**2. Problem Statement: OCR Coordination**

OCR is considered the most common apparatus for protection as applied in distribution systems. An OCR is used in measuring the current passing through it and also determines if a signal for opening a circuit breaker is to be sent or not [27]. Relays are of different types, some of which include definite-time OCRs and directional relays. However, one of the most preferred types is the inverse-time OCR since it is a protection relay with a time characteristic used for grading and, thus, can allow some loads to specifically draw more currents in a very short period of time [28]. OT is part of inverse-time OCRs, and is found to be in inverse proportionality with the fault current as indicated by the relay. OCRs are in two forms: electromechanical and digital. Electromechanical OCRs have dominated the market for the past two decades. This is because they were not expensive and had well-known performance, resulting from many years of application. The second form, which is the digital OCR, has several advantages over the electromechanical type and is more likely to be preferred in the future for the following reasons [29,30]: first, they are economically competitive, since they are cheap to acquire, similar to the electrochemical types. Second, they have increased reliability since they have properties that can detect and report any internal problems in the relay, thus avoiding any possible malfunction operation. Third, they have smart grid natives since they are compatible with the concept of the smart grid due to their digitalized nature. Further, they have a multifunctionality ability; thus, they can perform other added tasks such as measuring the current and voltage values as well as performing protection work. Finally, they have the flexibility ability, which arises from their capability to define TCCs, which are arbitrarily set by the user [28].

To achieve the maximum operation of OCRs, two parameters should be set. These consist of the TSM and plug setting multiplier (PSM). The former is determined based on the minimum load current as well as the maximum fault current. The required interrupting capacities of overcurrent protective devices can be determined by helping the maximum fault currents, while the minimum fault currents are utilized in overcurrent device coordination operations [31]. The maximum fault currents have ratings from 50 to 200%, at intervals

of 25%, whereas minimum fault currents have ratings from 0.05 to 1, with 0.05 intervals. TSM is calculated in such a way that the system for protection can disconnect easily from the power system's faulty part [28,32]. The digital relays, which are new in the market, however, are able to make these parameters be set at intervals of 0.01 [33]. In general, the structure of OCR coordination problems in MGs is complex and intensive. This is especially seen in linked distribution systems, in which the burden of computation increases as the size and the network intricacy also increases. Figure 2 illustrates the coordination constraint between primary and backup relays, in which is horizontal axis represents the fault location and the vertical axis represents the tripping time. As ordinarily in a coordination approach, the fault is isolated firstly by using the primary OCR (Rp). If the Rp does not operate, the fault will be isolated by using the backup OCR (Rb) after a particular time, called CTI, which is represented in Figure 2 between the green curve and black curve [3].

**Figure 2.** Coordination Constraint between Primary and Backup relay.

Most renewable energy sources such as wind turbines and photovoltaic systems have been used power electronic inverters for connecting to the MG system. In MGs, the inverterbased distributed generations (IDGs) have been used in protection; however, the inverters have a limitation of the generated fault current: 150% of the current rating [34]. This makes the conventional overcurrent devices either stop responding or respond at a much larger operating time [4]. New challenges and opportunities appeared due to the growing wind turbine energy share, which led to a preference in the use of doubly fed induction generators (DFIG) over fixed-speed wind turbine systems. The connection of the wind farm to the network contains a low-voltage ride-through (LVRT) ability that is the most important requirement. Furthermore, owing to the DFIG essentially working the same to synchronous distribution generations (SDGs), power factor control might be applied at a reduced cost [35]. The contribution of fault current from synchronous distribution generations (SDGs) can rise to about ten times the current rating [27]. In general, the fault currents in MGs are dependent on the ratio of ratings between SDGs and IDGs. Likewise, the fault current contribution ability (FCCA) of IDG is very low (110%). This means that the mode used is the islanding mode; the OCRs may be unsuccessful in the case the MG only has IDGs. This is because the ratio of IDGs to SDGs in the mixture causes difficulty in the protection coordination as well as low fault current. In general, the DOCR aims to deal with bidirectional flow of power failures evenly. In addition, if the MG is able to operate in the loop and radial topology, the relay coordination and the detection scheme under primary fault becomes complex using different types of DGs. In order to handle the protection challenges in MGs, a fast and robust optimal protection scheme is required. The main objective of this article is to present an optimal and fast coordination scheme that minimizes OCR operational times for all operation and fault scenarios in MGs.

#### *2.1. Problem Description: Illustration-Based Analysis*

MGs with sources of renewable energy must be protected to ensure that the operating conditions are reliable and safe [28–36]. The coordination of OCRs in distribution systems can be done easily, especially those in radial structures and weakly meshed MGs [37,38]. However, in interconnected and meshed systems, each of the given relays acts as backup relays, where several delays are set as a backup for one relay. Figure 3a shows the single-line diagram of DN with three distribution lines (*DL*1, *DL*2, *DL*3) protected by overcurrent relays (*R*1, *R*2, *R*3). Figure 3a illustrates the OCR coordination from the side of the load to the source. For instance, a fault at the F1 point results in *R*<sup>1</sup> primary relay and *R*<sup>2</sup> backup relay. If the *R*<sup>1</sup> is unable to detect the fault or tripping delays, the *R*<sup>2</sup> will have time delays. Figure 3b shows the effect of the fault location on the fault current. It can be noticed that the fault current is increased whenever closer to the main source. Figure 3c is an illustration of the OCR coordination curves in three modes, which are a conventional network (that has no sources of renewable energy), a power network with DG, and an islanding mode. Ordinarily, the OCR operating time, in the no-DG case, is high due to a CTI ranging from 0.2–0.5 s, stressing the network equipment, possibly causing the relay to fall into the precise time region. This is more so for maximal fault modes (when faults occur near the source) [29,34,36].

**Figure 3.** (**a**) The single-line diagram of the distribution grid with DG under three fault modes, (**b**) the relationship between fault current and fault location, and (**c**) the miscoordination between the primary and backup relays and fault characteristics with the absence and presence of DG.

The conventional protection scheme experiences more challenges because of the different fault characteristics between the distribution systems in the presence and absence of sources of renewable energy. The rising number of sources of renewable energy in the network has widened the variation range between the minimum and maximum levels of fault current. Consequently, the calculation of the traditional protection setting will not meet the system requirements of the main protection: sensitivity, speed, and selectivity [4,28]. For instance, when the fault occurs at the F2 point for a distribution grid that has DG as illustrated in Figure 3c, the fault current value will go up in the *R*<sup>2</sup> primary relay and go down in the *R*<sup>3</sup> backup relay, resulting in a time delay causing disconnection or OCR coordination failure [30,36]. Figure 3c illustrates the effect of the connection between the DG and the fault current, *If* . For *R*2, maximum fault current for DG power network, *Iwith DG <sup>f</sup>* ,*max* , goes up while it goes down for *R*<sup>3</sup> when comparing with the maximum fault current during the absence of DG, *Ino DG <sup>f</sup>* ,*max* . Generally, the fault characteristics of a DN with DG

have been altered because of loading/generation level changes, variations in the network typology, islanding, fault point resistance, and the location of the fault point in relation to the main relay. Alterations of the fault current will result in OCR miscoordination; for instance, the *R*<sup>3</sup> will possibly not operate once there is a failure on *R*<sup>2</sup> for the minimum fault current, *I islanding <sup>f</sup>* ,*max* , in which the current value will be reduced compared with the distribution network in the absence of DG, *Ino DG <sup>f</sup>* ,*min* . For the islanding case, the fault current is too low, making its detection based on the conventional scheme difficult [28,30]. As a result, the conventional coordination and protection scheme will not have the capacity to handle the issue, which makes the development of a new time–current characteristic to address the challenges of DG protection very important. Furthermore, a DG-equipped system of protection for DN is needed to respond to the faults in all modes of DG operation, e.g., grid and islanded. This study proposes and develops various quick and intelligent schemes of protection and coordination. The coordination schemes' main objective is calculating the PSM and coordination curves which reduce the OT.

#### *2.2. The Coordination of Overcurrent Relay*

The traditional coordination between OCRs is generally obtained with the assumption that the conditions and network parameters such as resistance, current, and voltage during a fault will remain constant [34,36]. Equation (1) describes the CTI between the OT for the primary relay and backup relay for short circuits that occurs, for instance, at the F2 point [34,36,39]. The CTI presentation is such that the time of coordination between the backup relay, *tR*3, and the main relay, *tR*2, is equal to or more than the designated CTI.

$$t\_{R3} - t\_{R2} \ge \text{CTI} \tag{1}$$

Figure 3b illustrates how DG addition to the distribution grid affects the scheme of coordination protection [40,41]. The fault current variations will result in a CTI between the primary relay and backup relay that is lesser than the chosen CTI, causing a miscoordination. Calculations of the OCR operating time, t, in traditional methods are based on constant fault currents as well as known fault currents, *Isc*, as shown in Equations (2) and (3).

$$t = \left[\frac{A}{\left(\frac{Isc}{I^p}\right)^B - 1}\right] \times \text{TMS} \tag{2}$$

$$t = \left[\frac{A}{\left(\frac{Isc}{I\_P^\*}\right)^B - 1} + C\right] \times \text{TMS} \tag{3}$$

where TMS represents the time multiplier setting, *Isc* represents the short-circuit current, and *I p* represents the pickup current. Parameters *A*, *B*, and *C* in Equations (2) and (3) are related to a variety of relay characteristics that are defined on the basis of the standard of relay [42,43]. In general, numerical OCRs have the capability to update and modify the time operating characteristics based on real-time measurements. In this paper, the numerical OCRs provide the ability to use different time operating characteristics, such as the standard characteristics (IEC, ANSI), or generate new nonstandard operating characteristics. The proposed nonstandard time characteristic in this paper, NSTCC, aims to minimize the total tripping time and improve the performance of power protection in terms of selectivity and sensitivity.
