Farnaz Feyli

This project applies chemical equilibrium modeling to investigate the thermal breakdown of per- and polyfluoroalkyl substances (PFAS). Using simulations in Cantera, I evaluated temperature, pressure, and oxygen conditions required for complete destruction of PFOA, PFOS, and GenX. My work involved developing a detailed species set, implementing Gibbs free energy minimization, and analyzing equilibrium products across a wide operating range. This research aims to guide the design of incinerators and combustors that can effectively eliminate PFAS emissions and reduce environmental persistence

Physics-Informed Neural Networks (PINNs) were developed to model unsteady flow and temperature fields around a circular cylinder at moderate Reynolds number. The model integrates Navier–Stokes and energy equations, boundary conditions, and sparse sensor data to reconstruct velocity, pressure, and temperature fields.

The study investigates:

• the effect of network depth, width, and training duration,

• loss-term balancing and physics enforcement,

• adaptive sensor placement using residual-based strategies, and

• uncertainty quantification using Monte Carlo dropout under noisy measurements.

Model predictions are validated against high-fidelity CFD data, highlighting both the strengths and current limitations of PINNs for complex coupled flow–thermal systems.

This project applies machine learning techniques to predict clinical symptoms in patients with Chiari Malformation Type I (CMI). Using morphometric features extracted from brain MRI data, I developed and evaluated predictive models to identify symptom patterns. My work involved data preprocessing, feature selection, and model training using methods such as Elastic Net, Random Forest, and XGBoost. This research aims to support more personalized diagnosis and improve patient outcomes by linking imaging features to clinical presentations.

This project investigates the craniospinal morphological differences between pediatric Chiari Malformation Type I (CMI) patients and adult healthy controls and CMI patients. Using brain MRI datasets, I analyzed structural variations across different pediatric age groups and compared them to adult populations. My work focused on extracting morphometric features and assessing age-related changes, aiming to better understand the developmental aspects of CMI. These findings contribute to improving early diagnosis and treatment strategies tailored to younger patients.

This project studies the motion of brainstem and cerebellar tissues during the cardiac cycle in patients with Chiari Malformation Type I (CMI) and healthy controls. Using cardiac-gated DENSE MRI, I quantified the displacement of key brain regions throughout the cardiac cycle. My work focused on processing displacement fields, identifying motion patterns, and comparing tissue dynamics between groups. These findings help advance the understanding of brain biomechanics in Chiari patients.

The array of experimental instrumentations (laser displacement sensor, force sensor) and data collecting systems (LabView) are utilized to test the dynamic reaction of the system. Other data analysis technologies such as signal processing are done. To study the system's response, I performed various flow visualization techniques such as the Hydrogen bubble method.

The effectiveness of air purifiers in two different orientations, horizontal and vertical, are investigated for use inside of a school bus. The optimal number of air purifiers is calculated.

These air purifiers are currently used in California school buses to improve air quality and filter against COVID 19.

An airloop is a passive air cleaning device that works based on induction and buoyancy mechanisms. In this study, geometric and qualitative properties of an airloop are studied for optimal performance.

​

The properties that have been investigated include diffuser diameter, length, placement, distance between two nozzles, filter grade, filter dimensions, nozzle angle and the duct dimensions.

​

Experimental studies were also conducted to validate the numerical results. The optimal geometry resulted in a higher entrainment ratio, reaching more than 60 percent entrainment.

​

An optimum shape should secure longitudinal stability of the block, maximize the velocity for high penetration into the seabed, and minimize the dissolving percentage (heat-transfer) of the CDV.

In this work, several shapes for CDVs are considered.to assure stability, accurately designed fins have been added to the tail of CDV.

A thermal model for evaluating the temperature and moisture distribution on the mucus surface covering the walls of the nasal airway was developed by adding User Defined Functions (UDF). The airflow pattern, temperature and moisture concentration contours are evaluated based upon the results of the simulation.

Several User Defined Functions (UDF) was developed to tackle different challenges arose from various length scales.

Results of this study suggest that the present technique is a more comprehensive approach for the assessment of convective heat transfer around the human body and also eye surface than assuming a constant convection coefficient.

Heat transfer of various types of grooved microchannel, under the effect of different applied pumping power was investigated.

Results show that the heat removal for this optimum case of grooved microchannel is 38 % more than that for the optimum simple microchannel. The increase of the COP for the same comparison is 21 %.