Access Official Technical PDF

Share Summary:LinkedInEmail

Nonlinear loads such as rectifiers and adjustable-speed drives can draw harmonic and reactive current from an AC supply. These currents can reduce power quality and affect other equipment connected to the system. A shunt active power filter (APF) can compensate for these effects, but sudden load changes also disturb the filter’s DC-link voltage. A proposed shunt active power filter control strategy combines proportional-integral (PI) and sliding mode control to improve voltage response during transients.

The hybrid approach, called sliding mode proportional-integral (SMPI) control, adapts the gains of a PI voltage controller according to the voltage-loop error and its rate of change. A transition rule switches to fixed PI gains near steady state, helping limit chattering associated with sliding mode control. The approach was evaluated in MATLAB/Simulink and compared with a conventional PI controller.

Why DC-link voltage matters in a shunt active power filter

A shunt APF uses a voltage source inverter (VSI), connected in parallel with the load through a filter impedance, to inject compensating current. Its DC-link capacitor provides the energy storage needed for inverter operation. The controller must maintain the capacitor voltage while the inverter responds to changing compensation demands.

When load power changes abruptly, the energy balance in the system changes as well. The DC-link voltage can deviate from its reference, and the controller must respond to restore it. A conventional PI controller is commonly used for this task, but its performance depends on its selected gains. The paper examines whether adapting those gains can improve the response to transient conditions.

Active filters can also be controlled using methods that extract harmonic components from load current, including synchronous reference-frame and instantaneous power approaches. The strategy discussed here instead uses a balanced-energy method. It regulates grid currents indirectly, without a separate load-harmonic detection scheme, and is designed to avoid measuring load currents for this control approach.

How the SMPI DC-link voltage controller works

The SMPI controller combines the familiar PI structure with sliding mode logic. The controller monitors the difference between the DC-link voltage reference and measured capacitor voltage, along with the change in that error. These signals define a sliding surface, which guides adjustments to the proportional and integral gains.

In practical terms, the gains are not held at one fixed setting throughout the response. Switching laws select gains based on the sliding surface, allowing the controller to respond differently as the voltage error evolves. The gains are designed using positive constants, with their values selected through standard PI design methods as described in the study.

Sliding mode switching can produce chattering: rapid changes in the control action around the switching boundary. The study addresses this with a transition rule based on a Gaussian function of the voltage error. The function is compared with a threshold to decide whether the controller uses switching SMPI behavior or fixed-gain PI behavior. This allows adaptive action during larger deviations and fixed gains as operation approaches steady state. The function’s smoothness parameter affects how sensitively the transition responds to voltage error.

Current regulation without a separate harmonic detector

The DC-link voltage loop generates a reference for the active component of grid current. A phase-locked loop provides the supply phase angle, which is used to form reference phase-current components. Separate current controllers then regulate the grid currents using a sequence-control approach for positive and negative sequences. The control signals are converted to three-phase references, and PWM determines the VSI switching duty ratio.

This arrangement is relevant to system designers because it separates the energy-balancing task from explicit load-harmonic extraction. The method does not mean that the APF operates without current control; rather, it regulates grid currents indirectly through the voltage-control and reference-generation process. The study notes that avoiding load-current measurement may reduce the number of current sensors required in an experimental implementation.

Simulation results during startup and load changes

The reported MATLAB/Simulink model included a three-phase supply, an uncontrolled rectifier nonlinear load, a VSI, filter inductors, and a DC-link capacitor. The comparison assessed startup, changes in DC-link reference voltage, load transients, and current total harmonic distortion (THD).

In the simulations, SMPI produced a smoother startup response, with less overshoot and a shorter response time than the conventional PI controller. It also performed better during upward and downward steps in the voltage reference. These reference changes were included for comparison; the paper states that a constant reference is used in normal APF operation.

For a simulated increase in load power, created by connecting an additional three-phase resistive load in parallel with the nonlinear load, the reported DC-link voltage undershoot was approximately 12% with SMPI and 23% with conventional PI. The study also reports improved performance with SMPI when the added load was disconnected. These figures describe the specific simulation conditions and should not be treated as universal performance guarantees.

The simulation reported current THD of 29.27% before compensation and 1.02% after compensation. This result illustrates the modeled filter’s harmonic-compensation performance; it is separate from the comparison of DC-link controller responses.

Practical implications for APF design

For engineers and integrators evaluating shunt APF control, the central design consideration is the balance between transient responsiveness and steady-state behavior. Adaptive gains can help the DC-link controller respond to abrupt changes, while a defined transition to fixed PI gains can reduce the unwanted switching activity associated with sliding mode control.

As with any simulation-based result, implementation decisions require attention to the actual system, including its supply, load, filter, and controller design. The study provides a control concept and comparative simulation results, not a guarantee that the same response will be achieved in every installation.

Conclusion

The SMPI strategy combines PI voltage regulation with sliding mode gain adjustment to address DC-link voltage deviations during load and reference transients. Its transition rule is intended to retain adaptive response when needed while limiting chattering near steady state. In the reported simulations, it outperformed conventional PI control on the tested transient responses, while the balanced-energy approach regulated grid currents without a separate harmonic-detection scheme. For plant and system designers, the practical value lies in considering both transient voltage regulation and current-sensing requirements when selecting an APF control architecture.

Access Official Technical PDF

About The Author