Projection of Safe Operation for Inverters Using Artificial Intelligence-Based Stability Criterion

This paper presents a new method to protect grid-interactive inverters against manipulated setpoints. The presented method applies an artificial intelligence-based stability criterion to detect harmful setpoints. The focus of this paper is on PQ-controlled inverters that can become unstable under weak grid conditions. This instability issue, which may unexpectedly happen in the nominal operating range of an inverter, is not observable under normal grid conditions. This paper formulates an embedded hybrid data-driven/analytical stability criterion to detect these hidden instability modes. The results demonstrate that the proposed method can significantly enhance the security of inverters by examining a setpoint through the hybrid stability criterion prior to engaging the setpoint to the local controller. The findings of this work are experimentally verified using a small-scale two-level 208 V, 5 kVA inverter feeding a 12 kW NHR 9410 power grid emulator.