The RaDA lab at the University of Texas promotes state of the art research and education related to AI and machine learning methods for engineering design, reliability, safety, failure resiliency and sustainability.
Digital Twin Technology
Digital twins have the potential to significantly accelerate scientific discovery and transform industries by offering a powerful tool for decision-making that integrates both models and data. Unlike traditional modeling and simulation, digital twins feature a dynamic, bidirectional relationship between a physical system and its virtual counterpart. This allows the digital twin to evolve alongside the real-world system, continually updating its virtual representation. By enabling predictions beyond existing data, digital twins provide valuable insights that drive more informed decision-making, whether automated or involving human input. This feedback loop ensures continuous interaction between the physical and virtual systems, enhancing the accuracy and effectiveness of decisions over time.
Physics-Informed Machine learning
Physics-informed machine learning techniques capitalize on domain knowledge and physics-based prior information to enhance machine learning methods, facilitating informed decisions regarding design choices, materials selection, and manufacturing processes. The integration of domain-specific knowledge and machine learning algorithms can yield predictions that are both accurate and interpretable, leading to improved design optimization and enhanced decision-making across diverse engineering domains.
Selected Publications
Xu, Y., Kohtz, S., Boakye, J., Gardoni, P., & Wang, P. (2023). Physics-informed machine learning for reliability and systems safety applications: State of the art and challenges. Reliability Engineering & System Safety, 230, 108900.
Kohtz, S., Xu, Y., Zheng, Z., & Wang, P. (2022). Physics-informed machine learning model for battery state of health prognostics using partial charging segments. Mechanical Systems and Signal Processing, 172, 109002.
Xu, Y., & Wang, P. (2023). Physics-Constrained Machine Learning for Reliability-Based Design Optimization. In 2023 Annual Reliability and Maintainability Symposium (RAMS) (pp. 1-6). IEEE.
Predictive Modeling & Uncertainty Quantification
Fundamental research of uncertainty quantification and management, especially in extremely high dimensional problems were conducted, which manages uncertainties associated with engineering systems and processes with new advanced techniques for uncertainty quantification and modeling to facilitate robust decision-making for system design, manufacturing, and system operation and maintenance.
Selected Publications
Xu, Y., Wu, H., Liu, Z., Wang, P., & Li, Y. (2024). Multi-Task Learning for Design under Uncertainty with Multi-Fidelity Partially Observed Information. Journal of Mechanical Design, 1-17.
Xu, Y., & Wang, P. (2023). Sequential Sampling-Based Asymptotic Probability Estimation of High-Dimensional Rare Events. Journal of Mechanical Design, 145(10), 101701.
Xu, Y., Renteria, A., & Wang, P. (2022). Adaptive surrogate models with partially observed information. Reliability Engineering & System Safety, 225, 108566.
Wu, H., Xu, Y., Liu, Z., Li, Y., & Wang, P. (2023). Adaptive machine learning with physics-based simulations for mean time to failure prediction of engineering systems. Reliability Engineering & System Safety, 240, 109553.
Semiconductor Design and Optimization
To meet the rapid expansion of the semiconductor industry, manufacturers need to embrace cutting-edge technologies like Digital Twins and Machine Learning methods. This research aims to solve the big issue in terms of digitalizing and designing the semiconductor system and applications on the sensors. Studies were conducted on designing semiconductor devices, such as Hall-effect sensors and HEMT devices.
Selected Publications
Xu, Y., Lalwani, A. V., Arora, K., Zheng, Z., Renteria, A., Senesky, D. G., & Wang, P. (2022). Hall-effect sensor design with physics-informed Gaussian process modeling. IEEE Sensors Journal, 22(23), 22519-22528.
Battery Health Management
Battery aging is one of the primary challenges hindering the widespread adoption in various fields. The state of health (SOH) of a battery, which reflects its ability to store and deliver energy relative to its initial state, is a key indicator of aging. It is essential to estimate the state of health of batteries to ensure safety, optimize better energy efficiency and enhance the battery life-cycle management. The proposed novel physics-informed architecture enables an accurate and efficient estimation of battery SOH with only a partial charging segment, and also allows for extrapolation of information from a partial charging segment.
Selected Publications
Kohtz, S., Xu, Y., Zheng, Z., & Wang, P. (2022). Physics-informed machine learning model for battery state of health prognostics using partial charging segments. Mechanical Systems and Signal Processing, 172, 109002.
Liu, Z., Xu, Y., Wu, H., Wang, P., & Li, Y. (2023, August). Data-Driven Control Co-Design for Indirect Liquid Cooling Plate With Microchannels for Battery Thermal Management. In IDETC-CIE (Vol. 87301, p. V03AT03A048). American Society of Mechanical Engineers.
Motor Failure Prognostic
Optimal sensor placement: a digital-twin assisted framework is proposed to determine optimal placement of sensors for fault detection within a PMSM. This study demonstrates that the magnetic field can be utilized as a signal for fault detection within a PMSM, for which there has been very few studies. It also demonstrates that sensor placement within the PMSM plays a critical role for fault detection, which has not been researched in this field.
Selected Publications
Kohtz, S., Zhao, J., Renteria, A., Lalwani, A., Xu, Y., Zhang, X., … & Wang, P. (2024). Optimal sensor placement for permanent magnet synchronous motor condition monitoring using a digital twin-assisted fault diagnosis approach. Reliability Engineering & System Safety, 242, 109714.
Offshore Energy Storage
Large-scale of Hydrogen underground storage in salt caverns: Hydrogen is taking a significant lead as a complementary energy carrier. One of the most significant structural challenges in the hydrogen supply chain is storing large volumes to ensure stability between generation, delivery, and utilization. Careful and rational design of the salt cavern are required to guarantee the structural stability and reliability of the cavern when in service. We proposed novel design methods to improve the reliability of the salt carven while considering the uncertainties. This research is one of the first to account for uncertainties within the salt cavern design process.
Selected Publications
Abreu, J. F., Costa, A. M., Costa, P. V., Miranda, A. C., Tassinari, C. C., Weber, N., … & Xu, Y. (2023, October). Large-Scale of Hydrogen Underground Storage in Salt Caverns: The Future of Sustainable Energy Storage. In ARMA/DGS/SEG International Geomechanics Symposium (pp. ARMA-IGS). ARMA.