Non-intrusive and semi-intrusive uncertainty quantification of a multiscale in-stent restenosis model

TL;DR

This paper compares non-intrusive and semi-intrusive uncertainty quantification methods for a multiscale in-stent restenosis model, demonstrating efficient and accurate surrogate modeling techniques, including a CNN-based blood flow surrogate.

Contribution

It introduces a CNN-based surrogate for blood flow, enhancing semi-intrusive uncertainty quantification accuracy and efficiency in a complex multiscale model.

Findings

Semi-intrusive method with CNN surrogate outperforms earlier models.

Both methods provide comparable uncertainty estimates to Monte Carlo simulations.

Semi-intrusive approach offers a good balance of efficiency and accuracy.

Abstract

Uncertainty estimations are presented of the response of a multiscale in-stent restenosis model, as obtained by both non-intrusive and semi-intrusive uncertainty quantification. The in-stent restenosis model is a fully coupled multiscale simulation of post-stenting tissue growth, in which the most costly submodel is the blood flow simulation. Surrogate modeling for non-intrusive uncertainty quantification takes the whole model as a black-box and maps directly from the three uncertain inputs to the quantity of interest, the neointimal area. The corresponding uncertain estimates matched the results from quasi-Monte Carlo simulations well. In the semi-intrusive uncertainty quantification, the most expensive submodel is replaced with a surrogate model. We developed a surrogate model for the blood flow simulation by using a convolutional neural network. The semi-intrusive method with the new surrogate model offered efficient estimates of uncertainty and sensitivity while keeping relatively high accuracy. It outperformed the result obtained with earlier surrogate models. It also achieved the estimates comparable to the non-intrusive method with similar efficiency. Presented results on uncertainty propagation with non-intrusive and semi-intrusive metamodeling methods allow us to draw some conclusions on the advantages and limitations of these methods.