Level Control of Coupled Conical Tank System using Adaptive Model Predictive Controller Muhammad Majid Gulzar 1 , Muhammad Munawar 1 , Zarak Dewan 1 , Muhammad Salman 2 , Sajid Iqbal 3 1 University of Central Punjab, Lahore, Pakistan, majidgulzar3@gmail.com 2 Kennesaw State University, Marietta, GA, USA. 3 University of Engineering and Technology (UET), Lahore, Pakistan. Abstract— The controlling techniques of liquid level in a coupled conical tank system is a challenging task owing to its continuous changing cross-section and non-linearity in the system. In this paper, an adaptive model predictive controller (AMPC) is presented to control the valve speed of conical shaped tank to maintain the liquid level. In AMPC, the plant model states are changed in every cycle along with the MPC controller to update the plant parameters in a precise manner, which is a major concern due to its non-linear behavior. Moreover, the comparative analysis of coupled conical tank system with other controllers like Fractional order PID (FOPID) controller and PID controller is carried out. The simulation results represent the superiority of the AMPC controller as compared to the other controlling methods in terms of response time and overshoot. Keywords— Coupled Conical Tank System, Adaptive Model Predictive Control (AMPC), PID, FOPID, MATLAB/ Simulink. I. INTRODUCTION It is well known that almost all the systems are practically non-linear. In different applications of the control system, the control method and tuning of such non-linear system is a major concern. So, the control technique is considered a common problem in process industries. The main issue in process industries is to control the level of liquid in the tank [1]. Most of the time, chemical and mixing features would be managed for the liquid in the tank such as milk processing and sugarcane industry [2]. The shape of the tank is conical which should have a constant fluid level and it consists of two variables: control variable and manipulated variable [3]. Fluid level is considered as supervised/control variable and fluid inflow to the tank is considered as manipulated variable. The dynamics of conical tanks system are non-linear due to conical shape which creates a challenge for the researchers to control the height level of conical tank [4]. To overcome this problem, a feedback loop component is contemplated as proportional- integral-derivative controller (PID). It utilizes PID controller in order to follow the fixed values and eliminate the noise in the activity. The tuning framework of PID controller is accomplished by relay feedback test [5-6]. Moreover, various controllers or mathematical models have been presented for the height level control of conical tank systems such as internal model control (IMC) [7], fuzzy logic controller (FLC) [8] and model predictive controller (MPC) [9-10]. In addition, for the level control of two and three tank system, the multi-input and multi-output (MIMO) [11] and the linear quadratic regulator (LQR) controllers have been studied [12]. Furthermore, the fractional-order proportional integral derivative (FOPID) has been simulated for conical tank [13]. This is used to control the parameters and tuning of controllers by using the conventional PID and proportional-integral (PI) and proportional-derivatives (PD) controllers. In which the derivative term or proportional term includes, where the power value of differentiator or integrator is considered as 0 or 1 [14]. Along with fractional IMC based PID controller has also been studied to control the height level of conical tank [15]. In this paper, for a non-linear system, modeling is presented for a conical tank system. Through which the height level of a conical tank is attained by using the AMPC. The results obtained by AMPC are further compared with the PID and FOPID. This is showing the superiority of proposed technique over comparative technique in the height level control of conical tank. Paper outlines are as follows; Brief overview of mathematical modeling of conical tank is discussed in section II. The control strategies of conical tank are given in section III. Simulation results and comparative analysis are presented in section IV. Finally, section V concludes the outcomes. II. MODELING OF CONICAL TANK SYSTEM The conical couple tank systems have maximum non- linearity at the bottom side as compare to the upper side. Moreover, the dynamics of tank are process gain and changing variable level and it is illustrated in Figure-1. Figure1. Conical Tank System Model The conical tank area which is given as  =  ! and height with relating angle tan θ is expressed in equations (1), (2). Furthermore, equation (3) is representing the mass of water in conical tank.  "# is the rate of inlet flow,  $%& is 2020 IEEE 17th International Conference on Smart Communities: Improving Quality of Life Using ICT, IoT and AI (HONET) 978-0-7381-0527-7/20/$31.00 ©2020 IEEE 236 2020 IEEE 17th International Conference on Smart Communities: Improving Quality of Life Using ICT, IoT and AI (HONET) | 978-0-7381-0527-7/20/$31.00 ©2020 IEEE | DOI: 10.1109/HONET50430.2020.9322842 Authorized licensed use limited to: UNIV OF ENGINEERING AND TECHNOLOGY LAHORE. Downloaded on February 11,2021 at 10:35:59 UTC from IEEE Xplore. Restrictions apply.