YUANQING WU

Dynamic Principal Applied Scientist with a focus on scientific machine learning and high-performance computing. Experienced in designing innovative Physics-Informed Neural Networks and directing complex R&D projects. Collaborates across disciplines to advance computational methods and foster graduate-level mentorship.

• Developed AI-driven surrogate models that significantly accelerated computationally intensive thermodynamic flash calculations.
• Designed scalable sparse-grid algorithms for high-dimensional nonlinear optimization problems.
• Independently designed and implemented multiple registered scientific simulation software systems.
• Led externally funded research programs totaling more than RMB 2.5 million.
• Published 20+ peer-reviewed papers in machine learning, computational physics, scientific computing, and numerical optimization.

• Y. Wu and S. Sun, Enhancing the Accuracy of Physics-Informed Neural Network Surrogates in Flash Calculations using Sparse Grid Guidance, Fluid Phase Equilibria, 2024
• Y. Wu and S. Sun, Removing the Performance Bottleneck of Pressure-Temperature Flash Calculations during Both Online and Offline Stages by Using Physics-Informed Neural Networks, Physics of Fluids, 2023
• Y. Wu and S. Sun, A Field-Based General Framework to Simulate Fluids in Parallel and the Framework's Application to a Matrix Acidization Simulation, PLOS ONE, 2022
• Y. Wu and Z. Chen, The Application of High-Dimensional Sparse Grids in Flash Calculations: From Theory to Realisation, Fluid Phase Equilibria, 2018
• Y. Wu, A. Salama, and S. Sun, Parallel Simulation of Wormhole Propagation with the Darcy-Brinkman-Forchheimer Framework, Computers and Geotechnics, 2015

• Physics-Informed Neural Network Accelerator, Designed PINN-based surrogate models to accelerate pressure–temperature flash calculations in compositional reservoir simulation. Improved computational efficiency by replacing expensive numerical solvers with machine learning surrogates while maintaining high predictive accuracy, i.e., most of the errors are in the order of 10-2., Python, PINNs, Deep Learning, Scientific Computing
• Sparse-Grid Simulation Framework, Developed high-dimensional sparse-grid surrogate methods for nonlinear thermodynamic calculations, reducing computational cost in large-scale simulation workflows by approximately threefold, decreasing the store requirement by about 2.2
• 10^12 times., FORTRAN, Sparse Grids, Numerical Optimization
• Parallel Scientific Simulation Platform, Designed and implemented parallel 2D/3D simulation software for compositional flow and matrix acidization, integrating scalable numerical algorithms with domain-specific physical models., FORTRAN, Parallel Computing, Scientific Software