ANFIS Based Inverse Controller Design for Liquid Level Control of a Spherical Tank

Abstract

In this study, an adaptive neuro fuzzy inference system (ANFIS) based inverse controller design is presented for liquid level control application of a spherical tank. First, an excitation signal is applied to the system and the corresponding output signal is obtained. ANFIS-based fuzzy model of the nonlinear spherical tank system is constructed by using this input-output data set. While constructing the fuzzy model, a fuzzy model structure with two inputs and one output is preferred considering design simplicity. The input-output data used for constructing the fuzzy model of the system are exchanged, and by using this new data set, an ANFIS based inverse controller is designed. To improve the control performance against disturbances and model mismatches, the inverse controller is used in an internal model control structure. The performance of the proposed controller is compared to that of classical PI and fuzzy PI controllers under set point variation and disturbance conditions. The results of comparisons reveal that the proposed inverse controller outperforms both the classical and fuzzy PI controllers.