State Affine Modeling and Observer Design for Hall-Héroult Process

Abstract Hall-Heroult process is a complex electrolysis procedure to produce aluminum. Due to the extreme operational conditions, it is difficult to continuously measure certain important values. This paper presents an approach to model and estimate the alumina concentration and anode-cathode distance based on the available measurements. Initially, a state affine model is obtained using a combination of physical-chemical relations and system identification. Then, a linear Kalman observer is designed to recover the desired signals. The proposed approach is validated on an industrial platform.