Why MPC Outperforms PI Controllers in AFE Rectifiers

Model predictive control outperforms conventional PI controllers in active front-end rectifiers because PI regulators are fundamentally ill-suited for tracking AC reference currents — particularly near the zero-crossing region — causing degraded power factor and elevated total harmonic distortion in line current. Research from Ajou University (2018) on single-phase Vienna rectifiers demonstrated that this sluggish dynamic response is a root cause of THD elevation in both continuous conduction mode (CCM) and discontinuous conduction mode (DCM), where classical PI control fails to follow the AC reference adequately. A model-based predictive controller was shown to improve average current tracking across both operating modes.

The Niroo Research Institute, Tehran (2019) corroborated this finding with a continuous control set MPC (CCS-MPC) implementation for three-phase controllable rectifiers, showing that a properly tuned CCS-MPC — with systematically chosen sampling time, prediction horizon, and control horizon — delivers acceptable THD and power factor under industrial operating conditions. Critically, the study showed that the tradeoff between computational burden and dynamic performance is manageable when these horizon parameters are selected with care.

MPC’s core advantage for unity power factor operation lies in its ability to simultaneously minimise active and reactive power errors within a single cost function. The Electric Power Research Institute of Guangdong (2021) proposed a three-vector model predictive direct power control (MPDPC) strategy for three-phase PWM rectifiers that reduces current harmonic content and suppresses instantaneous power pulsation under unbalanced grid conditions — a scenario where single-vector approaches fail. The duty cycle of selected voltage vectors is calculated by solving an optimal objective function based on instantaneous power error, yielding improved steady-state power quality.

The University of Seville’s landmark 2014 review — which remains a foundational reference across the field — identifies active front-end rectifiers as one of the primary application domains where MPC has been “successfully used” to handle multivariable cases, system constraints, and nonlinearities. The review catalogues both current control and power control variants, noting that MPC’s intuitive handling of constraints makes it particularly attractive for grid-interfaced converters where harmonic compliance standards must be enforced. This observation aligns with the requirements of standards published by IEC and the IEEE 519 harmonic limits framework.

“MPC’s intuitive handling of constraints makes it particularly attractive for grid-interfaced converters where harmonic compliance standards must be enforced — a capability that PI controllers structurally cannot replicate.”

Fixed Switching Frequency and Dead-Time Compensation: Translating Fast Dynamics into Low-THD Steady State

The primary challenge preventing FCS-MPC from achieving low-THD steady-state performance in practical rectifiers is its inherent variable switching frequency. Because FCS-MPC selects a single optimal voltage vector per sampling interval, the harmonic spectrum spreads unpredictably — complicating filter design and potentially violating IEEE 519 harmonic limits. Two research groups have developed complementary solutions to this problem that are now considered foundational to practical MPC deployment.

Henan Polytechnic University (2020) addressed the issue in a single-phase five-level PWM rectifier by combining FCS-MPC with deadbeat control technology to eliminate delay error and reduce harmonic distortion. The approach introduces a simplified objective function — replacing iterative optimisation — that reduces computational burden while retaining fast transient response and achieving accurate current loop tracking. The five-level topology provides additional degrees of freedom for harmonic shaping that are not available in conventional two-level rectifiers.

South China University of Technology (2021) proposed a complementary approach: combining an effective vector and two zero vectors to fix the switching frequency, alongside a revised current prediction equation that compensates for dead-time effects on switching vector action time. Dead-time distortion — a well-documented source of current waveform degradation in PWM converters — is directly targeted by this method, improving grid-side current THD in single-phase PWM rectifiers. The approach is significant because it addresses both the spectral spreading problem and the dead-time distortion problem within a unified predictive framework, rather than treating them as separate issues requiring separate compensators.

For multi-level topologies, the University of Seville’s 2019 work on multiphase drives identified the harmonic generation inherent in fixed-sampling-time FCS-MPC itself, proposing variable sampling times and observer-based approaches as natural remedies. This represents a deeper solution to the spectral spreading problem: rather than post-hoc correction, the control law is adapted to avoid generating harmonics in the first place. The weighting factor design challenge — balancing harmonic content, switching frequency, active power error, and reactive power error in a single cost function — remains an open research problem, with comparative studies from Shandong University (2023) documenting heuristic, offline tuning, sequential, and online optimisation methods tested on hardware-in-the-loop platforms.

Industrial Applications: From Medium-Voltage Cascaded Drives to Aerospace Power Systems

FCS-MPC applied to multicell active front-end rectifiers eliminates the need for bulky multipulse input transformers in medium-voltage AC drives operating at 2.3 kV to 13.8 kV, reducing system cost and weight while achieving harmonic cancellation equivalent to that of a multipulse transformer. This finding — established by a cluster of Chilean universities including Universidad de Talca (2022), Universidad del Bío-Bío (2021), and Universidad Católica de la Santísima Concepción (2021) — represents the most industrially targeted body of work in the dataset for high-power applications.

Induction Motor and Multi-Motor Drive Systems

For industrial induction motor drives, active front-end rectifiers with MPC deliver unity power factor and sinusoidal input current through voltage-oriented control (VOC), enabling bidirectional energy exchange between the load and the grid. Research from St. Petersburg Mining University (2021) reports that frequency converters equipped with active rectifiers maintain a power factor of 1 and sinusoidal input current — a prerequisite for regenerative braking energy recovery in industrial crane and elevator drives. The University of Niš (2020) extended this analysis to a full regenerative multi-motor drive system for an industrial crane using Siemens converters with an active front-end converter and multiple voltage-source inverters on a common DC bus, with LCL filter modelling and simulation-validated analysis of rectifier dynamics confirming the system’s influence on the distribution network in terms of higher harmonics and power factor.

Aerospace and Variable-Frequency Power Systems

Variable-frequency aerospace power systems present unique challenges for rectifier MPC because the AC side voltage frequency changes over a wide range, invalidating fixed-frequency model assumptions. Northwestern Polytechnical University (2018) addressed this with a Model Predictive Direct Power Control (MPDPC) approach incorporating Bayesian estimation to identify AC side impedance variations online — including filter inductance and equivalent series resistance — in Aircraft Alternating Current Variable Frequency (ACVF) power systems. This enables the predictive controller to maintain accurate power regulation despite parameter drift, directly preserving power factor and harmonic performance under airborne operating conditions. The challenge of wide-frequency operation is also relevant to electric vehicle applications, where Tiangong University (2022) identified that frequency and resistive parameter variations produce steady-state reactive power errors in deadbeat direct power control, proposing a repetitive control augmentation to compensate for both frequency-induced and temperature-induced parameter mismatches. Standards for aircraft electrical systems are defined by bodies including MIL-STD and SAE International.

Robustness Under Grid Disturbances and Unbalanced Conditions

Conventional model predictive direct power control produces high harmonic content in line currents and large instantaneous power pulsation under asymmetrical grid conditions — a practical limitation that restricts its deployment in grids subject to voltage unbalance or voltage dips. Two distinct research approaches have emerged to address this limitation, each targeting a different aspect of the robustness problem.

North China Electric Power University (2022) proposed model-free predictive power control (MFPPC), which extends the model-free predictive current control principle to power control and uses an extended finite control set of voltage vectors. By eliminating dependence on accurate system models, MFPPC maintains THD compliance and suppresses instantaneous power pulsation under asymmetrical grid conditions where conventional MPDPC fails. This approach is particularly relevant for industrial installations connected to weak or distorted grids — a growing concern as grid impedance increases with the penetration of distributed energy resources, as documented by the International Energy Agency.

For voltage dip scenarios, St. Petersburg Mining University (2018) characterised the mechanism by which three-phase voltage unbalance during short-circuit events distorts rectifier currents in voltage-oriented control, and proposed an improved control approach to restore sinusoidal current waveforms under unbalanced conditions. This is a critical requirement for maintaining grid compliance in heavy industrial installations — such as mining, steel, and cement production — where voltage dips are frequent and the cost of non-compliance with harmonic standards is significant.

“Model-free predictive power control eliminates dependence on accurate system models, maintaining THD compliance under asymmetrical grid conditions where conventional MPDPC produces high harmonic content in line currents and large instantaneous power pulsation.”

Innovation Landscape: Academic Leaders, Industrial Patents, and the Weighting Factor Frontier

The MPC-AFE rectifier field is concentrated in a small number of highly productive research clusters, with Chinese institutions constituting the largest single national grouping across the dataset. Contributors from Northwestern Polytechnical University, South China University of Technology, North China Electric Power University, Henan Polytechnic University, Tiangong University, and China University of Mining and Technology collectively span aerospace rectifiers, grid-connected PWM rectifiers, multi-level topologies, and AI-enhanced predictive controllers. Chilean universities — Universidad de Talca, Universidad del Bío-Bío, and Universidad Católica de la Santísima Concepción — form a visible cluster with work focused specifically on multicell AFE rectifiers for medium-voltage industrial drives using FCS-MPC, spanning 2021–2022.

Industrial Patent Activity: Rockwell Automation and CATL

Industrial patent filings confirm that MPC-based AFE rectifier control is transitioning from academic research into commercial products. Rockwell Automation Technologies holds patents describing the use of a switching rectifier modulated to indirectly control DC link current via motor flux adjustment, correcting power factor over a wide speed range — representing an early industrial implementation of coordinated rectifier-inverter power factor control in motor drives, with patents filed from 2008 to 2010. Contemporary Amperex Technology Co., Limited (CATL) filed a 2023 European patent integrating MPC with a Hamiltonian dissipation model for injection damping, targeting stable DC output current, enhanced system stability, and weakening of power transmission line impedance effects — a sophisticated convergence of passivity-based control theory with model predictive methods that signals the entry of large-scale battery manufacturer supply-chain equipment into this domain. Patent data from the European Patent Office confirms the growing international scope of these filings.

Weighting Factor Design: The Critical Unsolved Challenge

Across all MPC implementations, the design of cost function weighting factors is the most frequently cited barrier to broader industrial deployment. When simultaneously controlling active power, reactive power, switching frequency, and harmonic content, the relative weights assigned to each term in the cost function fundamentally determine control performance — yet no systematic design methodology has achieved consensus. Comparative studies from 2021 and from Shandong University (2023) document the difficulty of optimal weighting factor selection and compare heuristic, offline tuning, sequential, and online optimisation methods using hardware-in-the-loop experiments. This represents a maturing research front as MPC transitions from academic demonstration to industrial deployment, and is likely to be a focus of standardisation efforts by bodies such as IEEE as the technology matures.