VOL. 15, NO. 19, OCTOBER 2020 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences ©2006-2020 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 2070 MATHEMATICAL MODELING AND EXPERIMENTAL IDENTIFICATION OF THE CE 105MV TANK SYSTEM Faiber Robayo Betancourt 1 , Freddy Humberto Escobar 2 and Alvaro Javier Cangrejo Esquivel 3 1 Departamento de Ingeniería Electrónica, Facultad de Ingeniería, Universidad Surcolombiana, Neiva, Huila, Colombia 2 CENIGAA, Facultad de Ingeniería, Universidad Surcolombiana, Neiva, Huila, Colombia 3 Departamento de Matemáticas y Estadística, Universidad Surcolombiana, Neiva, Huila, Colombia E-Mail: faiber.robayo@usco.edu.co ABSTRACT In this work the modeling and identification of the CE105MV tanks system are performed. Mathematical modeling is achieved using the equations that describe the dynamic behavior of the system through its physical characteristics. As the system is not linear, experimental identification is also carried out supported by the MatLab software for the single tank system and coupled tanks; the results are compared with the mathematical model obtained. The linearization is achieved considering small variations on the desired level of the fluid in the tank, thus arriving at a transfer function that describes the behavior of the system. It can be concluded that, although the two methods approximate the real dynamics of the system, the identification performed with MatLab presents better results for both the single tank system and the coupled tank system. Keywords: mathematical modeling, identification, MatLab, tank system, linearization. 1. INTRODUCTION Level control in tanks and flow between tanks is common in industrial processes. In these processes the liquid must be stored and pumped to other tanks in such a way that the level of the liquid is controlled, and the flow regulated. It can be defined as control, the indirect manipulation of the magnitudes of a system called plant through another system called control system (Balcells and Romeral, 1997). The tanks can be configured in cascade or in a coupled manner, this is done with some type of connection between them, in addition to this feature, a system of coupled tanks allows interacting by controlling and manipulating one or more variables. However, in many control applications it is important to model this type of systems in such a way that their dynamics can be fully known, so that their performance can be represented and analyzed later. Good model representation is important for the success of any control strategy. Mathematical tools are commonly used to model and analyze the evolution of systems that vary over time (Edwards, 2013). The CE105MV multivariable coupled tank system is selected for this work. This training system has been worked in other universities generating important results. At the Higher Technical School of Industrial and Telecommunications Engineers in Pamplona, Spain, the liquid level of the CE 105 tanks was controlled and supervised by classical control methods using the LabVIEW programming language. Based on the physical equations that describe the system, the transfer function of the system was found. (Pérez, 2011). In the International Conference on Mechatronics, Electronics and Automotive Engineering a Model and Control for Coupled Tanks using LabVIEW was presented. The model identification is done by sinusoidal signal at different frequencies, as the result of the bode plot obtained is applied for designing the controller. Since the system is nonlinear the mathematical model is also presented to compare with the experimental model obtained. (Bastida, et al, 2013). In the University of Plymouth, the design of a novel nonlinear model predictive control (NMPC) strategy using a stochastic genetic algorithm (GA) to control highly nonlinear in CE105MV coupled tanks is presented. In order to derive a model for the plant, system identification is performed. The results show that the effectiveness of system identification. (Owa, 2014). The use of Proportional-Integral (PI) controller to monitor and control liquid level in an interconnected CE 105 model coupled tank is investigated (Hussein and Mishra, 2014). The PI SubVI already exists in the LabVIEW library that gives reasonable performance but to get a better system performance and monitor the liquid levels more accurately, another SubVI is derived from the PI controller mathematical equations. This article describes the mathematical modeling approach of the EC 105 MV tank system based on the physical and operational characteristics provided by the manufacturer (TecQuipment, 2018). As the system is not linear, the estimation of the model is proposed through MatLab software, using a Toolbox for identification in order to compare it with the mathematical model obtained (MathWorks, 2019). The visualization interface through the LabVIEW graphical environment allows observing the response of the variables in real-time. 2. MATERIALS AND METHODS 2.1 Hydraulic System This work is carried out in the CE105MV multivariable coupled tank system shown in Figure-1.