Key points are not available for this paper at this time.
This paper introduces an advanced control method for a Shunt Active Power Filter (SAPF), engineered specifically for the compensation of non-active current and power management in DC-powered systems. Non-active current components frequently arise in practical DC systems due to power electronics and dynamic loads. Their presence leads to increased current draw from the source, higher losses, and accelerated deterioration of DC energy providers, such as fuel cells and batteries. The proposed SAPF control strategy is based on the concept of an equivalent conductance signal, which dynamically reflects the load’s active power consumption and the SAPF’s internal losses. A key feature of this method is the derivation of the conductance signal primarily from the DC-link capacitor voltage, effectively eliminating the need for additional current or power sensors and thereby simplifying the control hardware and software. This methodology enables efficient buffering of energy flow through user-defined time constants, significantly reducing both the average value and the variability range of the current required to transmit the demanded power (as measured by the RMS parameter and standard deviation of the source current, respectively). As a result, the degradation process of energy sources can be mitigated. Furthermore, the conductance signal’s ability to assume negative values allows for effective management of generative loads, enabling power flow back into the system or directing it to specific loads. The flexibility of tuning the SAPF’s functionality—by adjusting the time constant and imposing limits on the conductance signal’s variation range—is demonstrated in the presented results. Simulation examples, including the potential for direct energy exchange with the DC-link capacitor without affecting the upstream source, validate the effectiveness and versatility of the proposed control method in improving power quality and extending the lifespan of DC energy storage systems.
Andrzej Szromba (Mon,) studied this question.