The statistics obtained from turbulent flow simulations are generally uncertain due to finite time averaging. Most techniques available in the literature to accurately estimate these uncertainties typically only work in an offline mode, that is, they require access to all available samples of a time series at once. In addition to the impossibility of online monitoring of uncertainties during the course of simulations, such an offline approach can lead to input/output (I/O) deficiencies and large storage/memory requirements, which can be problematic for large-scale simulations of turbulent flows. Here, we designed, implemented and tested a framework for estimating time-averaging uncertainties in turbulence statistics in an in-situ (online/streaming/updating) manner. The proposed algorithm relies on a novel low-memory update formula for computing the sample-estimated autocorrelation functions (ACFs). Based on this, smooth modeled ACFs of turbulence quantities can be generated to accurately estimate the time-averaging uncertainties in the corresponding sample mean estimators. The resulting uncertainty estimates are highly robust, accurate, and quantitatively the same as those obtained by standard offline estimators. Moreover, the computational overhead added by the in-situ algorithm is found to be negligible allowing for online estimation of uncertainties for multiple points and quantities. The framework is general and can be used with any flow solver and also integrated into the simulations over conformal and complex meshes created by adopting adaptive mesh refinement techniques. The results of the study are encouraging for the further development of the in-situ framework for other uncertainty quantification and data-driven analyses relevant not only to large-scale turbulent flow simulations, but also to the simulation of other dynamical systems leading to time-varying quantities with autocorrelated samples.

Wall-modeled large eddy simulation (WMLES) has been proven to be a cost-effective approach capable of resolving turbulence up to certain resolutions. Among the simplest wall models used are the equilibrium wall models, assuming the pressure gradient and convective terms balance out in the momentum equations. There is a lack of studies to assess the performance of these standard wall models in internal turbulent flows including separation regions with/without particles. Regarding this research gap, we have conducted WMLES of incompressible turbulent flows, to the authors’ knowledge, for the first time, in 90° bent pipes with and without particles using an algebraic equilibrium wall model (Spalding's function). A pipe flow simulation was conducted to confirm the simulation setup and assess the sensitivity with respect to the modeling parameters. In each case, comparisons are made with experiment or direct numerical simulation (DNS), and depending on the case, with other existing simulation methods in the literature: WMLES, standard (wall-resolving) LES, and Reynolds stress model (RSM) for Reynolds-averaged Navier-Stokes (RANS) simulations. Despite the controversy on the performance of equilibrium wall models in nonequilibrium flows, our results show acceptable accuracy of this type of wall models. Specifically in the bent pipe flow with particles, WMLES succeeded in predicting particle deposition efficiency at Stokes numbers greater than 0.5, but obtained less accurate results for smaller Stokes numbers. The WMLES errors were, however, on par with those of the standard LES employed with a tenfold higher grid cell count. Improved results would be expected if combined with auxiliary mechanisms such as stochastic models.

Well-resolved direct numerical simulations (DNS) have been performed of the flow in a smooth circular pipe of radius R and axial length 10 pi R at friction Reynolds numbers up to Re-tau = 5200 using the pseudo-spectral code OPENPIPEFLOW. Various turbulence statistics are documented and compared with other DNS and experimental data in pipes as well as channels. Small but distinct differences between various datasets are identified. The friction factor lambda overshoots by 2% and undershoots by 0.6% the Prandtl friction law at low and high Re ranges, respectively. In addition,. in our results is slightly higher than in Pirozzoli et al. (J. Fluid Mech., vol. 926, 2021, A28), but matches well the experiments in Furuichi et al. (Phys. Fluids, vol. 27, issue 9, 2015, 095108). The log-law indicator function, which is nearly indistinguishable between pipe and channel up to y(+) = 250, has not yet developed a plateau farther away from the wall in the pipes even for the Re-tau = 5200 cases. The wall shear stress fluctuations and the inner peak of the axial turbulence intensity - which grow monotonically with Re-tau - are lower in the pipe than in the channel, but the difference decreases with increasing Re-tau. While the wall value is slightly lower in the channel than in the pipe at the same Re-tau, the inner peak of the pressure fluctuation shows negligible differences between them. The Reynolds number scaling of all these quantities agrees with both the logarithmic and defect-power laws if the coefficients are properly chosen. The one-dimensional spectrum of the axial velocity fluctuation exhibits a k(-1) dependence at an intermediate distance from the wall - also seen in the channel. In summary, these high-fidelity data enable us to provide better insights into the flow physics in the pipes as well as the similarity/difference among different types of wall turbulence.

Multifidelity models (MFMs) can be used to construct predictive models for flow quantities of interest (QoIs) over the space of uncertain/design parameters, with the purpose of uncertainty quantification, data fusion and optimization. For numerical simulation of turbulence, there is a hierarchy of methodologies ranked by accuracy and cost, where each methodology may have several numerical/modelling parameters that control the predictive accuracy and robustness of its resulting outputs. Compatible with these specifications, the present hierarchical MFM strategy allows for simultaneous calibration of the fidelity-specific parameters in a Bayesian framework as developed by Goh et al. (Technometrics, vol. 55, no. 4, 2013, pp. 501-512). The purpose of the MFM is to provide an improved prediction, mainly interpolation over the range covered by training data, by combining lower-and higher-fidelity data in an optimal way for any number of fidelity levels; even providing confidence intervals for the resulting QoI. The capabilities of the MFM are first demonstrated on an illustrative toy problem, and it is then applied to three realistic cases relevant to engineering turbulent flows. The latter include the prediction of friction at different Reynolds numbers in turbulent channel flow, the prediction of aerodynamic coefficients for a range of angles of attack of a standard airfoil and the uncertainty propagation and sensitivity analysis of the separation bubble in the turbulent flow over periodic hills subject to geometrical uncertainties. In all cases, based on only a few high-fidelity data samples, the MFM leads to accurate predictions of the QoIs. The result of the uncertainty quantification and sensitivity analyses are also found to be accurate compared with the ground truth in each case.

Information theory (IT) provides tools to estimate causality between events, in various scientific domains. Here, we explore the potential of IT-based causality estimation in turbulent (i.e. chaotic) dynamical systems and investigate the impact of various hyperparameters on the outcomes. The influence of Markovian orders, i.e. the time lags, on the computation of the transfer entropy (TE) has been mostly overlooked in the literature. We show that the history effect remarkably affects the TE estimation, especially for turbulent signals. In a turbulent channel flow, we compare the TE with standard measures such as auto- and cross-correlation, showing that the TE has a dominant direction, i.e. from the walls towards the core of the flow. In addition, we found that, in generic low-order vector auto-regressive models (VAR), the causality time scale is determined from the order of the VAR, rather than the integral time scale. Eventually, we propose a novel application of TE as a sensitivity measure for controlling computational errors in numerical simulations with adaptive mesh refinement. The introduced indicator is fully data-driven, no solution of adjoint equations is required, with an improved convergence to the accurate function of interest. In summary, we demonstrate the potential of TE for turbulence, where other measures may only provide partial information.

Combining different existing uncertainty quantification (UQ) techniques, a framework is obtained to assess a set of metrics in computational physics problems, in general, and computational fluid dynamics (CFD), in particular. The metrics include accuracy, sensitivity and robustness of the simulator's outputs with respect to uncertain inputs and parameters. These inputs and parameters are divided into two groups: based on the variation of the first group (e.g. numerical/computational parameters such as grid resolution), a computer experiment is designed, the data of which may become uncertain due to the parameters of the second group (e.g. finite time-averaging). To construct a surrogate model based on uncertain data, Gaussian process regression (GPR) with observation-dependent (heteroscedastic) noise is used. To estimate the propagated uncertainties in the simulator's outputs from the first group of parameters, a probabilistic version of the polynomial chaos expansion (PCE) is employed Global sensitivity analysis is performed using probabilistic Sobol indices. To illustrate its capabilities, the framework is applied to the scale-resolving simulations of turbulent channel and lid-driven cavity flows using the open-source CFD solver Nek5000. It is shown that at wall distances where the time-averaging uncertainty is high, the quantities of interest are also more sensitive to numerical/computational parameters. In particular for high-fidelity codes such as Nek5000, a thorough assessment of the results' accuracy and reliability is crucial. The detailed analyses and the resulting conclusions can enhance our insight into the influence of different factors on physics simulations, in particular the simulations of high-Reynolds-number turbulent flows including wall turbulence.

Bayesian optimization (BO) based on Gaussian process regression (GPR) is applied to different CFD (computational fluid dynamics) problems which can be of practical relevance. The problems are i) shape optimization in a lid-driven cavity to minimize or maximize the energy dissipation, ii) shape optimization of the wall of a channel flow in order to obtain a desired pressure-gradient distribution along the edge of the turbulent boundary layer formed on the other wall, and finally, iii) optimization of the controlling parameters of a spoiler-ice model to attain the aerodynamic characteristics of the airfoil with an actual surface ice. The diversity of the optimization problems, independence of the optimization approach from any adjoint information, the ease of employing different CFD solvers in the optimization loop, and more importantly, the relatively small number of the required flow simulations reveal the flexibility, efficiency, and versatility of the BO-GPR approach in CFD applications. It is shown that to ensure finding the global optimum of the design parameters of the size up to 8, less than 90 executions of the CFD solvers are needed. Furthermore, it is observed that the number of flow simulations does not significantly increase with the number of design parameters. The associated computational cost of these simulations can be affordable for many optimization cases with practical relevance.

An automatic method is proposed for the removal of the initialization bias that is intrinsic to the output of any statistically stationary simulation. The general techniques based on optimization approaches such as Beyhaghi et al. [1] following the Marginal Standard Error Rules (MSER) method of White et al. [16] were observed to be highly sensitive to the fluctuations in a time series and resulted in frequent overprediction of the length of the initial truncation. As fluctuations are an innate part of turbulence data, these techniques performed poorly on turbulence quantities, meaning that the local minima was often wrongly interpreted as the minimum variance in the time series and resulted in different transient point predictions for any increments to the sample size. This limitation was overcome by considering the finite difference of the slope of the variance computed in the optimization algorithm. The start of the zero slope region was considered as the initial transient truncation point. This modification to the standard approach eliminated the sensitivity of the scheme, and led to consistent estimates of the transient truncation point, provided that the finite difference time interval was chosen large enough to cover the fluctuations in the time series. Therefore, the step size for the finite difference slope was computed using both visual inspection of the time series and trial and error. We propose the Augmented Dickey-Fuller test as an automatic and reliable method to determine the truncation point, from which the time series is considered stationary and without an initialization bias.

A framework is introduced for accurate estimation of time-average uncertainties in various types of turbulence statistics. A thorough set of guidelines is provided to adjust the different hyperparameters for estimating uncertainty in sample mean estimators (SMEs). For high-order turbulence statistics, a novel approach is proposed which avoids any linearization and preserves all relevant temporal and spatial correlations and cross-covariances between SMEs. This approach is able to accurately estimate uncertainties in any arbitrary statistical moment. The usability of the approach is demonstrated by applying it to data from direct numerical simulation (DNS) of the turbulent flow over a periodic hill and through a straight circular pipe.