DEGREES
Doctor of Philosophy, Mechanical Engineering | University of Alberta, Edmonton, Canada
Dissertation: Solid Oxide Fuel Cell State-of-Health Estimation and Remaining Useful Life Prediction using Physics-informed Machine Learning
Supervisors: Dr. Mahdi Shahbakhti and Dr. Charles Robert Koch
Abstract: Promising alternatives to fossil fuels in power generation are fuel cells. Solid Oxide Fuel Cells (SOFC), one of the common types of fuel cell, have high thermal efficiency and low emissions and can operate using different fuels. A large barrier to SOFC commercialization is degradation, its identification and mitigation. A real-time state-of-health (SOH) estimation and remaining useful life (RUL) prediction algorithm is needed to serve as a foundation for predictive maintenance. This predictive maintenance will help mitigate SOFC degradation improving the useful life resulting in SOFC cost reduction. The main focus of this work is integrating the SOFC voltage and impedance time-series data and knowledge of the physics into a real-time physics-informed machine learning (PIML) for faulty condition. This PIML model will be used as a basis of an online estimate of SOH and predict the RUL of SOFC over the operation lifetime. A temporal graph convolutional network (TGCN) was trained using voltage and impedance data collected from the cells under essential fault conditions of Redox cycling in SOFC. This graph-structured network is designed based on physics and can predict the SOH for the horizon of 6 hours (with a whole lifetime of 50 hours in an accelerated degradation test) with root mean square error (RMSE) of 0.084.
Master of Science, Mechatronics Engineering | K. N. Toosi University of Technology, Tehran, Iran
Dissertation: Dynamic Simulation of an Internal Combustion Engine and Vehicle Dynamics for Real-Time Usage in the Hardware-In-the-Loop Control System
Supervisors: Dr. Shahram Azadi and Dr. Amir Mousavinia
Abstract: This research presents the development and validation of a comprehensive powertrain model that includes subsystems such as the internal combustion engine, automatic transmission, driveline, and longitudinal vehicle dynamics for real-time usage within a hardware-in-the-loop (HIL) control system. The advantages of employing a sliding mode controller for air-to-fuel ratio (AFR) control are demonstrated through comparisons with a traditional PID controller. To enhance model realism, actuator and sensor models—including injector, electric throttle, oxygen sensor, and vehicle speed sensor—are incorporated. Furthermore, the control systems account for real-world conditions, including disturbances like road slope and wind speed change. AFR control performance is assessed using the New European Driving Cycle (NEDC).
Bachelor of Science, Aerospace Engineering | K. N. Toosi University of Technology, Tehran, Iran
Dissertation: Design, Fabrication, and Control of Reaction Wheel Inverted Pendulum Using Sliding Mode Controller
Supervisor: Dr. Alireza B. Novinzadeh
Abstract: This project presents the design, fabrication, and control of a reaction wheel inverted pendulum system. The system dynamics and governing equations were derived, and a controller was designed using a sliding mode control (SMC) approach. The control strategy was first implemented in a MATLAB simulation to validate performance. The dimensions and specifications of the inverted pendulum were determined using a specialized algorithm and then modeled in SOLIDWORKS. Following the design, the laboratory plant was fabricated, and control system was implemented using an Arduino microcontroller.
PROFESSIONAL EXPERIENCES
Machine Learning Researcher (2021 - Present)
Energy Mechatronics Laboratory
Department of Mechanical Engineering, University of Alberta, Edmonton, Canada
Machine Learning Researcher (2021 - Present)
Ph.D. project in collaboration with the company
Cummins Inc., Columbus, Indiana, United States
Graduate Teaching Assistant (2015 - Present)
Courses:
Feedback Control Design of Dynamic Systems (MECE 420) - Summer 2023, Fall 2023, Winter 2024, Fall 2024 - University of Alberta, Edmonton, Canada
Mechanical Engineering Laboratory I (MECE 301) - Summer 2024 - University of Alberta, Edmonton, Canada
Engineering Mechanics; Dynamics - Fall 2015, Fall 2016 - K. N. Toosi University of Technology, Tehran, Iran
Graduate Research Assistant (2017 - 2020)
Vehicle Dynamics and Control Laboratory
Department of Mechanical Engineering, K. N. Toosi University of Technology, Tehran, Iran
Graduate Research Assistant (2017 - 2020)
Smart Systems Laboratory
Department of Electrical Engineering, K. N. Toosi University of Technology, Tehran, Iran
Laboratory Assistant (2016 - 2017)
Guidance, Navigation and Control Laboratory
Department of Aerospace Engineering, K. N. Toosi University of Technology, Tehran, Iran
GRADUATE COURSES
Machine Learning Control for Engineering Applications (MECE 610) - University of Alberta
Advanced Neural Networks (ECE 626) - University of Alberta
Artificial Neural Networks - K. N. Toosi University of Technology
System Identification - K. N. Toosi University of Technology
Machine Learning - Stanford, Coursera
Sequence Models - DeepLearning.AI, Coursera
Iterative Learning Control - University of Alberta
Mechatronics 1 - K. N. Toosi University of Technology
Adaptive Control - K. N. Toosi University of Technology
Industrial Automation and Control - K. N. Toosi University of Technology
Linear Control - K. N. Toosi University of Technology
Convex Optimization - K. N. Toosi University of Technology
Digital Signal Processing - K. N. Toosi University of Technology
Digital Logic Circuits - K. N. Toosi University of Technology
Electronic Circuits - K. N. Toosi University of Technology
AVR and ARM Micro-controllers Programming - K. N. Toosi University of Technology
Continuum Mechanics (MECE 680) - University of Alberta
Internal Combustion Engine Fundamentals - K. N. Toosi University of Technology
Reciprocating Engine Modeling and Simulation - K. N. Toosi University of Technology
Fuels and Combustion - K. N. Toosi University of Technology
Computational Fluid Dynamics - K. N. Toosi University of Technology
SKILLS
Python (Scikit-learn, Keras, TensorFlow, GPy, Pytorch, Pandas, Numpy)
Version Control Systems (Git)
C/C++
Assembly
Arduino
Ladder Logic
Latex
MATLAB-Simulink
Automation Studio
COMSOL-Multiphysics
SOLIDWORKS
CATIA
Abaqus
ANSYS Fluent
Altium Designer
Microsoft Office
Teamwork
Problem-solving
Creativity
Time Management