I. Introduction
Significant penetration of inverter-based resource (IBR) (solar PV, wind generators, etc.) in the power system reduces total system inertia, which results in weak grid operation as the remaining alternators are not sufficient to provide frequency and voltage support to the system. At present, most of the grid-tied inverters are working in grid-following mode, and the voltage and frequency reference to these inverters is provided by the power system [1]. The major purpose of GFL inverter is to supply active and reactive power to the network, and they disconnect themselves in case of a fault. But a power system with a large penetration of IBR’s cannot afford the removal of the inverter during a fault as the total system inertia is low; therefore grid-forming inverter emerges as a possible solution in this scenario. The conceptual difference between GFL inverter and GFM inverter [1], [2] and different control algorithm of GFM inverter is explained in [3], [4]. The concept of a GFM inverter [5] is not new; it is already implemented in microgrids, where this inverter generates reference voltage and reference frequency for the GFL inverter. The fault current shared by the traditional synchronous machine is up to 10 p.u. [6], however, power electronics devices have certain thermal limitations and cannot supply such high currents. This makes the current limiter indispensable for GFM inverters. The current shared by GFM inverter is normally restricted to (1.2-2) p.u., which cannot be detected by the traditional protection system and this level of current can also be reached during phase jump and overload conditions. Therefore, characterization [6] of the fault current of GFM inverter is necessary for correct operation of protection system in hybrid power system comprising GFM inverter with GFL inverter and synchronous generator. During an overcurrent condition current can be controlled by converting the GFM inverter into a GFL inverter, but this is not advisable for the IBR’s dominated power system as it will lose its voltage control capability, and the transition between voltage and current control modes should be smooth and other than this, a backup PLL is also required. We can broadly divide current limiter into two parts, one is current based current limiter and other is voltage based current limiter. Current based current limiters such as d-q component current limiters, priority-based current limiters, and vector amplitude limiters [7] reduce the current reference and voltage based current limiter such as virtual impedance reduce the voltage reference to limit the current. Wind-up [8] is a common phenomenon that can be seen in the current based current limiters and this can be avoided by adopting clamping and back-calculation techniques. Virtual impedance does not introduce windup problems in the control loop; however, when the inductance value is increased to limit the current reference, dc bias is also added to the reference current [9]. Hence, current based current limiter with the anti-windup method is the simplest method. Virtual impedance can also be used for the post fault oscillation damping [8], [10] but virtual impedance leads to virtual power losses in addition to large reactive power which is require for fault recovery and this further leads to slow fault recovery and the risk of instability. Therefore, adaptive virtual impedance is proposed in [10]. Asymmetrical faults are more frequent than symmetrical faults, and these faults introduce negative sequence components in the response of GFM inverter, which result in double frequency components in d-q axis current and voltage. When a current saturator is used for limiting the current, these double-frequency components are clipped by the saturator, and the voltage and current responses are distorted. Therefore, if we are using a synchronously rotating frame, it is required to filter out the positive and negative sequence components of fault current [11], [12]. When the current is saturated, the power angle curve deviates from its previous curve [13] which further leads to a reduction in the clearing time or decreases the stability margin. During faults huge mismatch occur between nominal power and measured power, resulting in the rate of change of virtual power angle to increase, and if the virtual power angle crosses the critical clearing angle, then it leads to the loss of synchronism of GFM inverter with the grid. The current saturator not only decreases the maximum power transfer capability but also develops an angle deviation between the saturated VPA curve and the unsaturated VPA curve as shown in Fig.1 [13], [14]. Where the stable equilibrium point is , the unstable equilibrium point and is stable and is an unstable equilibrium point. For a stable equilibrium point [14], [15], there should be an intersection of nominal power with both unsaturated and saturated VPA curves. The transient stability limit of the network can be improved by increasing the maximum value of saturated power, and in [14], this is achieved by implementing a stability enhanced droop controller. The transient stability and slow recovery [16–17] can be improved by a power angle limiting controller. In [18], fast fault recovery and improved transient stability are achieved by an adaptive-droop coefficient with a virtual impedance. If the current is limited by reducing the current reference, the voltage reference cannot be used for the calculation of the adaptive droop coefficient because the value does not vary significantly during a fault condition. Therefore, in this work, to prove the effectiveness of the proposed method for a current based current limiter, a component current limiter is used, which is one type of current based current limiter, and the measured value of PCC voltage is used to calculate the droop coefficient modifier.