VOL. 10, NO. 4, MARCH 2015 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences © 2006-2015 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 1846 INTELLIGENT TUNING OF PI CONTROLLER IN AN AEROBIC BIOREACTOR S. Saravana Kumar 1 , K. Latha 1 and V. Rajinikanth 2 1 Department of IE, MIT Campus, Anna University, Chennai, India 2 Department of EIE, St. Joseph’s College of Engineering, Chennai, India ABSTRACT This paper proposes an intelligent controller tuning technique for a PI controller using a nature inspired metaheuristic algorithm, Firefly algorithm (FA). The PI controller is designed to maintain Dissolved Oxygen (DO) concentration in an aerobic biological reactor of a waste water treatment plant (WWTP). The proposed technique is compared with IMC based PID tuning method and PI parameters of Benchmark Simulation Model1 (BSM1). The PI controller tuned using the proposed technique produces satisfactory response and gives better results when subjected to set-point tracking and disturbance rejection test respectively. Keywords: firefly, PID, DO control, WWTP, ASM1, BSM1. 1. INTRODUCTION Wastewater treatment plants (WWTPs) are large non-linear systems, highly interactive and are uncertain due to the composition of the incoming wastewater. Activated Sludge in WWTPs is a biological process which uses bacteria and other micro organisms to remove contaminants from the wastewater. Some of these microbes grow only under aerobic conditions (dissolved oxygen present). It is very important to supply enough oxygen in the aeration tanks for the effective metabolism of microorganisms. Therefore, maintaining dissolved oxygen concentration (typically 1.0 - 3.0 mg/L) in the aerobic reactor is an important control problem [4]. Modeling and control of an activated sludge process is always a challenging task. Activated Sludge Models (ASMs) developed by International Water Association (IWA) task group includes ASM1, ASM2, ASM2d and ASM3 [2]. ASM1 is used most commonly to represent carbon oxidation, nitrification and de-nitrification processes [3]. In process industries, most of the controllers are PI/PID (Proportional plus Integral / Proportional plus Integral plus Derivative) type. The most important step in the application of PI/PID controller is tuning of its parameters. There are many procedures available for tuning the PI/PID controllers (for WWTPs) in the literature [5, 6]. It is necessary to find an advanced technique to tune the PI/PID controller which makes tuning procedure easier. The recent advancements in biologically inspired optimization algorithms optimize the controller parameter(s) based on the objective function(s) [10-12]. The firefly algorithm (FA) is a naturally inspired metaheuristic algorithm developed by Xin-She in 2007. The FA is based on the idealized behavior of the flashing characteristics of fireflies [7]. Being an effective optimizing algorithm, FA is used in many engineering applications [8, 9]. In control problems, FA and its variants are used to optimally tune the PI/PID controller [10, 11]. In this paper, a new method is proposed to tune the PI controller parameters for an aerobic bioreactor where the objective of FA is to minimize IAE. The paper is organized as follows. Section2 describes the process and its mathematical model. Section3 explains the controller design and section4 discusses the results obtained from the simulations carried out, followed by conclusion and references. 2. PROCEESS DESCRIPTION In WWTP, an aerobic biological reactor reduces the pollutant content present in the wastewater by assimilating them. The biological reactor is usually accompanied with the clarifier which separates the suspended solids from the treated wastewater (coming out from the reactor). For control purpose, only the aerobic reactor is considered. Figure-1 shows the schematic setup of a typical aerobic bioreactor used in WWTP. The aeration system supplies atmospheric air to the reactor via diffuser. To include the dynamics of the aeration system, it is modeled as a second order system (τ1= τ2=1.03min) with time delay of 4min [1] as shown in Figure-2. Figure-1. Schematic of an Aerobic Bioreactor.