Model-Based Control of Industrial Batch Crystallizers: Experiments on Enhanced Controllability by Seeding Actuation

Model-Based Control of Industrial Batch Crystallizers: Experiments on Enhanced Controllability by Seeding Actuation

Kalbasenka, A.N.

Bosgra, O.H. (promotor)
Kramer, H.J.M. (promotor)

Mechanical, Maritime and Materials Engineering

Delft Center for Systems and Control

2009-12-17

Crystallization is one of the oldest separation and purification techniques. Batch crystallizers are widely used in production of fine chemicals, food ingredients, specialty chemicals, and active pharmaceutical ingredients. Control of the crystalline material properties is a challenging task due to complexity and nonlinearity of batch crystallization processes and the lack of reliable measurement and actuation techniques. The goal of the research presented in this thesis was to design, validate, and evaluate different model-based control strategies for industrial batch crystallizers. A systematic approach was adopted in the development of control strategies that satisfy the requirements of industrial control systems. The followed approach included a number of steps such as process modeling, model reduction, controllability analysis, and control-system design. In the iterative design process, a particular attention was paid to experimental validation of the full-order and reduced crystallization models, state estimators, and model-based controllers by making use of pilot crystallization facilities of different scale and type. As a result of an extensive experimental study, the seeding technology was shown to be the pillar of the designed model-based control systems as it resulted in a reproducible operation and an improved quality of the end product. The yield of the crystalline product was increased by using model-based predictive optimizing controllers. For the given chemical system and batch crystallizers, linear control techniques were shown to be as suitable as nonlinear control methods for application in the designed strategies for control of the crystallization process.

batch crystallization
secondary nucleation
seeding
modeling
parameter estimation
controllability analysis
state estimation
model-based control
model predictive control
dynamic optimization

http://resolver.tudelft.nl/uuid:29a4aaf8-7d8c-4de0-b0a8-bd4f31365f09

GVO drukkers & vormgevers B.V. / Ponsen & Looijen

9789064643613

Institutional Repository

doctoral thesis

(c) 2009 Kalbasenka, A.N.