New dynamical observer for a batch crystallization process based on solute concentration

TL;DR

This paper introduces a novel infinite-dimensional observer for estimating the Crystal Size Distribution in batch crystallization processes using only solute concentration, temperature, and growth rate models, avoiding nucleation rate modeling.

Contribution

It presents a new Kazantzis-Kravaris/Luenberger observer combined with Tikhonov regularization for CSD estimation, differing from traditional moment-based methods.

Findings

Successfully estimates CSD from limited measurements.

Demonstrates exponential convergence of the observer.

Validated through numerical simulations.

Abstract

In this paper a new observer is introduced to estimate the Crystal Size Distribution (CSD) only from the measurements of the solute concentration, temperature and a model of the growth rate. No model of the nucleation rate is needed. This approach is based on the use of a Kazantzis-Kravaris/Luenberger observer which exponentially estimates functionals of the CSD. Then, the full state is estimated by means of a Tikhonov regularization procedure. Numerical simulations are provided. Our approach relies on an infinite-dimensional observer, contrarily to the usual moment based observers.