In industrial nickel and copper production, sulfur dioxide (SO₂) is generated from the combustion of sulfide ores. With increasingly tightened regulations on SO₂ emissions, a sulfuric acid plant has become a crucial part of industrial smelters. It converts environmentally harmful SO₂, which is generated in smelter furnaces, roasters, and Cu-reactors, into commercially beneficial sulfuric acid. This method is recognized as one of the most effective ways to ensure that smelters are able to satisfy the SO₂ emission regulations.
A sulfuric acid plant is primarily comprised of a central catalytic SO₂ converter, SO3 (sulfur trioxide) absorption towers and a series of interconnected heat exchangers. The catalytic SO₂ converter is the key component and the focus of this research. Both steady-state and dynamic models of the converter are developed in this thesis.
A steady-state model of the converter is established in accordance with steady-state mass and energy balances. The developed model provides an explicit relation between SO₂ conversion ratio and gas temperature, which is denoted as the heat-up path of the converter. By combining the heat-up path with the equilibrium curve of the SO₂ oxidation reaction, an equilibrium state for every converter stage can be obtained. Using the developed steady-state model, simulations are performed to investigate the effect of inlet SO₂ molar fraction and gas temperature on the equilibrium conversion ratio.
In an industrial SO₂ converter, the SO₂ concentration and conversion ratio out of each bed are important variables but are not measured in real time. To monitor these unmeasured variables in industrial operations, a soft sensor is proposed by combining the derived steadystate model with dynamic data analysis. The obtained soft sensor provides a real-time estimation of outlet SO₂ concentration and the conversion ratio from measured temperatures. For synchronization between the inlet SO₂ concentration and outlet temperature, a first-order exponential data filter is applied to the feed SO₂ data. With the filtered signal being used, the proposed soft sensors give a satisfactory estimation of both outlet SO₂ concentration and conversion ratio in the converter stages.
Dynamic modelling is carried out using two different model forms: ordinary differential equation (ODE) and partial differential equation (PDE) models. The ODE model is obtained by applying dynamic mass and energy conservation to the SO₂ converter. The resulting model can be used in industrial applications and describes the converter performance even if information of reaction kinetics is not available. A good fit with collected industrial data verifies the validity of the developed ODE model. The effect of process input variables is studied using simulations with the ODE model.
Dynamic modelling is performed by implementing mass and energy balances on both fluid and solid-phase gas flows. The proposed two-phase dynamic model, which takes the PDE form, is able to generate detailed profiles of the SO₂ converter within time and space. With the estimated parameters, this two-phase dynamic model generates a good fit between the simulated and measured outlet temperatures. Based on the PDE model, simulations are run to investigate the detailed mechanistic performance of the converter. The detailed PDE model provides useful explanation of, and prediction for the converter behaviour.