Modeling and simulation of water level system

The physical mechanism of three-tank water level system is analyzed, and an approximated linear model is deduced to represent the dynamic property of the water level system around an equilibrium point. On the other hand, Three different identification methods based on different model structures: linear model, BP neural network model, nonlinear Hammerstein ARMAX model are used to identify the dynamical character of water level system by using the measured input and output data. Based on Matlab Toolbox, a lot of simulation examples are presented to test the effectiveness of the proposed identification methods. The identified models and the corresponding methods will be a guidance for the design of control system.