505 Korean J. Chem. Eng., 36(4), 505-512 (2019) DOI: 10.1007/s11814-019-0225-y RAPID COMMUNICATION pISSN: 0256-1115 eISSN: 1975-7220 INVITED REVIEW PAPER † To whom correspondence should be addressed. E-mail: mynlee@ynu.ac.kr Copyright by The Korean Institute of Chemical Engineers. Control of a wastewater treatment plant using relay auto-tuning Saxena Nikita and Moonyong Lee † School of Chemical Engineering, Yeungnam University, Gyeongsan 38541, Korea (Received 1 August 2018 • accepted 30 December 2018) AbstractEfficient performance of a wastewater treatment plant largely depends on optimal process control. Owing to their complexity and nonlinearity, such processes are difficult to control. In this study, relay auto tuning method is analyzed to design of a proportional integral derivative controller for the activated sludge biological process. The pro- cess is estimated as a first-order process with time delay. The key control variable in wastewater treatment is the con- centration of dissolved oxygen during the aeration process. The influence of higher order harmonics on the system critical values is considered during the design using preload relay and a modified two-step relay. The system perfor- mance was evaluated for both servo and regulatory mechanisms. In addition, the designed controller was tested in the presence of noise for the robustness analysis. Keywords: ANAMMOX, Relay, Auto-tuning, Activated sludge process, PID control, DO control INTRODUCTION A control system is designed to regulate a process at the speci- fied operating conditions, along with the regulation of environmen- tal and product quality protocols, in order to increase the combined profitability and efficiency of the system. Optimal control of a wastewater treatment plant (WWTP) has high computational time and operating-maintenance cost requirements owing to the highly complex and integrated process units. The optimal control of any system largely depends on the selection of control valves, sensors, and transmitters. The proportional integral derivative (PID) con- troller is a crucial part of industrial control, owing to ease in imple- menting, tuning and availability along with the satisfactory per- formance. Efficient performance of the PID controller can only be ensured if the controller is tuned according to the process specifications. The dissolved oxygen (DO) concentration is considered the most critical control parameter for biological processes, as it directly affects the growth and inhibition of microbial species. The DO concentra- tion varies depending on the influent concentration, and it is directly related to the air flow rate, which in turn affects energy consump- tion and the operating cost of the process. Tuning the controller involves selection of appropriate proportional, derivative, and inte- gral gains. Tuning the controller is heuristic and time-consuming and numerous methods are available for controller tuning, includ- ing pole placement methods, internal model control (IMC), relay auto-tuning, and optimization based analytical methods [1,2]. A number of automated relay tuning methods have been developed recently. Various aspects of the auto-tuning method are reviewed by many researchers [3-10]. Auto-tuning, consisting of a conven- tional relay or modified relay, is widely used for linear and nonlin- ear process control. Saturation relay [11], asymmetric relay with an additional lead element [12], asymmetric relay [13], relay with mul- tiple switching [14], relay with hysteresis [15] and discrete time with general minimum variance [16] are used for the identification and control of stable systems. The implementation of auto-tuning tech- niques for wastewater treatment plants has not received much attention. For ease in industrial implementation of the relay auto- tuning technique, auto-tuners are available from several vendors, including ABB, Honeywell, and Siemens. Most of the methods presented in the literature for aeration control involve the design of advanced controllers. A nonlinear single-input single-output (SISO) model predictive control (MPC) based on oxygen dynamics for control of DO was developed [17] and improved using fuzzy predictive control [18]. The challenges faced during aeration control were addressed [19]. A rule-based feedback feed-forward controller to determine the set point for DO was widely used for the control purpose [20]. The feed-forward control determines the DO set point based on the nitrification capacity, while the feedback loop analyzes the influent ammonia concentration to determine the set point of the DO. A sub-space identification method was used to develop a DO model for the control of WWTP and was compared with MPC control- lers based on the overshoot and settling time [21]. A comparison method study for tracking oxygen concentration using a nonlinear model predictive control with an adaptive model reference con- trol [22] and model based optimization technique for determina- tion of DO concentration and controller parameter based on gain scheduling [23] was carried out. An investigation on methods for optimization of the aeration time and an MPC controller for set point tracking based on a linearized model was conducted [24]. Using the Ziegler-Nichols and the relay technique, a PID controller to regulate the air flow rate in a system in real-time was designed [25]. A combined model of a blower with MPC to regulate the con- centration in an activated sludge model [26] and fuzzy controller to control the DO concentration [27] was implemented. The con-