Population Balance Model Based Multi-objective Optimization of a Multi-segment Multi-addition (MSMA) Continuous Plug Flow Crystallizer

Bradley J. Ridder, Aniruddha Majumder, Zoltan K. Nagy (Corresponding Author)

Research output: Contribution to journalArticle

33 Citations (Scopus)

Abstract

Crystallization is a major separation process in the pharmaceutical industry. Most crystallizations are performed batchwise, but there is great incentive for converting them to continuous operations. This paper investigates the modeling, simulation, and optimization of a special antisolvent plug-flow crystallizer: the multisegmented, multiaddition plug-flow crystallizer (MSMA-PFC). The MSMA-PFC accepts multiple antisolvent flows along its length, permitting finer control of supersaturation. A steady-state population balance equation was applied for tracking the crystal size distribution, and a mass balance equation was used to track the depletion of dissolved solute (flufenamic acid). A multiobjective optimization framework was applied to determine the antisolvent flow rates into each segment that simultaneously maximize the average crystal size, and minimize the coefficient of variation. The set of coupled differential equations was solved, depending on circumstance, with either the method-of-moments (MOM), or the high-resolution finite-volume (FV) method. The significant nonconvexity in the objective functions motivated the use of the nondominated sorting genetic algorithm (NSGA-II) to calculate the Pareto frontiers for the two competing objectives. It was found that the optimal antisolvent profile provides better product crystals, compared to the cases with equal additions of antisolvent in 1–4 injection points by keeping the total amount of antisolvent the same. The sensitivity of the Pareto frontier to variation in the growth and nucleation kinetic parameters was investigated. In addition, a novel simultaneous design and control (SDC) approach was proposed, based on the optimization of the full crystallizer design, over not only antisolvent profile, but also the number of injections and total crystallizer length, providing the best crystallization design that can allow optimal product performance in conjunction with the multiaddition control approach.
Original languageEnglish
Pages (from-to)4387-4397
Number of pages11
JournalIndustrial & Engineering Chemistry Research
Volume53
Issue number11
DOIs
Publication statusPublished - 17 Feb 2014

Fingerprint

Crystallizers
Multiobjective optimization
Crystallization
Crystals
Flufenamic Acid
Supersaturation
Finite volume method
Method of moments
Sorting
Kinetic parameters
Drug products
Differential equations
Nucleation
Genetic algorithms
Flow rate
Acids
Computer simulation
Pharmaceutical Preparations
Industry

Cite this

Population Balance Model Based Multi-objective Optimization of a Multi-segment Multi-addition (MSMA) Continuous Plug Flow Crystallizer. / Ridder, Bradley J. ; Majumder, Aniruddha; Nagy, Zoltan K. (Corresponding Author).

In: Industrial & Engineering Chemistry Research, Vol. 53, No. 11, 17.02.2014, p. 4387-4397.

Research output: Contribution to journalArticle

@article{b8f194fb5def4775b151375cf70232dc,
title = "Population Balance Model Based Multi-objective Optimization of a Multi-segment Multi-addition (MSMA) Continuous Plug Flow Crystallizer",
abstract = "Crystallization is a major separation process in the pharmaceutical industry. Most crystallizations are performed batchwise, but there is great incentive for converting them to continuous operations. This paper investigates the modeling, simulation, and optimization of a special antisolvent plug-flow crystallizer: the multisegmented, multiaddition plug-flow crystallizer (MSMA-PFC). The MSMA-PFC accepts multiple antisolvent flows along its length, permitting finer control of supersaturation. A steady-state population balance equation was applied for tracking the crystal size distribution, and a mass balance equation was used to track the depletion of dissolved solute (flufenamic acid). A multiobjective optimization framework was applied to determine the antisolvent flow rates into each segment that simultaneously maximize the average crystal size, and minimize the coefficient of variation. The set of coupled differential equations was solved, depending on circumstance, with either the method-of-moments (MOM), or the high-resolution finite-volume (FV) method. The significant nonconvexity in the objective functions motivated the use of the nondominated sorting genetic algorithm (NSGA-II) to calculate the Pareto frontiers for the two competing objectives. It was found that the optimal antisolvent profile provides better product crystals, compared to the cases with equal additions of antisolvent in 1–4 injection points by keeping the total amount of antisolvent the same. The sensitivity of the Pareto frontier to variation in the growth and nucleation kinetic parameters was investigated. In addition, a novel simultaneous design and control (SDC) approach was proposed, based on the optimization of the full crystallizer design, over not only antisolvent profile, but also the number of injections and total crystallizer length, providing the best crystallization design that can allow optimal product performance in conjunction with the multiaddition control approach.",
author = "Ridder, {Bradley J.} and Aniruddha Majumder and Nagy, {Zoltan K.}",
year = "2014",
month = "2",
day = "17",
doi = "10.1021/ie402806n",
language = "English",
volume = "53",
pages = "4387--4397",
journal = "Industrial & Engineering Chemistry Research",
issn = "0888-5885",
publisher = "American Chemical Society",
number = "11",

}

TY - JOUR

T1 - Population Balance Model Based Multi-objective Optimization of a Multi-segment Multi-addition (MSMA) Continuous Plug Flow Crystallizer

AU - Ridder, Bradley J.

AU - Majumder, Aniruddha

AU - Nagy, Zoltan K.

PY - 2014/2/17

Y1 - 2014/2/17

N2 - Crystallization is a major separation process in the pharmaceutical industry. Most crystallizations are performed batchwise, but there is great incentive for converting them to continuous operations. This paper investigates the modeling, simulation, and optimization of a special antisolvent plug-flow crystallizer: the multisegmented, multiaddition plug-flow crystallizer (MSMA-PFC). The MSMA-PFC accepts multiple antisolvent flows along its length, permitting finer control of supersaturation. A steady-state population balance equation was applied for tracking the crystal size distribution, and a mass balance equation was used to track the depletion of dissolved solute (flufenamic acid). A multiobjective optimization framework was applied to determine the antisolvent flow rates into each segment that simultaneously maximize the average crystal size, and minimize the coefficient of variation. The set of coupled differential equations was solved, depending on circumstance, with either the method-of-moments (MOM), or the high-resolution finite-volume (FV) method. The significant nonconvexity in the objective functions motivated the use of the nondominated sorting genetic algorithm (NSGA-II) to calculate the Pareto frontiers for the two competing objectives. It was found that the optimal antisolvent profile provides better product crystals, compared to the cases with equal additions of antisolvent in 1–4 injection points by keeping the total amount of antisolvent the same. The sensitivity of the Pareto frontier to variation in the growth and nucleation kinetic parameters was investigated. In addition, a novel simultaneous design and control (SDC) approach was proposed, based on the optimization of the full crystallizer design, over not only antisolvent profile, but also the number of injections and total crystallizer length, providing the best crystallization design that can allow optimal product performance in conjunction with the multiaddition control approach.

AB - Crystallization is a major separation process in the pharmaceutical industry. Most crystallizations are performed batchwise, but there is great incentive for converting them to continuous operations. This paper investigates the modeling, simulation, and optimization of a special antisolvent plug-flow crystallizer: the multisegmented, multiaddition plug-flow crystallizer (MSMA-PFC). The MSMA-PFC accepts multiple antisolvent flows along its length, permitting finer control of supersaturation. A steady-state population balance equation was applied for tracking the crystal size distribution, and a mass balance equation was used to track the depletion of dissolved solute (flufenamic acid). A multiobjective optimization framework was applied to determine the antisolvent flow rates into each segment that simultaneously maximize the average crystal size, and minimize the coefficient of variation. The set of coupled differential equations was solved, depending on circumstance, with either the method-of-moments (MOM), or the high-resolution finite-volume (FV) method. The significant nonconvexity in the objective functions motivated the use of the nondominated sorting genetic algorithm (NSGA-II) to calculate the Pareto frontiers for the two competing objectives. It was found that the optimal antisolvent profile provides better product crystals, compared to the cases with equal additions of antisolvent in 1–4 injection points by keeping the total amount of antisolvent the same. The sensitivity of the Pareto frontier to variation in the growth and nucleation kinetic parameters was investigated. In addition, a novel simultaneous design and control (SDC) approach was proposed, based on the optimization of the full crystallizer design, over not only antisolvent profile, but also the number of injections and total crystallizer length, providing the best crystallization design that can allow optimal product performance in conjunction with the multiaddition control approach.

U2 - 10.1021/ie402806n

DO - 10.1021/ie402806n

M3 - Article

VL - 53

SP - 4387

EP - 4397

JO - Industrial & Engineering Chemistry Research

JF - Industrial & Engineering Chemistry Research

SN - 0888-5885

IS - 11

ER -