Generative Modeling of Turbulence

We present a mathematically well founded approach for the synthetic modeling of turbulent flows using generative adversarial networks (GAN). Based on the analysis of chaotic, deterministic systems in terms of ergodicity, we outline a mathematical proof that GAN can actually learn to sample state snapshots form the invariant measure of the chaotic system. Based on this analysis, we study a hierarchy of chaotic systems starting with the Lorenz attractor and then carry on to the modeling of turbulent flows with GAN. As training data, we use fields of velocity fluctuations obtained from large eddy simulations (LES). Two architectures are investigated in detail: we use a deep, convolutional GAN (DCGAN) to synthesise the turbulent flow around a cylinder. We furthermore simulate the flow around a low pressure turbine stator using the pix2pixHD architecture for a conditional DCGAN being conditioned on the position of a rotating wake in front of the stator. The settings of adversarial training and the effects of using specific GAN architectures are explained. We thereby show that GAN are efficient in simulating turbulence in technically challenging flow problems on the basis of a moderate amount of training date. GAN training and inference times significantly fall short when compared with classical numerical methods, in particular LES, while still providing turbulent flows in high resolution.



There are no comments yet.


page 14

page 17

page 20

page 21

page 22

page 24


Flows Succeed Where GANs Fail: Lessons from Low-Dimensional Data

Normalizing flows and generative adversarial networks (GANs) are both ap...

On catastrophic forgetting and mode collapse in Generative Adversarial Networks

Generative Adversarial Networks (GAN) are one of the most prominent tool...

Parameter-Conditioned Sequential Generative Modeling of Fluid Flows

The computational cost associated with simulating fluid flows can make i...

AM-DCGAN: Analog Memristive Hardware Accelerator for Deep Convolutional Generative Adversarial Networks

Generative Adversarial Network (GAN) is a well known computationally com...

Using Physics-Informed Super-Resolution Generative Adversarial Networks for Subgrid Modeling in Turbulent Reactive Flows

Turbulence is still one of the main challenges for accurately predicting...

Discovering state-parameter mappings in subsurface models using generative adversarial networks

A fundamental problem in geophysical modeling is related to the identifi...

Sounderfeit: Cloning a Physical Model using a Conditional Adversarial Autoencoder

An adversarial autoencoder conditioned on known parameters of a physical...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

1 Introduction

Turbulent flows are characterized by unsteadiness, chaotic-like flow states and high degree of non-linearity. The structures involved exhibit a wide range of spatial and temporal scales, with the ratio of largest to smallest structures scaling with the Reynolds number[19]. In order to capture all scales of fluid motion directly, very fine computational meshes and time steps are required, which makes the computational effort in the case of engineering-relevant (high Reynolds numbers) problems impossible to accomplish in reasonable time despite the rapidly increasing computer performance. To circumvent this problem, closures are used, which allow to model the structures that cannot be captured by the coarser numerical meshes. However, this advantage in computation time is paid for with a modeling error, which can be considerable depending on the chosen approach and the underlying flow case.

Recent developments in the field of machine learning (ML), which are largely driven by increased computational power as well as the availability of exceptionally large data sets, make it possible to address this issue, whereby different approaches can be taken. One obvious approach is ML-based improvement of the prediction quality of existing models, also known as ML augmented turbulence modeling. Here, one possibility is to calibrate the empirically determined constants of the respective models for the underlying use case by means of data-driven ML augmentation

[9, 14, 15, 73, 67, 66, 75, 74]. Pioneering publications in this area are the works of Ling et al. [42] and Jiang et al. [32]

who used deep neural networks (DNNs) to determine the model constants of nonlinear algebraic eddy viscosity models and were thus able to significantly improve the prediction of anisotropic turbulence effects. Another way is the correction of existing models with the help of additional source terms, which were successfully used in

[52, 60, 61, 27] for the augmentation of turbulence models and in [71] for the augmentation of transition models.

A completely different approach has been pursued recently, based on the generative adversarial networks (GAN) as introduced by Goodfellow [24], which allow a hierarchical identification and abstraction of features in images by means of deep neural networks (DNN). By the fact that also in the case of turbulent flows there is a complex superposition of different structures and scales suggests that these methods are well suited for learning the physical relationships in such flows. In [36, 37] it was shown that GAN are able to generate synthesizations of 2D flow fields after they have been previously trained based on DNS data. The reproductions even fulfilled some statistical constraints of turbulent flows such as Kolmogorov’s

law and the small scale intermittency of turbulence. Using a deep unsupervised learning approach and a combination of a GAN and a recurrent neural network (RNN), Kim & Lee

[34] were able to generate high-resolution turbulent inlet boundary conditions at different Reynolds numbers, which show a statistical similarity to real flow fields.

Another application of GAN is the field of super-resolution reconstruction of turbulent flows. With these methods it is possible to synthetically scale up flow fields which are low-resolution or noisy due to the measurement technique used or, in the case of numerical data, due to limited data storage capacity.

[20, 21, 43, 11, 70, 68, 62]

. These works assume a supervised learning approach, which means that labeled paired datasets of low-resolution and high-resolution images must be available. Here, the low-resolution data sets are usually generated by filtering the high-resolution data sets obtained, for example, from direct numerical simulations (DNS). In many practical situations, however, such high-resolution data sets are usually not available, which to a certain extent limits the range of applicability. A more general and therefore more practical approach is the unsupervised super-resolution reconstruction method. Here, pairwise data sets are no longer necessary, as Kim & Lee could show by successfully using an unsupervised GAN for the generation of boundary conditions for turbulent flow

[34]. Applications of such methods would be e.g. the augmentation or denoising of experimental data sets (PIV) or the derivation of subgrid-scale models for the application in the field of large-eddy simulation (LES).

In our work, we show the possibilty of synthesizing turbulence structures of a similar quality as predicted by of LES with GAN trained from scratch and completely unsupervised. Thus, we are able to produce data with help of the trained generator by only having a noise vector as input. Moreover, we show by investigation of conditional GAN that generators of synthetic turbulent flows can learn to cope with changes of the geometry of the flowpath, e.g. caused by a rotation wake. We also show that introducing generative learning to model turbulences finds its justification in the enormous reduction of computational time compared to LES, while maintaining the resolution. Lastly, besides the practical aspects, we prove, using the mathematical concept of ergodicity, that learning to generate states of chaotic systems using GAN is possible.


The paper is organised as follows. In section 2 we briefly summarize the concept of ergodicity, discuss the mathematical foundations behind GAN along with the learning theory for deterministic ergodic systems. Also, a survey of modern GAN architectures is given. The hierarchy of datasets used for our experiments, ranging from the Lorenz attractor and the flow around a cylinder to a perioic wake impinging on a low-pressure turbine stator blade, are described in section 3. This is followed by section 4, where we give details on the training of our various GAN models. In section 5 we discuss the results of our numerical experiments. Finally, in section 6 we present the conclusion and an short outlook.

2 Methodology

In this work we apply generative adversarial networks (GAN) to generate typical states of a deterministic chaotic dynamic system. This is made mathematically precise via the notion of ergodicity [55].

2.1 Ergodicity

Let be

a probability space, consisting of a state space

, a collection of events/subsets of the state space known as -algebra and a probability measure on that attributes the probability to the events . In our context, the state is chosen as the phase space of a dynamic system , , that fulfills and .

Frequently in this work, we need the concept of an image measure, i.e. the transformation of a measure by a mapping. To this purpose, let be a mesurable mapping with respect to the -algebra on and a second sigma algebra on , i.e. for all we have . The image probability measure of under , denoted by , is then defined by


In the following, without further mention, we assume all mappings to be measurable with respect to suitable -algebras.

In the case considered here, is a state space of a dynamic system. A dynamic system with the given state space consists of a collection of mappings that fulfill and , where . In many cases, like ours, the state of the dynamic system at time is obtained as a solution mapping

associated with a (discretized) ordinary or partial differential equation starting in the initial state

. E.g., for the case of the Lorenz attractor or with a large number of dimensions of the disrcetized state space of the fluid field in the case of the numerical simulation of turbulent fluids.

The probability measure is an invariant measure for the dynamic system defined by , if all solution mappings are measure preserving with respect to , i.e. for all .

We next turn to the space of physical observables on and define it as space of all square-integrable functions , i.e.


We next turn to to the notion of ergodicity, which equates the time average of a dynamic system with the ensemble average of its invariant measure. In mathematical notation, ergodicity of the dynamic system with respect to the invariant measure is defined as


Neumann [47] and Birkhoff [7] established quite general conditions, under which ergodicity holds. See also [16, 3] for extensive treatments of discrete and time continuous ergodic systems.

In some of our numerical experiments, we do not consider the entire state-space

, but reduce the degrees of freedom using a mapping

with the reduced state space. Let be the projected measure. Assuming the ergodictity of the original dynamics with respect to , we see that


whenever is square integrable with respect to . This easily follows from the general transformation formula and (3). Hence, ergodicity remains meaningful on the reduced state space , even if the dynamics can not be consistently formulated on .

2.2 Mathematical foundations of generative learning for ergodic systems

Generative Adversarial Networks (GAN) consist of two mappings - a generator and a discriminator . Here is a space of latent variables endowed with a probability measure that is easy to simulate, e.g. experiments uniform or Gaussian noise. The generator transforms the noise measure to the image measure . The goal of adversarial learning is, to learn a mapping from the feedback of the discriminator , such that is not able to distinguish synthetic samples from from real samples from the target measure . However, the discriminator

is a classifier that trained to assign real data a high probability of being real and synthetic data a low probability. If

has been so well trained, that even the best discriminator can not distinguish between samples from and , generative learning is successful, see also Fig. 1.

In practice, both the generator and the discriminator are realized by neural networks. The feedback of to is transported backwards by back-propagation [58] through the concatenated mapping in order to train the weights of the neural network . At the same time, the universal approximation property of (deep) neural networks guarantees that any mappings and can be represented with a given precision, provided the architecture of the networks is sufficiently wide and deep, see [25, 5, 56, 64, 31, 76, 65, 33] for qualitative and quantitative results.

The training of GAN is organized as a two-player minimax game between and . Mathematically, it is described by the min-max optimization problem


with the loss function, also known as binary cross-entropy



Here, the expected value is denoted by

, the random variable

with values in follows the distribution of the real world data and the latent random variable with values in follows the distribution of the noise measure . As has been observed in [25],


if the maximum is taken over a sufficiently large hypothesis space of discriminators. Here, stands for an information theoretic pseudo distance between the invariant measure and the generated measure known as the Jensen-Shannon divergence


with the Kulback-Leibler pseudo distance between the measures and with continuous probability densities and , respectively. Note that holds if and only if holds with -probability one and hence . Consequently, also measures the distance between and .

Figure 1: Architecture of the original GAN. According to: [25]. The generator produces fake samples by the random vector as its input. On the other hand the real-world data

representing the training data is given. The discriminator gets as input fake as well as real samples and estimates the probability that the given input sample comes from

than generated by . Thus, the output of is a single scalar value per sample in the range of

. The feedback of the discriminator reaches the generator when the weights of the GAN framework are updated by backpropagation

[58] during the training. Since both networks are fully differentiable and trained end-to-end the whole GAN framework can be backpropagated in one go using the same loss function for and . The optimum of the problem (5) is reached if the distribution of the real-world data is captured by the generator and the discriminator is not able to distinguish real from fake samples, so and .

2.3 Learning theory for deterministic ergodic systems

In this work, we show that it is possible to model turbulent flows with GAN in practice. In this section we outline a proof that generative learning for deterministic ergodic systems converges in the limit of large observation time .

As described in section 2.2 is the unknown invariant measure encoding the statistical properties of the dynamic system with the initial state. Our goal is to sample from but since it is unknown, we want to learn it from the data given by the observed trajectory . Thus, in context of generative learning a generator is searched for which holds , where is, e.g., the Lebesgue measure than corresponds to -dimensional uniform noise.

Let be the invariant measure of the dynamic system acting on the measurable space with the Borel--algebra and the sample space of state configurations with normalized state components in . It is assumed that with the continuous probability density in the space of -times differentiable -Hölder functions [1]. If this is not the case, one can easily regularize to achieve this. Moreover, we assume also lies in the space of --Hölder functions , . By the realizability theorem of [5] it follows that , such that


By knowing that is realizable in the hypotheses space


for sufficiently large, our goal is to estimate by based on the data given by the ergodic flow .

The estimation of is performed using an empirical loss function that is designed to approximate the theoretical loss function (6) and hence minimizing the difference between the measure of the ergodic system and the image measure of the synthesized images. Mathematically, we search the generator


with the discriminator hypotheses space such that an optimal choice of is feasible:


Here, stands for the continuous probability density associated with the probability measure . We propose


as empirical loss function for the ergodic system where denotes the rounding function.

Apparently, in the limit by ergodicity (3), the first term converges to the first term in (6

) whereas the second term converges almost surely by the law of large numbers. Therefore, the generator

that is learned from the empirical loss function (13) for large will approximately solve the minimax problem (5), which by (7) relates to the Jensen-Shannon distance between the estimated measure and the invariant measure of the ergodic system. In particular, we obtain the following:

Theorem 1.

Under the assumptions above it holds almost surely111w.r.t. the probability measure used for the sampling of the latent noise variables . that


Here we give a sketch of the proof. For a detailed argument in a related situation, see [5]. We introduce the following notation: is the discriminator solving and, likewise, , where we suppressed the suppressed the dependence of and to ease the notation. We obtain the estimate


In the first equality we used (7). In the third line, the definition of was used and in the fourth line we applied (11). In the sixth line, we used the definition of . In the seventh line, we again used (7) and in the final line we applied (9), which is possible under the given assumptions as proven in [5].

It remains for us to show that the sampling error on the right hand side of (15) vanishes as . Note that we can decompose


The second term on the right hand side vanishes by the uniform law of large numbers, as the hypothesis spaces and can be endowed with topologies that are, for , slightly little weaker than the -topology. Nevertheless, the hypothesis spaces under these topologies are compact, see [5] for the details. Consequently, the expression in the first term vanishes by the standard uniform law of large numbers, see e.g. [17].

For the first term, we have already seen that ergodicity implies that the expressions in the absolute value by ergodicity vanish in the limit . Also, with respect to the aforementioned -topologies the hypothesis spaces are compact. Last, it is easy to see that is equicontinuous in wrt. this topology (as is uniformly lower bounded away from zero in ). As for equicontinuous functions, pointwise convergence implies uniform convergence, the first term on the right hand side vanishes as well in the limit . ∎

We note that in practice, the Hölder generators and discriminators are replaced by deep neural networks. As such networks possess the universal approximation property, see e.g. [72], one can approximate the Hölder functions to arbitrary precision. Secondly, instead of solving the integral in (13) to compute the loss function, one uses a monte carlo approximation by sampling from the trajectory . Theorem 1 remains valid under this replacement, as one can see from one further application of the uniform law of large numbers.

Note however that these theoretical results do not guarantee the success of the numerical experiments. This is mostly due to the fact that the optimization problem (11) is highly non-convex and can not be solved exactly, as e.g. for neural nets this problem is NP-hard [59]. In practice, one rather finds sufficiently good local minima instead of a global optimum. Also, practical issues occur with the choice of the capacity and other elements of architecture of the neural networks.

2.4 Advanced GAN frameworks

After the introduction of the original GAN framework by Goodfellow figure 1, it became apparent that GAN are powerful models which can be applied to a wide variety of tasks by modifying or extending the architecture [51]. In this work three of these modified frameworks are investigated.

Wasserstein GAN (WGAN)

The Wasserstein GAN differs from the original GAN mainly in the change of the loss function and thus also in the change of the optimization problem [4]. For the WGAN framework the goal is not to minimize the Jensen Shannon divergence but the Wasserstein distance expressed by the Kantorovich-Rubinstein duality


with the supremum over all -Lipschitz functions and a compact metric set. Under the satisfaction of certain conditions the authors of [4] showed that the optimization problem


has a solution for and that the gradient of (17) exists.

In practice, the solution of (18) can be approximated by training a neural network parameterized by the weights with a compact space. This assumption implies that all parameterized functions are -Lipschitz for some . To ensure that all weights lie in a compact space and thus the Lipschitz constraint is preserved, the weights are clipped [8] to a certain range after each gradient update in the implementation.

Deep Convolutional GAN (DCGAN)

The deep convolutional GAN has the same base architecture as shown in figure 1, but the generator and the discriminator

are convolutional neural networks (CNNs)

[56]. These kind of neural networks are especially in the field of image processing successfully applicable [2, 35]. In order to be able to integrate CNNs into GAN the authors of [56] pointed out which guidelines are to follow to enable a stable training at higher resolution and with deeper architectures.

The stability of the training is ensured by applying batch normalization

[30] on the output layer of and the input layer of . To work with deeper architectures fully-connected layers [45]

should be avoided on top of convolutional features. Finally, the choice of the leaky rectified linear unit (LReLu) activation function

[49] for allows higher resolution modeling. Moreover, the generator captures faster the color space of the distribution by applying bounded activation functions in the last layer as the LReLu [49]. Finally, mentionable is that and are able to learn their own spatial up- or downsampling by replacing deterministic spatial pooling layers [13]

with (fractional-) strided convolutions.

Conditional GAN (cGAN)

By conditioning a GAN framework with additional information it is possible to take the control over the data production process performed by the generator [46]. Thereby, additional information can be represented for example by class labels or semantic segmentation masks [23]. As shown in figure 2 the conditioning can be realized by feeding the supplementary information to the discriminator and the generator as an extra input channel. During training, is sampled from a data model , where gives the distribution of in the data generation process. This extension of the architecture leads to the modified loss function

Figure 2: Architecture of a conditional GAN. According to: [46].

A special form of the cGAN investigated in this work is the so called pix2pixHD introduced by [64]. This conditional adversarial framework allows to generate high-resolution photo-realistic images from semantic segmentation masks. The pix2pixHD framework is based on its former version pix2pix [31] whose optimization problem is given as


with defined as in (19). To improve the photorealism and the resolution of the generated images the architecture was changed by introducing three innovations.

First, a coarse-to-fine generator was implemented. For this, the generator was decomposed into the two sub-networks having the role of a global generator and as a local enhancer. By this the global and local information can be aggregated effectively within the generator for the image synthesis task.

In order for the discriminator to distinguish between generated and real high-resolution images it needs a large receptive field. Therefore, the common discriminator was replaced by three multi-scale discriminators and which have an identical network architecture, but operate at three different image scales. Hence, the optimization problem (20) extended to


In particular, a pyramid of images is created during the training by downsampling the input image by factor two and four. Since the discriminator operating on the coarsest scale has the largest receptive field and hence a more global view it is possible to guide the generator producing globally consistent images. Whereas, the discriminator performing on the finest scale is able to make the generator pay attention to finer details during the data production.

Lastly, a feature matching loss [64] was added to (21) in order to stabilize the training of the pix2pixHD framework. By this, the complete optimization problem is defined as


with the weighting parameter for both terms.

3 Preparation of datasets

The datasets used for generative learning are described below. We proceed from the Lorentz attactor as a simple chaotic system to LES simulations of simple and complex turbulent flows.

3.1 Lorenz attractor

The Lorenz attractor is a non-periodic, non-linear and deterministic ergodic system which is given by the system of ordinary differential equations

[44] :


By [63] it has been proven that this dynamic system is representing a strange attractor. Within this hydrodynamic system describes the rate of convection, is proportional to the temperature variation between ascending and decreasing flow and represents the distortion rate of the vertical temperature profile from linearity [44].

The physical parameters are given by as the Prandtl number, as the relative Rayleigh number and representing the measure for the cell geometry. In this work we use the classic parameter values , and [40].

The training data for the generative learning is given by the points of the attractor’s trajectory within the three dimensional space computing the system (23) applying the odeint routine of the python package scipy.integrate which uses the lsoda algorithm [10]. In total data points of trajectories started from different initial points randomly sampled within the ranges , and .

3.2 Les

The computational fluid dynamics (CFD) results presented in this paper form the basis for GAN training. They were generated using large-eddy simulations (LES). In this approach, the spatially filtered variant of the Navier-Stokes equations is solved, with the computational grid designed to provide a resolution of at least 80% of the turbulent kinetic energy (TKE) of the flow. The effect of smaller turbulent structures, which are not captured by the grid, are represented using semiempirical models, the so-called subgrid scale models [28]. The spatial filter is thus implicitly given by the computational grid. The LES approach is reasonable, because it is the large vortex structures that transport the bulk of the energy [18] while the smaller structures can be considered to be mainly isotropic and homogeneous (not in the close vicinity of solid walls) by the assumption of local isotropy according to Kolmogorov [39], which simplifies their modeling considerably.

3.3 Test-cases & numerical setup

Two different test cases were chosen for training of GAN, which differ in the complexity of the resulting flow field. Both simulations were performed with the commercial flow solver ANSYS Fluent which was set up to solve the incompressible variant of the spatially filtered Navier-Stokes equations. For time integration, a non-iterative time advancement scheme is used in combination with a fractional step method for pressure-velocity coupling. The advective fluxes are treated by a bounded central scheme in order to introduce as low numerical dissipation as possible to avoid unphysical dampening of small turbulent structures [69].

(a) Flow around a cylinder
(b) T106 turbine stator
Figure 3: Numerical domains for both investigated test cases

3.3.1 Flow around a cylinder

The first test case is the flow around a cylinder at Reynolds number 3900. This is a widely used test case, which has been studied in great detail in the literature both experimentally [53, 48, 50] and numerically [53, 6, 22]. The flow field in this case is characterized by a Kármán vortex street, that forms in the wake region of the cylinder and consists of the typical coherent vortex system, where the axis of rotation of the individual vortices is parallel to the axis of the cylinder. A schematic representation of the numerical domain is shown in Fig. 2(a). The computational grid consists of a total of 15 million cells. The time step was chosen so that the CFL number was on the order of unity, and the simulation was run for a total of time steps after initial transient effects had disappeared, which corresponds to a total physical time period of approximately seconds.

3.3.2 T106 turbine stator under periodic wake impact

The second test case is an academic low-pressure turbine (LPT) stator under periodic wake impact. In this configuration, the wakes, which are comparable to those of the cylinder test case described above, are artificially generated by means of an upstream mounted rotating bar grid. The wakes are convected into the stator passages where deformation occurs as a consequence of the flow turning within the passage. Furthermore, a complex interaction between the wakes and the periodically detaching boundary layer takes place in the rear region of the suction side of the LPT stator, which in total makes this test case an interesting demonstrator for complex turbulent interaction phenomena. A schematic representation of the numerical domain is shown in Fig. 2(b). The computational grid consists of a total of approx. 72 million elements. The time step was chosen so that the CFL number was on the order of unity, and the simulation was run for a total of time steps after initial transient effects had disappeared, which corresponds to 10 bar passing periods or approx. .

3.4 Data sets and data production

The data sets used for training the GAN were generated by post-processing the transient LES velocity field data. In this process, grayscale images are generated via a projection mapping in the sense of (4). In the case of the flow around a cylinder experiment, the gray scale is showing the distribution of the absolute deviation of the local fluctuating velocity magnitude at the location from its time average


As the moving wake determines the turbulent flow field in the case of the LPT turbine, time averaging at a fixed point in this case does not make much sense. Therefore, a different representation of the turbulence (or projection mapping) is chosen, which simply depicts the velocity component perpendicular to the image, . Figure 4 shows an example image for each of the two test cases examined. The gray scale for is found in the upper left corner of the right panel. Negative values for are shown in lighter and positive values in darker grey.

Basic parameters of the generated data sets are summarized in table 1. The time step interval between two successive frames is chosen so that the respective snapshots are sufficiently far apart in time to minimize the correlation between the individual frames.

(a) Flow around a cylinder
(b) T106 turbine stator
Figure 4: Example snapshots for both investigated test cases extracted from the LES.
Sampling frequency Image resolution Number of files Total size
Cylinder 1000 x 600 px
Turbine 1000 x 625 px
Table 1: Summary of the main data set parameters.

3.5 Computational cost

At this point, the computational effort of the simulations presented in this paper should be briefly discussed, as this is the main criterion for the applicability of such scale-resolving simulations.

All simulations presented were performed on the in-house High-Performance Computing (HPC) cluster of the Chair of Thermal Turbomachines and Aero Engines, whose main specifications are summarized in table 2.

Partition Number of nodes Cores per node CPU type RAM Interconnect
#1 28 28 Intel Xeon "Skylake" Gold 6132 @2.6 GHz 96 GB Intel Omni-path
#2 8 40 Intel Xeon Scalable Gold 6248 @2.5 GHz 96 GB Intel Omni-path
TOTAL 36 1104 3.4 TB
Table 2: Summary of the main specifications of the HPC cluster.

In total 20 computational nodes of the #1 partition of the HPC cluster were allocated in both runs, resulting in a total number of 560 CPU cores. In the case of of the flow around a cylinder, this resulted in a total computation time of about one day for the output run consisting of iterations, which corresponds to about 72 core weeks. In the case of T106 LPT stator, the calculation time was approx. 8 days for the output run consisting of time steps, which corresponds to 10 bar passings, i.e. approx. 640 core weeks.

4 Setup and configuration of GAN training

The implementations details of the training with the GAN frameworks introduced in section 2.4 are summarized for the different datasets in the following. All GAN were set up and trained using the PyTorch [54].

4.1 Lorenz attractor

The Lorenz attractor was trained by a original GAN with a discriminator consisting of four fully connected hidden layers [26] with and neurons. Since the attractor is a deterministic ergodic system [44] Gaussian noise was added to the network of the discriminator as well as to the real input data to regularize the training and hence reduce overfitting [5, 12]. The real data representing the training data is given by the points of the attractor’s trajectory within the three dimensional space as described in section 3.1.

The generator is also given by a fully connected neural network composed of three hidden layers with and neurons. Its input is given by a random vector of dimension

whose elements come from the standard normal Gaussian distribution.

Both neural networks and

apply the ReLu activation function for the input and hidden layers. The activation of the output layer of the discriminator is given by a sigmoid function and for the generator by a linear function.

The GAN framework was trained for epochs with a batch size of . Hence, the trajectory consisting of data points was regarded during one epoch whereby the trajectory started from different randomly sampled initial points lying in the ranges , and .

The optimization problem was given as in (5). To update the weights of the neural networks and the Adam optimizer [38] was applied with the parameter , and a learning rate of . Here, only half of the batch size was used to update the weights of the discriminator.

4.2 Flow around a cylinder

Experiments have been performed on this dataset using the original GAN, WGAN and DCGAN framework. For the original GAN and WGAN the discriminator is given by a fully connected neural network with five layers in total whereby the hidden layers consist of and neurons. The generator of both GAN frameworks also consists of five fully connected layers in total with the number of and neurons for the hidden layers. In exception of the output layer the Leaky ReLu is applied as activation function. The last layer of the generator is activated by the hyperbolic tangent function. For the original GAN the discriminators last layer is activated by the sigmoid function and the linear activation function is used in case of the WGAN. For the training of the DCGAN the architecture suggested by [41] was adopted.

The three investigated GAN frameworks take images of size as input. In our experiments we investigated the training with . We trained all GAN for epochs with a batch size of using images of the dataset. For further investigations the DCGAN training was continued up to epoch . The input vector of the generator consists of elements randomly sampled of the standard Gaussian distribution.

For the update of the weights, the Adam optimizer is applied in case of the original GAN and DCGAN with the parameter settings and and a learning rate of

is used. For the WGAN the weight update is realized by the optimizer RMSProp

[57] with a learning rate of whereby the weights are clipped to the range .

4.3 T106 turbine stator under periodic wake impact

The DCGAN has been also trained for epochs and on the whole dataset of the wake disturbed turbine stator-row with the parameter settings described as in section 4.2.

Moreover, the pix2pixHD has been trained as second GAN framework with this dataset. As described in section 2.4 the pix2pixHD is a conditional GAN and hence incorporates additional information to the training. Here, this supplementary information is given by the binary segmentation masks shown in figure 5. In terms of conditional GAN-learning (19

), this corresponds to a uniform distribution

over the coordinate of the wake. For the experiments with the pix2pixHD the implementation of [64] has been used with small changes. To avoid the appearance of artifacts in the data synthesized by

we replaced the reflection padding with a replication padding and add a replication padding to the global generator before the convolution during the downsampling procedure.

Contrary to the DCGAN framework it is possible to train the pix2pixHD on images of size . The only important thing to take care of is that and are divisible by . For this reason, the images were resized for the training to size , such that the aspect ratio has been preserved.

Since the GAN is trained in a conditioned fashion the binary masks are also needed during the inference. For this reason, the dataset was split into a training- and test set. The training set contains the first images of the whole dataset and the test set consists of the remaining images.

The pix2pixHD has been trained for epochs with a batch size of . Analogous to the DCGAN the weights were updated by the Adam optimizer with the parameter , and a learning rate of .

Figure 5: An image of the training set (left) and its corresponding binary segmentation mask (right).

5 Results of experiments

The results of the numerical experiments are presented and discussed in this section. In the following, we refer to the process of applying a trained generator to the latent random vector as inference. At inference time, the latent vector also consists of

elements sampled from the standard normal distribution.

5.1 Lorenz attractor

As described in section 4.1 we trained a original GAN for epochs in order to synthesize three dimensional data points which come from a trajectory of the Lorenz attractor that has converged towards the strange attractor. For consistency, a trajectory of real data points is considered at inference time as in the training. To get a better overview of the results, data points produced by the trained generator are shown in figure 6. It can be observed that the generated data points are on or close to the true trajectory of the Lorenz attractor. For the points that do not seem to lie directly on the trajectory, it has to be taken into account that the trajectory shown here is also not very dense due to the small number of data points. Considering randomly sampled real data points of a trajectory consisting of one million data points as it must be noted that the distribution is similar to the one of the synthesized data points. Moreover, it can be seen from the rotated figure 7

that, apart from a few outliers, the generated data points are all located in the area of the trajectory in three-dimensional space.

Figure 6: The - plane perspective of a trajectory from the Lorenz attractor consisting of data points starting from the initial point (grey), synthesized data points (blue) and real data points randomly sampled from a trajectory consisting of one million data points (red).
Figure 7: Rotated perspective from the trajectory of real data points (grey), synthesized data points (blue) and randomly sampled real data points (red) given in figure 6.

5.2 Flow araound a cylinder

In order to generate the Kármáan vortex street, GAN frameworks with a simpler architecture have been considered first, namely the original GAN and the WGAN. As to observe in figure 8 the trained generators of both GAN are able to position the cylinder in the right place after epochs and that they try to synthesize the wake vortex. However, neither the original GAN nor the WGAN can capture the concrete structure of the vortex street. In addition, it is to observe that the color space has not been learned appropriate by the original GAN such that the generated images are significantly darker than the original images from the LES (see figure 3(a)). To address these issues, another GAN framework has been considered whose generator and discriminator are represented by convolutional neural networks. As already described in section 2.4, CNNs can be used particularly successfully in image processing. In our experiments, we also found that the DCGAN was able to capture the flow structures after epochs in contrast to the original GAN and the WGAN (see figure 8). To increase the quality of the synthesized images the DCGAN has been further trained out to epoch (see A for the training progress). Based on figure 9, it can be seen that the images produced by the generator of the DCGAN hardly differ from the real images from the LES after epochs of training.

Finally, it should be mentioned that the networks have been trained on images of size . It has been observed in our experiments that the quality of the generated images have been significantly better with increasing image resolution at inference time. Therefore, we present here the results for the training with images of size .

Figure 8: Comparison of images synthesized by the generator of a original GAN (left), WGAN (middle) and DCGAN (right) after epochs trained on images of size , .
Figure 9: Comparison of images from the LES (top) and synthesized Karman vortex streets produced by the generator of the DCGAN trained over epochs (bottom).

5.3 T106 turbine stator under periodic wake impact

Since we got impressive results from the DCGAN for the flow around a cylinder, we trained this GAN framework under the same parameter settings for the second test case. As we observe in figure 10, the LPT stator has been correctly positioned and the structure of the vortex flows has been also reasonably captured. However, at inference time, the generator has massive problems correctly capturing the position of the cylinder as it periodically slides from bottom to top over time. Especially by direct comparison in figure 11 we can observe, that the structures in the background are not properly captured and the synthesized images are significantly darker than the real images of the LES. To address these problems of the DCGAN we considered the pix2pixHD as another GAN framework. In order to have control over the position of the cylinder at inference, we feed binary segmentation masks shown in figure 5 as additional information to the GAN framework during training and at inference time (see section 2.4). These masks have the information about the position of the cylinder and the LPT stator. Moreover, we are allowed to generate high resolution images by the pix2pixHD framework such that the structure in the background of the images should also be preserved.

As shown in figure 12, using the generator from pix2pixHD we were able to generate images which again can be hardly qualitatively distinguished from the real image from the LES on a visual level after only epochs (see A for the training progress). It is also noticeable that the wake vortices do not look identical. Hence, the generator did not simply memorize the structure of the wake vortices at the respective positions and thus variation is given in the synthesized data.

Figure 10: Examples of images synthesized by of the DCGAN trained on images for epochs.
Figure 11: Comparison of a real images from the LES (left) and an images synthesized by of the DCGAN trained on images for epochs (right).
Figure 12: Comparison of images from the LES (left) and synthesized turbulences under periodic wake impact produced by the generator of the pix2pixHD trained over epochs (right).

5.4 Comparison of Computational Costs

Finally, the computational costs of the training and inference performed on a GPU of type Quadro RTX 8000 with GB of the successful GAN frameworks are reported in this section.

The training of the DCGAN with images of the dataset showing the flow around a cylinder has taken minutes per epoch. The computational time of pure inference is given by seconds per image. Thus, the production of a dataset containing images would take with the beforehand trained generator about minutes. This leads to a tremendous amount of time saved compared to one day needed for the generation of the images by the LES.

Since the pix2pixHD has a much more complex architecture than the DCGAN the training of one epoch with images has taken minutes. However, the computational time of pure inference is also given by only seconds per image. Hence, the production of images of the LPT stator under periodic wake impact would take about at inference. Thus, the saved computational time for the data production is very significant in comparison to days for the LES.

6 Conclusion and Outlook

We introduced generative adversarial networks as another way to model turbulence. In doing so, we showed that through generative learning it is possible to synthesize turbulence that matches the quality of LES images on a visual level while dramatically reducing computational time. Unlike previous work, we trained the GAN from scratch and only require a randomly samples noise vector for the data production in the unconditional case. For training and inference of conditional GAN, we also need binary segmentation masks which can be created manually and do not necessarily need to be obtained by simulations. Using conditional GAN, we have found a solution for generating visually high-quality turbulence when solid objects as the rotation wake change position in space. Thus, we have provided a first approach to generalization with respect to spatial changes. Finally, we have also demonstrated that generative learning of ergodic systems also works at the theoretical level.

So far, we have evaluated our experiments at the visual level. In our future research, numerical experiments are of interest to measure also quantitatively the similarity between the LES images and the images generated by the GAN. Attention will also be given to exploring appropriate methods for making these comparisons. In addition, we have so far ignored the physics involved. Therefore, the next step is to feed the GAN with physical parameters so that turbulent flows can also be captured by the GAN in a physically correct manner. Having provided a first approach to generalization in terms of changes in turbulence space, in future work we will also consider how generalization can be realized in terms of geometries and further boundary conditions.


C.D. and H.G. thank Matthias Rottmann and Hayk Asatryan for discussion and useful advice. The authors also thank Pascal Post for valuable hints for the literature research.

Appendix A Training history of the GAN frameworks

The training progress of the experiments with the DCGAN discussed in section 5.2 is described in figure 13. Since we trained the GAN framework on images of size we also got images of this size as output during the training. It can be observed that the synthesized images already show a quite good quality after epochs. However, on closer inspection, it is noticeable that the structures of the vortex street become finer with an increasing number of training epochs and that the color space is also captured much better after . In figure 14 the development of the synthesized images during the training is illustrated for the pix2pixHD whose results are discussed in section 5.3. Similar to the DCGAN we can observe that the results improve significantly with increasing number of training epochs.

Figure 13: Development of training results after , , and epochs for the DCGAN.
Figure 14: Development of training results after , , and epochs for the pix2pixHD framework.


  • [1] R. A. Adams and J. J. Fournier (2003) Sobolev spaces. Elsevier. Cited by: §2.3.
  • [2] S. Albawi, T. A. Mohammed, and S. Al-Zawi (2017) Understanding of a convolutional neural network. In 2017 International Conference on Engineering and Technology (ICET), pp. 1–6. External Links: Document Cited by: §2.4.
  • [3] H. Arbabi and I. Mezić (2017) Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the koopman operator. SIAM J. Appl. Dyn. Syst. 16, pp. 2096–2126. Cited by: §2.1.
  • [4] M. Arjovsky, S. Chintala, and L. Bottou (2017-06–11 Aug) Wasserstein generative adversarial networks. In Proceedings of the 34th International Conference on Machine Learning, D. Precup and Y. W. Teh (Eds.), Proceedings of Machine Learning Research, Vol. 70, pp. 214–223. External Links: Link Cited by: §2.4.
  • [5] H. Asatryan, H. Gottschalk, M. Lippert, and M. Rottmann (2020) A convenient infinite dimensional framework for generative adversarial learning. arXiv preprint arXiv:2011.12087. Cited by: §2.2, §2.3, §2.3, §2.3, §4.1.
  • [6] P. B. Beaudan (1995) Numerical experiments on the flow past a circular cylinder at sub-critical reynolds number. Ph.D. Thesis, Stanford University. Cited by: §3.3.1.
  • [7] G. D. Birkhoff (1931) Proof of the ergodic theorem. Proceedings of the National Academy of Sciences 17 (12), pp. 656–660. External Links: Document, ISSN 0027-8424, Link, Cited by: §2.1.
  • [8] X. Chen, S. Z. Wu, and M. Hong (2020)

    Understanding gradient clipping in private sgd: a geometric perspective

    In Advances in Neural Information Processing Systems, H. Larochelle, M. Ranzato, R. Hadsell, M. F. Balcan, and H. Lin (Eds.), Vol. 33, pp. 13773–13782. External Links: Link Cited by: §2.4.
  • [9] S. H. Cheung, T. A. Oliver, E. E. Prudencio, S. Prudhomme, and R. D. Moser (2011) Bayesian uncertainty analysis with applications to turbulence modeling. Reliability Engineering & System Safety 96 (9), pp. 1137–1149. Note: Quantification of Margins and Uncertainties External Links: ISSN 0951-8320, Document, Link Cited by: §1.
  • [10] T. S. community (2008-2021) SciPy documentation. Note: Accessed: 04.12.2021 External Links: Link Cited by: §3.1.
  • [11] Deng,Zhiwen, He,Chuangxin, Liu,Yingzheng, and K. Chun (2019)

    Super-resolution reconstruction of turbulent velocity fields using a generative adversarial network-based artificial intelligence framework

    Physics of Fluids 31 (12), pp. 125111. External Links: Document, Link, Cited by: §1.
  • [12] T. Dietterich (1995) Overfitting and undercomputing in machine learning. ACM computing surveys (CSUR) 27 (3), pp. 326–327. Cited by: §4.1.
  • [13] V. Dumoulin and F. Visin (2018)

    A guide to convolution arithmetic for deep learning

    External Links: 1603.07285 Cited by: §2.4.
  • [14] W.N. Edeling, P. Cinnella, and R.P. Dwight (2014) Predictive rans simulations via bayesian model-scenario averaging. Journal of Computational Physics 275, pp. 65–91. External Links: ISSN 0021-9991, Document, Link Cited by: §1.
  • [15] W. Edeling, P. Cinnella, R. Dwight, and H. Bijl (2014-02) Bayesian estimates of parameter variability in the turbulence model. Journal of Computational Physics 258, pp. 73–94. External Links: Document Cited by: §1.
  • [16] T. Eisner, B. Farkas, M. Haase, and R. Nagel (2015) Operator theoretic aspects of ergodic theory. Springer International Publishing, Cham. External Links: Link Cited by: §2.1.
  • [17] T. S. Ferguson (2017) A course in large sample theory. Routledge. Cited by: §2.3.
  • [18] J.H. Ferziger and M. Perić (2008) "Computational Methods for Fluid Dynamics"., Berlin. External Links: Document Cited by: §3.2.
  • [19] U. Frisch and A. N. Kolmogorov (1995) Turbulence: the legacy of an kolmogorov. Cambridge university press. Cited by: §1.
  • [20] K. Fukami, K. Fukagata, and K. Taira (2019-05) Super-resolution reconstruction of turbulent flows with machine learning. Journal of Fluid Mechanics 870, pp. 106–120. External Links: ISSN 1469-7645, Link, Document Cited by: §1.
  • [21] K. Fukami, K. Fukagata, and K. Taira (2020) Machine learning based spatio-temporal super resolution reconstruction of turbulent flows. External Links: 2004.11566 Cited by: §1.
  • [22] K. G. and Moin,Parviz (2000) Numerical studies of flow over a circular cylinder at red=3900. Physics of Fluids 12 (2), pp. 403–417. External Links: Document Cited by: §3.3.1.
  • [23] A. Garcia-Garcia, S. Orts-Escolano, S. Oprea, V. Villena-Martinez, and J. Garcia-Rodriguez (2017) A review on deep learning techniques applied to semantic segmentation. External Links: 1704.06857 Cited by: §2.4.
  • [24] I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio (2014) Generative adversarial networks. External Links: 1406.2661 Cited by: §1.
  • [25] I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio (2014-06) Generative adversarial networks. Advances in Neural Information Processing Systems 3, pp. . External Links: Document Cited by: Figure 1, §2.2, §2.2.
  • [26] M. T. Hagan, H. B. Demuth, and M. Beale (1997) Neural network design. PWS Publishing Co.. Cited by: §4.1.
  • [27] C. He, Y. Liu, and L. Gan (2018-10) A data assimilation model for turbulent flows using continuous adjoint formulation. Physics of Fluids 30, pp. 105108. External Links: Document Cited by: §1.
  • [28] C. Hirsch (2007-01) "Numerical Computation of Internal and External Flows: The Fundamentals of Computational Fluid Dynamics". . External Links: Document Cited by: §3.2.
  • [29] Y. Ho and S. Wookey (2020) The real-world-weight cross-entropy loss function: modeling the costs of mislabeling. IEEE Access 8, pp. 4806–4813. External Links: Document Cited by: §2.2.
  • [30] S. Ioffe and C. Szegedy (2015-07–09 Jul) Batch normalization: accelerating deep network training by reducing internal covariate shift. In Proceedings of the 32nd International Conference on Machine Learning, F. Bach and D. Blei (Eds.), Proceedings of Machine Learning Research, Vol. 37, Lille, France, pp. 448–456. External Links: Link Cited by: §2.4.
  • [31] P. Isola, J. Zhu, T. Zhou, and A. Efros (2017-07) Image-to-image translation with conditional adversarial networks. pp. 5967–5976. External Links: Document Cited by: §2.2, §2.4.
  • [32] C. Jiang, J. Mi, S. Laima, and H. Li (2020-01) A novel algebraic stress model with machine-learning-assisted parameterization. Energies 13, pp. 258. External Links: Document Cited by: §1.
  • [33] T. Karras, S. Laine, and T. Aila (2019) A style-based generator architecture for generative adversarial networks. In

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    pp. 4401–4410. Cited by: §2.2.
  • [34] J. Kim and C. Lee (2020) Deep unsupervised learning of turbulence for inflow generation at various reynolds numbers. Journal of Computational Physics 406, pp. 109216. External Links: ISSN 0021-9991, Document, Link Cited by: §1, §1.
  • [35] P. Kim (2017) Convolutional neural network. In MATLAB Deep Learning: With Machine Learning, Neural Networks and Artificial Intelligence, pp. 121–147. External Links: ISBN 978-1-4842-2845-6, Document, Link Cited by: §2.4.
  • [36] R. King, P. Graf, and M. Chertkov (2017-11) Creating Turbulent Flow Realizations with Generative Adversarial Networks. In APS Division of Fluid Dynamics Meeting Abstracts, APS Meeting Abstracts, pp. A31.008. Cited by: §1.
  • [37] R. King, O. Hennigh, A. Mohan, and M. Chertkov (2018) From deep to physics-informed learning of turbulence: diagnostics. External Links: 1810.07785 Cited by: §1.
  • [38] D. P. Kingma and J. Ba (2014) Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980. Cited by: §4.1.
  • [39] A. N. Kolmogorov (1991) "The Local Structure of Turbulence in Incompressible Viscous Fluid for Very Large Reynolds Numbers". Mathematical and Physical Sciences 434 (1890), pp. 9–13. External Links: ISSN 09628444, Document Cited by: §3.2.
  • [40] N. V. Kuznetsov, T. N. Mokaev, O. A. Kuznetsova, and E. V. Kudryashova (2020) The lorenz system: hidden boundary of practical stability and the lyapunov dimension. Nonlinear Dynamics 102, pp. 713–732. Cited by: §3.1.
  • [41] E. Linder-Norén Pytorch-gan. Note: 12.11.2021 Cited by: §4.2.
  • [42] J. Ling, A. Kurzawski, and J. Templeton (2016) Reynolds averaged turbulence modelling using deep neural networks with embedded invariance. Journal of Fluid Mechanics 807, pp. 155–166. External Links: Document Cited by: §1.
  • [43] Liu,Bo, Tang,Jiupeng, Huang,Haibo, and Lu,Xi-Yun (2020) Deep learning methods for super-resolution reconstruction of turbulent flows. Physics of Fluids 32 (2), pp. 025105. External Links: Document, Link, Cited by: §1.
  • [44] E. N. Lorenz (1963) Deterministic nonperiodic flow. Journal of atmospheric sciences 20 (2), pp. 130–141. Cited by: §3.1, §4.1.
  • [45] W. Ma and J. Lu (2017) An equivalence of fully connected layer and convolutional layer. External Links: 1712.01252 Cited by: §2.4.
  • [46] M. Mirza and S. Osindero (2014) Conditional generative adversarial nets. External Links: 1411.1784 Cited by: Figure 2, §2.4.
  • [47] J. v. Neumann (1932) Proof of the quasi-ergodic hypothesis. Proceedings of the National Academy of Sciences 18 (1), pp. 70–82. External Links: Document, ISSN 0027-8424, Link, Cited by: §2.1.
  • [48] C. Norberg (1994) An experimental investigation of the flow around a circular cylinder: influence of aspect ratio. Journal of Fluid Mechanics 258, pp. 287–316. External Links: Document Cited by: §3.3.1.
  • [49] C. Nwankpa, W. Ijomah, A. Gachagan, and S. Marshall (2018) Activation functions: comparison of trends in practice and research for deep learning. External Links: 1811.03378 Cited by: §2.4.
  • [50] L. Ong and J. M. Wallace (1996) The velocity field of the turbulent very near wake of a circular cylinder. Experiments in Fluids 20, pp. 441–453. Cited by: §3.3.1.
  • [51] Z. Pan, W. Yu, X. Yi, A. Khan, F. Yuan, and Y. Zheng (2019) Recent progress on generative adversarial networks (gans): a survey. IEEE Access 7, pp. 36322–36333. External Links: Document Cited by: §2.4.
  • [52] E. J. Parish and K. Duraisamy (2016) A paradigm for data-driven predictive modeling using field inversion and machine learning. Journal of Computational Physics 305, pp. 758–774. External Links: ISSN 0021-9991, Document, Link Cited by: §1.
  • [53] Parnaudeau,Philippe, Carlier,Johan, Heitz,Dominique, and Lamballais,Eric (2008) Experimental and numerical studies of the flow over a circular cylinder at reynolds number 3900. Physics of Fluids 20 (8), pp. 085101. External Links: Document Cited by: §3.3.1.
  • [54] A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala (2019) PyTorch: an imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems 32, pp. 8024–8035. External Links: Link Cited by: §4.
  • [55] O. Peters (2019-12) The ergodicity problem in economics. Nature Physics 15, pp. 1216–1221. External Links: Document Cited by: §2.
  • [56] A. Radford, L. Metz, and S. Chintala (2016) Unsupervised representation learning with deep convolutional generative adversarial networks. In 4th International Conference on Learning Representations, ICLR 2016, San Juan, Puerto Rico, May 2-4, 2016, Conference Track Proceedings, Y. Bengio and Y. LeCun (Eds.), External Links: Link Cited by: §2.2, §2.4.
  • [57] S. Ruder (2016) An overview of gradient descent optimization algorithms. arXiv preprint arXiv:1609.04747. Cited by: §4.2.
  • [58] D. E. Rumelhart, G. E. Hinton, and R. J. Williams (1986) Learning representations by back-propagating errors. Nature 323, pp. 533–536. Cited by: Figure 1, §2.2.
  • [59] S. Shalev-Shwartz and S. Ben-David (2014) Understanding machine learning: from theory to algorithms. Cambridge university press. Cited by: §2.3.
  • [60] A. Singh and K. Duraisamy (2016-04) Using field inversion to quantify functional errors in turbulence closures. Physics of Fluids 28, pp. 045110. External Links: Document Cited by: §1.
  • [61] A. Singh, S. Medida, and K. Duraisamy (2016-08) Machine-learning-augmented predictive modeling of turbulent separated flows over airfoils. AIAA Journal 55, pp. . External Links: Document Cited by: §1.
  • [62] A. Subramaniam, M. L. Wong, R. D. Borker, S. Nimmagadda, and S. K. Lele (2020) Turbulence enrichment using physics-informed generative adversarial networks. External Links: 2003.01907 Cited by: §1.
  • [63] W. Tucker (1999) The lorenz attractor exists. Comptes Rendus de l’Académie des Sciences - Series I - Mathematics 328 (12), pp. 1197–1202. External Links: ISSN 0764-4442, Document, Link Cited by: §3.1.
  • [64] T. Wang, M. Liu, J. Zhu, A. Tao, J. Kautz, and B. Catanzaro (2018) High-resolution image synthesis and semantic manipulation with conditional gans. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vol. , pp. 8798–8807. External Links: Document Cited by: §2.2, §2.4, §2.4, §4.3.
  • [65] X. Wang, K. Yu, S. Wu, J. Gu, Y. Liu, C. Dong, Y. Qiao, and C. Change Loy (2018) Esrgan: enhanced super-resolution generative adversarial networks. In Proceedings of the European conference on computer vision (ECCV) workshops, Cited by: §2.2.
  • [66] J. Weatheritt and R.D. Sandberg (2017)

    The development of algebraic stress models using a novel evolutionary algorithm

    International Journal of Heat and Fluid Flow 68, pp. 298–318. External Links: ISSN 0142-727X, Document, Link Cited by: §1.
  • [67] J. Weatheritt and R. Sandberg (2016) A novel evolutionary algorithm applied to algebraic modifications of the rans stress–strain relationship. Journal of Computational Physics 325, pp. 22–37. External Links: ISSN 0021-9991, Document, Link Cited by: §1.
  • [68] M. Werhahn, Y. Xie, M. Chu, and N. Thuerey (2019-07) A multi-pass gan for fluid flow super-resolution. Proceedings of the ACM on Computer Graphics and Interactive Techniques 2 (2), pp. 1–21. External Links: ISSN 2577-6193, Link, Document Cited by: §1.
  • [69] B. Winhart, M. Sinkwitz, A. Schramm, P. Post, and F. di Mare (2020) Large eddy simulation of periodic wake impact on boundary layer transition mechanisms on a highly loaded low-pressure turbine blade. In Turbo Expo: Power for Land, Sea, and Air, Vol. 84102, pp. V02ET41A013. Cited by: §3.3.
  • [70] Y. Xie, E. Franz, M. Chu, and N. Thuerey (2017-07) Data-driven synthesis of smoke flows with cnn-based feature descriptors. ACM Transactions on Graphics 36 (4), pp. 1–14. External Links: ISSN 1557-7368, Link, Document Cited by: §1.
  • [71] M. Yang and Z. Xiao (2020-06) Improving the transition model by the field inversion and machine learning framework. Physics of Fluids 32, pp. . External Links: Document Cited by: §1.
  • [72] D. Yarotsky (2017) Error bounds for approximations with deep relu networks. Neural Networks 94, pp. 103–114. Cited by: §2.3.
  • [73] J. Zhang and S. Fu (2018) An efficient bayesian uncertainty quantification approach with application to transition modeling. Computers & Fluids 161, pp. 211–224. External Links: ISSN 0045-7930, Document, Link Cited by: §1.
  • [74] Cited by: §1.
  • [75] Y. Zhao, H. D. Akolekar, J. Weatheritt, V. Michelassi, and R. D. Sandberg (2020) RANS turbulence model development using cfd-driven machine learning. Journal of Computational Physics 411, pp. 109413. External Links: ISSN 0021-9991, Document, Link Cited by: §1.
  • [76] J. Zhu, T. Park, P. Isola, and A. A. Efros (2017) Unpaired image-to-image translation using cycle-consistent adversarial networks. In Computer Vision (ICCV), 2017 IEEE International Conference on, Cited by: §2.2.