Preface

Like its first edition, the second Workshop on CFD Uncertainty Analysis is meant to bring people together who are concerned about the confidence in the reliability of their computational results. It is gratifying to see that the interest in participating has grown: we have 10 groups delivering results this time.

Two years have passed since the successful first edition of the Workshop on CFD Uncertainty Analysis. In those two years there has been, no doubt, some progress in CFD uncertainty analysis. Also a broader acceptance of the relevance of the subject can be observed. Nevertheless, organizing this second workshop was considered worthwhile for a specific reason. Last time we focused on verification, clearly distinguishing it from validation, but not fully recognizing the dual nature of verification, viz. code verification (error evaluation) and calculations verification (error estimation). So when the error bands estimated for the two test cases did not overlap, it was unclear whether the uncertainty estimation procedures were too optimistic or code implementation errors played a role. Therefore a Manufactured Solution has been constructed to serve as the first test case of this second workshop.

Nevertheless, the focus of the Workshop remains error estimation. Therefore the flow over a backward facing step (included as test C-30 in the ERCOFTAC database) is retained from the first edition as the second test case. But there is a new aspect here. While in the first edition common grid sets were provided, the participants for the present workshop were left free to select the grids to their own preference. This means that an impression of the effect of grid lay-out on the results is obtained.

We are very happy to have once again Patrick Roache in Lisbon, presenting an invited lecture and leading the discussions during the workshop. With his inspiring assistance, we sincerely hope that the workshop will evoke thorough discussions on the subject of CFD uncertainty analysis and broaden the awareness of the problems still to be solved. We trust that at the end of the workshop we have the confidence that a reliable error estimate can be made for calculations of the flow over a backward-facing step. On the other hand, we foresee that several challenges will remain to extend this confidence to CFD predictions for more complicated flow problems.

Welcome at IST in Lisbon to contribute to new achievements.