Framework
The Risk, Reliability and Uncertainty Quantification (R2UQ) research framework was conceptualised in the fall of 2023 as a result of a series of reflections on my research journey while authoring and revising my PhD Thesis. There are two aspects to the R2UQ framework - the Philosophical and the Technical aspects.
The research philosophy is based upon the three non-negotiable values pertaining to research conduct:
On the Technical aspect, the R2UQ framework encompasses the following disciplines:
Risk analysis;
Reliability engineering; and
Uncertainty quantification.
There are three research objectives which the R2UQ framework seeks to achieve:
To develop robust stochastic model updating framework towards inferring key parameters and model calibration under limited information;
To study and develop appropriate metrics to quantify model performance under model form uncertainty; and
To develop robust probabilistic bounds analysis framework for forward propagation under uncertain dependencies towards risk quantification.
With these objectives in mind, the research focus is classified into four distinct themes:
Theme 1: Data-driven stochastic model updating and calibration
Research focus: Development of robust numerical techniques towards characterising the uncertainty of the input parameter(s) and to calibrate models under limited information.
Key publications:
Adolphus Lye, and Luca Marino (2023). An investigation into an alternative transition criterion of the Transitional Markov Chain Monte Carlo method for Bayesian model updating. In Proceedings of the 33rd European Safety and Reliability Conference, Southampton. doi: 10.3850/978-981-18-8071-1_P331-cd
Adolphus Lye, Alice Cicirello, and Edoardo Patelli (2022). An efficient and robust sampler for Bayesian inference: Transitional Ensemble Markov Chain Monte Carlo. Mechanical Systems and Signal Processing, 167, 108471. doi: 10.1016/j.ymssp.2021.108471
Adolphus Lye, Alice Cicirello, and Edoardo Patelli (2021). Sampling methods for solving Bayesian model updating problems: A tutorial. Mechanical Systems and Signal Processing, 159, 107760. doi: 10.1016/j.ymssp.2021.107760
Theme 2: Data-informed model selection and averaging
Research focus: Development numerical techniques towards performing model selection and model averaging under limited information.
Key publications:
Michael McGurk, Adolphus Lye, Ludovic Renson, and Jie Yuan (2024). Data-Driven Bayesian Inference for Stochastic Model Identification of Nonlinear Aeroelastic Systems. AIAA Journal. doi: 10.2514/1.J063611
Adolphus Lye, Luca Marino, Alice Cicirello, and Edoardo Patelli (2023). Sequential Ensemble Monte Carlo Sampler for On-Line Bayesian Inference of Time-Varying Parameter In Engineering Applications. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems Part B: Mechanical Engineering, 9, 031202. doi: 10.1115/1.4056934
Adolphus Lye, Alice Cicirello, and Edoardo Patelli (2022). On-line Bayesian Model Updating and Model Selection of a Piece-wise model for the Creep-growth rate prediction of a Nuclear component. In Proceedings of the 8th International Symposium on Reliability Engineering and Risk Management, Hannover. doi: 10.3850/978-981-18-5184-1_MS-02-208-cd
Theme 3: Forward uncertainty propagation and interval analysis
Research focus: Development of numerical methods towards propagating imprecise model inputs through a model and yielding meaningful uncertain outputs to quantify the system performance/reliability.
Key publications:
Adolphus Lye, Ander Gray, Marco de Angelis, and Scott Ferson (2023). Robust Probability Bounds Analysis for Failure Analysis under Lack of Data and Model Uncertainty. In Proceedings of the 5th International Conference on Uncertainty Quantification in Computational Sciences and Engineering, Athens. doi: 10.7712/120223.10345.19797
Adolphus Lye, Masaru Kitahara, Matteo Broggi, and Edoardo Patelli (2022). Robust optimisation of a dynamic Black-box system under severe uncertainty: A Distribution-free framework. Mechanical Systems and Signal Processing, 167, 108522. doi: 10.1016/j.ymssp.2021.108522
Theme 4: Model Verification and Validation
Research focus: Development and the study of appropriate statistical distance functions to quantify how well a model agrees with the physics (verification) and the actual phenomenon (validation).
Key publications:
Adolphus Lye, Scott Ferson, Sicong Xiao (2024). Comparison between distance functions for Approximate Bayesian Computation towards Stochastic model updating and Model validation under limited data. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems Part A: Civil Engineering, 10, 03124001. doi: 10.1061/AJRUA6.RUENG-1223