Life Science Journal 2013;10(12s) http://www.lifesciencesite.com 761 Control of DFIG for improvement of voltage regulation in a power system using recurrent neural networks Ali Asghar Shojaei*, Mohd Fauzi Othman, Rasoul Rahmani, Masoud Samadi Centre for Artificial Intelligence and Robots, Universiti Teknologi Malaysia, 54100 Kuala Lumpur, Malaysia Shojaei2012@gmail.com Abstract: This article focuses on the voltage control of Doubly Fed Induction Generator (DFIG) wind turbines using Recurrent Neural Network (RNN). The paper also compares the performance of Static Synchronous Compensator (STATCOM) and DFIG systems, subject to the line to ground fault. The RNN is used in two main parts which are RNN Identifier (RNNI) and RNN Controller (RNNC). Performance of the DFIG is simulated and analyzed with and without the RNN controller. In this study, voltage regulation on Recurrent Neural Network is designed to control for a standard multi-machine power system. The results demonstrated significant improvement in the voltage regulation using the RNN controller for DFIG in the power system. [Ali Asghar Shojaei, Mohd Fauzi Othman, Rasoul Rahmani, Masoud Samadi. Control of DFIG for improvement of voltage regulation in a power system using recurrent neural networks. Life Sci J 2013;10(12s):761-769]. (ISSN:1097-8135). http://www.lifesciencesite.com . 122 Keyword: Recurrent neural network, control, DFIG, voltage regulation, STATCOM 1. Introduction Nowadays, power systems play important roles in human’s life. Reliability and stability of such systems are vital issues in power engineering. A fault occurrence in transmission lines may lead to damages in power equipments or consumer devices. Therefore, control of sudden load changes or fault occurrence in a power system is essential [1]. Today, different types of induction generators are used in wind energy technology. The two major types are fixed speed and variable speed generators. Variable speed equipment includes wind production in which the wind turbine power coefficient is optimum for the general use of the speed collection. The two common types of wind turbine are DFIG and synchronous generator which are driven by generators converter [2]. DFIG has a rotor winding which changes voltage source coupled to slip rings of the rotor. The stator winding is related in a straight line to the network and the rotor winding is associated with the network via an electronic power regulator. In [3], a detailed description is presented to show how a DFIG works. For constancy studies, network scheming of a DFIG should be taken into consideration in the dynamic analysis of a serious disturbance [4]. In the past, wind turbines were primarily used to protect the system from the poor performance of the turbine oriented and impact of the functioning on the network, because the penetration of wind turbine was very low. However, this situation has changed dramatically, and achieved high penetration of wind energy [5]. Independent of providing electrical energy, it is as well essential to contribute in voltage regulation. Some scientific studies have been dedicated to challenges in proliferation of wind energy. The function has no negative influence on the stability of a small network. The study in [6] is focused on the variable speed turbines which can improve the transient stability near conservative power plants. It is suggested in [7] that some types of wind turbines may participate in the energy vibration. By increasing the penetration of wind energy, some requirements are defined for the wind turbines connectivity requirements to the grid. The requirements are different around the world and are related to issues such as expansion, preservation and operation in synchronized, dependability and effectiveness [8-10]. The voltage control is compulsory, especially for the cases that the power system includes DFIG and STATCOM systems. This article proposes the use of wind DFIG controlled by RNN to improve voltage regulation of integrated utilities. The RNN is utilized in two parts including RNN Identifier (RNNI) and RNN Controller (RNNC). The RNNI’s function is detecting irregular outputs of power system plant through a feedback mechanism. The RNNC’s function is controlling and regulating outputs of power system. The results demonstrate that by applying the new control strategy for integrated systems, stability of voltage regulation can be improved significantly. The structure of this paper is as follows. In section 2, recurrent neural network control strategy is presented. Section 3 contains power system model. Voltage regulation approaches is explained in section 4. In section 5, power system simulation is introduced. In addition, the results analysis to explained in section 6. Finally, the paper is concluded in Section 7.