Abstract
Many industrial processes involve the handling of liquids, which necessitate precise control of variables such as flow and level. In engineering, managing the levels in coupled tanks is particularly challenging for control engineers due to the liquid turbulence or bubbles. Those disturbances could produce incorrect reading in the instruments and the controller can face difficulty achieving the desired setpoint This project explores several key aspects of this control problem: linearizing the nonlinear relationships between the levels in a dual-tank system, using the linear model to train a neural network for system modeling, designing a controller with pole placement to achieve a specified settling time, and developing a self-tuning Proportional, Derivative, and Integral (PID) controller using the gradient descent algorithm.