Biological wastewater treatment processes are essential in the sustainable management of water resources, offering an efficient method for removing contaminants and pollutants, such as ammonium, from wastewater to protect both public health and the environment. Among various treatment methods, submerged aerated biofilters stand out for their efficiency in converting high ammonium concentrations into nitrate. This process stimulates the growth of specific microorganisms on filtering materials, aiding in efficient pollutant conversion.
However, the complexity of biological wastewater treatment processes presents significant modeling challenges, especially under varying operational conditions. Linear Parameter-Varying (LPV) models have emerged as a promising solution to accurately represent these nonlinear systems. Despite their potential, constructing LPV models remains complex, especially for intricate biological treatment processes like wastewater treatment.
This paper presents a novel methodology within the global approach framework for estimating continuous-time LPV models. The proposed approach addresses the challenge of initializing iterative procedures due to the lack of prior knowledge about LPV model parameters. By extending the reinitialized partial moment approach to LPV models, the methodology provides an effective pre-estimate for initializing parameter estimation algorithms. Validation of the proposed methodology through simulation examples establishes a robust foundation for extending the approach to real-world applications, such as estimating LPV models for the nitrification process in wastewater treatment plants.