Jialin Zheng

Jialin Zheng

Purdue University

Welcome to my homepage!

 

I am now a postdoctoral fellow at Purdue University , working with Xiaonan Lu. I received Ph.D. degree from the Department of Electrical Engineering at Tsinghua University, China, under the supervision of Professor Zhengming Zhao, and received my B.S. from the Department of Electrical Engineering at Beijing Jiaotong University.

 

Curriculum Vitae (Update Nov 2024)

 

My research focuses on the intersection of machine learning（ML）, digital twins(DT), and power electronics-dense grids （PEDGs）. Through advanced computational techniques and rigorous mathematical methods to achieve high-fidelity DTs and reliable ML, I work to advance the digital transformation of PEDGs, from individual power electronic converters to large-scale smart grids. By developing diverse application paradigms based on DTs and ML, I aim to shift PEDGs development from human-centered to computer-centered approaches, reducing the need for manual intervention while ensuring performance that matches or exceeds that achieved by skilled human operators.

 

Currently, my research focuses on embodied intelligence within power electronic converters in PEDGs, enhancing both individual and collective performance across the grid. This involves the integration combining power electronics control, large-scale network analysis, embedded AI, and edge computing. Prior to this, my Ph.D. research focused on real-time modeling and numerical integration algorithm design for power electronic converters. I have published these efforts in top journals, including IEEE Transactions on Industrial Electronics (TIE), IEEE Transactions on Power Electronics (TPEL), and IEEE Transactions on Transportation Electrification (TTE), along with top international conferences such as ECCE and PEDG, with a total of 25+ papers authored. A detailed list can be found here.

 

I particularly enjoy collaborating and exchanging ideas with researchers from diverse professional backgrounds. My collaborators and I have achieved significant results in the past, which I believe is largely due to our varied perspectives and efficient collaboration. They include Han Xu at Caltech, Haoyu Wang at UT Austin, Di Mou at Cambridge University, and Yangbin Zeng at the South China University of Technology.If you are interested in any kind of academic collaboration, please feel free to contact me at zjl1724030403@gmail.com.

 

Outside of research, I’m an amateur outdoor camping enthusiast with a goal of camping in every national park in the United States. I’m also a collector of Steam games and an amateur street photographer.

🔥 News

  • 2024.09: Congratulations on our paper being recognized as a highly cited article by Web of Science!
  • 2024.08: 🎉 I joined Purdue University as a postdoctoral fellow!
  • 2024.06: 🎉 I got my PhD degree from Tsinghua University! Thanks to those who helped me along the way!
  • 2024.05: 🎉 I won the Outstanding Doctoral Dissertation of Tsinghua University!
  • 2019.09: I’m starting my PhD at Tsinghua University.

📝 Publications

So far, my research has focused on Power Electronic Dense Grids (PEDGs) and categorized into five topics. In the future, I hope to do more interesting work in physics-informed machine learning for modeling, control and optimization of PEDGs.

📈 Digital Twins for PEDG Critical Devices

IEEE Trans Ind. Informat.
sym

Cognitive Digital Twins for Model Predictive Control of High-Frequency Power Converters (Under Review)
Jialin Zheng, Haoyu Wang, Yangbin Zeng, Di Mou, et al.

  • First use of digital twins for model predictive control of high-frequency power converters.
  • Get real operating modes and system parameters of the physical converter at the converter edge.
  • The cognitively obtained information is computed to enhance the dynamic performance of the converter.
IEEE Trans. Neural Networks Learn. Syst.
sym

Neural ODE-based Online Modeling of Unknown Power Electronic Systems(Under Review)
Jialin Zheng, Haoyu Wang, Yangbin Zeng, Di Mou, et al.

  • Directly neuralizing the converters’ ordinary differential equations (ODEs)，ensuring physical constraints and compatibility with existing ODE-based design and control methods.
  • Modeling unknown systems on local edge devices to guarantee privacy
  • Use adjoint states to obtain nearly constant training cost, independent of model depth
IEEE Trans. Ind. Electron.
sym

Real-Time Digital Mapped Method for Sensorless Multitimescale Operation Condition Monitoring of Power Electronics Systems
Y. Zeng, J. Zheng(Corresponding Author), Z. Zhao, W. Liu, S. Ji and H. Li

  • Non-invasive acquisition of all states inside the converter without any kind of sensors
  • Across multiple time scales, from subtle switching dynamics to second-level voltage fluctuations
  • Signal-level interaction for safety, convenience, and efficiency

🎙 Deep Machine Learning in PEDGs

IEEE Trans. Smart Grid
sym

A Scalable Network-Aware Multi-Agent Reinforcement Learning Framework for Decentralized Inverter-based Voltage Control
Han Xu, Jialin Zheng, Guannan Qu

  • Using Localized observation for critics in the CTDE framework，enhanceing scalability for large-scale PEDGs.
  • Truncating Q-function inputs based on network structure，reducing communication costs during training significantly.
  • Theoretical guarantees ensure robust approximation, compatible across diverse multi-agent actor-critic algorithms.

📚 Edge Computing for PEDG Application

IEEE Trans. Ind. Electron.
sym

A Semi-Implicit Parallel Leapfrog Solver With Half-Step Sampling Technique for FPGA-Based Real-Time HIL Simulation of Power Converters
J. Zheng, Y. Zeng, Z. Zhao, W. Liu, H. Xu and S. Ji

  • A semi-implicit numerical integration scheme, combining the high parallelism of explicit methods with the numerical stability of implicit methods.
  • Leapfrog operation with half-step increments, naturally doubling the sampling frequency, applicable across any switching frequency.
  • A hierarchical modeling approach that translates topology changes caused by switching actions into input variations, maintaining a constant system matrix.
IEEE Trans. Transp. Electrif.
sym

MPSoC-Based Dynamic Adjustable Time-Stepping Scheme With Switch Event Oversampling Technique for Real-Time HIL Simulation of Power Converters
Jialin Zheng; Yangbin Zeng; Zhengming Zhao; Weicheng Liu; Han Xu; Haoyu Wang and Di Mou

  • An oversampling method where the sampling frequency depends solely on hardware performance, unaffected by system scale or algorithm execution time.
  • Dynamically adjusts computation step-size, performing calculations only at critical points.
  • Capable of accurately characterizing all switching actions in multi-active bridge converters at frequencies up to 400 kHz, far exceeding the current average of 20 kHz.
IEEE Trans. Ind. Electron.
sym

Topology-Aware Matrix Partitioning Method for FPGA Real-Time Simulation of Power Electronics Systems
Han Xu, Jialin Zheng(Corresponding Author), Yangbin Zeng, Weicheng Liu, et al.

  • Divides the system matrix into blocks with explicit topological significance according to the topology structure.
  • Constructs a constant iterative matrix using its constant diagonal blocks, achieving a fixed iterative matrix and iteration count.
  • Utilizes a fully implicit numerical integration method without the need to store all matrices or perform indefinite iterative calculations, as required by traditional methods.

🎼 Real-Time Simulation and HIL Simulation

IEEE Trans. Power Electron
sym

An Event-Driven Parallel Acceleration Real-Time Simulation for Power Electronic Systems Without Simulation Distortion in Circuit Partitioning
Jialin Zheng, Zhengming Zhao, Yangbin Zeng, et al.

  • Simulates the modulation process to avoid sampling, achieving high-precision switching information at a lower hardware cost.
  • Designs a clever coordination among parallel computing processes by partitioning state variables, thereby avoiding errors typically introduced by traditional decoupling methods.
  • Achieves implementation on standard personal computers, reducing simulation hardware costs by over 30 times.
IEEE Trans. Ind. Electron.
sym

An Event Driven Synchronization Framework for Physical Controller Co-Simulation of Megawatt-Level Power Electronic Systems
Jialin Zheng, Yangbin Zeng, Zhengming Zhao, Weicheng Liu, Han Xu, et al.

  • A heterogeneous computing architecture based on CPU and FPGA, leveraging the high flexibility of CPU computation and the extensive data interfaces of FPGA.
  • A hardware-software co-simulation that achieves performance consistent with real-time HIL, yet without real-time constraints.
  • First-ever testing of a physical controller for a megawatt-scale power electronic transformer with over 576 switches, conducted without any simplifications or equivalencies.
IEEE Trans. Ind. Electron.
sym

A Self-Restoring Fault-Tolerant Method for Controller Cooperation Simulation of Power Electronics Systems
Yangbin Zeng, Zhengming Zhao, Jialin Zheng(Corresponding Author), et al.

  • Establishes a RTOS on a PC to enable timely responses to controller hardware via the PCIe interface.
  • Sets up an interlocking flag-based communication protocol between subsystems on different clocks to monitor and handle random data transmission delays.
  • Designs a self-restarting simulation stepping strategy with memory of previous states for flexible backtracking.

🧑‍🎨 Design and Control of PEDG Converter

🎖 Honors and Awards

  • 2024.06 Outstanding Doctoral Thesis Award，Tsinghua University (Top 1%)
  • 2023.10 First Class Scholarship for General Excellence，Tsinghua University (Top 1%)
  • 2023.10 Second Delta Scholarship，Delta Company (Top 1%)
  • 2022.10 Siyuan Electric Scholarship，Tsinghua University (Top 1%)
  • 2019.06 Beijing Outstanding Graduates Award, Beijing Municipal Education Commission (Top 1%)
  • 2019.06 BJTU Outstanding Graduates Award，Beijing Jiaotong University (Top 1%)
  • 2018.10 National Scholarship (Top 1%)
  • 2017.10 National Scholarship (Top 1%)

📖 Educations

  • 2024.08 - 2025.06, Postdocal fellow, Purdue University, West Lafayette, IN, USA.
  • 2019.06 - 2024.06, Phd, Electrical Engineering, Tsinghua University, Beijing, China.
  • 2015.09 - 2019.06, Undergraduate, Electrical Engineering, Beijing Jiaotong Univeristy, Beijing,China.

💬 Teaching and Industrial Experience

  • 2020.09-2021.01, Teaching Assistant, Design and Analysis of Electrical Machine Systems (40220682) with Prof. Zhengming Zhao.
  • 2024.06-2024.07, Teaching Assistant, Modeling and Control of Power Converters with Prof. Dushan Boroyevich and Prof. Kai Sun.
  • 2021.06-2021.08, Research Assistant, Artificial Intelligence in Power Electronic Systems，Delta Electronics, Inc., Shanghai,China.

💻 Acadamic Service

Reviewer

  • IEEE Transactions on Power Electronics (TPEL)
  • IEEE Transactions on Industrial Electronics (TIE)
  • IEEE Journal of Emerging and Selected Topics in Power Electronics (JESTPE)
  • IEEE Transactions on Power System (TPWRS)
  • IEEE Open Journal of the Industrial Electronics Society (OJIES)
  • IEEE Open Access Journal of Power and Energy (OAJPE)
  • IEEE Energy Conversion Congress and Exposition (ECCE)
  • IEEE International Symposium on Power Electronics for Distributed Generation Systems (PEDG)