Marta S. Basualdo
GIAIQ. UTN . FRR
Departamento de Electrónica. FCEIyA. UNR
FAX: 54-41-821772. E-mail basualdo@ifir.edu.ar
Rosario. Argentina
Simulation is the construction and use of a computer- based representation or rigorous model, of some part of real world as a substitute vehicle for experimental and behavior prediction. Digital simulation is recognized as a powerful tool for solving the equations that represent a real system.
Over the years specialist simulation software has been developed and is extensively marketed by commercial organizations. In theory, these simulation languages relieve the engineer of knowing anything about numerical integration. Therefore, these packages make it easier for the engineer to set up and solve the problems. In practice, however, these simulation languages have limited utility. In their push for generality, they usually have become inefficient. Certainly most of engineer undergraduates know some computer programming language. Hence, it is much better that the future engineer can develop a specific program for the problem at hand. Not only it is more computationally efficient, but it guarantees that the future engineer knows what is in the program and what the assumptions and techniques are. This makes debugging when it doesn't work and modifying it to handle new situations much easier. Finally, when the program is working correctly, the fun part of using the computer is getting useful results in a fast way. However, it should be noted that simulation does not, of itself, generate a solution but provides information enabling a user to determine a solution.
The central components of the simulation process are building the model (modeling) and running the model (i.e. the experiment). Many text [2], [3], present a great number of chemical process models which are implemented in programs written in FORTRAN. These software's serve as a guide to make an owner program with a low cost. In addition, several technical journals report simulation in a wide range of applications, based on that major decisions appear to be made using simulation and substantial benefits are claimed.
This paper commences by examining the simulationarea of application in linear identification and control design. A feedback structure, to control the top column distillation composition, in a single input-single output (SISO) way, is proposed as a practical work for the students. Hence, they evaluate the impact of changes to support the decision making process and have a better understanding of what design is more adequate. Experimentation on a model is more practical than experiment on the real plant. A pre-requisite, necessarily, is that the model is a good representation of reality and therefore an adequate predictor of what will happen.
Many times electrical or electronic engineers are incorporated to work in control of chemical plants and also they must talk the same language of chemical engineers. Chemical process generally presents important control problems such as great delays, highly nonlinear behavior and, for multivariable systems, the interaction among the loops can be turned in a serious troubles. Therefore, once the students have been learned the basic concepts about classical (or advanced) control, good practical works can be done via digital simulation. In this way the future electronic engineers are encouraged to learn the most common equipment functionality of chemical industry and its corresponding terminology.
In this work the phenomena consisting on heat and mass transfer occurring in a real column distillation process is translated into a quantitative mathematical model. Because of the very large number of equations needed for the rigorous description the calculations are made, with the help of a personal computer, by using some integration method. Therefore, the simulator can mimic the nonlinear behavior of that system so the undergraduate students can use it as an unknown plant which need to be controlled. Since the program allows to introduce step changes in the manipulated and disturbance variables, the pupils begin the study by generating the transfer function (identification). The software is flexible to simulate either SISO or MIMO control strategies, from classical feedback P, PI or PID to more sophisticate structures such as feedforward designs can be implemented easily by selecting the options in an iterative form.
Hence the students can observe rapidly the difference among diverse tuning criteria, or different control designs. In addition they learn that the design is made over a linear approximation but the controller is applied over the real plant (the simulator). In other words the software serves as an authentic pilot plant.
The simulator was firstly designed for scientific investigation purposes and is extensively described in [1]. In the next a brief description of the mathematical model is presented.
