Data-Driven Modeling and Design of Multivariable Dynamic Sliding Mode Control for the Underground Coal Gasification Project Thar

Abstract

The energy output per unit time is an important performance metric to determine the potential of an underground coal gasification (UCG) site for electricity production. The energy output per unit time is a function of heating value and flow rate of syngas, and therefore, it is essential to devise a multivariable closed-loop system to enhance the efficiency of the UCG process. In this work, a model-based, multivariable dynamic sliding mode control (DSMC) has been designed for the cavity simulation model (CAVSIM), parameterized with the operating parameters and coal properties of the UCG Project Thar (UPT) field. The model-based control of CAVSIM is not possible due to its complex and multidimensional dynamics, and thus, a simple linear multivariable model is identified by employing a subspace (N4SID) system identification technique. The regular form of the linear model is formulated to design the model-based DSMC. Moreover, the stability of zero dynamics is shown on the approximate model of CAVSIM. Consequently, the designed controller is implemented on CAVSIM, and simulation results are compared with conventional sliding mode control (SMC). It has been observed that both the controllers have achieved the tracking objectives in the presence of input disturbance and modeling uncertainties. However, DSMC utilizes lesser control energy to achieve the desired objectives. Furthermore, the continuous control inputs in DSMC significantly reduce chattering.