64
5 Conclusion
depends on stretching and folding of fluid streams for efficient mixing. Chaotic advection can be generated in a variety of ways, and the geometrical modifications using
surface patterning, serpentine channels, obstacles on channel walls, split and recombination, and three-dimensional channel structures, were discussed. Due to limited
length of the book, it was not possible to cover all active and passive micromixers,
but some review papers focussed on active and passive micromixers were cited for
the readers who wish to look into more detailed comparison of micromixers.
Mixing in micromixers needs to be quantified for the evaluation of performance,
and the derived performance matrices can be used for the purpose of design optimization. Different numerical techniques to find the solutions of the governing differential
equations for flow dynamics and mixing in micromixers were introduced in Chap. 3.
The solutions for the velocity and concentration fields can be manipulated to derive
performance measures. Since flow and mixing analyses are based on both the Eulerian
and Lagrangian approaches, both the approaches were discussed separately in detail
for better understanding and clarity to the readers. An overview on commercial CFD
packages relevant to micromixers were presented. The challenges and limitations of
numerical schemes, computational grid requirements, and the associated errors in the
numerical solutions were also discussed. Chapter 3 concluded with the technique for
mixing characterization using velocity and concentration fields, and forms the basis
for performance evaluation of different micromixer designs. The chapter not only
provides the fundamentals of numerical techniques, but also discusses their relative
advantages and disadvantages for different types of mixing problems encountered in
practice. Also, details of numerical models in popular commercial CFD packages,
viz. CFD-ACE+
® , ANSYS-CFX
® , ANSYS-Fluent
® , were briefed for end users who
wish to conduct research on micromixers.
Chapter 4 was dedicated to design optimization of micromixers. Several optimization techniques were presented, but the major emphasis was laid on surrogatebased optimization as it was found to be the most widely used method for design
optimization of micromixers. In case of micromixers, the optimization problems
can be classified as either single-objective (e.g., to maximize mixing performance
of the micromixer) or multi-objective (e.g., to maximize mixing performance and
minimize pressure drop in the micromixer, simultaneously) problem. Conventional
optimizations generally employ gradient-based methods which tend to get trapped
in local optima rather than finding the global optimum solution, or population based
methods like genetic algorithm which require impractical computational cost to
obtain global solutions. Since simulation models used for design of micromixers
are generally expensive, surrogate models are good alternative to reduce the computational burden with acceptable approximation of the simulation data. Surrogate
model(s) for the objective function(s) coupled with an optimization algorithm can
be used to find optimal geometric configuration(s) subjected to a set of design
constraints. This chapter focused on the algorithm for single-objective and multiobjective optimizations using surrogate models for the design and development of
micromixers.
5 Conclusion
depends on stretching and folding of fluid streams for efficient mixing. Chaotic advection can be generated in a variety of ways, and the geometrical modifications using
surface patterning, serpentine channels, obstacles on channel walls, split and recombination, and three-dimensional channel structures, were discussed. Due to limited
length of the book, it was not possible to cover all active and passive micromixers,
but some review papers focussed on active and passive micromixers were cited for
the readers who wish to look into more detailed comparison of micromixers.
Mixing in micromixers needs to be quantified for the evaluation of performance,
and the derived performance matrices can be used for the purpose of design optimization. Different numerical techniques to find the solutions of the governing differential
equations for flow dynamics and mixing in micromixers were introduced in Chap. 3.
The solutions for the velocity and concentration fields can be manipulated to derive
performance measures. Since flow and mixing analyses are based on both the Eulerian
and Lagrangian approaches, both the approaches were discussed separately in detail
for better understanding and clarity to the readers. An overview on commercial CFD
packages relevant to micromixers were presented. The challenges and limitations of
numerical schemes, computational grid requirements, and the associated errors in the
numerical solutions were also discussed. Chapter 3 concluded with the technique for
mixing characterization using velocity and concentration fields, and forms the basis
for performance evaluation of different micromixer designs. The chapter not only
provides the fundamentals of numerical techniques, but also discusses their relative
advantages and disadvantages for different types of mixing problems encountered in
practice. Also, details of numerical models in popular commercial CFD packages,
viz. CFD-ACE+
® , ANSYS-CFX
® , ANSYS-Fluent
® , were briefed for end users who
wish to conduct research on micromixers.
Chapter 4 was dedicated to design optimization of micromixers. Several optimization techniques were presented, but the major emphasis was laid on surrogatebased optimization as it was found to be the most widely used method for design
optimization of micromixers. In case of micromixers, the optimization problems
can be classified as either single-objective (e.g., to maximize mixing performance
of the micromixer) or multi-objective (e.g., to maximize mixing performance and
minimize pressure drop in the micromixer, simultaneously) problem. Conventional
optimizations generally employ gradient-based methods which tend to get trapped
in local optima rather than finding the global optimum solution, or population based
methods like genetic algorithm which require impractical computational cost to
obtain global solutions. Since simulation models used for design of micromixers
are generally expensive, surrogate models are good alternative to reduce the computational burden with acceptable approximation of the simulation data. Surrogate
model(s) for the objective function(s) coupled with an optimization algorithm can
be used to find optimal geometric configuration(s) subjected to a set of design
constraints. This chapter focused on the algorithm for single-objective and multiobjective optimizations using surrogate models for the design and development of
micromixers.