Unified adaptive neuro-fuzzy inference system control for OFF board electric vehicle charger

This paper illustrates the control algorithm design for a two stage off board bidirectional smart electric vehicle (EV) charger architecture. The proposed EV charger is controlled to perform four quadrant operation (i.e. vehicle-to-grid (V2G) and grid-to-vehicle (G2V)) while compensating the load harmonics, simultaneously. Here, the 3-phase AC-DC converter and DC-DC converter are two main components, where the first one is controlled to regulate DC-link voltage as well as reactive power and current harmonics of nearby non-linear load while the second one regulates the exchange of active power. Generally, the AC-DC converter is controlled via two control loops, i.e. outer loop and inner loop. However, setting the gains of almost four proportional integral (PI) controllers and determining decoupling terms for inner control loops is very difficult, especially under the dynamic operating conditions. Therefore, an adaptive neuro-fuzzy inference system (ANFIS) has been designed to estimate the direct and quadrature axis reference currents directly while regulating two different quantities in single step only and hence, named as unified ANFIS controller. The proposed EV charger is simulated in MATLAB/Simulink and controller performance is validated with scaled down hardware model in real time.

Automated summary

This paper presents the control algorithm design for a two stage off board bidirectional smart electric vehicle charger architecture, utilizing an adaptive neuro-fuzzy inference system to estimate various parameters for effective charger control.