(IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 10, No. 10, 2019 316 | Page www.ijacsa.thesai.org Selection of Sensitive Buses using the Firefly Algorithm for Optimal Multiple Types of Distributed Generations Allocation Yuli Asmi Rahman 1 Department of Electrical Engineering Tadulako University Palu, Indonesia Salama Manjang 2 , Yusran 3 Department of Electrical Engineering Hasanuddin University Makassar, Indonesia Amil Ahmad Ilham 4 Department of Informatics Hasanuddin University Makassar, Indonesia Abstract—Power loss is one aspect of an electric power system performance indicator. Loss of power can have an impact on poor voltage performance at the receiving end. DG integration in the network has become one of the more powerful methods. To get the maximum benefit from synchronizing the system with DG, it is necessary to ascertain the size, location, and type of DG. This study aims to determine the capacity and location of DG connections for DG type I and type II. To address the aim of this paper, a metaheuristic solution based on a firefly algorithm is used. FA can cover up the lack of metaheuristic algorithms that require a long computational time. To ensure that the load bus location solution is selected as the best DG connection location, the input of the load bus candidate has been filtered based on stability sensitivity. The proposed method is tested on IEEE 30 buses. The optimization results show a decrease in power loss and an increase in bus voltage, which affects an increase in system stability by integrating three DG units. FA validation of the evolution-based algorithm shows a significant reduction in computational time. Keywords—Firefly algorithm; time computation; real power loss index; voltage profile index; multi-type DG I. INTRODUCTION The use of distributed generators (DG) on a scale of capacity and various types of innovation is becoming increasingly prevalent used in electric power systems. Technically, the benefits obtained from DG integration have been mentioned in previous studies, including reducing power losses, increasing reliability, improving stability, and improving the voltage profile [1]-[4]. However, the implementation of the addition of DG to the existing system will cause a new problem that is synchronizing the work between the old generating system and the new generating system. Therefore it is essential to do a thorough analysis of the technical factors related to the placement of DG, including the size, location of the connection, and the resulting impact. DG units can inject and absorb active and reactive power in the distribution network. They can maintain and enhance the voltage profile at different power factors. The power factor is a factor determining the type of DG that governs the type and size of the injected power. Based on the kind of power injected, DG is divided into four classes which are presented in Table I [5]. The maximum result of DG integration is achieved by the size and installed location of DG. Therefore, further analysis is needed regarding this matter. Previous research has examined various DG allocation techniques, precisely Analytical techniques developed by becoming techniques based on artificial intelligence. Analytical technique is a form of mathematical settlement in the form of numerical equations that are expressed as objective functions. This analytical technique represents the target to be achieved after DG placement. Some goals that generally want to be performed are the minimization of power losses, and increasing the voltage profile, achieving system stability [6]-[8]. In [7], the DG optimization process uses a combination of genetic algorithms with analytical techniques. DG optimization is done to reduce power losses in the system. In line with research [7], reference [8] has proposed a method for determining the location and size of DG type combinations using the efficient analytical method (EA) integrated power flow study. The target function is only to prioritize the power loss function, which is described by the real power loss (RPL) formula without assessing the bus voltage value after DG placement. Artificial intelligence techniques using metaheuristics now also dominate DG research [9]-[13]. This is due to metaheuristic success to resolve cases in an extensive system range. Metaheuristic algorithms mimic natural processes in matters such as biological systems and chemical processes. Met[[aheuristics can be divided into search points, namely one point search and population search. Population-based algorithms have the advantage of being able to explore effectively in the search space, making it suitable for global search. This is due to the ability of local exploitation and comprehensive exploration to avoid convergent locality. One approach using population metaheuristics is the firefly algorithm. In this study, using techniques developed by Dr. Xin-She Yang named the firefly algorithm (FA)[14]. FA imitated the information of fireflies in solving problems where in previous studies have been proven to resolve technical issues by solving global solutions [15]-[16]. The flash produced by fireflies is formulated based on objective functions. The brightest firefly determined as the most optimal solution.