FDA’s Nozzle Numerical Simulation Challenge: Non- Newtonian Fluid Effects and Blood Damage Miquel Trias, Antonio Arbona*, Joan Masso ´ , Borja Min ˜ ano, Carles Bona The Institute of Applied Computing & Community Code (IAC 3 ), University of the Balearic Islands, Palma, Spain Abstract Data from FDA’s nozzle challenge–a study to assess the suitability of simulating fluid flow in an idealized medical device–is used to validate the simulations obtained from a numerical, finite-differences code. Various physiological indicators are computed and compared with experimental data from three different laboratories, getting a very good agreement. Special care is taken with the derivation of blood damage (hemolysis). The paper is focused on the laminar regime, in order to investigate non-Newtonian effects (non-constant fluid viscosity). The code can deal with these effects with just a small extra computational cost, improving Newtonian estimations up to a ten percent. The relevance of non-Newtonian effects for hemolysis parameters is discussed. Citation: Trias M, Arbona A, Masso ´ J, Min ˜ ano B, Bona C (2014) FDA’s Nozzle Numerical Simulation Challenge: Non-Newtonian Fluid Effects and Blood Damage. PLoS ONE 9(3): e92638. doi:10.1371/journal.pone.0092638 Editor: Timothy W. Secomb, University of Arizona, United States of America Received September 10, 2013; Accepted February 25, 2014; Published March 25, 2014 Copyright: ß 2014 Trias et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This research has been partially funded by the European Commission (TSI-020400-2010-84), the Ministerio de Educacio ´ n y Ciencia (CSD2007-00042, PI12/02565), and the Balearic Islands government (53/2011). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: aarbona@iac3.eu Introduction Computational Fluid Dynamics (CFD) is commonly used by engineering teams in the design and development of many cardiovascular medical devices. However, its applicability for both demonstrating device safety and predicting potential problems based on patient-specific data is still very limited. In most cases, the efficacy of these techniques has not been fully proven. The main reason inhibiting the use of computational methods for such purposes within a regulatory review is the lack of reliable standarized methods. In order to meet this need, the U.S. Food and Drug Administration (FDA) recently completed a computa- tional interlaboratory study [1] to determine the suitability and methodology for simulating fluid flow in an idealized medical device. In particular, the goal of the FDA’s challenge was to establish a set of experimentally validated benchmark computa- tional models applicable to cardiovascular medical devices. As the FDA’s challenge article [1] states: ‘‘It is imperative to undertake and openly publish high quality validation cases relevant to the biomedical community (for both fluid dynamics and blood damage) to help improve the quality of biomedical CFD simulations’’. The FDA’s CFD challenge covers a whole range of Reynolds numbers, involving different physiological scenarios and moments of the cardiac cycle. This paper focuses on the low Reynolds (laminar) regime. We intend to contribute to the question of whether in such regime the simpler Newtonian models are enough or if, on the contrary, one must consider non-Newtonian models from the start. Non-Newtonian models are only potentially relevant in the laminar regime, as their effects tend to diminish with increasing Reynolds number. Different regimes will appear at different moments of the cardiac cycle and in different locations. Therefore a single code aiming at simulating full cardiac cycles, or dealing with devices that create around them different regimes should handle both turbulent, transitional, and laminar regimes. The development of such codes is far from trivial. Usually, each regime is targeted with quite different physical models (variations of the basic Navier- Stokes equations) and discretization techniques. If, additionally, non-Newtonian models are an ingredient we need to fit in, the whole picture is even a bit more complex. We intend to develop one of such codes, with an incremental strategy. In this sense, our approach consists in taking as starting point a Direct Numerical Simulation (DNS) approach for the laminar regime, with a view to extend later its validity to the transitional regime by means of either adopting Adaptive Mesh Refinement (AMR) capabilities or by including models used in Large Eddy Simulations (LES) when faced with the transition to turbulence. Our simulation code was automatically generated by Simflowny [2], a general-purpose platform for the management of physical models and simulation problems developed by ourselves at the IAC 3 . In this first paper, we address the laminar regime, limiting ourselves to the proposed Re t ~500 case. This allows us to investigate the role of non-Newtonian effects (non-constant viscosity). Even if we confirm previous claims [3,4] limiting the non-Newtonian effect in physiological parameters up to a ten percent, let us note that our code can incorporate such effects with virtually no computational cost. In this sense, we could also say that we are improving the results up to a ten percent without a significant extra computational cost. On the other hand, our results point out that even a few percent on the flow variables can translate into something bigger in crucial parameters, such as blood damage. PLOS ONE | www.plosone.org 1 March 2014 | Volume 9 | Issue 3 | e92638