Anti-Islanding Protection of Distributed Generation Based on Social Spider Optimization Technique

Anti-islanding protection is one of the most important requirements for the connection of Distributed Generators in power systems. This paper proposes a Social Spider Optimization (SSO) algorithm to detect unintentional islanding in power systems with distributed generation. The SSO algorithm is employed to differentiate frequency oscillations in synchronous generator those caused by non-islanding events. The SSO algorithm is based on the forging strategy of social spiders, which generated vibrations spread over the spider web to determine the positions of preys or any other disturbances. The vibrations from the spider are used to detect the occurrence of islanding in the synchronous generator. The SSO algorithm has superior performance when tested with IEEE 34 bus distribution system. The taken test system is evaluated for different scenarios and load distribution. The proposed SSO algorithm detects the islanding and prevents the system from undue tripping and outages. Furthermore, this technique may apply to prevent the system from islanding and maintains the future Indian Distributed Generation (DG) system reliability.


I. INTRODUCTION
The growing power demand and increasing concern for the use of fossil fuels in conventional power plants are increasing nowadays. The new paradigm of distributed generation is gaining greater commercial and technical importance. Distributed Generation (DG) involves the interconnection of small-scale, on-site Distributed Energy Resources (DER) with the main power utility at distribution voltage level [1]. Distributed Energy Resources mainly constitute nonconventional and renewable energy sources like solar PV, wind turbines, fuel cells, small-scale hydro, tidal and wave generators, micro-turbines etc. These generation technologies are being preferred for their high-energy efficiency and low environmental impact. Their applicability as uninterruptible power supplies to power quality sensitive loads. Electric energy market reforms and developments in electronics and use of anti-islanding protection are justified by the operational requirements of the utilities [2]. Anti-islanding systems are used to ensure personnel safety at the grid end and to prevent the generator out of synchronism. The islanding condition is a situation in which a part of an electric power system is solely energized and separated from the rest of the system. Failure to islanding detection [3] have several negative impacts for generators and connected loads. Imported one is the islanded grid because it cannot effectively control its frequency and voltage. This results in damage of equipment. Due to these damages, it causes safety hazards to utility workers and customers.
To avoid these problems, many power utilities using reclosers with transferred trip in the DG connection point. Other utilities request dedicated feeders with transfer trip. The detection methods are local techniques and Communication based techniques. These communicationbased methods are more effective than local techniques. The local methods are proposed as alternatives to methods based on communication and it is divided into three categories. The methods are active and passive methods. The active methods [4] inject small signals in the distribution system or force the DG to an abnormal situation, while the connection to the system keeps it under normal conditions. The disturbances inserted in the distribution system may cause power quality deterioration. The passive method uses wavelet [5]. The wavelets extract voltage and current features and use a decision tree to identify the islanding. The method uses a very large data set for training. The group living phenomenon has been studied intensively in animal behavior ecology. One of the reasons that an animal gather and live together is to increase the possibility of successful foraging and reduce the energy cost in this process. In order to facilitate the analysis of social foraging behavior [6], researchers proposed two foraging models: Information Sharing (IS) model [7] and Producer-Scrounger (PS) model [8]. The individuals under the IS model perform individual searching and seek for opportunity to join other individuals simultaneously. In the PS model, the individuals are divided into leaders and followers. Since there is no leader in social spiders, it seems the IS model is more suitable. The Bacterial Swarm Optimization (BSO) [9]- [10] is proposed for scheduling generating system. Even though it solves the problem, a step involved in solving this algorithm is large. Inspired by geographical elements, Biogeography-Based Optimization (BBO) [11]-[12] introduced to solve numerous problems in DG formulated micro grid. Sometimes the BBO struck in local optima leads towards worst optimal solution. The Ant Colony Optimization (ACO) [13] is introduced to solve the optimal power flow SUREOHP WKH $QW ¶V SDWK WDNHV PRUH WLPH WR ILQG WKH optimal path. The Bat Motivated Optimization (BMO) [14] involves the inspiration from social facts leads to poor solution. In this paper, the Social Spider Optimization (SSO) [15] algorithm is introduced to solve Anti-islanding protection of DGs. It is inspired by the social behavior of the social spiders, especially their foraging behavior. The foraging behavior of the social spider is described as the cooperative movement of the spiders towards the food source position. The spiders receive and analyze the vibrations propagated on the web to determine the potential direction of a food source [16]. In this process, the spiders cooperate with each other to move towards the prey. The natural behavior is utilized to perform optimization over the search space in SSO. The proposed algorithm has less iteration and fast ability to find optimal solution when compared with other techniques.

II. PROBLEM FORMULATION 2.1 Formulation of Synchronous machine models
The synchronous machine operating in steady state, the relative position between rotor and resulting magnetic field remain almost constant. When a sudden disturbance occurs, the angle between them oscillates dynamically according to the swing equation given by (1).  (1) Where G is the relative rotor angle, t is the time, H is the generator inertia constant, D is the damping coefficient, 0 Z is the DG synchronous speed, m P , e P are mechanical input and electric power output of the DG, respectively.

Frequency variation during non-islanding events
When a small disturbance occurs in the electrical system, the DG oscillates and returns to its original state after some time. The electrical power injected by DG in the distribution system is written as (2). G sin max P P e (2) A small perturbation G ' in G from the initial operating Due to this perturbation, the swing equation (1) is linearized and rewritten as 0 2H s P is known as the synchronizing power coefficient and is defined by the equation Solving the differential equation shown in (4), shows that the frequency deviation from nominal synchronous speed is given by (6).
From (6), the frequency is given by a damped sinusoidal waveform.

Frequency variation during islanding events
During an islanding event, the DG loses connection with the main system and, therefore, the synchronizing coefficient is 0. In this way, (4) is rewritten as (10).
p ' is the power variation due to the islanding; in other words, the transmitting power in the electrical system split point. In this case, p ' is assumed as constant during the islanding. p ' is assumed positive when the electrical power in the split point is flowing from the main system to DG. Since the rotor angle is synchronized with the stator magnetic field before islanding, the two initial conditions Comparing (6) to (11), it is observed that the frequency of the DG behaves differently. During DG parallel operation with the system, the frequency tends to oscillate at the damped natural frequency d Z . Disregarding the voltage controllers, governors, and the load dynamic during islanding, which may change due to voltage and frequency variation, the frequency does not oscillate during an islanding, but it is given by an exponential response.

III. SOCIAL SPIDER OPTIMIZATION TECHNIQUE
The Social Spider Optimization is the one of the nature inspired optimization technique and it is developed from behavior of social spiders. Swarm intelligence is a research field that models the collective behavior in swarms of insects or animals. Several algorithms arising from such models have been proposed to solve a wide range of complex optimization problems. The SSO algorithm is based on the simulation of cooperative behavior of social-spiders. In the proposed algorithm, individuals emulate a group of spiders, which interact to each other based on the biological laws of the cooperative colony.

3.1
Algorithm for SSO Step 1: Considering N as the total number of ndimensional colony members, define the number of male Where rand is random number between [0, 1] whereas floor(.) maps real number into integer number.
Step 5: Move the male spiders according to the male cooperative operator. According to this, change of positions for the male spider is modeled as follows: Step 6: If the stop criteria is met, the process is finished; otherwise, go back to Step 3. Here the vibration specifies the synchronous generators frequency oscillations that causes due to the sudden inclusion of unexpected load. (6) If that huge vibration is identified, then the spider separates the Zone of DG units (Islanding). i.e., it prevents the system from islanding (Anti-islanding). If step 3 is not satisfied, then randomly place the DG unit (Spider). The following steps are repeated until the optimal solution (less vibration) is found.

V. TEST RESULTS AND DISCUSSIONS
To evaluate the performance of the proposed method, it has been used on the IEEE 34 node distribution test system (Online Available) presented in Figure 1.

Fig.1: IEEE 34 Node Test Feeder
The transformer data are given in Table 1. The diesel generator controls the power factor to 0.98 inductive its data are presented in Table 2, and DG voltage and frequency regulators are given in [18]. The excitation system model used in a static excitation equivalent and the governor is the same used in [19].   Table 3 ROCOF 3 operates if voltage and reactive remains greater than 0.8p.u. ROCOF 1 and 2 do not use any voltage restriction. This temporization is important because of the high sensitivity of ROCOF protection, and helps to avoid unwanted Trips for short time transients in the distribution system, especially short circuits.  Table 4, which presents the line switched, the load condition, the DG generated power, the active switching interrupted power, and the protections tripping time. It is possible to see that the proposed method did not fail in any of the simulated cases. Table 5 shows the methods performance during a singlephase to ground short circuit sustained in the system for 350ms. After this time, the fault line is disconnected, thus causing the DG islanding. Table 5 shows the shortcircuited bus and the fault resistance. The islanding detection time is the difference between the protection trip times and 350ms; in this way, negative times represent protection trips before DG islanding, i.e., they represent failed trips. The proposed method did not fail in any simulated case presented Table 5. ROCOF 1 failed once and had some detection times greater than 500ms. ROCOF 2 failed in almost all cases, presenting negative islanding detection times. It detected the islanding during the short circuit in four times and did not trip during real islanding in three cases.

International Journal of Advanced Engineering Research and Science (IJAERS)
[  Not det. Table 6 shows the algorithms performance for temporary phase to ground short circuit. The fault remains during 350ms and disappears spontaneously without any switching.
The proposed algorithm as well as ROCOF 1 and ROCOF 3 worked well in all simulated cases. ROCOF 2 failed in 12 and 3 cases, respectively. Due to frequency pattern recognition, the proposed method avoids the nuisance tripping that would happen in other frequency-based relays such as ROCOF and Under/Over frequency. This is an advantage since, for instance, in case of a big generation trip in a large DG penetration scenario, the DG may help the system in the recovering process. However, a large perturbation on the generation or transmission system may cause frequency variations similar to those present in case of islanding, producing an undesirable tripping.  Operating characteristic of the system Islanding detection time (s) Therefore, the Standard IEEE 1547 allows the system operator to specify the frequency setting and time delay for under frequency trips down to 57 Hz. In these cases, the settings of the proposed method should take this recommendation in to account.

VI. CONCLUSION This paper proposes the Social Spider Optimization
Algorithm technique for islanding detection. During islanding, the synchronous generator oscillates at very VORZ IUHTXHQF\ GXH WR JRYHUQRU ¶V DFWLRQV RU WKH IUHTXHQF\ growth exponentially when the governors are unable to correct it. However, while connected to the main grid, the DG oscillates at a higher frequency. The method uses the communal web vibrations methodology that detects the frequency oscillation during islanding and sends a trip signal to the synchronous generator-operating switch. The suggested algorithm takes less convergence that seek to estimate the frequency of oscillation and damping coefficient, providing faster tripping compared to other optimization techniques. The main advantage of the proposed algorithm is conceptually simple and relatively easy to implement, which is clear from the presented result.