Ph. D. Candidate | Villanova University, United States
Zihan Liu is a Ph.D. candidate in Mechanical Engineering at Villanova University, specializing in nonlinear dynamics and machine learning. His interdisciplinary research focuses on integrating physics-based modeling with neural networks to enhance diagnostic and predictive capabilities in complex mechanical systems. With a solid academic background from Harbin Engineering University, he has developed dynamic modeling solutions for systems like electromagnetic bearings and underwater robotics. Liu has co-authored multiple high-impact journal articles and presented his work at prestigious international conferences such as NODYCON and PHM Society. He is also an accomplished teaching assistant, supporting courses in Machine Learning, Vibration, and Nonlinear Dynamics. In addition to his technical expertise, Liu demonstrates strong communication skills and a passion for teaching. Beyond academia, he enjoys calligraphy, playing the bamboo flute, hiking, and cooking. His career trajectory reflects a unique blend of engineering innovation, deep analytical thinking, and creative exploration across mechanical and data-driven domains.
Profile
Scopus
🎓 Education
Zihan Liu began his academic journey at Harbin Engineering University, earning a B.Eng. in Machine Design & Manufacturing and Automation (2012–2016) with a thesis on underwater robot attitude control. He continued at the same university for his M.Eng. in Mechanical Engineering (2016–2019), where his thesis focused on controller design for electromagnetic bearings in flywheel energy storage systems. Currently, he is pursuing a Ph.D. in Mechanical Engineering at Villanova University (2019–present). His doctoral research integrates nonlinear dynamic information with machine learning for advanced system modeling, aiming to create hybrid data-driven and physics-informed approaches for fault detection and dynamic analysis. His educational path has been marked by a continuous focus on advanced mechanical systems, modeling, and diagnostics, underpinned by strong theoretical grounding and practical application. His progression from foundational engineering principles to cutting-edge machine learning integration highlights his commitment to innovation and academic rigor in mechanical systems engineering.
🧪 Experience
Zihan Liu has rich academic and research experience centered on dynamic systems, diagnostics, and intelligent modeling. At Villanova University, he served as a teaching assistant for Machine Learning (2023), Vibration (2023), and Nonlinear Dynamics (2024), instructing programming modules and supporting coursework. He has collaborated on several high-level research projects focusing on rotating dynamics, nonlinear systems, and electro-hydraulic servo valve diagnostics. Liu has presented at notable conferences like NODYCON (2021, 2023) and the PHM Society (2023), where he also became a finalist in the Three Minute Thesis Competition. He has authored and co-authored papers published or under review in top-tier journals such as Knowledge-Based Systems, Environmental Modelling & Software, and Mechanical Systems and Signal Processing. His software proficiency includes Abaqus, SolidWorks, Python, Matlab, and more, showcasing his technical versatility. This diverse experience underscores his capability in both theoretical research and applied system diagnostics across mechanical and mechatronic domains.
🏅 Awards and Honors
Zihan Liu has earned recognition for his research and communication skills. In 2023, he was a Finalist in the Three Minute Thesis Competition, highlighting his ability to convey complex ideas succinctly. His paper presentations at top conferences like NODYCON and the PHM Society Annual Conference reflect high academic impact and peer acknowledgment. Several of his co-authored papers are accepted or under peer review in leading journals, signifying rigorous and innovative research in physics-informed machine learning and mechanical diagnostics. His role as a teaching assistant across courses like Machine Learning, Vibration, and Nonlinear Dynamics at Villanova University also attests to his academic excellence and instructional abilities. Moreover, Liu has contributed to research featured in multi-disciplinary collaborations, including environmental modeling, indicating cross-domain appreciation of his skills. These accolades and experiences collectively demonstrate his excellence in research communication, technical instruction, and interdisciplinary collaboration within engineering and computational sciences.
🔬 Research Focus
Zihan Liu’s research sits at the intersection of mechanical dynamics and machine learning, with a particular emphasis on physics-infused neural networks for system modeling, diagnostics, and prognostics. His doctoral work at Villanova University explores integrating nonlinear dynamic data with machine learning to develop hybrid adaptive models that preserve physical insights while enhancing data-driven accuracy. Key focus areas include rotating dynamics, vibration analysis, fault detection, and signal decomposition techniques like POD and SOD for early crack detection in gear-train systems. He has proposed Hamiltonian and Routhian neural network models for rigid disk systems and investigated uncertainty quantification in parameter estimation. His work extends into fluid-mechanical system diagnostics, such as fault identification in electro-hydraulic servo valves. Liu’s interdisciplinary approach bridges mechanical engineering with AI, offering new methodologies for modeling complex systems efficiently and accurately. This fusion of physics-based features and data science defines his research trajectory toward intelligent, explainable system modeling.
✅ Conclusion
Zihan Liu is a promising candidate for the Young Scientist Award. His impactful research at the intersection of physics, machine learning, and mechanical systems, combined with solid communication and academic contributions, makes him a deserving nominee with significant potential for future leadership in his field.
Publications
-
Title: Hybrid Adaptive Modeling using Neural Networks Trained with Physics-Based Features
Year: Accepted
Authors: Zihan Liu, Prashant N. Kambali, C. Nataraj
-
Title: Exploiting POD and SOD Methods for Diagnostic of Multiple Defects in Gear Train Systems
Year: Under Reviewing
Authors: Zihan Liu, T. Haj Mohamad, Shahab Ilbeigi, C. Nataraj
-
Title: Using Diagnostic Signal to Extract Dynamics-based Features for Fault Identification in Electro-Hydraulic Servo Valves
Year: Under Reviewing
Authors: Zihan Liu, Prashant N. Kambali, C. Nataraj
-
Title: Spatiotemporal Flood Depth and Velocity Dynamics Using a Convolutional Neural Network Within a Sequential Deep-Learning Framework
Year: 2025
Authors: Mohamed M. Fathi, Zihan Liu, Anjali M. Fernandes, Michael T. Hren, Dennis O. Terry, C. Nataraj, Virginia Smith
-
Title: An Integrated Deep Learning Framework Enables Rapid Spatiotemporal Morphodynamic Predictions Toward Long-Term Simulations
Year: Under Reviewing
Authors: Mohamed M. Fathi, Zihan Liu, Anjali M. Fernandes, Michael T. Hren, Dennis O. Terry, C. Nataraj, Virginia Smith
-
Title: Hamiltonian Neural Network Model of Rigid Disk Rotating System with Damping and Unbalance
Year: Under Preparation
Authors: Zihan Liu, Prashant N. Kambali, C. Nataraj
-
Title: Exploring Gyroscopic Effect of Rigid Disk Rotating System by Routhian Neural Network
Year: Under Preparation
Authors: Zihan Liu, Prashant N. Kambali, C. Nataraj