**5. Discussion**

An adaptive OC protection for the AC microgrid using the IEC 61850 communication standard and LVRT capability of DGs is presented in this paper. Previously, an adaptive OC protection was presented in [48]; the scheme updated the inverse-time OC relay curve by changing the pick-up current with respect to DG infeed, but the focus was mainly on medium impedance faults at the end of the radial distribution network detected by a single-substation relay. The similar type of adaptive OC protection was proposed in [49] for distribution networks with DGs using local data and two-setting groups for the inverse-time OC relay. However, the scheme does not use remote data by communication systems and is prone to nuisance and slow trips, resulting in DG loss. A directional adaptive inverse-time relay was presented more recently for HIL (Hardware-In-Loop) testing, and real-time simulations in [50,51]. In these papers, an FCL was used for limiting the wind turbine generator fault contribution, and its e ffects on relay settings were observed. This directional adaptive inverse-time relay using a multiagent system does not consider/mention the e ffects of communication delays on protection coordination. Moreover, it involves tedious calculations to generate various inverse time setting groups for changing the network configurations and those for all relays in the networks. Although, traditionally, inverse-time OC performs better than definite-time OC in terms of the minimum operation time close to the power source, but in the AC microgrid environment with many DG sources, this may not be completely true. The inverse-time OC relays are a ffected the most with the increased penetration level of DGs [4], and it is the usual practice to limit the fault contribution from DGs to overcome the adverse e ffects, as it was also done in [51]. An adaptive OC protection for distribution networks was proposed in [52] that calculated and applied the new settings of OC relays directly whenever any significant change in the network occurred. The algorithm presented did not use precalculated settings and was initiated either by the monitoring block in the coordination layer or energy managemen<sup>t</sup> system during topology changes using the communication link. The scheme was verified using the real-time digital simulator (RTDS) and IEC 61850 GOOSE messaging. However, the adaptive OC scheme was implemented using a centralized approach; the type of DG unit was not specifically described, and higher coordination delays were used. Moreover, the proposed adaptive OC scheme also gave slower tripping times compared with the traditional OC relay in some cases. In this paper, the communication-based definite-time OC relays with only two predefined setting groups are suggested. These two setting groups can be changed adaptively and quickly by each IED autonomously after receiving the status of DGs (on/off), CB status change (open/close), fault current magnitude and fault detection GOOSE signal from other IEDs. Moreover, the simple method of calculating the magnitude of the current at their locations, comparing it with the predefined threshold 1.2 p.u. of the max current and sharing this information with other IEDs will be useful for quick detection of the fault location after knowing the status of each DG. The nuisance tripping can also be avoided with fault current magnitude sharing between IEDs and with the careful use of trip block/release signals. The proposed current magnitude comparison method also avoids the additional measurement of voltage for the detection of the fault current direction. However, a voltage magnitude measurement can be used as a local backup protection for OC function. The proposed scheme in this paper can also be extended to low impedance single-phase ground faults and other asymmetrical faults. The method proposed in this paper will be evaluated with a real-time digital simulator of OPAL-RT for hardware-in-the-loop (HIL) simulations using actual Ethernet-based GOOSE communication between IEC 61850-based IEDs from di fferent vendors for its practical implementation, and this will be presented in a separate research article in the future.
