Postgraduate research opportunities Stochastic model updating with novel quantification approaches for multi-source uncertainties in aerospace engineering

Computational models of large-scale structural systems with acceptable precision, robustness, and efficiency are critical, especially for applications where a large amount of experimental data is hard to obtain, such as aerospace engineering.

Model updating has been developed as a typical technique to reduce the discrepancy between the numerical simulations and the experimental measurements. The typical deterministic updating is performed under the single-test-single-simulation strategy, where the emphasis is placed on pursuing the maximum fidelity regarding the single measurement.

Recently, it tends to consider the inevitable uncertainties involved in both simulations and experiments. A better understanding of the discrepancy between them, in the background of uncertainty, would achieve a better outcome of model updating. The involvement of uncertainty motivates the development of the so-called Stochastic Model Updating, which possesses the most interests of the current research in model updating.

In stochastic model updating, uncertainties are regarded as the reason of the discrepancy between the numerical simulations and the experimental measurements. The deterministic updating approaches assume the model parameters as unknown scalars, and calibrate them such that the model predictions are tuned towards the measurement. These approaches take an arbitrary treatment of the parameter and modelling uncertainties together while neglecting experimental uncertainty.

The stochastic model updating is hence required to provide model predictions with considerable robustness according to multi-source uncertainties; to integrate novel Uncertainty Quantification (UQ) approaches, e.g. the imprecise probability, interval, and fuzzy, capable of capturing uncertainty information from limited experimental data; and finally, to develop efficient algorithms to inversely calibrate the input parameters with the aid of advanced sampling and optimisation algorithms and the AI technologies.