(4dl) Understanding and Controlling Multi-Scale Complex Fluid Flows

Complex fluid flows are routinely encountered in a broad range of industrial processes. Additionally, many novel materials developed for energy and biomedical applications exhibit non-Newtonian behavior under processing flows for device fabrication. In order to improve existing processes and guide the design of new ones, it is essential to have both a fundamental and practical understanding of the origin of these behaviors and how they can be controlled. An inherent challenge is that the relevant time and length scales of these flows span orders of magnitude, from the deformation of the microstructure and individual molecules to the interaction of the fluid with the boundaries of the flow geometry. Thus, they are often difficult to understand and intractable to simulate.

I intend to integrate computational and theoretical tools to bridge the gap between molecular and continuum scales. The physical details of the microstructure can then be directly linked to macroscopically observed behaviors. To be useful in application, these tools must be sufficiently flexible that they can accurately describe the chemical diversity of complex fluids and the geometry and operating conditions of the processing method. Using machine learning, we can develop constitutive models informed by molecular conformations such that the evolution of field variables such as stress and concentration can be accurately reproduced at continuum scales. These fields can then be input to a molecular simulation for prediction of the conformational dynamics at a significantly reduced computational cost. The mechanical and electronic properties of the material originate from the microstructural orientation and deformation, meaning we can inform device design by predicting the effects of input microstructure and processing conditions.

My PhD research focused on molecular simulation of the microscopic dynamics of polymer solutions in flow and the effects of changing chemical details of the polymer architecture. Through this I developed expertise in the capacity and limitation of molecular simulation methods, as well as opportunities for extending them. In my postdoc, I am investigating complex fluid flows at the continuum level using machine learning and continuum simulation methods. I plan to combine these skill sets as described above in order to realize the capabilities of molecular simulations which are routinely promoted but less often achieved.

Teaching Interests

Our teaching approaches must develop to match the changes in chemical engineering practice and research. Computational proficiency is now essential for routine tasks such as data processing and automation. It is even more important in applying numerical methods for process modeling and simulation. Chemical engineers are also increasingly find careers in controls and data science, underscoring the need to complement traditional core topics with computing expertise. Programming is often not a fundamentally integrated component of many undergraduate engineering curricula, and as a result the material that is taught does not have lasting benefits after students graduate.

Computational tools can also aid in the pedagogical process. Students are increasingly inclined towards on-screen learning. Simulations provide a means to easily visualize the effect of variable parameters. The underlying code demonstrates a step-by-step process of implementing theorems learned in lecture. Thus, computation and simulation can aid in student understanding and their ability to put their education into practice.

I have extensive experience in developing customized code for specific applications as well as utilizing generic software packages for engineering applications. Throughout my educational process, approaching problems computationally often improved my understanding by opening another pedagogical approach which is not used in traditional classroom settings. I am particularly interested in teaching fluid mechanics and thermodynamics, where computation has been invaluable in research and industry.