Properties of reactive current injection by AC power electronic systems for loss minimization

This paper derives the mathematical basis for analysing optimal injection of reactive current from power electronic interfaces, considering loss minimization in an AC system. The power electronic interface, working as load or generation, is modelled as constant power on the AC side, assuming a high bandwidth controller maintaining the DC-link voltage constant. Under these considerations the mathematical condition for optimal share of both active and reactive currents are derived for a system composed of a single load connected to a network represented by its Thevenin equivalent. The presented analysis is based on the active and reactive current components of the converter, which better suits a system in which the variables under control are the currents, in contrast to the classical power system approach in which active and reactive power are the variables under control. The derived mathematical expressions illustrate the properties of reactive current control for system loss minimization under changing grid impedances and grid voltage. An interesting property to be highlighted is the relation between power factor and minimum losses of the system: for the realistic case of non-zero grid impedance it is demonstrated how the widely accepted unity power factor approach is clearly not the optimal point of operation, neither from the grid perspective nor from the power electronics perspective.