Developed a digitalization and prediction tool that predicts the probabilistic remaining useful life (RUL) of an engineered system through model-based prediction and uncertainty quantification.Developed a sensor error detection model that detects and corrects multiple types of sensor errors (e.g., noise, outlier) within multiple engineered systems.Developed diagnostics and prognostics models that identify anomalies and prevent future failures within engineered systems through feature… Show more Developed a digitalization and prediction tool that predicts the probabilistic remaining useful life (RUL) of an engineered system through model-based prediction and uncertainty quantification.Developed a sensor error detection model that detects and corrects multiple types of sensor errors (e.g., noise, outlier) within multiple engineered systems.Developed diagnostics and prognostics models that identify anomalies and prevent future failures within engineered systems through feature extraction of different signals and machine learning algorithms.Developed a digital-twin framework that replaces computationally expensive FEA models and generates virtual sensor data for engineered systems.Deployed the models to production through Amazon Web Service (AWS) and set up model visualization dashboards in Grafana and Tableau. Show less

Investigated new surrogate-based approaches for design of experiment (DoE), reliability assessment and design under uncertainty.Investigated data-driven and model-based approaches for the diagnostics and prognostics of the engineered systems.

Developed a Python package with comprehensive user guide for reliability analysis of a component or system based on sequential DoE.Developed a Python-ABAQUS interface to facilitate the sequential DoE with computer simulation.

Led the product development team for the design, development and maintenance of the outside rear view mirror (OSRVM) in three global development projects (Buick Verano, Chevey Cruze and Malibu).Proposed and implemented eleven engineering changes on the OSRVM in the three projects to optimize the cost and performance.Established an OSRVM configuration guideline for new projects by collecting benchmark conguration data among all the models in Chinese market.