Graduate Thesis Or Dissertation
 

Empirical Characterization and Data-Driven Optimization of Water-Energy Systems

Public Deposited

Downloadable Content

Download PDF
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/rj430c47v

Descriptions

Attribute NameValues
Creator
Abstract
  • The accelerated development in water-energy systems calls for advanced modeling and optimization tools to improve technologies’ performance. Water desalination is a manifestation of water-energy system that is challenged by different operational constraints. The scarcity and uneven distribution of freshwater around the world makes water desalination an indispensable process to address the ever-rising demand. Most desalination processes, however, are plagued by high energy consumption and detrimental environmental impacts. To devise an efficient optimization scheme for these challenges, deep understanding of some key subsystems and how they behave with different inlet conditions is critical. This dissertation identifies spray evaporation and cyclonic salt separation in thermal desalination as key research areas. It, thereafter, provides a generalized global modeling and optimization digital twin (DT) framework that can be used to improve the efficiency of water-energy systems. Spray evaporation rate is a critical parameter to design and optimize humidification-dehumidification water-energy systems. A novel spray evaporation measurement technique was developed to replace temperature measurements. Conventional humidity measurement methods are either very costly or cannot accurately operate in a spray environment because of interference from liquid droplets. A new method of analysis was developed to obtain local humidity ratios from local pressure readings. Pressure changes due to atomization pressure loss and spray expansion were investigated and incorporated in the analysis. Empirical models were developed for the local humidity ratios and were accurate to within 10% of the experimental data. These models are only dependent on inlet information to obtain saline water spray spatial evaporation profiles. Using only inlet conditions is important for overall cycle optimization. A cyclone separator was developed to transform conventional humidification-dehumidification (HDH) desalination to treat highly saline streams and produce only freshwater and solid salt, as opposed to current technologies which discharge brine concentrates back to the environment. The cyclone separator was designed analytically, numerically, and tested experimentally for different humidity ratios, relative humidities, and feed stream’s salinities. The cyclone separator was integrated into an HDH cycle and was capable to treat water with salinities in the range of 3.5%-81% down to freshwater salinity (< 500 ppm) in a once-through operation. Salt scales on the cyclone’s inner walls and builds up until the temperature of the inner wall reach the dew point of the carrier humid air. After that, flow inertia overwhelms salt buildups and cleans the cyclone separator. This self-cleaning behavior allows the cyclone separator to treat extreme salinities, such as in brines discharged from other desalination technologies. After empirically testing and characterizing the subsystems mentioned above, the dissertation takes a broader perspective and presents a data-driven framework for energy-water systems. Although prevalent in many applications, data-driven artificial intelligence modeling techniques for water-energy systems are still in early research stage. A generalized digital twin framework was developed to fully characterize parametric relationships, optimize performance, and have the flexibility to be implemented in different complex and non-linear systems, as is the case in most energy-water systems. As part of developing a novel HDH technology, the framework was used to predict energy consumption based on 16 different design parameters and was accurate to within 1% of the data. That was achieved using two orders of magnitude fewer training data points than required by common AI methods. The parameters used include temperatures, flowrates, pressures, salinities, and different psychrometric properties. Pairwise parameters correlations were determined by the DT framework for two different case studies. The first one for the novel HDH cycle. The second case is for a spray evaporation subsystem. Data-efficient Bayesian methods were incorporated in the framework and were used to optimize the technology’s performance with minimal number of tests.
Contributor
License
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Rights Statement
Publisher
Peer Reviewed
Language
Embargo reason
  • Pending Publication
Embargo date range
  • 2022-06-02 to 2023-01-03

Relationships

Parents:

This work has no parents.

In Collection:

Items

Thumbnail Title Date Uploaded Visibility Actions
file details Mohammed_Elhashimi_-_Dissertation_v6.pdf 2022-06-02 Public Download