M Ferreira Dos Santos et al Int. Journal of Engineering Research and Applications www.ijera.com ISSN : 2248-9622, Vol. 4, Issue 5( Version 6), May 2014, pp.146-150 www.ijera.com 146 | Page Identification Method by Least Squares Applied On a Level Didactic Plant Viafoundation Fieldbus Protocol Murillo Ferreira Dos Santos*, Amanda Knaipp Badaró **, Marlon José Do Carmo*** *(CEFET-MG – Campus III Leopoldina, MG, Brazil, Professor) ** (CEFET-MG – Campus III Leopoldina, MG, Brazil) *** (CEFET-MG – Campus III Leopoldina, MG, Brazil, Professor) ABSTRACT The industrial field is always considered a growing area, which leads some systems toimprove the techniques used on its manufacturing. By consequence of this concept, level systems became an important part of the whole system, showing that needs to be studied more specific to get the optimal controlled response. It’s known that the good controlled response is gotten when the system is identified correctly. Then, the objective of this paper is to present a didactic project of modeling and identification method applied on a level system, which uses a didactic system with Foundation Fieldbus protocol developed by SMAR ® enterprise, belonging to CEFET MG- Campus III –Leopoldina, Brazil. The experiments were implemented considering the least squares method to identify the system dynamic, which the results were obtained using the OPC toolbox from MATLAB/Simulink ® to establish the communication between the computer and the system. The modeling and identification results were satisfactory, showing that the applied technic can be used to approximate the system’s level dynamic to a second order transfer function. Keywords – System Identification, System Modeling, Level Plant, Methodological Experimentation. I. INTRODUCTION Due to the globalized world, which every day new technologies arise, the industries seek to produce more and more, with enhanced quality and quickness, lower cost and faults. Then, it’s required to train professionals which can provide a better control analysis of their processes, which are becoming increasingly complex. To assist in the design and analysis of the control systems functioning, it’s necessary to obtain the mathematical model that represents the actual physical process. This model is a mathematical equation used to answer questions about the system, such as the temporal variation and/or spatial variables of this, without conducting trials. With a good mathematical model, it is possible to analyze and predict the behavior of a system under various operating conditions, and adjust the performance of the same, if he didn’t show satisfactory. Thus, it permits to perform simulations of the system safely with low cost [1]. They are considered good models if theycan describe the phenomena of interest with considerable accuracy [2]. To determine the mathematical model of a system, it’s made the system’s modeling and identification which can represent its main features for diagnosis, monitoring, optimization and control. Within the context of mathematical modeling, it arises two types, phenomenological modeling and modeling by identification. Phenomenological modeling is based on the physics of the process, in other words, it addresses the phenomena involved through differential equations[3]. Modeling by identification is based on techniques that seek to describe the relations of cause and effect between the input and output variables, as the resulting models and techniques used, associated with the different phenomenological modeling. Thus, the modeling by identification becomes a very useful tool, advised to obtain the approximate mathematical equation of any system’s loop [4]. This paper is divided as it follows: The Section II shows the system used to explain the method; The Section III comments the OPC and Foundation Fieldbus protocols; The Section IV demonstrates the non-recursive least squares estimator; The Section V explains the procedure of algorithm’sexecution; The Section VI concludes the paper’s results. II. SMAR ® DIDACTIC PLANT The SMAR ® Didactic Plants were developed to simulatefaithfully some industrial processes in smaller scale. Due to the system be a didactic plant, it performs multithreaded, allowing the simulation of various processes commonly found in the industry and using the same tools and their configurations used in real industrial processes. It is shown in Fig. (1). RESEARCH ARTICLE OPEN ACCESS