complex.
• High number of switching components and diodes; complexity increases significantly beyond three levels.
• Difficult to maintain balanced capacitor voltages; active balancing methods are required.
• Commonly used in medium-voltage industrial motor drives and HVDC systems.

• Moderate modularity, but requires many capacitors per phase.
• Does not require clamp diodes; voltage balancing is achieved internally within each phase.
• Capacitors must be pre-charged and their voltages actively regulated.
• Used in limited applications such as UPS systems and low-power PV installations.

• High modularity; individual control of each cell is possible.
• Moderate number of switching components; scalability is easy due to modularity.
• Asymmetric configurations can achieve higher output levels with fewer cells by using different voltage levels.
• Commonly used in grid-connected solar energy systems, wind turbines, and distributed energy applications.

• Modularity and structural properties vary depending on the topology combination.
• Number of switching components is reduced compared to conventional topologies.
• Control complexity may increase, particularly regarding capacitor voltage balancing and modulation synchronization.
• Preferred in high-power or customized industrial systems.