A task based design procedure and modelling approached for industrial crystallization processes

Abstract

A synthesis-based approach to the design of crystallizers and industrial crystallization processes is introduced in this thesis. An ontology for a task-based design procedure has been developed which breaks the crystallization process into a subset of basic functions (physical tasks) which transform the physical state of matter to a final desired state. These tasks are connected in a network to accomplish the transformation of the feed into the product within quality specifications. This approach can facilitate the creative process of arriving at novel crystallizer configurations, with the aim to reduce cost of resources and capital, and so arrive at process intensification. Experimental investigations into the effect of actuators on process dynamics and subsequent model validation and parameter estimation studies in fed-batch processes is also studied. The study reveals a significant influence of the actuators on the initial start-up and dynamic behavior of the process, especially the effect of the propeller frequency and the fines removal rate. Modelling of growth rate dispersion (using multi-dimensional population balance models) due to strain in secondary nuclei via attrition provides an interesting insight into the broadening of crystal size distribution and the healing of strained crystals. The sustained cyclic behavior observed in DTB crystallizers for the continuous crystallization of ammonium sulphate from water is captured very well using a two-population balance model approach. Here, secondary nucleation mechanisms with a strong dependency on supersaturation are studied and the nucleation bursts could exclusively be located into the boiling zone of the crystallizer using a compartmental modelling approach.