Robust Optimisation of Microfluidic Flow Systems

Current designs of the vast array of microfluidic flow systems used throughout biology, chemistry and engineering are based on processing small volumes of liquid in arrays of simple fluidic channels formed from combinations of regular channel geometries. Wider adoption of microfluidic technologies will require much more flexible and robust channel geometry optimisation methods which can deliver a step-change in functional performance whilst accounting for variations in manufacturing tolerances and operating conditions. This project will use experimental and computational methods to explore the performance of two different approaches to the optimisation of microfluidic channels for practical applications.

The first optimisation approach will be based on the robust shape optimisation of micro-channels, using 3-D conjugate heat transfer CFD and a relatively small number (~10) of design variables. The efficiency of various methods for generating the statistical moments of performance (mean and standard deviation), such as Monte Carlo and Polynomial Chaos, will be compared. Use will be made of NAG library software (possibly including pre-release software) to allow optimization routines from their library to be applied and assessed on this challenging class of problem. The second approach will use Topology Optimisation to optimise the heat sink and enables size, shape and topology to be optimised simultaneously.

These methods will be applied first to liquid-cooled serpentine channel heat sinks for electronics cooling and then applied to the effective design of microfluidic channels design for the use of Polymerase Chain Reactions to replicate DNA, requiring precise temperature control of the liquids being manipulated.

The optimized designs from each approach will be manufactured using the School of Mechanical Engineering’s 3-D metal printers and their heat transfer and hydraulic performance will be validated against experimental measurements.