Transformerless inverters are being used for integrating photovoltaic (PV) sources into the grid. Highly efficient and reliable inverter concept (HERIC) inverter is one such topology, which is derived from an H-bridge by adding a bypass leg on the ac side using two back-to-back insulated gate bipolar transistors (IGBTs). Open-circuit fault in the bypass leg would result in conduction of current through antiparallel diodes of the main switches of the inverter, effectively making the inverter operate in bipolar mode. As a result of the fault, there would not be an issue of leakage current, and the operation would continue at the expense of increased conduction losses and dc-offset in the grid current. This article proposes an online noninvasive model-based technique to detect switch faults in the bypass leg of the HERIC inverter. The grid current at the peak of the carrier signal is predicted based on the measured grid current at the start of the zero states. By comparing the estimated and measured grid currents at the peak of the carrier signal, the faulty switch in the bypass leg is localized. Once the fault is localized, the inverter's modulation scheme is changed to conventional bipolar mode to improve the performance of the converter during the post-fault stage. A detailed simulation study is carried out to verify the effectiveness of the proposed approach. Experimental tests on the prototype converter show that the proposed algorithm can reliably detect the open-circuit fault in switches of the bypass leg, and once a fault is detected, the algorithm would change the modulation scheme to bipolar mode to minimize the inverter loss and dc-offset in the grid current. © 2013 IEEE.