A Computational Framework for Design and Optimization of Dynamic Membrane Processes

It is estimated that over half of the world’s population is affected by fresh-water scarcity and that this figure will continue to rise because of the Earth’s changing climate, unbalanced socio-economic development, and continued population growth. The same can be said for the geographic imbalances of energy availability, which are driven by parallel factors.

Desalination offers a viable solution to clean-water production from seawater and brackish groundwater. Among desalination techniques, reverse osmosis (RO) membrane separators are the most widely used due to their energy efficiency relative to those based on distillation. However, even advanced multistage RO processes suffer from inefficiencies and inherent operational challenges including membrane fouling and scaling, transient operation due to the need for periodic flushing of the membrane, the computational challenges posed in model-based optimization of RO networks, and the need for high water recovery levels in inland areas where costs are associated with disposal of brine waste.

This research program will advance the fundamental understanding of dynamic (batch, semi-batch, and cyclic modes) membrane-based separation processes and will optimize their design and performance during dynamic operation. Specifically, this proposal aims to (i) investigate salt retention, fouling, scaling, and concentration polarization (a measure of solute concentration gradient at the membrane surface) under transient conditions and internal separator geometry design using computational fluid dynamics (CFD), (ii) explore system dynamics and apply optimal control theory to enhance system performance, (iii) develop novel networked process designs, and (iv) validate results using pilot-scale data from local water districts.

These objectives will ultimately lead to transformative innovations in dynamic membrane processes to help address the pressing issues in water and energy availability. The project also will provide research opportunities to a diverse group of undergraduate and high school students. Educational modules on advanced RO system design and operation will be developed to benefit students, researchers, and industrial practitioners in the field.

This proposal presents a vision for developing dynamic reverse osmosis (RO) processes to overcome the practical limitation of the infinite number of membrane stages and inter-stage booster pumps that would be required for (theoretically) optimal desalination system performance under steady operation, the current nominal mode of operation. The challenge to be addressed is the concurrent optimal design of the multistage networked membrane separator system together with determining the optimal time-periodic mode of operating the entire system.

Dynamic 3-dimensional computational fluid dynamic (CFD) techniques are required for accurate RO module simulations – however, using these types of simulations in the task of RO network design and optimization is computationally intractable at this time. Therefore, a reduced-order 1-dimensional model with parameters fitted from the detailed CFD simulations and validated against actual desalination plant data will be developed.

This reduced model can be efficiently discretized by orthogonal collocation and subsequently will be used to define a rigorous optimal control problem that seeks to minimize energy use and maximize clean water recovery. This research program will leverage current knowledge on the design of pressure-swing adsorption (PSA) processes, a mature industrial technology for gas separations, to guide initial designs for the networked RO desalination systems.

Like the planned RO systems, PSA plants operate under transient conditions and feature spatial concentration gradients within the separation units. The dynamic modeling, optimization, and design tools for RO networks inspired by PSA systems will be extended to green power generation systems that effectively operate as RO in reverse: osmotic pressure generated by the permeation of water through a membrane to a saline solution increases the latter’s pressure and volumetric flowrate which is subsequently harvested for power.

Overall, this research program will fundamentally advance the set of systems engineering tools available for producing reduced-order simulations from detailed, spatiotemporally distributed transport models needed for the model-based design and optimization of transient chemical processes outside of RO systems. Coupled with this research plan, a range of education and outreach plans focusing on undergraduate researcher support and a unique student exchange program with UCLA is planned.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.