Computer Control of Non Linear Spherical tank system using Firefly and Bat Algorithm

Abstract

The key element of most process industries are tanks. Tanks are used for various mixing, blending and chemical treatments in different parts of the plant. Generally cylindrical tanks and cylindrical tanks with conical bottoms are used extensively owing to its linear nature. Use of spherical tanks in the place of the conventional tanks increases the efficiency of the processes as spherical tanks provide thorough mixing, better disposal of the liquid and has no deposition of chemicals on the walls of the tank. Level control in such a system is difficult due to the non-linear nature of the tank. In this work, a two-tank interacting spherical system is chosen as the non-linear system. Hence, tuning the controller should be efficient enough to overcome the effects of non-linearity and obtain a desired response from the plant. This work suggests the use of meta heuristic approaches to tune the PID controller. Meta heuristic approaches are gaining wide acceptance in the field of controller tuning as they provide optimum tuning parameters with few initial conditions and minimized cost function. Here, Firefly Algorithm (FA) and Bat Algorithm (BA) are used for tuning the PID controller. The merits of the methods are contrasted using the performance measures such as error indices and quality indices.