Literature Review: Harmonic Injection for Power Transfer and Modulation Optimization (From 3rd-Harmonic Injection to the General n-th Harmonic Framework)

Harmonic injection has a long and interdisciplinary history spanning electric machines, power electronics modulation, and grid-connected power quality control. Across these domains, a recurring theme is the purposeful addition of selected harmonic components to a fundamental sinusoid to either (i) reshape the waveform envelope for improved DC-link utilization and reduced modulation distortion, or (ii) compensate for harmonic pollution and impedance-induced effects in power systems. A canonical single-harmonic formulation used throughout the literature is
\[
y_n(\theta)=\sin(\theta)+A_n\sin(n\theta),\qquad \theta=\omega t,
\]
where \(n\ge 2\) and \(A_n\) is selected to meet a specified objective (e.g., minimizing peak amplitude for fixed fundamental, harmonic compensation at a point of common coupling, or enhancing converter linear modulation range).

Roots in machine drives: from SVM issues to harmonic mitigation

Early motivation for systematic harmonic shaping is strongly linked to the evolution of high-performance AC drives. Space Vector Modulation (SVM), originally advanced to address torque pulsations (notably 6th-harmonic torque ripple) in multi-phase induction motor systems, revealed practical side effects including elevated harmonic currents in machine components under certain operating conditions. These observations accelerated interest in harmonic mitigation strategies that predate modern digital control: harmonic compensation methods were already discussed in mid-20th-century contexts such as analog amplifiers, analog computing, and thyristor converters. Within this early body of work, triplen harmonics—especially the third—were repeatedly observed to have “useful” waveform-shaping properties in polyphase systems, contributing to improved utilization and efficiency in specific drive and converter settings.

By the late 1990s, third-harmonic injection was widely recognized as a practical technique in polyphase induction machine control. The appeal was not simply “adding a harmonic,” but doing so in a way that preserves the fundamental component while reshaping peak values (i.e., reducing crest factor-like constraints) under amplitude saturation limits. This allowed improved operation margins and reliability in a range of induction machine implementations reported across the drive literature.

Power electronics modulation: third-harmonic injection to avoid overmodulation

A parallel track emerged in PWM inverter design. In early inverter generations, pure sinusoidal modulation could limit achievable line voltage and power delivery at rated motor speeds unless designers either oversized hardware (increasing cost and volume) or entered overmodulation regimes that introduced significant waveform distortion and undesirable harmonics. In this context, the Houldsworth–Duncan method (HDM) became a reference approach: the inverter reference signal is intentionally “distorted” by adding a third-harmonic component so that the composite signal’s peak is reduced without sacrificing the fundamental magnitude. A classic third-harmonic injected form is:
\[
y_3(\theta)=\sin(\theta)+A_3\sin(3\theta).
\]
The design goal in HDM is to select \(A_3\) so that \( \max_{\theta}|y_3(\theta)| \) is minimized, enabling a larger scaling factor before saturation (or equivalently enabling the same fundamental performance with less risk of overmodulation).

The HDM-style optimization typically begins by locating critical points via a stationarity condition:
\[
\frac{\partial y_3}{\partial \theta}=\cos(\theta)+3A_3\cos(3\theta)=0,
\]
and then selecting \(A_3\) to extremize the peak value at those critical points. Historically, a key practical observation strengthened adoption in three-phase systems: triplen harmonics are co-phasal across phases, so they vanish in line-to-line voltages, mitigating concerns about directly increasing line voltage harmonic content in balanced three-phase applications. This “line-to-line transparency” property of triplen harmonics is frequently cited as a major reason why the third harmonic became the default injection choice in three-phase inverter modulation.

Modern power systems: harmonic compensation in microgrids and distributed generation

In the last decade, harmonic injection also became a mainstream tool in grid-connected power electronics, particularly in microgrids and distributed generation. The motivation shifted from “DC-link utilization” toward “power quality and stability under nonlinear loading.” With rising penetration of renewable sources and inverter-interfaced assets, harmonic content in currents and voltages is increasingly shaped by converter control laws rather than passive network behavior. Harmonic injection in this context commonly targets: (i) mitigation of harmonics introduced by nonlinear loads, (ii) control of harmonic emissions of distributed generators, and (iii) harmonic current compensation to improve stability or reduce distortion at the point of common coupling. These applications typically embed harmonic injection inside current controllers, repetitive controllers, resonant controllers, or observer-based compensators, where injected terms may be tuned for specific harmonic orders observed in situ.

Despite the broadened application scope, a persistent pattern remains: the literature overwhelmingly favors odd harmonics, and in many practical implementations the third harmonic dominates. While the 5th and 7th are occasionally included (sometimes as multi-harmonic strategies), even-order harmonic injection is rarely contemplated in mainstream designs, and systematic mathematical explanations for this asymmetry are often absent or scattered across domain-specific arguments.

The “odd-only” tradition and the gap in comprehensive analysis

Several engineering reasons have been cited—implicitly or explicitly—for prioritizing odd harmonics. In balanced three-phase systems, certain harmonic families have structurally different impacts on line-to-line quantities and neutral currents, making triplen components particularly convenient in modulation contexts. In grid-tied control, even harmonics are frequently associated with asymmetry, DC offsets, measurement bias, or unbalanced operating conditions; they can interact with synchronization subsystems (e.g., PLL dynamics) and may be subject to stricter compliance constraints in certain standards or utility requirements. Still, these reasons tend to appear as “rules of thumb” or application-specific warnings rather than as a unified mathematical result tied to the waveform-shaping objective.

The open questions highlighted by the 2024 brief—“why is the third harmonic the only one widely used?” and “why are even harmonics rarely considered?”—are therefore best understood as gaps between (i) practical conventions established by decades of drive and converter design and (ii) a generalized, order-\(n\) analytic framework that formally explains when harmonic injection can reduce peak amplitude while preserving the fundamental.

Generalized harmonic injection as a unifying lens

A generalized formulation,
\[
y_n(\theta)=\sin(\theta)+A_n\sin(n\theta),
\]
enables a direct comparison across harmonic orders under a shared objective: reduce the composite waveform’s peak magnitude while maintaining the fundamental component. In HDM-like approaches, critical points are found by enforcing
\[
\frac{\partial y_n}{\partial \theta}=\cos(\theta)+nA_n\cos(n\theta)=0,
\]
and then selecting amplitudes based on extremal conditions at those critical angles. This generalized lens is significant for two reasons:

• It bridges domains. The same mathematics underlies “DC-link utilization” in PWM inverters and “waveform shaping” in certain compensation and test signal synthesis tasks.
• It clarifies harmonic-order effects. Instead of treating third-harmonic injection as a special trick, it frames it as the \(n=3\) member of a family, allowing rigorous comparison of achievable peak reduction and associated trade-offs.

The literature has historically included attempts to extend third-harmonic logic to other orders, but without a consolidated proof structure, results were often presented empirically (plots, simulations) or constrained to specific operating scenarios. The 2024 brief’s contribution is to formalize the order-\(n\) landscape and connect observed trends—such as even-order injection failing to reduce peak amplitude and higher odd orders delivering diminishing returns—to explicit analytic conditions derived from the generalized model.

Toward multi-harmonic injection: an emerging research direction

Beyond single-harmonic injection, there is a growing interest (though still limited in consolidated theory) in multi-harmonic waveform synthesis, where the injected signal becomes
\[
y(\theta)=\sin(\theta)+\sum_{k\in\mathcal{K}}A_k\sin(k\theta),
\]
with \(\mathcal{K}\) a chosen set of harmonic orders. In practice, multi-harmonic strategies are motivated by either tighter constraints (e.g., peak minimization under additional spectral limits) or by system-level objectives (e.g., shaping impedance interaction, compensating multiple dominant distortion components). However, multi-harmonic optimization is inherently higher-dimensional and typically requires either numerical methods or structured analytic reductions. This makes a generalized single-harmonic theory valuable as a baseline and as a building block for future multi-harmonic design methodologies.

Summary

In summary, harmonic injection literature spans (i) drive control and SVM-related mitigation, (ii) PWM inverter modulation for improved utilization and reduced overmodulation distortion, and (iii) grid-connected power quality control in microgrids and distributed generation. Across these areas, the third harmonic became dominant due to a blend of practical system properties (notably in three-phase line-to-line behavior) and effective peak-reduction capability for reference waveforms. Yet the broader question of “which harmonic orders can fundamentally achieve peak reduction while preserving the fundamental” has not always been answered in a unified analytic way. The generalized order-\(n\) framework directly addresses this gap and provides a principled foundation for comparing odd and even harmonics and for motivating future multi-harmonic extensions.

References

Use the original paper’s reference list numbering to keep consistency. Below is a placeholder structure aligned with the citations mentioned in the excerpt:

• [1] Gopakumar et al., Space Vector Modulation development for multi-phase induction machines.
• [2] Follow-up work noting rotor harmonic currents under SVM in dual three-phase systems.
• [3]–[5] Early harmonic mitigation techniques in analog amplifiers, analog computers, and thyristor converters (1950s).
• [6] Houldsworth and Duncan, Third-harmonic injection method for PWM inverter modulation.
• [7]–[8] Harmonic injection adoption in polyphase induction machine control (1990s onward).
• [9] Harmonic injection in power electronics and inverter modulation literature.
• [10] Harmonic mitigation for nonlinear loads in grid contexts.
• [11] Harmonic emission control for distributed generators.
• [12] Harmonic current compensation for improving grid stability.
• [13] Multi-harmonic injection attempts (limited coverage in literature).
• [14] Notes on stricter restrictions/concerns for even harmonics (e.g., asymmetry, PLL impacts, compliance constraints).